Contents:Volume 112, Issue 1; July 5, 2005 Issue Highlights:

Issue Highlights

Circulation 2005 112: 1 Editor's Note:

Editor’s Note Joseph Loscalzo Circulation 2005 112: 2, Editorials:

Endothelial Progenitor Cells, Neointimal Hyperplasia, and Hemodialysis Vascular Access Dysfunction: Novel Therapies for a Recalcitrant Clinical Problem Prabir Roy-Chaudhury Circulation 2005 112: 3 - 5, Cell Death and Adenosine Triphosphate: The Paradox Jutta Schaper and Sawa Kostin Circulation 2005 112: 6 - 8, N-Terminal-Pro-B–Type Natriuretic Peptide: Universal Marker of Cardiovascular Risk? A. Mark Richards and Christopher M. Frampton Circulation 2005 112: 9 - 11, Original Articles:

Cardiovascular Surgery:

gfedc In Vivo Cell Seeding With Anti-CD34 Antibodies Successfully Accelerates Endothelialization but Stimulates Intimal Hyperplasia in Porcine Arteriovenous Expanded Polytetrafluoroethylene Grafts Joris I. Rotmans, Jan M.M. Heyligers, Hence J.M. Verhagen, Evelyn Velema, Machiel M. Nagtegaal, Dominique P.V. de Kleijn, Flip G. de Groot, Erik S.G. Stroes, and Gerard Pasterkamp Circulation 2005 112: 12 - 18; published online before print June 27 2005,

Coronary Heart Disease: gfedc Periodontal Disease and Coronary Heart Disease: A Reappraisal of the Exposure James D. Beck, Paul Eke, Gerardo Heiss, Phoebus Madianos, David Couper, Dongming Lin, Kevin Moss, John Elter, and Steven Offenbacher Circulation 2005 112: 19 - 24; published online before print June 27 2005, gfedc C-Reactive Protein and the 10-Year Incidence of Coronary Heart Disease in Older Men and Women: The Cardiovascular Health Study Mary Cushman, Alice M. Arnold, Bruce M. Psaty, Teri A. Manolio, Lewis H. Kuller, Gregory L. Burke, Joseph F. Polak, and Russell P. Tracy Circulation 2005 112: 25 - 31; published online before print June 27 2005,

Epidemiology: gfedc Obesity, Insulin Resistance, and the Metabolic Syndrome: Determinants of Endothelial Dysfunction in Whites and Blacks A.A. Lteif, K. Han, and K.J. Mather Circulation 2005 112: 32 - 38; published online before print June 27 2005,

Health Services and Outcomes Research: gfedc Adoption of Spironolactone Therapy for Older Patients With Heart Failure and Left Ventricular Systolic Dysfunction in the United States, 1998–2001 Frederick A. Masoudi, Cary P. Gross, Yongfei Wang, Saif S. Rathore, Edward P. Havranek, JoAnne Micale Foody, and Harlan M. Krumholz Circulation 2005 112: 39 - 47; published online before print June 27 2005,

Heart Failure: gfedc Effects of Candesartan on the Development of a New Diagnosis of Diabetes Mellitus in Patients With Heart Failure Salim Yusuf, Jan B. Ostergren, Hertzel C. Gerstein, Marc A. Pfeffer, Karl Swedberg, Christopher B. Granger, Bertil Olofsson, Jeffrey Probstfield, John V. McMurray on behalf of the Candesartan in Heart Failure—Assessment of Reduction in Mortality and Morbidity Program (CHARM) Investigators Circulation 2005 112: 48 - 53; published online before print June 27 2005, gfedc -Myosin Heavy Chain: A Sarcomeric Gene Associated With Dilated and Hypertrophic Phenotypes of Cardiomyopathy Elisa Carniel, Matthew R.G. Taylor, Gianfranco Sinagra, Andrea Di Lenarda, Lisa Ku, Pamela R. Fain, Mark M. Boucek, Jean Cavanaugh, Snjezana Miocic, Dobromir Slavov, Sharon L. Graw, Jennie Feiger, Xiao Zhong Zhu, Dmi Dao, Debra A. Ferguson, Michael R. Bristow, and Luisa Mestroni Circulation 2005 112: 54 - 59, Hypertension: gfedc Elevated Blood Pressure Linked to Primary Hyperaldosteronism and Impaired Vasodilation in BK Channel–Deficient Mice Matthias Sausbier, Claudia Arntz, Iancu Bucurenciu, Hong Zhao, Xiao-Bo Zhou, Ulrike Sausbier, Susanne Feil, Simone Kamm, Kyrill Essin, Claudia A. Sailer, Usamah Abdullah, Peter Krippeit-Drews, Robert Feil, Franz Hofmann, Hans-Günther Knaus, Chris Kenyon, Michael J. Shipston, Johan F. Storm, Winfried Neuhuber, Michael Korth, Rudolf Schubert, Maik Gollasch, and Peter Ruth Circulation 2005 112: 60 - 68; published online before print May 2 2005,

Imaging: gfedc Risk of Embolism and Death in Infective Endocarditis: Prognostic Value of Echocardiography: A Prospective Multicenter Study Franck Thuny, Giovanni Disalvo, Olivier Belliard, Jean-François Avierinos, Valeria Pergola, Valerie Rosenberg, Jean-Paul Casalta, Joanny Gouvernet, Geneviève Derumeaux, Diana Iarussi, Pierre Ambrosi, Raffaello Calabro, Alberto Riberi, Frédéric Collart, Dominique Metras, Hubert Lepidi, Didier Raoult, Jean-Robert Harle, Pierre-Jean Weiller, Ariel Cohen, and Gilbert Habib Circulation 2005 112: 69 - 75; published online before print June 27 2005,

Molecular Cardiology: gfedc Bcl-xL Gene Transfer Inhibits Bax Translocation and Prolongs Cardiac Cold Preservation Time in Rats Jianhua Huang, Kiminori Nakamura, Yoshinori Ito, Takeshi Uzuka, Masayuki Morikawa, Sachie Hirai, Kei Tomihara, Toshihiro Tanaka, Yukari Masuta, Keiji Ishii, Kazunori Kato, and Hirofumi Hamada Circulation 2005 112: 76 - 83; published online before print June 27 2005, gfedc Augmented Cardiac Hypertrophy in Response to Pressure Overload in Mice Lacking the Prostaglandin I2 Receptor Akiyoshi Hara, Koh-ichi Yuhki, Takayuki Fujino, Takehiro Yamada, Koji Takayama, Shuhko Kuriyama, Osamu Takahata, Hideji Karibe, Yuji Okada, Chun-Yang Xiao, Hong Ma, Shuh Narumiya, and Fumitaka Ushikubi Circulation 2005 112: 84 - 92; published online before print June 27 2005, gfedc Neuronal Nitric Oxide Synthase Mediates Statin-Induced Restoration of Vasa Nervorum and Reversal of Diabetic Neuropathy Masaaki Ii, Hiromi Nishimura, Kengo F. Kusano, Gangjian Qin, Young-sup Yoon, Andrea Wecker, Takayuki Asahara, and Douglas W. Losordo Circulation 2005 112: 93 - 102; published online before print June 27 2005,

Pediatric Cardiology:

gfedc Early Structural and Functional Changes of the Vasculature in HIV-Infected Children: Impact of Disease and Antiretroviral Therapy Marietta Charakida, Ann E. Donald, Hannah Green, Clare Storry, Margaret Clapson, Muriel Caslake, David T. Dunn, Julian P. Halcox, Diana M. Gibb, Nigel J. Klein, and John E. Deanfield Circulation 2005 112: 103 - 109; published online before print June 27 2005,

Stroke:

gfedc Prediction of Myocardial Infarction by N-Terminal-Pro-B-Type Natriuretic Peptide, C-Reactive Protein, and Renin in Subjects With Cerebrovascular Disease Duncan J. Campbell, Mark Woodward, John P. Chalmers, Samuel A. Colman, Alicia J. Jenkins, Bruce E. Kemp, Bruce C. Neal, Anushka Patel, and Stephen W. MacMahon Circulation 2005 112: 110 - 116; published online before print June 27 2005,

Vascular Medicine:

gfedc Statin Treatment After Onset of Sepsis in a Murine Model Improves Survival Marc W. Merx, Elisa A. Liehn, Jürgen Graf, Annette van de Sandt, Maren Schaltenbrand, Jürgen Schrader, Peter Hanrath, and Christian Weber Circulation 2005 112: 117 - 124, Contemporary Reviews in Cardiovascular Medicine:

gfedc Valvular Heart Disease: Aortic Regurgitation Raffi Bekeredjian and Paul A. Grayburn Circulation 2005 112: 125 - 134, New Drugs and Technologies:

Frontiers in Cardiovascular Magnetic Resonance Valentin Fuster and Raymond J. Kim Circulation 2005 112: 135 - 144, Special Reports:

A Vision for the Future: Opportunities and Challenges: Notes From the Director of the National Heart, Lung, and Blood Institute Elizabeth G. Nabel Circulation 2005 112: 145 - 146, Images in Cardiovascular Medicine:

Detection of Luminal-Intimal Border and Coronary Wall Enhancement in Intravascular Ultrasound Imaging After Injection of Microbubbles and Simultaneous Sonication With Transthoracic Echocardiography Manolis Vavuranakis, Ioannis A. Kakadiaris, Sean M. O’Malley, Christodoulos Stefanadis, Sophia Vaina, Maria Drakopoulou, Ioannis Mitropoulos, Stephane Carlier, and Morteza Naghavi Circulation 2005 112: e1 - e2, Detection of Carotid Atherosclerotic Plaque Ulceration, Calcification, and Thrombosis by Multicontrast Weighted Magnetic Resonance Imaging Baocheng Chu, Marina S. Ferguson, Hunter Underhill, Norihide Takaya, Jianming Cai, Michel Kliot, Chun Yuan, and Thomas S. Hatsukami Circulation 2005 112: e3 - e4, Primary Lymphoma of the Heart Jeffrey Kuvin, Nisha Parikh, Robert Salomon, Arthur Tischler, Philip Daoust, Yevgeniy Arshanskiy, Karl Coyner, Philip Carpino, Natesa G. Pandian, Carey Kimmelstiel, Caroline Foote, John Erban, and Hassan Rastegar Circulation 2005 112: e5 - e6, Correspondence:

Letter Regarding Article by McNair et al, "SCN5A Mutation Associated With Dilated Cardiomyopathy, Conduction Disorder, and Arrhythmia" • Response W.A. Groenewegen, A.A.M. Wilde, William P. McNair, Lisa Ku, Matthew R.G. Taylor, Pam R. Fain, Eugene Wolfel, and Luisa Mestroni Circulation 2005 112: e9 - e10, Letter Regarding Article by Galbreath et al, "Long-Term Healthcare and Cost Outcomes of Disease Management in a Large, Randomized, Community-Based Population With Heart Failure" • Response Ariel Linden, Thomas Wilson, Autumn Dawn Galbreath, Gregory L. Freeman, Brad Smith, Richard A. Krasuski, Karl C. Stajduhar, Michael D. Kwan, and Robert Ellis Circulation 2005 112: e11, Book Reviews:

Computed Tomography of the Coronary Arteries Frederick L. Ruberg Circulation 2005 112: e7 - e8, Acknowledgment of Reviewers:

Acknowledgment of Reviewers

Circulation 2005 112: e12 - e26,

Circulation JOURNAL OF THE AMERICAN HEART ASSOCIATION IIssssuuee HHiigghhlliigghhttss Vol 111, No 1, July 5, 2005

PERIODONTAL DISEASE AND CORONARY HEART STATIN TREATMENT AFTER ONSET OF SEPSIS IN DISEASE: A REAPPRAISAL OF THE EXPOSURE, by A MURINE MODEL IMPROVES SURVIVAL, by Merx Beck et al. et al. There currently is great interest in understanding the role of certain The clinical benefits achieved with HMG-CoA reductase inhibi- chronic infections as risk factors for cardiovascular disease. In this tors have been shown to extend well beyond the recognized regard, prior studies have suggested that chronic periodontal lipid-lowering effects of these agents. The pleiotropic effects disease is associated with increased cardiovascular risk, although attributed to statins include increased bioavailable nitric oxide, it remains possible that the confounding effects of smoking and increased antioxidant properties, inhibition of inflammatory re- other classical risk factors explain the association. In this issue of sponses, and improvement of endothelial dysfunction. Interest- Circulation, Beck and colleagues investigated the relation be- ingly, investigators have recognized that these pleiotropic effects tween periodontal disease and prevalent coronary heart disease in may be exploited to offer a therapeutic advantage in clinical 5002 participants in the fourth examination of the Atherosclerosis situations other than atherothrombotic vascular disease. For ex- Risk in Communities Study. They observed that elevated serum ample, Merx et al have previously shown that the antiinflamma- IgG antibodies to several oral pathogens were associated with tory effects of statins mediate responses to sepsis; pretreatment coronary heart disease, but findings on oral examination were not. with statins improved survival in a murine model of sepsis. In this Although further prospective studies of this question are needed, issue of Circulation, Merx et al extend their previous work and the present study suggests that the host response to oral infection examine the effects of initiating statin therapy after sepsis has been may be more important for cardiovascular risk than the local established. These studies may have implications for how clini- extent of periodontal disease. See p 19. cians treat sepsis in the future. See p 117.

EFFECTS OF CANDESARTAN ON THE Visit http://www.circ.ahajournals.org: DEVELOPMENT OF A NEW DIAGNOSIS OF DIABETES MELLITUS IN PATIENTS WITH HEART Images in Cardiovascular Medicine FAILURE, by Yusuf et al. Detection of Luminal-Intimal Border and Coronary Wall En- Angiotensin-converting enzyme inhibitors and angiotensin recep- hancement in Intravascular Ultrasound Imaging After Injec- tor blockers (ARBs) have been associated with a lower incidence tion of Microbubbles and Simultaneous Sonication With of diabetes mellitus in clinical trials of hypertensives and other Transthoracic Echocardiography. See p e1. high-risk patients. However, it is less clear if use of renin-angio- Detection of Carotid Atherosclerotic Plaque Ulceration, Cal- tensin axis inhibitors lowers the risk of diabetes in patients with cification, and Thrombosis by Multicontrast Weighted Mag- heart failure. In this issue of Circulation, Yusuf and colleagues netic Resonance Imaging. See p e3. analyze data from the CHARM trial to assess the risk of devel- Primary Lymphoma of the Heart. See p e5. oping new-onset diabetes (a prespecified secondary outcome) in more than 5000 heart failure patients randomized to candesartan (an ARB) or placebo. The investigators report that candesartan use was associated with a 22% reduction in incidence of diabetes compared with placebo. The reduction in incidence of diabetes was consistent across clinical subgroups (age, sex, body mass index) and was particularly striking in patients with a relatively preserved ejection fraction. The authors emphasize that use of ARBs in heart failure patients provides added value by lowering the future risk of diabetes. See p 48.

Book Review Computed Tomography of the Coronary Arteries. See p e7.

Correspondence See p e9.

1 Editor’s Note

eginning in this issue of Circulation, we present two new features designed to enhance the transfer of information to our clinician readers. First, you will notice that the cover Bincludes a new box in the lower right corner listing articles we believe comprise a clinical cardiology curriculum. These articles include those in our regular review series as well as online information directly relevant to clinical practice. Second, you will find at the end of original articles with a basic focus a text box that includes a short clinical perspective on the information contained within that article. Articles published with an accompanying “Clinical Perspective” will be noted in the Table of Contents. We trust that these changes will enhance the appeal of the journal to our clinical readership, and we welcome other suggestions you may have for improving the quality of Circulation. Joseph Loscalzo, MD, PhD Editor-in-Chief, Circulation

2 Editorial

Endothelial Progenitor Cells, Neointimal Hyperplasia, and Hemodialysis Vascular Access Dysfunction Novel Therapies for a Recalcitrant Clinical Problem

Prabir Roy-Chaudhury, MD, PhD

ascular stenosis as a result of neointimal hyperplasia hyperplasia, as expected. Rather, the endothelialized grafts is a major clinical problem that has an impact on demonstrated a paradoxical increase in neointimal hyperpla- Vmultiple and diverse disciplines, including cardiolo- sia at the graft-vein anastomosis, which is the end point for gy (coronary restenosis), cardiothoracic and vascular surgery this model. (saphenous vein and polytetrafluoroethylene [PTFE] graft The article by Rotmans and colleagues comes at a partic- failure), neurology (carotid stenosis), nephrology (dialysis ularly opportune moment, as it addresses 3 issues of current access dysfunction), and transplant medicine (chronic allo- interest: (1) it recognizes that the combination of recent graft rejection in hearts and kidneys). The traditional response advances in biomedical engineering, drug delivery, and mo- to injury hypothesis on the pathogenesis of neointimal hyper- lecular biology has resulted in a fertile substrate for investi- plasia focuses on the migration of medial smooth muscle cells gators in this field; (2) it challenges conventional wisdom from the media into the intima.1 Recently, there has been a about the inverse association between endothelial repair great deal of excitement about the role of circulating smooth (endothelialization) and neointimal hyperplasia; and (3) it muscle progenitor cells in the pathogenesis of neointimal draws attention to the huge but often ignored clinical problem hyperplasia. These cells have been identified in a variety of of hemodialysis vascular access dysfunction. This editorial experimental models of vascular injury,2 and interventions addresses each of these issues in turn. that reduce the number of these cells can attenuate neointimal There have been great advances in the past 5 to 10 years in hyperplasia. In marked contrast to the deleterious effects of the in vivo application of both experimental and clinical smooth muscle progenitor cells on neointimal hyperplasia, therapies for vascular stenosis and neointimal hyperplasia. circulating endothelial progenitor cells (EPCs) are believed to These advances have been made possible through a fusion of play an important role in vascular repair and in the inhibition advances in biomaterials, drug delivery techniques, and 3 of neointimal hyperplasia. molecular and cell biology. A few examples of such technol- Seep12 ogies include a multitude of drug-eluting coronary stents,6 the use of perivascular drug-releasing or cell-containing poly- In this issue of Circulation, Rotmans and colleagues have mers in experimental models of neointimal hyperplasia,7 and attempted to achieve the “holy grail” for a vascular access exciting new advances in the generation of nitric oxide– procedure: rapid and complete endothelialization.4 Their releasing polymers.9 All of these approaches share a common experiments are based on a newly developed technique that thread in that they are local therapies applied directly to the can coat the surface of stent and graft material with antibodies site of vascular injury. In particular, the remarkable success against CD34 (Orbus Medical Technologies). CD34 is a of the drug-eluting coronary stents against the background of marker for hematopoietic stem cells, and previous research multiple unsuccessful clinical trials of systemic therapies for has demonstrated that coronary stents coated with anti-CD34 neointimal hyperplasia suggests that local therapy could be have a significant increase in endothelialization as early as 1 hour after deployment because of the binding of circulating the delivery mode of choice in the setting of vascular stenosis. ϩ For many years the unattainable goal for vascular access CD34 cells, which then differentiate into endothelial cells.5 Placement of anti-CD34–coated grafts in an arteriovenous interventions has been the complete endothelialization of the model of PTFE graft stenosis by Rotmans et al resulted in region of vascular injury after surgery or angioplasty to almost complete endothelialization of the grafts (both at 3 prevent both thrombosis and stenosis. The recent identifica- days and at the end of the experiments at 28 days). Endothe- tion of EPCs has allowed us to come another step closer to lialization, however, did not result in a decrease in neointimal this goal. EPCs were initially identified by Asahara and colleagues,10 who demonstrated that a subset of CD34ϩ cells could be differentiated ex vivo into an endothelial phenotype The opinions expressed in this article are not necessarily those of the (which expressed both the stem cell marker CD34 and the editors or of the American Heart Association. From the University of Cincinnati Medical Center, Cincinnati, Ohio. endothelial marker protein VEGFR2). Other groups have Correspondence to Dr Prabir Roy-Chaudhury, Division of Nephrolo- since demonstrated enhanced endothelial coverage and a gy, MSB G-258, University of Cincinnati, 231 Albert Sabin Way, reduction in neointimal hyperplasia after an infusion of EPCs Cincinnati, OH 45267-0585. E-mail [email protected] 11,12 (Circulation. 2005;112:3-5.) in animal models of carotid angioplasty injury. In all of © 2005 American Heart Association, Inc. these experiments, an increase in endothelialization of the Circulation is available at http://www.circulationaha.org region of vascular injury invariably translated into a reduction DOI: 10.1161/CIRCULATIONAHA.105.548651 of neointimal hyperplasia. In the article by Rotmans et al, 3 4 Circulation July 5, 2005 however, endothelialization of PTFE graft with anti-CD34 number that is expected to double by the year 2020.8 The 2 antibodies was accompanied by an increase rather than a main forms of dialysis access are the arteriovenous fistula decrease in neointimal hyperplasia. What could be the pos- (32% prevalence rate in the United States) and the PTFE sible reasons for this paradox? arteriovenous graft (50% prevalence rate in the United Previous studies have demonstrated that it is possible to States), both of which have dismal survival rates as compared obtain endothelial coverage of prosthetic stent or graft mate- with other vascular procedures—between 50% and 75% rial by isolating or expanding EPC cultures ex vivo and then 1-year survival at best.17 The single most important reason for using these cells to coat prosthetic material.13–15 Unfortu- hemodialysis vascular access dysfunction is a stenosis at the nately, these approaches tend to be time consuming, labor graft-vein anastomosis of PTFE dialysis access grafts or in intensive, and expensive. In marked contrast, the ability to the proximal vein of an arteriovenous fistula as a result of store anti-CD34–coated graft or stent material in the cardiac venous neointimal hyperplasia. At a histological level, we catheterization laboratory, interventional radiology suite, or and others have demonstrated that venous neointimal hyper- operating room clearly lends itself to current clinical practice. plasia in the setting of dialysis access grafts and fistulae The Rotmans group, very appropriately, did not attempt an comprises smooth muscle cells, myofibroblasts, microvessels ex vivo endothelialization of their prosthetic graft material within the neointima, and a peri-graft macrophage layer.18 In but instead tried to achieve this in vivo through the use of marked contrast to arterial stenoses, venous neointimal hy- anti-CD34–coated grafts. As alluded to by the authors, the perplasia is difficult to treat with angioplasty (nonthrombosed reason for the dichotomy between endothelialization and PTFE grafts have a 50%, 6-month primary patency, whereas intimal hyperplasia in these experiments could be linked to thrombosed grafts have a 40%, 3-month primary patency the fact that CD34 is a hematopoietic stem cell marker. The 17 ϩ after angioplasty ). The reasons for this poor survival rate are antibodies to CD34, therefore, could have attracted CD34 unclear, but they include (1) anatomic and physiological cells that still had the potential to transform into smooth differences between veins and arteries, (2) the presence of muscle cells and myofibroblasts, as has been previously continuous hemodynamic stress in the form of turbulence and 16 reported. Why was there no evidence of smooth muscle possibly low shear stress at the graft-vein or artery-vein cells on the actual PTFE graft material, which was covered anastomosis, (3) recurrent needle damage during the dialysis only by cells that had an endothelial phenotype? One expla- procedure, and (4) the presence of uremia, which could nation could be that the transformation of CD34ϩ cells predispose a patient to endothelial dysfunction.19 preferentially into smooth muscle cells may require addi- Regardless of the cause, the aggressive natural history of tional stimuli such as turbulence or low shear stress, which neointimal hyperplasia in the setting of hemodialysis vascular would be present only at the graft-vein anastomosis. Alter- access translates into a financial cost of Ϸ$1 billion per year. natively, true EPCs/endothelial cells themselves may have the More important, it results in tremendous morbidity and is plasticity to dedifferentiate into smooth muscle cells at the responsible for 20% of all hospital admissions in the hemo- graft-vein anastomosis in response to hemodynamic and dialysis population.19 Despite the huge clinical, economic, surgical stress. Finally, it needs to be mentioned that these and social impact of hemodialysis vascular access dysfunc- studies were performed in a model of venous (rather than arterial) stenosis, which is characterized by continuous ongo- tion, there are no truly effective therapies available for this ing hemodynamic stress and compliance mismatch at the clinical problem. Percutaneous angioplasty and surgical revi- graft-vein anastomosis. Whether the results would have been sion are commonly used, but in marked contrast to the cardiac different in an arterial interposition or angioplasty model is literature, these are not accompanied by interventions to open to speculation. prevent restenosis. Even more disappointing, when compared Despite its surprising results, the study by Rotmans et al is with the multitude of clinical trials in the setting of coronary critically important because it is an important step forward stenosis, is that there are very few ongoing clinical trials toward the application of EPC technology to the problem of targeting hemodialysis vascular access dysfunction. The lack vascular stenosis in a manner that is clinically relevant. of focused clinical research in this area is all the more Further refinements to this approach could include an anti- surprising because dialysis access grafts and fistulae could be body coating that has the specificity to bind to circulating the ideal clinical model for testing novel local therapeutic cells that only have the potential to differentiate into lining interventions for vascular stenosis in general. This is because endothelial cells. To do this, however, we will need to learn (1) the aggressive nature of vascular stenosis in hemodialysis more about the biology and plasticity of EPCs. For example, patients could result in clinical trials being conducted with a it is likely that what we consider to be EPCs are in fact a smaller sample size and in a shorter time; (2) dialysis access diverse population of different cell types with different grafts and fistulae are superficially located and away from functions.3 Clearly, many more studies are needed for us to important anatomic structures, making these patients ideal learn how to tweak this fascinating cell so that it can be candidates for the delivery of local therapies either through effectively used to prevent vascular stenosis in the clinical the percutaneous approach or at the time of surgical place- setting. ment; and (3) patients on hemodialysis have large-bore Finally, by using an arteriovenous model of PTFE graft needles placed within 3 to 6 cm of the site of venous stenosis stenosis, the authors draw attention to the huge clinical 3 times per week for dialysis, which could allow for the problem of hemodialysis vascular access dysfunction. There repeated delivery of novel therapies (including endothelial are Ϸ300 000 hemodialysis patients in the United States, a progenitor cells) during the dialysis procedure itself. Roy-Chaudhury EPCs for Hemodialysis Vascular Access Dysfunction 5

In summary, the article by Rotmans et al brings out both 8. US Renal Data System. USRDS 2002 Annual Data Report: Atlas of the advantages and the pitfalls of using EPCs to reduce End-Stage Renal Disease in the United States. Bethesda, Md: National Institutes of Health, National Institute of Diabetes and Digestive and vascular stenosis. Regardless of its final outcome, this study Kidney Diseases; 2002. has brought us one step closer to the use of EPCs in the 9. Frost MC, Reynolds MM, Meyerhoff ME. Polymers incorporating nitric setting of clinical vascular stenosis. Perhaps most important, oxide releasing/generating substances for improved biocompatibility of blood-contacting medical devices. Biomaterials. 2005;26:1685–1693. this article draws attention to hemodialysis vascular access 10. Asahara T, Murohara T, Sullivan A, Silver M, van der Zee R, Li T, dysfunction, a huge clinical problem that is lacking in Witzenbichler B, Schatteman G, Isner JM. Isolation of putative progenitor effective treatments but at the same time could be ideally endothelial cells for angiogenesis. Science. 1997;275:964–967. 11. Werner N, Junk S, Laufs U, Link A, Walenta K, Bohm M, Nickenig G. suited to the application of novel local therapeutic Intravenous transfusion of endothelial progenitor cells reduces neointima interventions. formation after vascular injury. Circ Res. 2003;93:e17–e24. 12. Griese DP, Ehsan A, Melo LG, Kong D, Zhang L, Mann MJ, Pratt RE, Mulligan RC, Dzau VJ. Isolation and transplantation of autologous cir- Acknowledgments culating endothelial cells into denuded vessels and prosthetic grafts: This work was supported by National Institutes of Health grant implications for cell-based vascular therapy. Circulation. 2003;108: DK-61689 and by a grant from Satellite Dialysis Clinics. 2710–2715. 13. He H, Shirota T, Yasui H, Matsuda T. Canine endothelial progenitor References cell-lined hybrid vascular graft with nonthrombogenic potential. J Thorac Cardiovasc Surg. 2003;126:455–464. 1. Ross R. The pathogenesis of atherosclerosis: a perspective for the 1990s. 14. Bhattacharya V, McSweeney PA, Shi Q, Bruno B, Ishida A, Nash R, Nature. 1993;362:801–809. Storb RF, Sauvage LR, Hammond WP, Wu MH. Enhanced endotheli- 2. Sata M, Saiura A, Kunisato A, Tojo A, Okada S, Tokuhisa T, Hirai H, alization and microvessel formation in polyester grafts seeded with Makuuchi M, Hirata Y, Nagai R. Hematopoietic stem cells differentiate CD34(ϩ) bone marrow cells. Blood. 2000;95:581–585. into vascular cells that participate in the pathogenesis of atherosclerosis. 15. Shirota T, Yasui H, Shimokawa H, Matsuda T. Fabrication of endothelial Nat Med. 2002;8:403–409. progenitor cell (EPC)–seeded intravascular stent devices and in vitro 3. Urbich C, Dimmeler S. Endothelial progenitor cells: characterization and endothelialization on hybrid vascular tissue. Biomaterials. 2003;24: role in vascular biology. Circ Res. 2004;95:343–353. 2295–2302. 4. Rotmans JI, Heyligers JMM, Verhagen HJM, Velema E, Nagtegaal MM, 16. Yeh ET, Zhang S, Wu HD, Korbling M, Willerson JT, Estrov Z. Trans- de Kleijn DPV, de Groot FG, Stroes ESG, Pasterkamp G. In vivo cell differentiation of human peripheral blood CD34ϩ-enriched cell popu- seeding using anti-CD34 antibodies successfully accelerates endotheli- lation into cardiomyocytes, endothelial cells, and smooth muscle cells in alization but stimulates intimal hyperplasia in porcine arteriovenous vivo. Circulation. 2003;108:2070–2073. expanded polytetrafluoroethylene grafts. Circulation. 2005;112:12–18. 17. Schwab SJ, Harrington JT, Singh A, Roher R, Shohaib SA, Perrone RD, 5. Ong AT, Aoki J, Kutryk MJ, Serruys PW. How to accelerate the endo- Meyer K, Beasley D. Vascular access for hemodialysis. Kidney Int. thelialization of stents. Arch Mal Coeur Vaiss. 2005;98:123–126. 1999;55:2078–2090. 6. Morice MC, Serruys PW, Sousa JE, Fajadet J, Ban Hayashi E, Perin M, 18. Roy-Chaudhury P, Kelly BS, Miller MA, Reaves A, Armstrong J, Colombo A, Schuler G, Barragan P, Guagliumi G, Molnar F, Falotico R. Nanayakkara N, Heffelfinger SC. Venous neointimal hyperplasia in poly- A randomized comparison of a sirolimus-eluting stent with a standard tetrafluoroethylene dialysis grafts. Kidney Int. 2001;59:2325–2334. stent for coronary revascularization. N Engl J Med. 2002;346:1773–1780. 19. Roy-Chaudhury P, Kelly BS, Zhang J, Narayana A, Desai P, Melham M, 7. Golomb G, Fishbein I, Banai S, Mishaly D, Moscovitz D, Gertz SD, Gazit Duncan H, Heffelfinger SC. Hemodialysis vascular access dysfunction: A, Poradosu E, Levitzki A. Controlled delivery of a tyrphostin inhibits from pathophysiology to novel therapies. Blood Purif. 2003;21:99–110. intimal hyperplasia in a rat carotid artery injury model. Atherosclerosis. 1996;125:171–182. KEY WORDS: Editorials Ⅲ stenosis Ⅲ grafting Ⅲ hyperplasia Ⅲ endothelium Editorial

Cell Death and Adenosine Triphosphate The Paradox

Jutta Schaper, MD, PhD; Sawa Kostin, MD

he inhibition of cell death lies at the heart of the both modes of cell death employ different mechanisms of interesting publication in this issue of Circulation by initiation and cellular execution.5 Apoptosis is a prepro- THuang et al.1 The authors’ major message is that grammed (suicidal), mostly caspase-driven and energy- Bcl-xL gene transfer prolongs the cold preservation time of dependent process, whereas oncotic cell death is accidental rat hearts destined for transplantation. Bcl-xL belongs to the (because ischemic injury is not preprogrammed), is indepen- large Bcl-2 family and has been reported to inhibit Bax dent of caspase activation, and occurs subsequent to ATP translocation to the mitochondria and to reduce cytochrome c depletion.6 In addition, a third type of cellular demise has release, thereby interrupting the apoptotic cascade and reduc- been described, autophagic cell death, but this may not be of ing the number of cells dying by apoptosis.2,3 This is precisely great importance in the present animal model because it what Huang et al1 have shown—namely, that inhibition of the occurs mostly in tissue exhibiting chronic degeneration, such apoptotic pathway can partially prevent the deleterious ef- as Alzheimer’s disease in the brain and failure of the human fects of long-term ischemia by reducing the rate of apoptosis. heart.7,8 In addition, these authors show that “infarct size” was Huang et al1 showed that Bcl-xL gene transfer prevented reduced in treated hearts and that the rate of polymorphonu- Bax loss from the cytosol and decreased cytochrome c release clear cell infiltration was minimal as compared with hearts from the mitochondria, whereas Bax was translocated from without gene transfer. Infarct size was determined in Lange- the cytosol to the mitochondria and caused massive cyto- ndorff buffer–perfused hearts without any coronary artery chrome c release in untreated hearts.1 These data are in line ligation and describes the tissue area with necrotic cells as with accumulating evidence suggesting an intimate interrela- demonstrated by triphenyltetrazolium (TTC) staining. To the tionship of apoptosis and mitochondrial function that occurs uninitiated reader it appears, therefore, that an intervention at the interface between Bcl-2 family proteins and the outer aimed at reducing apoptosis is also able to interfere with the mitochondrial membrane protein, namely voltage-dependent process of ischemic cell death, ie, with oncosis (this is the preferred, anion channel (VDAC). Although Bcl-xL was shown to modern term for necrosis; necrosis proper is the process of cell stimulate VDAC closure, thereby preventing mitochondrial demise after any type of cell death4). The number of cardiomyo- alterations in response to death stimuli, Bax protein promotes cytes dying by either cell death mechanism, apoptotic or oncotic, its opening to release cytochrome c and promote apoptosis.9 was reduced: Only 6.5% of all cardiomyocytes were terminal In the accompanying article by Huang et al,1 however, deoxynucleotidyl transferase biotin-dUTP nick end labeling there is no clear-cut explanation of the simultaneous occur- (TUNEL) positive in hearts with Bcl-xL gene transfer, as rence of apoptosis and oncosis and their inhibition by Bcl-xL compared with 18.9% in the untreated hearts, and infarct size gene transfer. Because, unfortunately, myocardial ATP levels was 23% in the treated group versus 47.7% in the hearts without were not measured, the explanation of these puzzling results gene transfer. Creatine kinase (CK) activity measured in the must be speculative. Referring to the work of Tatsumi et al10 coronary outflow of Langendorff-perfused hearts was likewise and Kuznetsov et al,11 Huang et al believe that the inhibition reduced in hearts with Bcl-xL gene transfer. of cytochrome c release and preservation of mitochondrial Seep76 function are decisive in preventing myocardial injury. Kuznetsov et al reported that cytochrome c release from the The cell death paradox is here represented by the astound- mitochondria, usually held to be a major proapoptotic event, ing phenomenon that apparently 2 modes of cell death, caused necrosis by reducing ATP levels, and that the hetero- apoptotic and oncotic, were influenced by Bcl-xL gene geneity and extent of cytochrome c release regulate the transfer, which should exclusively interfere with the apoptotic switch between apoptosis and necrosis.11 Tatsumi et al10 pathway. This is surprising given the well-known fact that reported that intracellular ATP plays an important role in the execution of apoptosis and not of necrosis (or oncosis), The opinions expressed in this article are not necessarily those of the confirming earlier results by Leist et al.12 Tatsumi and editors or of the American Heart Association. colleagues demonstrated in neonatal cardiomyocytes a dis- From the Department of Experimental Cardiology, Max-Planck- Institute, Bad Nauheim, Germany. tinct correlation between the rates of occurrence of apoptotic Correspondence to Jutta Schaper, MD, Max-Planck-Institute, cell death and ATP levels (ie, ATP is necessary for apoptosis Benekestrasse 2, D-61231 Bad Nauheim, Germany. E-mail to take place and more apoptosis occurs at higher ATP [email protected] (Circulation. 2005;112:6-8.) levels). Furthermore, they showed that necrosis appeared © 2005 American Heart Association, Inc. exclusively in the presence of total glucose deprivation Circulation is available at http://www.circulationaha.org accompanied by total loss of ATP. They also demonstrated DOI: 10.1161/CIRCULATIONAHA.105.551846 that constitutively expressed caspase-3 is not activated in the 6 Schaper and Kostin Cell Death Paradox 7 absence of ATP; however, cleavage of the procaspase (32 gave false-positive results. The mechanism of action of kDa) resulting in the activated form of caspase-3 (17 kDa) Bcl-xL gene transfer on endothelial cells, however, remains was observed in the presence of higher ATP levels that unknown and its likelihood most probably is low. resulted in apoptotic cell death.10 Thus, it appears that oncosis Unsolved technical problems in determining the rate of occurs at low to zero ATP levels and apoptosis can only occur apoptosis may also play a role.15,16 The TUNEL reaction has in the presence of higher intracellular energy stores. been shown to label not only apoptotic but also necrotic When transferring this knowledge to the data presented by cells,17 as well as those undergoing DNA replication18 and Huang et al,1 it must be concluded that ATP levels in both this may result in false-positive numbers. For a positive treated and untreated hearts must have been different in control of this technique, the use of tissue with a known rate different areas of the myocardium, resulting in either necrotic of apoptosis, such as small intestine, is strongly recom- or apoptotic cell death. Their results furthermore suggest that mended. The possibility that TUNEL and TTC reaction may ATP levels in the myocardium treated with Bcl-xL gene be potentially misleading should be kept in mind to exclude transfer were elevated because infarct size (ie, the area of that the data presented may be subject to artifacts. necrotic cells) was reduced in this situation. Because the In conclusion, the work by Huang et al1 is not only number of TUNEL-positive cells was lower in hearts with interesting but also thought provoking because of the aston- Bcl-xL gene transfer, ATP levels should have been high ishing effect that antiapoptotic treatment is able to reduce the enough to preserve the tissue and to prevent the occurrence of rate of apoptotic and oncotic (ischemic) cell death. Mecha- apoptosis. Because ATP values are unknown and caspase-3 nisms of action are unknown at present, and an explanation is and its degree of activation were not determined, however, hampered by the lack of data on local ATP content and the this attempt to explain the data presented by Huang et al activation of caspase-3. The discussion of the role of apo- remains purely speculative and more work is needed to ptotic, necrotic, and autophagic cell death in ischemia/ clarify this issue. reperfusion is ongoing, and before any therapeutic principles Furthermore, in the work of Tatsumi et al there is no are recommended or used in human studies, the apparent cell evidence-based reason that allows the conclusion that necro- death paradox should be explored and clarified in more detail sis and apoptosis share common initial mechanisms.10 In our in future studies. opinion, this proposal is not very likely because one of the critical differences between these 2 modes of cell death is that damage to the sarcolemma resulting in cell membrane leak- References 1. Huang J, Nakamura K, Ito Y, Uzuka T, Morikawa M, Hirai S, Tomihara iness is one of the major and early hallmarks of oncotic cell K, Tanaka T, Masuta Y, Ishii K, Kato K, Hamada H. Bcl-xL gene transfer death, whereas in apoptosis the membrane remains intact and inhibits Bax translocation and prolongs cardiac cold preservation time in functional.5 The authors noticed neither a simultaneous oc- rats. Circulation. 2005;112:76–83. currence of apoptosis and necrosis in the same culture dish 2. He L, Perkins GA, Poblenz AT, Harris JB, Hung M, Ellisman MH, Fox DA. Bcl-xL overexpression blocks bax-mediated mitochondrial contact nor the transition of necrosis to apoptosis or vice versa in the site formation and apoptosis in rod photoreceptors of lead-exposed mice. same cell. For this reason it must be concluded from their Proc Natl Acad Sci U S A. 2003;100:1022–1027. work that cells die either by apoptosis or by necrosis, 3. Green DR, Kroemer G. The pathophysiology of mitochondrial cell death. according to the intracellular energy level. A common mech- Science. 2004;305:626–629. 4. Levin S, Bucci TJ, Cohen SM, Fix AS, Hardisty JF, LeGrand EK, anism of initiation is questionable. Maronpot RR, Trump BF. The nomenclature of cell death: recommen- There is ongoing discussion of whether apoptosis or dations of an ad hoc Committee of the Society of Toxicologic Pathol- necrosis is the primary form of cell death after myocardial ogists. Toxicol Pathol. 1999;27:484–490. injury, especially in the setting of ischemia/reperfusion.13,14 5. Majno G, Joris I. Apoptosis, oncosis and necrosis. Am J Pathol. 1995; 146:3–15. This may be the result of varying experimental design as well 6. Kostin S. Pathways of myocyte death: Implications for development of as unsolved technical issues in determining the mode and rate clinical laboratory markers. Adv Clin Chem. In press. of cell death. 7. Kostin S, Pool L, Elsässer A, Hein S, Drexler HC, Arnon E, Hayakawa The problem of the no-reflow phenomenon is inherent to Y, Zimmermann R, Bauer EP, Klövekorn WP, Schaper J. Myocytes die models of long-term ischemia followed by reperfusion. Is- by multiple mechanisms in failing human hearts. Circ Res. 2003;92: 715–724. chemia causes injury not only to the cardiomyocytes but also 8. Hein S, Arnon E, Kostin S, Schönburg M, Elsässer A, Polyakova V, to capillaries and the arteriolar endothelium so that microvas- Bauer EP, Klövekorn WP, Schaper J. Progression from compensated cular edema is common in ischemic myocardium. This cell hypertrophy to failure in the pressure-overloaded human heart: structural swelling obliterates the vascular lumen and prevents tissue deterioration and compensatory mechanisms. Circulation. 2003;107: 984–991. perfusion during reperfusion. The TTC reaction used to show 9. Shimizu S, Narita M, Tsujimoto Y. Bcl-2 family proteins regulate the the occurrence of necrotic cells may then show false-positive release of apoptogenic cytochrome c by the mitochondrial channel results. In other words, dead tissue, instead of being TTC VDAC. Nature. 1999;399:483–487. negative, shows the brick red positive stain because NADH 10. Tatsumi T, Shiraishi J, Keira N, Akashi K, Mano A, Yamanaka S, Matoba S, Fushiki S, Fliss H, Nakagawa M. Intracellular ATP is required for (or NAD) has not been washed out. A reperfusion marker mitochondrial apoptotic pathways in isolated hypoxic rat cardiac such as fluorescein should have been used to show that tissue myocytes. Cardiovasc Res. 2003;59:428–440. perfusion is reestablished and uniform in the areas with 11. Kuznetsov AV, Schneeberger S, Seiler R, Brandacher G, Mark W, positive TTC reaction. It cannot be excluded (but again, this Steurer W, Saks V, Usson Y, Margreiter R, Gnaiger E. Mitochondrial defects and heterogeneous cytochrome c release after cardiac cold ische- is purely speculative) that endothelial edema was reduced by mia and reperfusion. Am J Physiol Heart Circ Physiol. 2004;286: Bcl-xL gene transfer and that the TTC reaction to some extent 633–641. 8 Circulation July 5, 2005

12. Leist M, Single B, Castoldi AF, Kühnle S, Nicotera P. Intracellular 17. Ohno M, Takemura G, Ohno H, Misao J, Hayakawa Y, Minatoguchi S, adenosine triphosphate (ATP) concentration: a switch in the decision Fujiwara T, Fujiwara H. “Apoptotic” myocytes in the infarct area in between apoptosis and necrosis. J Exp Med. 1997;185:1481–1486. rabbit hearts may be oncotic myocytes with DNA fragmentation: analysis 13. Jennings RB, Reimer KA. The cell biology of acute myocardial ischemia. by immunogold electron microscopy combined with in situ nick end Annu Rev Med. 1991;42:225–246. labeling. Circulation. 1998;98:1422–1430. 14. Jugdutt BI, Idikio HA. Apoptosis and oncosis in acute coronary syn- 18. Kanoh M, Takemura G, Misao J, Hayakawa Y, Aoyama T, Nishigaki K, dromes: assessment and implications. Mol Cell Biochem. 2005;270: Noda T, Fujiwara T, Fukuda K, Minatoguchi S, Fujiwara H. Significance 177–200. of myocytes with positive DNA end-labeling (TUNEL) in hearts with 15. Elsässer A, Suzuki K, Schaper J. Unresolved issues regarding the role of dilated cardiomyopathy. Circulation. 1999;99:2757–2764. apoptosis in the pathogenesis of ischemic injury and heart failure. J Mol Cell Cardiol. 2000;32:711–724. 16. Rodriguez M, Schaper J. Apoptosis: measurement and technical issues. J Key Words: Editorials Ⅲ apoptosis Ⅲ necrosis Ⅲ ischemia Ⅲ Mol Cell Cardiol. 2005;38:15–20. proto-oncogene proteins c-bcl-2 Editorial

N-Terminal-Pro-B–Type Natriuretic Peptide Universal Marker of Cardiovascular Risk?

A. Mark Richards, MD, PhD, DSc, FRACP; Christopher M. Frampton, BSc, PhD

lasma brain-type natriuretic peptide (BNP) and amino- frequent peripheral arterial disease, more frequent valvular terminal proBNP (NTproBNP) provide prognostic in- disease, and higher cholesterol, the 2 groups did differ Pformation on cardiovascular morbidity and mortality significantly for 2 of the biomarkers. NTproBNP and beyond that provided by standard risk factors. Clinical C-reactive protein in the infarct patients were, respectively, applications of B-type peptides under ongoing research in- 1.6- and 1.3-fold higher than in controls. Renin was not clude their use in diagnosing acute heart failure (HF), in risk significantly elevated in the MI group. Cases were then stratification in both acute and established HF, in acute combined with controls, creating a selective population coronary syndromes (ACS), in asymptomatic populations at within which the risk of incurring MI was compared between cardiovascular risk (older adults and people with hyperten- marker quartiles generated from the 618 patients. This inter- sion), and as part of a screening strategy for detection of left quartile comparison was conducted with 2 models: the first ventricular impairment and prediction of cardiovascular risk unadjusted except for the previously mentioned matched in the general population.1,2 In this issue of Circulation, variables, and the second adjusted for baseline variables that Campbell and colleagues3 assess the ability of NTproBNP to showed significant univariate associations with MI, ie, sys- predict myocardial infarction (MI) in subjects who have tolic blood pressure, cholesterol status, and presence or experienced a cerebrovascular event. NTproBNP (reflecting absence of a history of coronary heart disease, valvular heart cardiac distension) is compared with C-reactive protein (a disease, and peripheral arterial disease. In both models, systemic marker of inflammation) and renin (a marker of NTproBNP in the upper quartile was associated with a 2- to sodium status regulated by renal perfusion and delivery of 3-fold increase in risk of MI as compared with the first sodium to the renal glomerulus). quartile. C-reactive protein exhibited a similar 2-fold increase See p 110 in risk in the first model, but in the second, CRP had no significant association with increased risk of MI. Renin, not The nested case-control study is from the 6105 participants seen to be significantly different in a simple univariate in the Perindopril Protection Against Recurrent Stroke Study comparison between cases and controls, became significant (PROGRESS), a placebo-controlled study of converting en- when considered as a putative independent predictor in the zyme inhibitor–based therapy in patients with previous cere- adjusted models. Upper-quartile renin was associated with a 4 brovascular events. Within PROGRESS, 206 subjects in- 1.7- to 1.9-fold increase in risk of MI as compared with the curred an MI during 3.9 years of follow-up. The investigators bottom quartile. Combining NTproBNP with renin added matched those incurring an MI with control PROGRESS predictive power, and individuals with both NTproBNP and patients avoiding MI from time of randomization to time of renin in the highest quartiles had 4.5 times the risk of MI as case ascertainment. Cases and controls were matched for age, compared with both in the lowest quartile. gender, treatment allocation, region, and cerebrovascular These findings add to reports indicating the predictive qualifying event. The form of matching meant that individual power of B-type peptides for all-cause mortality and cardio- patients may have been controls initially and subsequently vascular mortality and/or morbidity.1 Clinical application became cases on incurring an MI during further follow-up. may be difficult, however. The clinician does not seek the Matching in this fashion may confuse the interpretation of the relative risk of an MI in a given patient who has experienced nonconditioned analysis of baseline variables. a cerebrovascular event as compared with another individual Comparing the 206 cases with 412 controls at randomiza- with cerebrovascular disease of the same gender and similar tion, the investigators report that in addition to higher systolic age. Rather, clinicians require the best possible assessment of blood pressure, more frequent known coronary disease, more the absolute risk for an individual patient, taking all risk factors into account as far as possible. The clinician will note The opinions expressed in this article are not necessarily those of the gender, age, ethnic background, blood pressure, lipid status, a editors or of the American Heart Association. range of possible ECG abnormalities, presence or absence of From the Department of Medicine, Christchurch School of Medicine diabetes, renal function, and the presence or absence of a and Health Sciences, Christchurch, New Zealand. Correspondence to Prof A. Mark Richards, Dept of Medicine, history of adverse arterial events (coronary, cerebrovascular, Christchurch School of Medicine and Health Sciences, PO Box 4345, renovascular, or other peripheral vascular disease) and may Riccarton Ave, Christchurch, New Zealand. E-mail mark.richards@ also be privileged to have cardiac imaging and/or stress test cdhb.govt.nz (Circulation. 2005;112:9-11.) data in hand. In this regard, it would have been an informative © 2005 American Heart Association, Inc. corroborative exercise for the authors to have undertaken a Circulation is available at http://www.circulationaha.org multivariate time-to-event analysis of all 5918 subjects from DOI: 10.1161/CIRCULATIONAHA.105.551853 PROGRESS who had blood samples available and to have 9 10 Circulation July 5, 2005 provided readers with absolute event rates and hazard ratios ing a comprehensive multivariate model that included ele- for the risk of MI (and other morbid events) adjusted not only ments that were absent from the Campbell et al analysis. in the manner conducted in the present nested case-control Troponin T, creatinine clearance, and ST-segment depression study but, as far as possible, for all of the above-mentioned on the ECG all were independently related to risk of future established indicators of coronary risk. MI after adjustment for age, smoking status, angina, previous Curiously, in the analysis by Campbell and colleagues, MI, heart failure, diabetes, ECG changes, heart rate, and cases and controls have a similar prevalence of smoking, creatinine clearance, but NTproBNP was not. An exception to diabetes, and levels of plasma creatinine—powerful estab- the general reported absence of independent prediction of MI lished predictors of MI within the general population and by BNP/NTproBNP in coronary populations is the report by within populations with overt arterial disease. The later Richards et al.10 This study is one of a very few that include multivariate analyses conducted by Campbell et al are not ventricular imaging. The authors found that NTproBNP therefore subject to any adjustment for these powerful pre- independently predicted subsequent MI only in subjects with dictors, which may individually and collectively confound reduced left ventricular function after their index MI. This their conclusions. result may reflect the greater vulnerability of the dilated, It is notable that in suggesting NTproBNP independently adversely remodeled ventricle (with areas of increased wall predicts MI, this article is at odds with several existing stress, a reduced coronary perfusion pressure gradient, and a reports. Within the Framingham offspring study, BNP levels, more adverse neurohumoral milieu) to undergo detectable though associated with increased risk of death and first myocyte necrosis for any given acute reduction in coronary cardiovascular event, did not predict coronary heart disease perfusion. In this regard, it would be of interest to know the events.5 Campbell et al suggest this reflects the difference prevalence of previous unrecognized left ventricular dysfunc- between the Framingham offspring participants and tion in the MI cases in the PROGRESS trial. The almost PROGRESS patients, who were older and had established universal finding that BNP/NTproBNP predict all-cause mor- vascular disease as evidenced by a preceding cerebrovascular tality may reflect the independent associations of the pep- event. If this explanation held true, BNP and NTproBNP tides, not only with cardiac function but also with age and should have independent predictive power of MI in coronary renal function.1,6–10 Hence, the plasma BNP/NTproBNP level disease; however, this is not the case. De Lemos et al integrates several powerful independent predictors of conducted a substudy in 2525 participants in the Orbofiban in mortality. Patients with Unstable coronary Syndromes, Thrombolysis In The additive effect of combining NTproBNP with renin is Myocardial Infarction 16 (OPUS-TIMI 16) trial of an oral of interest and reminds us of early reports of the relationship glycoprotein IIb/IIIa inhibitor in ACS.6 BNP had a powerful between renin and coronary risk.11,12 Again, however, the univariate association with the risk of MI at 30 days and 10 manner in which this multimarker approach may be applied months; however, on multivariate analysis, though predictive in clinical practice is unclear. The renin measurements of all-cause mortality independent of age, troponin, HF, renal employed in the analysis by Campbell et al3 were undertaken insufficiency, ST-segment changes or left bundle-branch in the absence of an angiotensin converting enzyme inhibitor block on the ECG, Killip class, and diabetes, BNP was not an or angiotensin II receptor blocker. Therefore, clinicians independent predictor of MI. Similarly, Jernberg et al, in 755 would have to apply these measurements (in a combined patients admitted with chest pain, found BNP predicted marker strategy) in the absence of what is now a commonly all-cause mortality independent of age, gender, diabetes, prescribed drug class in the population experiencing cerebro- hypertension, past MI, treatment, HF, ECG changes, tropo- vascular events. The analyses conducted by Campbell et al nin, or renal function, but it was not reported to independently are somewhat inconsistent in retaining renin, which did not predict MI.7 Sabatine and colleagues inspected BNP and differ between cases and controls by univariate comparison, C-reactive protein in Ͼ2000 ACS patients from OPUS-TIMI for examination in the 2 multivariate analyses comparing 16 and Treat angina with Aggrastat and determine Costs of marker quartiles within the combined groups. Meanwhile, Therapy with Invasive or Conservative Strategies—Throm- they discarded a previous diagnosis of hypertension, smoking bolysis In Myocardial Infarction 18 (TACTICS-TIMI 18) status, use of certain drugs, diabetes, and renal function, all of trials.8 They found both to be predictors of the composite end which likewise did not differ between cases and controls but point of mortality, MI, and/or HF independent of age, all of which are well-established determinants of cardiovas- diabetes, previous MI, HF, and ECG changes but did not cular risk. Furthermore, unlike NTproBNP assays, renin report BNP as an independent predictor of MI as a lone end assays remain widely variable in their methodology and point. James et al measured NTproBNP, troponin T, and reference ranges throughout the world, and clinicians would C-reactive protein in 6809 patients with non–ST-segment have to pay careful attention to the methods, reference ranges, elevation ACS participating in the Global Utilization of and quality control within their local laboratories to enable Streptokinase and tPA for Occluded arteries IV (GUSTO-IV) correct interpretation of their locally provided renin results. trial of the glycoprotein IIb/IIIa inhibitor abciximab.9 Increas- Renin is influenced by medications, being elevated by diuret- ing quartiles of NTproBNP were powerfully related to short- ics and converting enzyme inhibitors and suppressed by and long-term mortality. Levels of troponin T and C-reactive ␤-blockers. It is also responsive to sodium status, physical protein and creatinine clearance were also independently activity, and posture via concurrent changes in renal perfu- related to 1-year mortality. These authors carefully assessed sion pressure, intraglomerular sodium flux, and sympathetic the possible association of NTproBNP with later MI, employ- traffic to the juxtaglomerular apparatus. Renin levels differ Richards and Frampton NTproBNP and Cardiovascular Risk 11 substantially between ethnic groups. This vulnerability to the and left ventricular hypertrophy: a LIFE study. J Hypertens. 2004;22: influence of ethnicity, commonly prescribed drugs, common- 1597–1604. 3. Campbell DJ, Woodward M, Chalmers JP, Colman SA, Jenkins AJ, place changes in sodium and volume status, posture, and Kemp BE, Neal BC, Patel A, MacMahon SW. Prediction of myocardial activity suggests that the application of renin as a risk marker infarction by N-terminal-pro-B-type natriuretic peptide, C-reactive in the individual patient will remain fraught with difficulties. protein, and renin in subjects with cerebrovascular disease. Circulation. 2005;112:110–116. The unimpressive performance of C-reactive protein in the 4. PROGRESS Collaborative Group. Randomised trial of a adjusted models included in the Campbell et al analysis is of perindopril-based blood-pressure–lowering regimen among 6,105 indi- interest and, as the authors point out, is concordant with viduals with previous stroke or transient ischaemic attack. Lancet. 2001; recent publications suggesting that C-reactive protein may be 358:1033–1041. 5. Wang TJ, Larson MG, Levy D, Benjamin EJ, Leip EP, Omland T, Wolf a more modest predictor of coronary heart disease than PA, Vasan RS. Plasma natriuretic peptide levels and the risk of cardio- previously reported.13,14 They rightly point out that the vascular events and death. N Engl J Med. 2004;350:655–663. presence of drugs, including aspirin, may well have attenu- 6. De Lemos JA, Morrow DA, Bentley JH, Omland T, Sabatine MS, McCabe CH, Hall C, Cannon CP, Braunwald E. The prognostic value of ated the predictive power of C-reactive protein within this B-type natriuretic peptide in patients with acute coronary syndromes. population. The increasingly widespread prescription of st- N Engl J Med. 2001;345:1014–1021. atins may also reduce C-reactive protein levels and may 7. Jernberg T, Stridsberg M, Venge P, Lindahl B. N-terminal pro brain natriuretic peptide on admission for early risk stratification of patients modify its predictive power for later events in arteriopathic with chest pain and no ST-segment elevation. J Am Coll Cardiol. 2002; patients. Nevertheless, the findings of Campbell et al remain 40:437–445. at variance to those from studies in larger groups with 8. Sabatine MS, Morrow DA, de Lemos JA, Gibson CM, Murphy SA, Rifai coronary disease in which such drugs are widely used.8,9 N, McCabe C, Antman EM, Cannon CP, Braunwald E. Multimarker approach to risk stratification in non–ST elevation acute coronary syn- NTproBNP stands out as the strongest of the 3 markers in dromes: simultaneous assessment of troponin I, C-reactive protein, and the current comparison; however, its independent association B-type natriuretic peptide. Circulation. 2002;105:1760–1763. with subsequent MI is at odds with the bulk of existing 9. James SK, Lindahl B, Siegbahn A, Stridsberg M, Venge P, Armstrong P, Barnathan ES, Califf R, Topol EJ, Simoons ML, Wallentin L. N-terminal reports, including some derived from large coronary disease pro-brain natriuretic peptide and other risk markers for the separate cohorts subjected to analysis incorporating comprehensive prediction of mortality and subsequent myocardial infarction in patients multivariate models. The report by Campbell et al analyzes a with unstable coronary artery disease. A Global Utilization of Strategies To Open occluded arteries (GUSTO)-IV Substudy. Circulation. 2003; selective group of patients and requires confirmation in 108:275–281. further cerebrovascular cohorts with analyses that directly 10. Richards AM, Nicholls MG, Espiner EA, Lainchbury JG, Troughton RW, lend themselves to potential application for risk stratification Elliott J, Frampton C, Turner J, Crozier IG, Yandle TG. B-type natriuretic in “real-life” clinical settings. peptides and ejection fraction for prognosis after myocardial infarction. Circulation. 2003;107:2786–2792. 11. Alderman MH, Madhavan S, Ooi WL, Cohen H, Sealey JE, Laragh JH. Disclosure Association of the renin-sodium profile with the risk of myocardial Dr Richards has received a research grant from, has served on the infarction in patients with hypertension. N Engl J Med. 1991;324: speakers’ bureau of and/or received honoraria from, and has con- 1098–1104. sulted for and/or served on the advisory board of Roche Diagnostics. 12. Meade TW, Cooper JA, Peart WS. Plasma renin activity and ischemic heart disease. N Engl J Med. 1993;329:616–619. Dr Richards has also consulted for and/or served on the advisory 13. Danesh J, Wheeler JG, Hirschfield GM, Eda S, Eiriksdottir G, Rumley A, board of Synex. Lowe GD, Pepys MB, Gudnason V. C-reactive protein and other circu- lating markers of inflammation in the prediction of coronary heart References disease. N Engl J Med. 2004;350:1387–1397. 1. Rademaker MT, Richards AM. Cardiac natriuretic peptides for cardiac 14. Tall AR. C-reactive protein reassessed. N Engl J Med. 2004;350: health. Clin Sci. 2005;108:23–36. 1450–1452. 2. Olsen MH, Wachtell K, Tuxen C, Fossum E, Bang LE, Hall C, Ibsen H, Rokkedal J, Devereux RB, Hildebrandt P. N-terminal pro-brain natriuret- KEY WORDS: Editorials Ⅲ cardiovascular diseases Ⅲ coronary disease Ⅲ ic peptide predicts cardiovascular events in patients with hypertension natriuretic peptides Ⅲ renin Acknowledgment of Reviewers

The Editors express appreciation to the following referees who served from April 1, 2004, to December 31, 2004.

Einari Aavik Peter Alexandersen David Antoniucci Robert J. Bache Nicola Abate Khaled Alfakih Jovan P. Antovic Jean E. Bachet Amr E. Abbas Ottavio R. Alfieri Charles Antzelevitch Walter Backes Antonio Abbate Francois Alhenc-Gelas Piero Anversa Peter Backx Kevin C. Abbott Ziad A. Ali L.J. Appel Larry Baddour Robert D. Abbott Etienne M. Aliot Ian Appleton David Badesch Koji Abe Hussein R. Al-Khalidi Andrew E. Arai Juan J. Badimon E. Dale Abel Lindsey D. Allan Maria Rosario G. Araneta Lina Badimon George S. Abela Yves Allemann Stephen L. Archer Cornel Badorff Benjamin S. Abella Maurits A. Allessie Moshe Arditi Stephen F. Badylak Aiden Abidov Kristina Allikmets Ross Arena Alexei Y. Bagrov Alexandre Abizaid Matthew A. Allison Thomas Arentz Ajay Bahl M. Roselle Abraham Thomas G. Allison Laurent Argaud Ferdinand H. Bahlmann Pierre Abraham Kevin C. Allman Gary C. Armitage Colin Baigent Theodore P. Abraham Laura Almasy Paul W. Armstrong Steven Bailey Dan Abramov Jesus Almendral Donna K. Arnett Donald S. Baim Charles S. Abrams Carlos Alonso-Villaverde Leonard F. Arnolda Alison E. Baird Jerome L. Abramson Joseph S. Alpert Wilbert S. Aronow S. Paul Bajaj Hugues Abriel Martin A. Alpert Marie Arsenault Arvind Bakhru Stephan Achenbach David A. Alter Margaret A. Arstall Patricia F. Bakker Michael Acker Peter Alter Michael Artman Stephan J.L. Bakker Michael J. Ackerman Guy Alvarez Yujiro Asada George L. Bakris Stamatis Adamopoulos John A. Ambrose Takayuki Asahara Prabhakaran Balagopal Robert J. Adams Pierre Ambrosi Masanori Asakrua Stephan Baldus Volker Adams Peter Ammann Raimondo Ascione Christie M. Ballantyne Ian Adatia Ezra Amsterdam Alexzander Asea Jean-Luc Balligand Philip A. Ades Ping An Arlene S. Ash Scott W. Ballinger Jennifer Adgey Inder S. Anand Euan A. Ashley Enzo Ballotta Marina Afanasyeva G.M. Anantharamaiah Muhammad Ashraf James A. Balschi Vahid Afshar-Kharghan Burt Anderson Nick Ashton Ko Bando Stefan Agewall Gitte Andersen Gregory K. Asimakis José R. Banegas Tetsuro Ago H. Vernon Anderson Samuel J. Asirvatham Mary A. Banerji Piergiuseppe Agostoni Jeffrey L. Anderson Gerd Assmann Adrian P. Banning Pietro Agricola Kelley P. Anderson Birgit Assmus David A. Baran David Aguilar Mark E. Anderson Brad Astor Eddy Barasch Seyedhossein Aharinejad Neil Anderson Bela F. Asztalos Giuseppe Barbaro Amrita Ahluwalia Page A.W. Anderson Dan Atar John C. Barbato Ismayil Ahmet Peter G. Anderson Vasilios G. Athyros Paule C. Barbeau Enrico Aidala Robert H. Anderson Dianne L. Atkins Jean T. Barbey Ken-ichi Aihara Todd J. Anderson Larry D. Atwood Robert Bard Masanori Aikawa Stefan Andreas Andrew M. Atz Edit Bardi Barbara Ainsworth Arne K. Andreassen Johann W. Auer Philip M. Barger William C. Aird Felicita Andreotti Pal Aukrust Stephen G. E. Barker Masazumi Akahoshi Douglas Andres Angelo Auricchio S. Serge Barold Olakunle O. Akinboboye Vicente Andres Gerard P. Aurigemma Alain D. Baron Masahiro Akishita Ramaroson Andriantsitohaina Melissa A. Austin Lili A. Barouch Junya Ako Rajesh K. Aneja Richard C. Austin Jose A. Barrabes Christine M. Albert Giovanni Anfossi Michael V. Autieri Elizabeth Barrett-Connor Jeffrey Albert Annalisa Angelini Pablo Avanzas William H. Barry Michelle Albert Gianni D. Angelini Abraham Aviv Robyn J. Barst Gabriel Aldea Dominick J. Angiolillo Philip E. Aylward Thomas Bartel Michael H. Alderman E. Angles-Cano Michel Azizi Philip Barter Alexey N. Aleshin Stefan D. Anker Angelo Azzi Jürgen Barth Marie-Christine Alessi Brian H. Annex Jan Baan John R. Bartholomew John H. Alexander Jack Ansell Vladimir R. Babaev Matthias Barton M. Yvonne Alexander Riitta L. Antikainen Gerard Babatasi Benico Barzilai Mark E. Alexander Tarek F. Antonios Fritz H. Bach Theodore A. Bass e12 Acknowledgment of Reviewers e13

Jean-Pierre Bassand Lars Berglund Peter F. Bodary Sorin J. Brener Craig T. Basson Martin W. Bergmann Christoph Bode Kate M. Brett Shari S. Bassuk Steven R. Bergmann William E. Boden Christoph Brezinka Eric R. Bates Bradford C. Berk Gerd Bodlaj J. Timothy Bricker Anjan Batra Lisa Berkman Manfred Boehm Charles R. Bridges Robert Bauernschmitt Javier Bermejo Michael Boehm Regina Brigelius-Flohe Johann Bauersachs Jose M. Bernal Jolanda M. Boer Michele Brignole Kenneth L. Baughman Michael C. Berndt Guido Boerrigter Ralph G. Brindis Iris Baumgartner Gerald J. Berry Eric Boersma Charles Brink Ralf W. Baumgartner Mark F. Berry Eric Boerwinkle Eliot A. Brinton William Baumgartner Donald M. Bers Rainer H. Boger Michael R. Bristow Christophe Bauters Giuseppe S. Berton Richard G. Bogle Steven L. Britton Jeroen J. Bax Alain G. Bertoni Frank Bogun Bruce R. Brodie B. Timothy Baxter Michel E. Bertrand David F. Bohr Sergey V. Brodsky Gary F. Baxter Charles I. Berul William Boisvert Ulrich Broeckel Kirk W. Beach Patricia J.M. Best Joerg Bojunga Alessandra Brofferio M. Flint Beal Reinaldo B. Bestetti Paula M. Bokesch Robert D. Brook Alvarez Beatriz Christer Betsholtz Thomas Boland Maria M. Brooks Jean-Louis Beaudeux Saroja Bharati Roberto Bolli James M. Brophy Christoph R. Becker Aruni Bhatnagar Victoria Bolotina M. Julia Brosnan Diane M. Becker Deepak Bhatnagar Marvin O. Boluyt K. Bridget Brosnihan Lance B. Becker Deepak L. Bhatt Massimo Bonacchi Brigitta C. Brott Lewis C. Becker Italo Biaggioni Raoul Bonan Margaret E. Brousseau Richard C. Becker Cesario Bianchi Nikolaos Bonaros Jan-Leendert P. Brouwer Frank Beckers Marco E. Bianchi Lawrence I. Bonchek Gregory L. Brower Joshua A. Beckman Giorgio M. Biasi Meredith Bond B. Greg Brown Bettina Beech Hans K. Biesalski Diana Bonderman David L. Brown Juerg H. Beer D.H. Biesma Hendrik Bonnemeier Kathleen K. Brown Philipp Beerbaum Erik A.L. Biessen Enzo Bonora Nancy J. Brown Richard H. Behrman J. Thomas Bigger Robert O. Bonow Warren S. Browner Berthold Bein Nick R. Bijsterveld Maria R. Bonsignore Michael Brownlee Sean C. Beinart Diane E. Bild Piet W. Boonstra Ian N. Bruce Alexa Beiser Jacques Billette Jeffrey Boord Christian Bruch Romualdo Belardinelli George Billman George W. Booz Josep Brugada Robert M. Bell Feng B. Bin Nicolas Borenstein Ramon Brugada Jonathan N. Bella Philip F. Binkley Jeffrey S. Borer Lyndia Brumback George A. Beller John D. Birkmeyer Martin Borggrefe Corinna B. Brunckhorst Michelle P. Bendeck Yochai Birnbaum John Boscardin Frank Martin Brunkhorst David G. Benditt Eva Biro René M. Botnar Eric J. Brunner Martin Bendszus Nanette H. Bishopric Michiel L. Bots Hans R. Brunner Athanase Benetos John A. Bittl Chantal M. Boulanger Helle Bruunsgaard R. Benezra Vera Bittner Anne Bouloumie Robert M. Bryan Frank M. Bengel Edward O. Bixler Henri M. Bounameaux Michael Bryer-Ash Jean-Pierre Bénitah Daniel J. Blackman Pierre Boutouyrie Rosemary S. Bubien Ivor J. Benjamin Eugene H. Blackstone Daniel F. Bowen-Pope Paolo Bucciarelli Ralf Benndorf Steven N. Blair Neil E. Bowles Timothy G. Buchman Joel S. Bennett Jean-Jacques Blanc Penelope A. Boyden John Buckwalter Martin R. Bennett William S. Blaner Mark R. Boyett Matthew J. Budoff William M. Bennett Stefan Blankenberg Biykem Bozkurt Werner Budts Rondelet Benoit W. Matthijs Blankesteijn David J. Bradley Arno Buecker Neal Benowitz Andrew D. Blann T. Douglas Bradley Raffaele Bugiardini D. Woodrow Benson Alex Blatt Randy W. Braith L. Maximilian Buja Lee N. Benson Andrew D. Blaufox Frederick Brancati Jens Bulow Merrill D. Benson Erwin Blessing Ralf P. Brandes Burkhard D. Bültmann Aloys Berg Peter C. Block Ruediger C. Braun-Dullaeus Cecil M. Burchfiel Robert A. Berg David A. Bluemke Eugene Braunwald Lora Burke Knut E. Berge Friedrich C. Blumberg Alan C. Braverman John C. Burnett Alan K. Berger Roger S. Blumenthal Molly S. Bray Jane C. Burns Peter B. Berger Alex Bobik Claudia Bregonzio Paul B. Burton Ronald Berger Edimar A. Bocchi Roger E. Breitbart Ivo Buschmann Rudolf Berger Jorge B. Boczkowski Ole A. Breithardt David W. Busija e14 Acknowledgment of Reviewers

Rudi Busse Paola Casanello Jina Choo Domenico Corrado Javed Butler Juan P. Casas Tz-Chong Chou Javier Corral Alfred E. Buxton Wayne E. Cascio Robin P. Choudhury Dalila Corry Brian F. Buxton Ivan P. Casserly Benjamin J.W. Chow Alberto Corsini Peter H. Byers Lisa A. Cassis Judith C. Chow Marshal A. Corson Benjamin F. Byrd, III Agustin Castellanos Timothy F. Christian Roberto Corti Melissa J. Byrne Edoardo Castelli Mina Chung Joseph S. Coselli Christopher H. Cabell Alessandro Cataliotti Taylor Chung Francesco Cosentino Candido Cabo Marco Cattaneo Mariantonietta Cicoira Francisco G. Cosio Howard Cabral David S. Celermajer Marilyn J. Cipolla Marco A. Costa Kevin S. Cahill Bojan Cercek Francesco Cipollone Lisa C. Costello-Boerrigter Hua L. Cai Antonio Ceriello Kieran Clarke William G. Cotts Michael E. Cain Manuel Cerqueira Robert Clarke Paul J.M. Coucke Paolo Calabro Matteo Cesari William R. Clarke Thierry Couffinhal Antonio M. Calafiore Claudia U. Chae Catherine M. Clase David Couper S.C. Calaghan Mohamed Chahine John G.F. Cleland Adrian Covic Angelino Calderone Alan Chait Paula R. Clemen Dermot Cox Mary Caldwell Bernard R. Chaitman Ton J. Cleophas Jill P. Crandall David A. Calhoun Aravinda Chakravarti Aldo Clerico Michael H. Crawford Robert M. Califf Lorraine Chalifour Angela Clerk Sybil Crawford Hugh Calkins Angel Chamorro Alexander Clowes Filippo Crea David J. Callans Hunter C. Champion William T. Clusin Mark A. Creager Francois A. Cambien Bysani Chandrasekar Ronald I. Clyman Harry J.G.M. Crijns Richard P. Cambria Anthony C. Chang Andrew J. Coats Janet B. Croft Paolo G. Camici Chih-Jen Chang Leonard A. Cobb Kevin D. Croft Vito M. Campese Ruey-Kang R. Chang William A. Coetzee Rachelle H. Crosbie Umberto Campia Kevin S. Channer Thomas M. Coffman Carroll E. Cross Antonio C. Campos de Keith M. Channon David J. Cohen John R. Crouse Carvalho M. John Chapman Eric A. Cohen Richard S. Crow Paul Canner Nora M. Chapman Hillel W. Cohen J. Kennedy Cruickshank Christopher P. Cannon Israel F. Charo Howard A. Cohen Bruce F. Culleton Richard O. Cannon John C. Chatham Marc Cohen Juraj Culman Charles E. Canter Lee-Young Chau Michael V. Cohen R. William Currie John G. Canto Sarwat Chaudhry Richard A. Cohen Jesse W. Currier Warren J. Cantor Gautam Chaudhuri Jay N. Cohn Anne B. Curtis John M. Canty Aurea J. Chaves Lawrence H. Cohn Mary Cushman Noel M. Caplice Melvin D. Cheitlin William E. Cohn Daniele M. Cusi Maurizo C. Capogrossi Alex F. Chen Steven D. Colan Jeffrey A. Cutler Riccardo Cappato Horng H. Chen Jamie Coleman Maria-Cristina Cuturi Thomas P. Cappola Ian Y. Chen Désiré Collen Myron I. Cybulsky Alessandro M. Capponi Jiu-an Chen Barry S. Coller Tillmann Cyrus Joseph A. Caprini Peng-Sheng Chen Robert W. Colman Sébastien Czernichow Sonia Caprio Shih-Ann Chen Antonio Colombo Mat J.A.P. Daemen Alessandro Capucci Edouard Cheneau Maria Giovanna Colombo Michael Daffertshofer Massimo Caputo Debbie Cheng David Colquhoun Ralph B. D’Agostino Blase A. Carabello James H. Chesebro Catherine Communal Mohamed R. Daha Brian J. Carey Mordechai Chevion Gianluigi Condorelli Hiroyuki Daida Stephane G. Carlier Derek P. Chew William E. Connor James E. Dalen Mark D. Carlson Elena Chiarpotto Robert V. Considine Ronald L. Dalman Edward Carmeliet Masaaki Chiku C. Richard Conti Edward R. Damiano Pizzi Carmine John S. Child Elena Conti Patricia D’Amore Mercedes R. Carnethon William M. Chilian David J. Cook Nicolas Danchin Robert M. Carney Kazuo Chin George A. Cook Paresh Dandona Clara Carpeggiani Michael T. Chin Nancy R. Cook Haim D. Danenberg Oscar A. Carretero Marcello Chinali John P. Cooke George Dangas John D. Carroll Randolph W. Chitwood, Jr Joshua M. Cooper Peter G. Danias Joseph Carrozza Ray C-J Chiu Leslie T. Cooper Werner G. Daniel Jeffrey L. Carson Leslie Cho Mark E. Cooper Jean-Marie Daniel Lamaziere Andrew J. Carter Aram V. Chobanian Richard S. Cooper Stephen R. Daniels Lori Carter-Edwards Anand Chockalingam Josef Coresh A.H. Jan Danser Wayne Carver Eric T. Choi James Coromilas Victor Darley-Usmar Acknowledgment of Reviewers e15

Anthony M. Dart Ranjan Deka Torsten Doenst Mark J. Eisenberg Dipak K. Das Jacqueline M. Dekker Pieter A. Doevendans Mickey S. Eisenberg Undurti N. Das Federica del Monte Hisham Dokainish Graeme Eisenhofer Jean C. Daubert Etienne Delacrétaz Anna F. Dominiczak Daniel T. Eitzman Harold L. Dauerman Joris Delanghe William T. Donahoo Mikael Elam Alan Daugherty Claude Delcayre J. Kevin Donahue John A. Elefteriades Anthony P. Davenport Christophe Delclaux Rosario Donato Michael T. Eliasziw Tirone E. David Thomas E. Delea Peter Doris Uri Elkayam Sandra T. Davidge Jose A. Delgado Alves Gerald W. Dorn, II Kenneth A. Ellenbogen Charles J. Davidson E. Patchen Dellinger David E. Dostal Myrvin Ellestad Karina W. Davidson R. Phillip Dellinger Pamela S. Douglas Oyvind Ellingsen Barry R. Davis Louis J. Dell’Italia James M. Downey Patrick Ellinor Patricia H. Davis Kevin C. Dellsperger Stephen W. Downing Justine A. Ellis Roger A. Davis Mario Delmar Ramona Doyle Stephen G. Ellis Roger B. Davis Michael D. Delp Kathleen Dracup Kristin E. Ellison Russell C. Davis Anthony N. DeMaria Luciano F. Drager Nabil El-Sherif Robin L. Davisson Yi Deng Mark H. Drazner Mary Emond Vedat Davutoglu Donald R. Dengel Wim Drenthen Noriaki Emoto Kevin P. Davy Christophe Depre Helmut Drexler Masao Endoh Dana Dawson Dimitri Deserranno Daniel L. Dries Matthias Endres Jonathan R.S. Day Alessandro Desideri Ahmed Ben Driss Richard M. Engelman Sharlene M. Day Christopher A. DeSouza Jie Du David M. Engman Jacques de Bakker George Despotis Terry W. Du Clos Gunnar Engstrom Fred De Beer Jean-Pierre Despres Dayue Duan Andrew E. Epstein Bernard De Bruyne Zeruesenay Desta Raghvendra K. Dubey Frederick H. Epstein Raffaele De Caterina Anita DeStefano Anne M. Dubin Laurence M. Epstein Ulf de Faires Robert C. Detrano Anique Ducharme Stephen E. Epstein Pim J. de Feyter Katherine M. Detre Samuel C. Dudley Raimund R. Erbel Giovanni de Gaetano Tobias Deuse Stephen J. Duffy John M. Erikson Peter de Jaegere Mohan Devbhandari Dirk J. Duncker Einar E. Eriksson Paul E. de Jong Richard B. Devereux Daniel Duprez Wolfgang Erl Marlies de Lange Mieke Dewerchin Jocelyn Dupuis Thomas P. Erlinger Peter W. de Leeuw Marc Dewey Josee Dupuis Thomas Eschenhagen James A. de Lemos Mandeep Dhadly David T. Durack Charles T. Esmon Michel de Lorgeril Naranjan S. Dhalla Carlos M.G. Duran Ricardo J. Esper Giuseppe De Luca Stefan Dhein William Durante Christine Espinola-Klein Moniek P.M. de Maat Gerald F. Di Bona J. Larry Durstine Katherine Esposito Ebo D. de Muinck Marcelo F. Di Carli Firat Duru N.A. Mark Estes Ramon de Nooijer Carlo Di Mario Susan Duval Zeev Estrov Anne De Paepe Mario Di Napoli Kim A. Eagle Masato Eto Giovanni De Pergola Pietro Di Pasquale Elaine D. Eaker Tanenao Eto Dominique de Prost Michaela Diamant Robert T. Eberhardt Paulo Roberto B. Evora Giovanni de Simone George A. Diamond Franz R. Eberli Justin A. Ezekowitz Johan H. De Sutter David A. Dichek Lynn E. Eberly Michael D. Ezekowitz Robbert J. de Winter Wolfgang Dichtl Steven N. Ebert John W. Fabre Dick de Zeeuw Kenneth Dickstein Shah Ebrahim Rosalind Fabunmi Barbara J. Deal Sean P. Didion Dwain L. Eckberg Gianpaolo Fadini John E. Deanfield Andre M. Diedrich Robert H. Eckel Erling Falk Arjun Deb Jutta Dierkes Jay M. Edelberg Rodney H. Falk Robert F. DeBusk Javier Diez Elazer R. Edelman Bonita Falkner G. William Dec Wolfgang H. Dillmann Robert R. Edelman James C. Fang William M. DeCampli Joseph S. Dillon Thomas S. Edgington James I. Fann Jeanne M. DeCara Vasken Dilsizian L. Henry Edmunds Frank Faraci Robert S. Decker John P. DiMarco Igor R. Efimov Andrew Farb Gordon R. DeFoe Stefanie Dimmeler Kensuke Egashira Harrison W. Farber Curt G. DeGroff Anh Tuan Dinh-Xuan Satoru Eguchi Jawed Fareed Gregory J. Dehmer Donald J. DiPette Marek P. Ehrlich William B. Farquhar Jan T. Deichen Michael Diringer Benjamin W. Eidem Vladimir G. Fast Elisabeth Deindl Abhinav Diwan John F. Eidt Khalil Fattouch Carolin Deiner Sanjay Dixit John W. Eikelboom Elda Favari Elisabetta Dejana Douglas W. Dockery Howard J. Eisen David P. Faxon e16 Acknowledgment of Reviewers

William P. Fay John S. Floras Mingui Fu A. Martin Gerdes Zahi A. Fayad James Floyd Shmuel Fuchs Marie D. Gerhard-Herman Jocelyne Fayn Martin Fluck Bianca Fuhrman Lucie Germain Franz Fazekas Joseph Flynn Masashi Fujita Guido Germano Sergio Fazio Robert Fogel Naomi Fukagawa Bernard J. Gersh William F. Fearon Alan M. Fogelman Tohru Fukai M. Eric Gershwin Paul W.M. Fedak Antonio F. Folino Keiichi Fukuda Myron C. Gerson Eugenia Fedoseyeva Franco Folli Kesiuke Fukuo Welton M. Gersony Jeffrey A. Feinstein Gregg C. Fonarow Pino Fundarò Edward P. Gerstenfeld Mary F. Feitosa Guo-Hua Fong John W. Funder Robert E. Gerszten Robert Felder Ignatius W. Fong Colin D. Funk S. David Gertz G. Michael Felker Vivian A. Fonseca Curt D. Furberg Leonard S. Gettes Michael P. Feneley JoAnne M. Foody Raffaello Furlan Tal Geva Qingping Feng Thomas Force Mark I. Furman Henry Gewirtz Peter Ferdinandy Earl S. Ford Masato Furuhashi Michael Gewitz James J. Ferguson, III George D. Ford Valentin Fuster Jalal K. Ghali T. Bruce Ferguson Myriam Fornage William H. Gaasch William A. Ghali Francisco Fernandez-Aviles James S. Forrester Christian Gachet Mihai Gheorghiade Jose M. Fernandez-Real Ulrich Forstermann David R. Gagnon Lorenzo Ghiadoni Olivier Feron Trudy M. Forte James V. Gainer Hossein A. Ghofrani Markus Ferrari Elyse Foster Fiorenzo Gaita Carlo Giansante Victor A. Ferrari F. Gerald Fowkes Patrick J. Gallagher Gary H. Gibbons Victor A. Ferraris Caroline S. Fox Claudio Galli Raymond J. Gibbons Paolo Ferrazzi Ervin Fox Augusto Gallino C. Michael Gibson Robert E. Ferrell Alain Fraisse James M. Galloway Samuel S. Gidding Andreas Festa Mark W. Frampton Jonas B. Galper Stephan Gielen Stephan Fichtlscherer Silvia Franceschi Apoor S. Gami Martine Gilard Anette Fiebeler Charles W. Francis Santhi Ganesh Ian C. Gilchrist Loren J. Field Gary S. Francis Peter Ganz Thomas D. Giles David S. Fieno Veronica Franco Mario J. Garcia Wayne R. Giles Michael A. Fifer Dignat-George Françoise Julius M. Gardin Linda D. Gillam Hans R. Figulla Nikolaos G. Frangogiannis Helena M. Gardiner Jonathan Gillard Janos G. Filep Markus H. Frank Sheila M. Gardiner A. Marc Gillinov Antonio P. Filipe, Jr. Stanley S. Franklin David G. Gardner Anne M. Gillis Jeffrey R. Fineman Wayne J. Franklin Martin J. Gardner Matthew W. Gillman Mitchell S. Finkel Michael R. Franz Roy S. Gardner Richard F. Gillum Toren Finkel Robert P. Frantz Alan Garfinkel Robert F. Gilmour Dianne M. Finkelstein Stefan Frantz Philippe Garot Larry C. Gilstrap Marcus Fischer Maria Grazia Franzosi Peter Garred Frank J. Giordano Thorsten Fischer Nancy Frasure-Smith Jean-Michel T. Gaspoz Domenico Girelli Richard D. Fish Robert W.M. Frater Michael A. Gatzoulis Cynthia J. Girman Michael C. Fishbein David S. Freedman Kimberlee Gauvreau Anselm K. Gitt Edward A. Fisher Bruce A. Freeman Timothy P. Gavin Dario Giugliano John D. Fisher Balz Frei Haralambos P. Gavras Gregory R. Giugliano John T. Fisher John K. French Irene Gavras Alexandre Giusti-Paiva Patrick W. Fisher Michael P. Frenneaux Meinrad Gawaz Michael M. Givertz Steven A. Fisher Ulrich H. Frey Steffen Gay David Gjertson Glenn I. Fishman Matthias Frick J. William Gaynor Mark T. Gladwin Desmond J. Fitzgerald Linda F. Fried J. Michael Gaziano Stanton A. Glantz Garret A. FitzGerald Susan K. Fried Raul J. Gazmuri Stephen P. Glasser Peter J. Fitzgerald R.P. Friedland Carmine Gazzaruso Stephen J. Glatt Frank A. Flachskampf Deborah M. Friedman Robert L. Geggel Nicola Glorioso Greg C. Flaker Paul A. Friedman Bruce D. Gelb Donald D. Glower Scott D. Flamm Jefferson Frisbee Jacques Genest Charles J. Glueck Marcus D. Flather Victor Froelicher Yong-Jian Geng Robert J. Glynn Jerome L. Fleg Jiri J. Frohlich Alfred L. George Alan S. Go Kirsten E. Fleischmann Alberto Froio Sarah J. George Ulrich Göbel Ingrid Fleming Peter C. Frommelt Demetrios Georgiou Andrew Gogbashian Richard M. Fleming Andrea Frustaci Alexander Geppert Noyan Gokce Markus Flesch Robert L. Frye Lior Gepstein Diane R. Gold Gerald F. Fletcher Ryan M. Fryer Hanspeter Gerber Jeffrey P. Gold Acknowledgment of Reviewers e17

Michael R. Gold Eugene A. Grossi Anthony J. Hanley Keith Henry Ira J. Goldberg Ehud Grossman James A. Hanley Timothy D. Henry Ronald B. Goldberg Blair P. Grubb Edward L. Hannan Heike M. Hermanns Jeffrey J. Goldberger Eberhard Grube William H. Hansen Ramon C. Hermida Sherita H. Golden Scott M. Grundy Goran K. Hansson Adrian F. Hernandez Ira D. Goldfine Gary L. Grunkemeier Akiyoshi Hara Miika Hernelahti Joshua I. Goldhaber Eliseo Guallar Shuntaro Hara Victoria L.M. Herrera Samuel Z. Goldhaber Maurizio D. Guazzi Joshua M. Hare David M. Herrington Lee Goldman Vilmundur G. Gudnason Robert A. Harrington Howard C. Herrmann Pascal J. Goldschmidt Peter G. Guerra William S. Harris Ray E. Hershberger Steven R. Goldsmith Gerard M. Guiraudon David G. Harrison Charles A. Herzog Larry B. Goldstein Martha Gulati Alison L. Harte David C. Hess Michael S. Goligorsky Giosue Gulli Paula J. Harvey Otto M. Hess Paolo Golino Hakan Gullu Rodrigo Hasbun Gerd F. Heusch Jonathan Golledge Julian Gunn David Hasdai Karsten Heusser Celso E. Gomez-Sanchez Mahesh P. Gupta Naoyuki Hasebe Steven B. Heymsfield Philimon Gona Narendra K. Gupta Gerd Hasenfuss William R. Hiatt Mario D. Gonzalez Paul A. Gurbel J. Michael Hasenkam Yukihito Higashi Elizabeth Goodman Enrique P. Gurfinkel Naotake Hashimoto Charles B. Higgins Lawrence T. Goodnough Geoffrey C. Gurtner Paul M. Hassoun Denise Hilfiker-Kleiner Theodore A. Gooley Swaminatha V. Gurudevan Thomas S. Hatsukami Joseph A. Hill John Gorcsan Arjun Gururaj Daniel C. Hatton Gerhard Hindricks Neil Gordon Matthias Gutberlet Richard N.W. Hauer Thomas H. Hintze Joel M. Gore David D. Gutterman Paul J. Hauptman Shuji Hinuma Tommaso Gori Przemyslaw Guzik Elizabeth R. Hauser Masayasu Hiraoka Mark W. Gorman Tomasz J. Guzik Richard J. Havel Loren F. Hiratzka Robert C. Gorman Stefano Guzzetti Axel Haverich Karen K. Hirschi Shinya Goto Katrina Gwin-Hardy Edward P. Havranek Valeria Hirschler Roberta A. Gottlieb Donald C. Haas Robert A. Haworth John W. Hirshfeld Stephen S. Gottlieb Felix Haas Ilan Hay Keiichi Hishikawa Antonio M. Gotto Helmut Habazettl Nissim Hay Mark A. Hlatky K. Lance Gould Robert H. Habib Junichiro Hayano Donald Hnatowich Luis Henrique W. Gowdak Rory Hachamovitch David L. Hayes Carolyn Ho Kristof Graf Walter E. Haefeli Sharonne N. Hayes Helen H. Hobbs Patricia M. Grambsch Judith Haendeler Daniel Hayoz Didier Hober Juan F. Granada Hans U. Haering Stanley L. Hazen Robert W. Hobson Christopher B. Granger Steven M. Haffner Jiang He Judith S. Hochman D. Neil Granger David E. Haines Ka He Hanoch Hod Augustus O. Grant William D. Haire Tongrong He Julien I. Hoffman Henk L. Granzier Michel Haissaguerre Anthony M. Heagerty Udo Hoffmann Guido Grassi Roger J. Hajjar Harvey S. Hecht Peter Höglund David Gray Charles A. Hales Susan R. Heckbert Thomas Hohlfeld Paul A. Grayburn Kathleen J. Haley Peter S. Heeger Stefan H. Hohnloser J. Thomas Grayston Michael E. Halkos Timothy Heeren Brian D. Hoit Daniel J. Green Jennifer L. Hall Christopher Heeschen John E. Hokanson Darren C. Greenwood Par Hallberg Linda J. Heffner Fernando Holguin Edward W. Gregg Hermann Haller Robert A. Hegele Judd E. Hollander Michelle Ann Grenier Jozsef Haller Paul A. Heidenreich Morley Hollenberg Kathy K. Griendling Kevin J. Hallock Jörg Heierhorst Thomas A. Holly Brian P. Griffin Perry V. Halushka Albert Heim William L. Holman Helen R. Griffiths Rainer Hambrecht Robert J. Heine David R. Holmes Clarence E. Grim Mohamed H. Hamdan Gerardo Heiss Paul Holvoet Cindy L. Grines Pavel Hamet Alan W. Heldman Michael Holzer Steven K. Grinspoon John A. Hamilton Gary V. Heller Shunichi Homma Jean Ann Grisso Christian W. Hamm Joan Heller Brown Myeong-Ki Hong Francine Grodstein H. Kirk Hammond Barbara L. Hempstead Yuling Hong William J. Groh Zhong C. Han Robert Henderson Rocio S. Honigmann Marie-Louise M. Gronholdt Wayne W. Hancock Marc Hendrikx Jane L. Hoover-Plow Robert Gropler Diane E. Handy Peter Henke Richard Hopkins Garrett J. Gross Claude E. Hanet Charles H. Hennekens William E. Hopkins Oliver Gross Graeme J. Hankey Michael Hennerici Uta C. Hoppe e18 Acknowledgment of Reviewers

Masatsugu Hori Masahiro Ito Shuichi Jono Philipp A. Kaufmann Lisa K. Hornberger Wulf D. Ito Jens Jordan Sanjay Kaul Benjamin D. Horne Toshiyuki Itoi Jacob Joseph Sanjiv Kaul Burkhard Hornig Susan L. Ivey Mark E. Josephson Koji Kawahito John D. Horowitz D. Dunbar Ivy Kaumudi J. Joshipura Chuichi Kawai Lawrence D. Horwitz Yuichi Iwaki Janna Journeycake David M. Kaye Steven R. Houser Tohru Izumi Pekka Jousilahti Teruhisa Kazui Barbara V. Howard Christopher L. Jackson Aleksandar Jovanovic Elsadig Kazzam George Howard EdwinK. Jackson Ian R. Jowsey Mark T. Keating T. Howard Howell Graham Jackson Michael J. Joyner Craig A. Keebler Henry H. Hsia Shaun P. Jackson Bodh I. Jugdutt Beate E. Kehrel Frank B. Hu Alice K. Jacobs J. Wouter Jukema Aaron S. Kelly Peifeng Hu David R. Jacobs David N. Juurlink Daniel P. Kelly Paul L. Huang Donald W. Jacobsen Stefan Kaab Ralph A. Kelly Sally A. Huber Paul Jacques Jens J. Kaden Malte Kelm Whady A. Hueb Tazeen H. Jafar Alan H. Kadish Anita M. Kelsey Joerg Huelsken Allan S. Jaffe Yutaka Kagaya Byron W. Kemper Chris C. Hughes Thomas Jahnke Henry S. Kahn Richard D. Kenagy David Y. Hui Mukesh K. Jain Richard Kahn Thomas A. Kent Heikki V. Huikuri Rajan Jain Fumihiko Kajiya Richard E. Kerber P.P. Hujoel Pierre Jaïs Gabor Kaley Dean J. Kereiakes Russell D. Hull Jose Jalife Christoph Kalka Karl B. Kern Per M. Humpert Jorge E. Jalil Klaus Kallenbach Morton J. Kern Karin H. Humphries Ik-Kyung Jang Jonathan M. Kalman William S. Kerwin Stephen E. Humphries Joseph S. Janicki Lalit Kalra Steven J. Keteyian Thomas Hund Warren R. Janowitz Grzegorz L. Kaluza Paul Khairy Joseph Hung Michiel J. Janse Balaraman Kalyanaraman Bijoy Khandheria Kelly J. Hunt Ian Janssen Vaijinath S. Kamanna Ashwani Khanna Steven C. Hunt Stefan P. Janssens Timothy J. Kamp Stefan Kiechl Patrick R. Hunziker Craig T. January Junji Kanda Jan T. Kielstein Winston L. Hutchinson James L. Januzzi Takeshi Kanda Shinji Kihara Adolph M. Hutter Rudolf Jarai David E. Kandzari Tatsuya Kiji Guido Iaccarino Mikko J. Jarvisalo Laura B. Kane Dae Jung Kim Mark Iafrati Patrick Y. Jay Nicholas Kang Hyo-Soo Kim Sahoko Ichihara Goy Jean-Jacques William B. Kannel InKyeom Kim Raymond E. Ideker David J. Jenkins Norman M. Kaplan Jason K. Kim Richard G. Ijzerman Rolf Jenni Tomas Kara Raymond J. Kim Uichi Ikeda Allen Jeremias Richard H. Karas Thomas R. Kimball Katsunori Ikewaki Michael Jerosch-Herold Johan Karlberg Stephen E. Kimmel John S. Ikonomidis Paula Jerrard-Dunne Joel S. Karliner Carey D. Kimmelstiel Sabino Iliceto Xavier Jeunemaitre Morris Karmazyn Akinori Kimura Armin Imhof Ashish K. Jha Aly Karsan George L. King Akihiro Inazu Ishwarlal Jialal Karl R. Karsch Spencer B. King Sandro Inchiostro Canwen Jiang Soji Kasayama John G. Kingma Ciro Indolfi Huang Jianhua Carlos S. Kase Scott Kinlay Julie R. Ingelfinger Bernd Jilma Juan-Carlos Kaski Ulrich Kintscher David A. Ingram, Jr. Hanjoong Jo Edward K. Kasper Kevin E. Kip Joanne S. Ingwall Mark A. Jobling David A. Kass Charles J.H.J. Kirchhof Nobutaka Inoue Edward J. Johns Robert S. Kass James Kirklin Teruo Inoue B. Delia Johnson Ghassan Kassab Lorrie A. Kirshenbaum Cecilia Invitti Bruce D. Johnson John J.P. Kastelein Chiharu Kishimoto D.P. Inwald Jason L. Johnson Adnan Kastrati Brett M. Kissela Dan-Dominic G. Ionescu Richard J. Johnson Naoto Katakami Toru Kita Kaikobad J. Irani Robert L. Johnson Sekar Kathiresan Akira Kitabatake Mitsuhiro Isaka Daniel W. Jones Masahiko Kato Masafumi Kitakaze Shun Ishibashi Gregory T. Jones Tomohiro Katsuya Kazuo Kitamura Junnichi Ishii Peter L. Jones Hugo A. Katus Richard N. Kitsis Kikuo Isoda Robert H. Jones Zvonimir S. Katusic Andre G. Kleber Eric M. Isselbacher Steven P. Jones Arnold M. Katz Robert Kleemann Takaaki Isshiki W. Keith Jones Stuart D. Katz Neal S. Kleiman Hiroshi Ito Habo J. Jongsma Marc P. Kaufman Allan L. Klein Acknowledgment of Reviewers e19

George J. Klein Irving L. Kron Torben B. Larsen Jerrold H. Levy Lloyd W. Klein Marvin W. Kronenberg Warren K. Laskey Wilbur Y. Lew Paul D. Kligfield Itzhak Kronzon Robert D. Lasley Martin M. LeWinter James R. Klinger Florian Krotz Larry A. Latson Alan B. Lewis Elizabeth S. Klings Paul Kubes Jo-Dee L. Lattimore Klaus F. Ley Francis J. Klocke Nils Kucher Joseph Lau Andrew C. Li Robert A. Kloner Karl-Heinz Kuck Wei C. Lau Jennifer S. Li Bradley P. Knight Nino Kuenzli Michael S. Lauer Jian Li John L. Knight Hartmut Kuhn Ulrich Laufs Jianhua Li Anne A. Knowlton Michaela Kuhn Jari A. Laukkanen Na Li Kirk U. Knowlton Helena Kuivaniemi Stephane Laurent Shengxu Li Sarah S. Knox Marrick L. Kukin Kenneth R. Laurita Yan C. Li Merril L. Knudtson Rakesh C. Kukreja Debbie A. Lawlor Bruce T. Liang Juhani Knuuti Lewis H. Kuller Daniel A. Lawrence Chang-seng Liang Dennis T. Ko Iftikhar J. Kullo Lesley Lawrenson James K. Liao Yoshio Kobayashi Premkumari Kumarathasan Louise Lawson Ronglih Liao Colleen G. Koch Richard E. Kuntz Jennifer S. Lawton Youlian Liao Walter J. Koch Calvin J. Kuo Harold L. Lazar Peter Libby Itsuo Kodama Lih Kuo Ronald M. Lazar Joseph R. Libonati Wolfgang Koenig Christian Kupatt Dominique Le Guludec David S. Liebeskind Theo Kofidis Dhandapani Kuppuswamy Alexander Leaf Philip R. Liebson Kwang K. Koh Masahiko Kurabayashi Sam D. Leary Choong-Chin Liew Frank Kolodgie Tobias Kurth Robert J. Lederman Stephen B. Liggett Issei Komuro Theodore Kurtz Amanda J. Lee Kathleen C. Light Takahisa Kondo Kengo F. Kusano Hon-Chi Lee Yean L. Lim Marvin A. Konstam Jeffrey T. Kuvin Richard T. Lee Chee Chew Lim Stavros V. Konstantinides William A. Kuziel Thomas H. Lee Joao A. Lima Igor E. Konstantinov Tatiana Kuznetsova C.P.M. Leeson Marian C. Limacher Anatol Kontush Kevin F. Kwaku David J. Lefer Ming T. Lin Marianne Eline Kooi Raymond Y. Kwong Jean-Francois Legare Shien-Fong Lin Willem J. Kop Michael Kyller Jacopo M. Legramante Michael Lincoff Stephen L. Kopecky Maria T. La Rovere Michael H. Lehmann Bertil Lindahl Bruce A. Koplan David E. Laaksonen Stephan E. Lehnart JoAnn Lindenfeld Ran Kornowski Arthur Labovitz Leslie Leinwand Marshall D. Lindheimer Mikhail Kosiborod Louis M. Labrousse Norbert Leitinger Jonathan R. Lindner Andreas Koster Roger J. Laham Thierry H. LeJemtel Volkhard Lindner Rudolph W. Koster Shenghan Lai Paul LeLorier Jerry B. Lingrel Sawa Kostin John G. Lainchbury Giuseppe Lembo MacRae F. Linton Theodore A. Kotchen Edward G. Lakatta Pedro A. Lemos Gregory Y.H. Lip Hans Kottkamp Hanna-Maaria Lakka Steven R. Lentz William C. Little Nicholas T. Kouchoukos Timo A. Lakka David A. Leon Sheldon E. Litwin Petri T. Kovanen Jules Y.T. Lam Antonio Maria Leone Jun Liu Peter R. Kowey Stephen C.T. Lam Jonathan Leor Kiang Liu Jun Koyama Benoit Lamarche Amir Lerman Peter P. Liu Andrew D. Krahn John J. Lamberti Bruce B. Lerman Simin Liu Aldi Kraja Rachel Lampert Edward J. Lesnefsky Yongge Liu Jonathan Krakoff Katja H. Lampinen Philippe F. Lesnik Eng H. Lo Christopher M. Kramer Kathryn G. Lamping Heather S. Lett Amanda Lochner Evangelia G. Kranias Hui Y. Lan Donald Y. Leung James E. Lock William E. Kraus Gary Landreth Michael C. H. Leung Warren E. Lockette Ronald M. Krauss Donald W. Landry Marcel M. Levi Ian M. Loftus Andrew J. Krentz Michael J. Landzberg Roberto Levi Anne-Marie Lompre Nancy R. Kressin David A. Lane Adeera Levin Barry London Reinhold Kreutz Roberto M. Lang Benjamin D. Levine Gérard M. London Jorg Kreuzer Jonathan J. Langberg Glenn N. Levine Carlin S. Long P.A. Krieg Bas Langeveld Robert A. Levine Eva M. Lonn Murali C. Krishna B. Lowell Langille Sidney Levitsky Gary D. Lopaschuk Rajesh Krishnamurthy David Langleben Bernard I. Levy John J. Lopez Eswar Krishnan Alexandra J. Lansky Daniel Levy Patricio López-Jaramillo Leonard Kritharides Harris Lari Robert J. Levy Francisco Lopez-Jimenez Michael H. Kroll John C. LaRosa Andrew P. Levy Christine H. Lorenz e20 Acknowledgment of Reviewers

David J. Loskutoff Michael J. Mann William M. McClellan Michael Miller Douglas W. Losordo Stewart Mann Seth McClennen Nancy H. Miller Eric B. Loucks Peter B. Manning Michael V. McConnell Todd D. Miller Charles J. Lowenstein Warren J. Manning James McCord Virginia M. Miller Gerald Luc Pier M. Mannucci Brian W. McCrindle Tohru Minamino Lee Lucas Teri A. Manolio Peter A. McCullough Erich Minar Benedict R. Lucchesi Moussa Mansour James D. McCully Gary S. Mintz Pamela A. Lucchesi Michael S. Marber Mary M. McDermott Israel Mirsky Andreas Luchner Keith L. March Theresa A. McDonagh Yoshio Misawa John Ludbrook Simona Marchesi Daniel McGee Seema Mital David Ludwig Francis E. Marchlinski John C. McGiff Brett M. Mitchell Russell V. Luepker Frank I. Marcus Henry C. McGill Gary F. Mitchell Friedrich C. Luft Maurizio Margaglione Michael McGoon Jere H. Mitchell Esther Lutgens Ali J. Marian Thomas M. McIntyre Richard N. Mitchell Aernout L. Luttun Allyn L. Mark William J. McKenna R. Scott Mitchell Robert L. Lux Daniel B. Mark Timothy A. McKinsey Arnold Mitnitski Bruce W. Lytle Andrew R. Marks Tracey L. McLaughlin Suneet Mittal Christoph Maack Barry J. Maron Vallerie V. McLaughlin Murray A. Mittleman David M. Maahs Luc Maroteaux Julie R. McMullen Kunio Miyatake Renke Maas Michel Marre John J.V. McMurray Kohei Miyazono Peter S. Macdonald Mario B. Marrero Elizabeth M. McNally Emile R. Mohler, III Christopher K. Macgowan Oscar C. Marroquin Coleen A. McNamara Nicanor I. Moldovan Guy A. MacGowan Philip A. Marsden Patrick H. McNulty Peter Molenaar Francois Mach Audrey C. Marshall Tim C. McQuinn Ernesto Molina Stella M. Macin Steven P. Marso Charles F. McTiernan David J. Moliterno Christopher Mack Fabio Martelli Gary E. McVeigh Jeffery D. Molkentin Michael J. Mack Douglas Martin Roger Mee Tom E. Mollnes Wendy J. Mack Jack L. Martin Mandeep R. Mehra Kevin M. Monahan Isla S. Mackenzie Paul T. Martin Roxana Mehran Laurent Monassier Rachel H. Mackey Wim Martinet Jawahar L. Mehta Gilles Montalescot William R. MacLellan Yukio Maruyama Rajendra H. Mehta Joan Montaner Michal Maczewski Thomas H. Marwick Shamir R. Mehta Nicola Montano Paolo Madeddu Gerald R. Marx James B. Meigs Alan R. Moody Mohammad Madjid Nikolaus Marx Cynthia J. Meininger James C. Moon Joren C. Madsen Steven O. Marx Gerhard W. Meissner David F. Moore Koji Maemura Attilio Maseri Jan R. Mellembakken Phillip Moore Aldo P. Maggioni Peter J. Mason Philippe Menasche Martin Morad Kenneth W. Mahaffey Robert J. Mason Michael E. Mendelsohn Fred Morady Michael C. Mahaney Frederick A. Masoudi Carlos F. Mendes de Leon Christine S. Moravec Lynn Mahony Joseph M. Massaro Armando J. Mendez Henning Morawietz Heimo Mairbaurl Barry M. Massie Maurizio Menichelli Kerrie L. Moreau Bernhard Maisch Bashir M. Matata George A. Mensah Pierre Moreau Alan S. Maisel Ellisiv B. Mathiesen James O. Menzoian Pedro R. Moreno William H. Maisel Hiroaki Matsubara Jean-Jacques Mercadier Raul Moreno Mark W. Majesky Hikaru Matsuda Anwar T. Merchant Thomas M. Morgan Amy S. Major Reiko Matsui Yahye Merhi Peter M. Morganelli Robert T. Mallet Kanji Matsukawa Ilse L. Mertens Anthony P. Morise Ziad Mallat Akira Matsumori Franz H. Messerli Ryuichi Morishita Alberto Malliani Hidehiro Matsuoka Ruben Mestril Toshisuke Morita Giuseppe Mancia Hiroaki Matsuoka Luisa Mestroni Nicholas W. Morrell G.B. John Mancini Rumiko Matsuoka Heiko Methe Joel Morrisett Pitchaiah Mandava Masunori Matsuzaki Philippe Meurin John A. Morrison Olivia Manfrini Christian M. Matter Martijn Meuwissen Sean J. Morrison Dennis T. Mangano Ray V. Matthews Theo E. Meyer David A. Morrow Arduino A. Mangoni Kimmo J. Mattila Evangelos D. Michelakis Jason D. Morrow Venkatesh Mani Clive N. May Holly R. Middlekauff Richard F. Mortensen Calin V. Maniu Bongani M. Mayosi Michele Mietus-Snyder Lori J. Mosca Douglas L. Mann Melanie Maytin Richard V. Milani Ralph S. Mosca Giovanni E. Mann Todor N. Mazgalev D. Craig Miller Mauro Moscucci Johannes F. Mann Nathalie M. Mazure D. Douglas Miller Jeffrey W. Moses Kenneth G. Mann Eileen McCall Leslie W. Miller Arthur J. Moss Acknowledgment of Reviewers e21

Richard L. Moss Andrea Natale Christopher J. O’Donnell Richard L. Page Evangeline D. Motley Rama Natarajan Erwin N. Oechslin Ramdas G. Pai Karen S. Moulton Viswanathan Natarajan Patrick T. O’Gara Rosario Palacios Jean-Jacques Mourad Hendrik Nathoe Hisao Ogawa Wulf Palinski Issam D. Moussa Stanley Nattel Yoshihiro Ogawa Julio C. Palmaz Gilbert H. Mudge Matthew T. Naughton Toshio Ogihara Lyle J. Palmer Christian Mueller Mohamad Navab Jae K. Oh Sebastian Palmeri Thomas Muenzel Frank Naya Ann M. O’Hare Hui-Lin Pan Alessandro Mugelli Krassen Nedeltchev Takayoshi Ohkubo Demosthenes Panagiotakos Andreas Mugge Ilka Nemere Tomoko Ohkusa Natesa G. Pandian Joseph B. Muhlestein Dario Neri Erik M. Ohman James S. Pankow Debabrata Mukherjee Shawna D. Nesbitt Veronica Ojetti Julio A. Panza Rupak Mukherjee Aleksandar N. Neskovic Akinlolu O. Ojo Nicholas F. Paoni Barbara J.M. Mulder Paul J. Nestel Peter M. Okin Carlo Pappone James E. Muller Stefan Neubauer Katashi Okoshi Gilles Paradis Jochen Muller-Ehmsen Ellis J. Neufeld Jeffrey E. Olgin Patrick Parfrey Janet M. Mullington Gishel New Jobien K. Olijhoek Michael Parides Michael J. Mulvany Andrew C. Newby Jeffrey W. Olin Stephen Paridon Neal I. Muni David E. Newby Brian Olshansky Paolo Parini Jorg Muntwyler L. Kristin Newby Timothy M. Olson Jeong-Euy Park Joanne Murabito Anne B. Newman Patrick G. O’Malley Seung-Jung Park David Murdoch John H. Newman Reed A. Omary Donna Parker Toyoaki Murohara Gary E. Newton Jeffrey H. Omens John D. Parker Elizabeth Murphy Christopher H. Newton-Cheh Torbjorn Omland Ira A. Parness Philip M. Murphy Ludwig Neyses Steve R. Ommen Juan C. Parodi Timothy P. Murphy Graham Nichol Altan Onat Alessandro Parolari Charles E. Murry Stephen J. Nicholls Marie S. O’Neill Steve W. Parry Timothy I. Musch Wilmer W. Nichols William W. O’Neill Vincenzo Pasceri Rene J. Musters Andrew C. Nicholson Takayuki Ono Ares D. Pasipoularides Steven E. Mutsaers Georg Nickenig Koji Onoda Gerard Pasterkamp Bulent Mutus Martin J. Nicklin Henry Ooi Ayan Patel Robert J. Myerburg Pascal H. Nicod Suzanne Oparil Lisa Patel Daniel D. Myers Christoph A. Nienaber Tobias Opthof Rakesh P. Patel Jonathan Myers Michael R. Nihill Hakan Oral Vickas V. Patel Elizabeth G. Nabel Seppo T. Nikkari John F. Oram David J. Paterson Christoph K. Naber Dimitar Nikolov E. John Orav Paola Patrignani Bernardo Nadal-Ginard Richard M. Niles Trevor J. Orchard Carlo Patrono Zurab G. Nadareishvili Rick A. Nishimura Jose M. Ordovas Richard D. Patten Koonlawee Nademanee Mari K. Nishizaka Donald Orlic Cam Patterson Abraham Nader Steven E. Nissen John A. Ormiston Peter M. Pattynama Vinay Nadkarni Tianhua Niu Joseph P. Ornato Walter J. Paulus Ryozo Nagai Koichi Node Brian O’Rourke Jeffrey M. Pearl Hideaki Nagase Constance T. Noguchi Michael F. O’Rourke Justin D. Pearlman Noritoshi Nagaya Eisei Noiri Robert A. O’Rourke Mary A. Peberdy Eike Nagel Georg Noll Tetsuya Oshima Ole D. Pedersen Sherif F. Nagueh Borge G. Nordestgaard Clive Osmond Susanne S. Pedersen Hiroshi Nakagawa Mikael Norman Jan Ostergren Patrick Peeters Hajime Nakamura Kari E. North David Ott Antonio Pelliccia Takeshi Nakano Gavin R. Norton Fillipo Ottani Patricia A. Pellikka Kanji Nakatsu Michel Noutsias Catherine M. Otto Theo Pelzer Gilles Nalbone Gian M. Novaro Feifan Ouyang Michael Pencina Brahmajee K. Nallamothu Ulrike Nowak-Gottl Michel Ovize Marc S. Penn Byung-Ho Nam Evgeny Nudler Mehmet C. Oz Dudley J. Pennell Navin C. Nanda William C. Nugent Susan E. Ozanne William Penny Manasi Nandi Carole Ober Pal Pacher Carl J. Pepine Raffaele Napoli Martin Oberhoff Chris J. Packard Mark B. Pepys Girish Narayan Edward R. O’Brien Douglas L. Packer Mark A. Pereira Sanjiv M. Narayan Christopher J. Occleshaw Francis D. Pagani Francisco Perez-Vizcaino Craig R. Narins Ira S. Ockene Massimo Pagani Emerson C. Perin Krzysztof Narkiewicz Christopher M. O’Connor Patrick J. Pagano Harris Perlman Jagat Narula Gerald T. O’Connor Pierre Page Joseph K. Perloff e22 Acknowledgment of Reviewers

Eduardo R. Perna Wendy S. Post K. Srinath Reddy Campbell Rogers Thomas V. Perneger Robert S. Poston Vivek Y. Reddy Ariel Roguin Francesco Perticone Tina S. Poulsen Margaret M. Redfield John Rogus Arkadii M. Pertsov Neil Poulter Andrew N. Redington Mary J. Roman Inga Peter Janet T. Powell Judith G. Regensteiner Mats Rönnback Annette Peters Andrew J. Powell Enrique Regidor Dieter Ropers Nicholas S. Peters William S. Powell Jalees Rehman Emilio Ros Eric D. Peterson Scott K. Powers Johan Reiber Wayne D. Rosamond Kirk L. Peterson Henry J. Pownall Nathaniel Reichek Jonathan Rosand J.Thomas Peterson Abhiram Prasad Muredach Reilly Noel R. Rose Eva Petkova Francesco Prati Sharon C. Reimold Michael R. Rosen Patricia A. Peyser Domenico Pratico Steven E. Reis David S. Rosenbaum Marc A. Pfeffer Josef Prchal Peter J. Reiser Frits R. Rosendaal Ivan Philip Stephen M. Prescott Michael J. Reiter Clive Rosendorff George J. Philippides Russell L. Prewitt P.H. Reitsma Michael E. Rosenfeld Bradley G. Phillips Beth F. Printz Jian-Fang Ren Bruce R. Rosengard Christopher O. Phillips Frits W. Prinzen Jun Ren Todd K. Rosengart Richard P. Phipps Silvia G. Priori Helaine E. Resnick David N. Rosenthal Francesco Piarulli Kirkwood A. Pritchard Ariel J. Reyes Anthony Rosenzweig Philippe Pibarot Linda L. Pritchard Dwight Reynolds Bernard Rosner Eugenio Picano Eric N. Prystowsky Matthew R. Reynolds Allan M. Ross Michael H. Picard Bruce M. Psaty Shereif H. Rezkalla John Ross J. Geoffrey Pickering William Pu Jonathan Rhodes Robert Ross Galen M. Pieper Vladimir Pucovsky Flavio Ribichini Andrea Rossi Luc A. Pierard John D. Puskas Paul M. Ribisl Gian Paolo D. Rossi Grant N. Pierce Pirkko J. Pussinen Romeo Ricci Ranieri Rossi Burkert Pieske Reed Pyeritz Peter A. Rice Thomas Rostock Bill A. Pietra Kalevi Pyorala Lawrence Rice Michael Roth Gabriele Piffaretti Stuart F. Quan Michael W. Rich Richard B. Rothman Frank A. Pigula Thomas Quaschning Vincent Richard Steven A. Rothman Nico H.J. Pijls Thomas Quertermous A. Mark Richards Peter M. Rothwell Louise Pilote Miguel A. Quiñones Paul M. Ridker Philippe Rouet Ileana Pina Ton J. Rabelink Barbara Riegel Jean-Lucien Rouleau Theodore Pincus Marlene Rabinovitch Walter F. Riesen Anna V. Roux David J. Pinsky Miriam T. Rademaker Nader Rifai Anne H. Rowley Duane S. Pinto Daniel J. Rader Giorgio Rigatelli Prabir Roy-Chaudhury Yigal M. Pinto Marek W. Radomski Eric B. Rimm Alan Rozanski Tobias Pischon Shahin Rafii Gilles Rioufol Hong Ruan Federico Piscione Paolo Raggi Rebecca H. Ritchie Melvyn Rubenfire Cristina Pislaru Shahbudin H. Rahimtoola James M. Ritter Frederick L. Ruberg Bertram Pitt Elaine W. Raines Eberhard Ritz Lewis J. Rubin Maria V. Pitzalis Olli T. Raitakari Alain Rivard Terrence D. Ruddy Manel Pladevall Satish R. Raj Jeffrey Robbins Lawrence L. Rudel Jonathan F. Plehn Sanjay Rajagopalan Robert C. Robbins Neil B. Ruderman Johannes Pleiner Nalini M. Rajamannan Robert Roberts Yoram Rudy Jorge Plutzky Sumathi Ramachandran David Robertson Marc Ruel A. Graham Pockley Kenneth S. Ramos Sander J. Robins Wolfram Ruf Stuart J. Pocock J. Scott Rankin Richard B. Robinson Zaverio M. Ruggeri Bruno K. Podesser Dabeeru C. Rao Simon C. Robson Jean-Bernard Ruidavets Philip J. Podrid L.Vijay Rao Albert P. Rocchini Luis M. Ruilope Paul Poirier Elliot Rapaport Luc Rochette John A. Rumberger Roberto Pola Tienush Rassaf Howard E. Rockett John S. Rumsfeld Don Poldermans Saif S. Rathore Howard A. Rockman Marschall S. Runge Victoria Polyakova Peter B. Raven Dan M. Roden Heinz Rupp Philip A. Poole-Wilson Ursula Ravens David Rodman Frank Ruschitzka Clive A. Pope Katya Ravid Beatriz L. Rodriguez James W.E. Rush Jeffrey J. Popma Chester A. Ray Leonardo Rodriguez Jeremy N. Ruskin Richard L. Popp Reza S. Razavi Fernando Rodriguez-Artalejo Kerry S. Russell J. David Port Fabio A. Recchia Alicia Rodriguez-Pla Mary E. Russell Francesco Portaluppi Rita F. Redberg Marco Roffi Raymond R. Russell Thomas R. Porter Alluru S. Reddi Veronique L. Roger Wolfgang Rutsch Acknowledgment of Reviewers e23

Martin K. Rutter Saul Schaefer Miki L. Schwartzman Allan A. Shor Peter N. Ruygrok Juergen R. Schaefer Robert A. Schweikert Angela C. Shore Thomas J. Ryan Klaus P. Schafers Robert H.G. Schwinger Linda Shore-Lesserson Jack Rychik Hartzell V. Schaff Juerg Schwitter Ashfaq Shuaib Lars Ryden Martin J. Schalij Alan Scott Robert J. Siegel Tobias Saam Jutta Schaper Russell S. Scott Hans-Hinrich Sievers Manel Sabate Wolfgang Schaper Miran Sebestjen Ulrich Sigwart Marc S. Sabatine Christoph Scharf Udo Sechtem Michael J. Silka Roger A. Sabbadini Gina Schatteman Artyom Sedrakyan Marc A. Silver Hani N. Sabbah Robert G. Schaub Ellen W. Seely Donald S. Silverberg Joseph F. Sabik Andre J. Scheen Harry Segall David I. Silverman Luigi Sacca James M. Scheiman Pravin B. Sehgal Gregg J. Silverman Ralph L. Sacco Dierk Scheinert Christine E. Seidman Norman H. Silverman Michael N. Sack Melvin M. Scheinman Jonathan G. Seidman Jean-Sebastien Silvestre Jonathan D. Sackner-Bernstein Heinrich R. Schelbert Christian Seiler Robert D. Simari Frank M. Sacks Benjamin J. Scherlag Frank W. Sellke R. John Simes H. Mehrdad Sadeghi Ralph T. Schermuly Joseph B. Selvanayagam Paolo Simioni Junichi Sadoshima Urs Scherrer Luyi Sen Daniel I. Simon Michel E. Safar Marielle Scherrer-Crosbie Laureen Sena Scott I. Simon Jeffrey E. Saffitz Deborah A. Scheuer Shoichi Senda Orlando P. Simonetti Kiran B. Sagar James Scheuer Roxy Senior Leon A. Simons David J. Sahn Ernesto L. Schiffrin Thomas D. Sequist Michael Simons Yoshifumi Saijo Nelson B. Schiller Susan M. Sereika Maarten L. Simoons Yoshihiko Saito Mark D. Schluchter Patrick W. Serruys Paul C. Simpson Tomohiro Sakamoto Klaus-Dieter Schluter William C. Sessa John E. Sims Ichiro Sakuma Alvin Schmaier Howard D. Sesso Alan R. Sinaiko Tomas A. Salerno John F. Schmedtje Peter S. Sever Jürgen R. Sindermann Veikko Salomaa Chris Schmid Robert E. Shaddy Pawan K. Singal Koen J. Salu Holger Schmid Ajay M. Shah Krishna Singh Carlo Salvarani Ann-Marie Schmidt Dipen C. Shah Sanjay Singh Flora Sam Carsten B. Schmidt-Weber Pravin M. Shah Lawrence I. Sinoway Habib Samady Gerd Schmitz Prediman K. Shah Albert J. Sinusas Afshin Samali David J. Schneider Robin Shandas Karin Sipido Nilesh J. Samani Michael D. Schneider Richard P. Shannon David S. Siscovick Gianmario Sambuceti Andreas Schober Oz M. Shapira Samuel C. Siu Jonathan M. Samet Gabriele Schoedon Arya M. Sharma Deborah A. Siwik Willis K. Samson Albert Schoemig Frank R. Sharp Carsten Skurk Prashanthan Sanders Frederick J. Schoen Norman Sharpe Cornelis J. Slager John E. Sanderson David A. Schoenfeld A. Richey Sharrett Mara Slawsky David C. Sane Paul Schoenhagen Michael J. Shattock Peter Sleight Anthony J. Sanfilippo Peter M. Scholz Philip W. Shaul Marvin J. Slepian L. Fernando Santana Uwe Schonbeck Leslee J. Shaw Karen Sliwa Massimo Santini Ronald Schondorf Amanda M. Shearman Gregory Sloop Marisa Santos Wilhelm Schoner Michael Shechter Joost P.G. Sluijter Maria-Jesus Sanz Rolf Schroeder Imad Sheiban Richard W. Smalling John L. Sapp Stephen Schroeder James Shepherd Eric J. Smart Maurice E. Sarano Valarie Schroeder Warren Sherman Otto A. Smiseth Ian J. Sarembock Karsten Schrör Mark V. Sherrid Alberto Smith Mark J. Sarnak Joerg B. Schulz Sanjay Shete Felicity B. Smith Masataka Sata Rainer Schulz Weibin Shi George D. Smith Toshiaki Sato Richard Schulz Rei Shibata Gordon C.S. Smith Naveed Sattar P. Christian Schulze Mei-Chiung Shih Jonathan Smith J. Philip Saul Paul T. Schumacker Kazuyuki Shimada Nicholas L. Smith Elijah Saunders Holger J. Schunemann Wataru Shimizu Scott A. Smith Kurt W. Saupe Markus Schwaiger Hiroaki Shimokawa Sidney C. Smith Bernhard Sauter Lee H. Schwamm Ken Shinmura Steven R. Smith Motoji Sawabe Gregory G. Schwartz Satoshi Shintani Ulf Smith Tatsuya Sawamura Ketty Schwartz Ichiro Shiojima Warren M. Smith Leslie A. Saxon Peter J. Schwartz Kalyanam Shivkumar William M. Smith James W. Sayre Stephen M. Schwartz Michael G. Shlipak Pieter C. Smits Angelo M. Scanu David S. Schwartzman Ralph V. Shohet Allan D. Sniderman e24 Acknowledgment of Reviewers

Marieke B. Snijder Christoph Stellbrink Istvan Szokodi Rong Tian Burton E. Sobel Kurt R. Stenmark Roman F. Sztajzel Uwe J.F. Tietge Kenji Sobue David W. Stepp Ira A. Tabas Laurence Tiret Stefan Soderberg Andrew Steptoe Stefano Taddei Marc D. Tischler Kyoko Soejima David M. Stern Heinrich Taegtmeyer Susan Tiukinoy Constantinos T. Sofocleous Naftali Stern Peter Taggart Jonathan M. Tobis Raija Soininen Lynne Warner Stevenson Kazuhiro Takahashi Geoffrey H. Tofler Minna Soinio Duncan J. Stewart Masato M. Takahashi Stevan P. Tofovic John Solaro Kerry J. Stewart Bonpei Takase Naoki Tokita Steven J. Sollott Jim Stewart Hiroshi Takayama Eran Toledo Scott D. Solomon Julian M. Stewart Yoshiyu Takeda Douglas M. Tollefsen Prem Soman Ralph A. H. Stewart Satoshi Takeo Robert J. Tomanek John Somberg Roland Stocker Akira Takeshita Gordon F. Tomaselli Virend K. Somers Jean-Claude Stoclet Renato Talamini Marcello Tonelli Paul D. Sorlie Katarzyna Stolarz William T. Talman Peter Tontonoz Farzaneh Aghdassi Sorond Claudia Stollberger Rasa Tamosiuniene Eric Topol J. Eduardo Sousa Gregg W. Stone Chee Eng Tan Jan H.M. Tordoir P.C. Souverein Neil J. Stone Walter A. Tan Per Tornvall James R. Sowers Peter H. Stone Toshihiro Tanaka Olga H. Toro-Salazar Madison S. Spach George A. Stouffer Weihong Tang Christian Torp-Pedersen Rainer Spanbroek Vibeke Strand W.H. Wilson Tang Guillermo Torre-Amione Carl P. Sparrow Timo E. Strandberg Yao Liang Tang Jan Torzewski Christian M. Spaulding John R. Stratton Rajendra K. Tangirala Tor D. Tosteson J. David Spence Bodo E. Strauer Ataru Taniguchi Peter P. Toth William H. Spencer Arnold Strauss Laszlo B. Tanko Florence Toti John A. Spertus William B. Strawn Felix C. Tanner Arturo G. Touchard Philip Spevak S. Adam Strickberger Jean-Claude Tardif Rhian M. Touyz Lukas E. Spieker Jack P. Strong Robert B. Tate Jeffrey A. Towbin Francis G. Spinale Allan D. Struthers Hideki Tatewaki Dwight A. Towler David H. Spodick Matthias Stuber Allen J. Taylor Jonathan N. Townend David Spragg Jorg Stypmann Andrew M. Taylor Paul A. Townsend Joachim Spranger Ding-Feng Su Anne L. Taylor Maurizio Trevisan Deepak Srivastava Isabella Sudano Joan M. Taylor Richard W. Troughton Martin G. St. John Sutton Peter H. Sugden W. Robert Taylor Nathan A. Trueblood Eugenio Stabile Galina K Sukhova James E. Tcheng Shen K. Tsai Austin Stack Yao Sun Guillermo J. Tearney Min-Fu Tsan Jan A. Staessen Zhonghua Sun Alain Tedgui Teresa S.M. Tsang Diana M. Stafforini Thoralf M. Sundt, III Usha Tedrow Philip S. Tsao Gregory L. Stahl Ruey J. Sung Paul S. Teirstein Hung-Fat Tse Anton F.H. Stalenhoef H. Robert Superko David F. Teitel Etsuko Tsuda Bruce S. Stambler Howard K. Surks George Tellides Gozoh Tsujimoto Jonathan S. Stamler Mark A. Sussman Marije ten Wolde Katsuhiko Tsujioka Meir J. Stampfer Thomas M. Suter Koon K. Teo Hiroyuki Tsukui Kenneth Stanley Fraser W.H. Sutherland Oren M. Tepper Jack V. Tu William C. Stanley George R. Sutherland Gail R. ter Haar Michael L. Tuck Alice V. Stanton John L. Sutko Hiroki Teragawa Rubin M. Tuder Randall C. Starling Richard Sutton Enrique Teran Paul A. Tunick Brian L. Stauffer Kim Sutton-Tyrrell Dellara F. Terry Zoltan G. Turi Charles Steenbergen Hiroshi Suzuki Daniel Teupser Craig D. Turnbull Philippe G. Steg Ken Suzuki David Thaler Stephen T. Turner Coen D. Stehouwer Alan F. Sved Pierre Theroux Alexander G.G. Turpie Evan A. Stein Lars G. Svensson Aravinda Thiagalingam Katherine R. Tuttle Kenneth M. Stein Madhav Swaminathan Perumal Thiagarajan E. Murat Tuzcu Paul D. Stein Lorna Swan Chris Thiemermann Marcel Twickler Francene M. Steinberg Karl Swedberg Gaetano Thiene Suresh C. Tyagi Helmut O. Steinberg G. Sweeney Victor L.J.L. Thijssen Toshimitsu Uede Susan F. Steinberg Michael O. Sweeney Anita C. Thomas Per M. Ueland Julia Steinberger Charles D. Swerdlow James D. Thomas Renan Uflacker Gustav Steinhoff Bernard Swynghedauw MaryLou Thompson Cuno S.P.M. Uiterwaal Robin H. Steinhorn Christer Sylven Paul D. Thompson Gudrun Ulrich-Merzenich Steve R. Steinhubl Zoltán Szabó Kent Thornburg Shin-Ichiro Umemura Acknowledgment of Reviewers e25

Paul M. Underwood Renaud Vincent Mark W.I. Webster Heinrike Wilkens Roger H. Unger Jakob Vinten-Johansen William Weglicki Ian B. Wilkinson Thomas Unger Francesco Violi Chiming Wei Bruce L. Wilkoff Zoltan Ungvari Maria L. Virella Li Wei Walter C. Willett Joseph L. Unthank Renu Virmani Max H. Weil David O. Williams Gilbert R. Upchurch Sami Viskin Hartmut Weiler Kevin J. Williams Zsolt Urbán Eric Vittinghoff Janice Weinberg Mark A. Williams Fumitaka Ushikubi Barbara Voetsch Tanja Weinbrenner Paul T. Williams Viola Vaccarino Michael Vogel Andrew R. Weintraub Roberta G. Williams Marco Valgimigli Robert A. Vogel William S. Weintraub R. Sanders Williams Patrick J.T. Vallance Paul G.A. Volders Michael Weis Allison E. Willing Jesus G. Vallejo Stefano Volpato Richard D. Weisel Scott R. Willoughby Eric Van Belle Klaus von Bergmann Mary C. Weiser-Evans Emily Wilson Gerald van Belle Jan H. von der Thüsen Myron L. Weisfeldt Peter W. Wilson Marc van Bilsen Arnold von Eckardstein Daiana Weiss Gayle L. Winters Luc M. Van Bortel Robert Voswinckel Guenter Weiss Andrew L. Wit Frans J. Van de Werf Atsuyuki Wada James N. Weiss Hanspeter Witschi Johanna G. van der Bom Carol Wadham Robert G. Weiss Maarten Witsenburg Willem J. van der Giessen Bernard Waeber Neil J. Weissman Jacqueline C.M. Witteman Bernd van der Loo Lynne Wagenknecht Jeffrey I. Weitz Rochus Witthaut Freek J. van der Meer Andreas H. Wagner Babette B. Weksler Fred H.M. Wittkampf Irene M. van der Meer Denisa D. Wagner Hein J. Wellens Joseph L.Witztum Yvonne T. van der Schouw Galen S. Wagner Ian J. Welsby Stephen D. Wiviott Jolanda van der Velden Louis K. Wagner Frederick G. Welt J. Frederick Woessner Ger J. van der Vusse Peter D. Wagner Francine K. Welty Wojciech Wojakowski Miranda Van Eck Shawn Wagner Stephen E. Welty Philip A. Wolf George F. Van Hare Ron Waksman Nanette K. Wenger Michael S. Wolin C. Heleen van Ommen Albert L. Waldo Bruce M. Wentworth Robert Wolk Niels van Royen Brian R. Walker Jolanda J. Wentzel Kai C. Wollert Dirk J. van Veldhuisen Lars Wallentin Rene R. Wenzel Ernst Wolner David R. Van Wagoner B. Gunnar Wallin Volker Wenzel Cheuk-kit Wong Anne M. VanBuskirk John Wallwork Gerald S. Werner LennieWong Mani A. Vannan Edward P. Walsh Rainer Wessely Nathan D. Wong Nerea Varo Peter N. Walsh Malcolm West John C. Wood Sudesh Vasdev Dirk H. Walter Rudi G. Westendorp Mark A. Wood Giuseppe Vassalli Thomas Walther Justin Westhuyzen Hermann Wrigge Theodoros Vassilakopoulos Bingcheng Wang Charles V. Wetli Jackson T. Wright, Jr. Guy Vassort Donna H.Wang Glenn T. Wetzel R. Scott Wright Stephen F. Vatner Thomas J. Wang Lewis Wexler Ed X. Wu Matteo Vatta Wei Wang Cornelia M. Weyand Gordon D. Wu Douglas E. Vaughan Xiaohong Wang Arthur E. Weyman Joseph C. Wu William K. Vaughn Yibin Wang Andrew S. Weyrich Kenneth K. Wu Mark A. Veazie Zhiguo Wang Christopher J. White D. George Wyse James L. Velianou Carole A. Warnes C. Roger White Guohua Xi Richard C. Venema Karl Wasserman Halina White Lei Xiao Paolo Verdecchia David D. Waters Richard H. White Rui-Ping Xiao Pieter D. Verdouw Hugh Watkins William B. White Chengjie Xiong Stefan Verheye Steve P. Watson Patrick L. Whitlow Magdi H. Yacoub Petra Verhoef Wendy A. Wattigney J. Lindsay Whitton Jay S. Yadav Jean Philippe Verhoye Gerald F. Watts Mark H. Wholey Yoshiji Yamada Subodh Verma Sergio Waxman Lawrence Wickerham Kazuhiro Yamamoto Richard L. Verrier W. Douglas Weaver Samuel A. Wickline Yoshiharu Yamamoto Francesco Versaci Catherine Webb Petr Widimsky Atsushi Yamashita Giorgio A. Vescovo David J. Webb Susan E. Wiegers Gan-Xin Yan George W. Vetrovec Gary Webb Wouter Wieling Xinhua Yan Aristidis Veves Steven A. Webber FrankWiesmann Clyde W. Yancy G. Wesley Vick Christian Weber William Wijns Qiong Yang Neill Videlefsky Karl T. Weber David J. Wilber Xiao-Ping Yang Flordeliza S. Villanueva Michael A. Weber Arthur A.M. Wilde Zhihong Yang Francisco Villarreal Nina C. Weber Rachel P. Wildman Hirofumi Yasue D. Geoffrey Vince Keith A. Webster Markus J. Wilhelm Frank Yatsu e26 Acknowledgment of Reviewers

Richard Ye Pierre Y. Youinou Marc Zee Felix Zijlstra Jerry Yee James B. Young Robert Y. Zee Michael R. Zile Edward T.H. Yeh Lawrence H. Young Kenton J. Zehr Peter J. Zimetbaum Mao-Hsiung Yen Martin E. Young Andreas M. Zeiher Marc Zimmermann Midori A. Yenari Pampee P. Young Darryl C. Zeldin Jean-Marc Zingg Shaw-Fang Yet Chun Yuan Andrey G. Zenovich Douglas P. Zipes Alan C. Yeung Sun Yuhua Uwe Zeymer Carmine Zoccali Seppo Yla-Herttuala Salim Yusuf Cuihua Zhang William A. Zoghbi Agneta Yngve Susanne Zadelaar Yingyi Zhang Ai-Ping Zou Paul G. Yock Kenneth G. Zahka Guixiang Zhao Ming Zou Junji Yodoi Osama O. Zaidat Zhi-Jie Zheng Irving H. Zucker Young-sup Yoon Alberto Zanchetti Guangming Zhong Bram D. Zuckerman Chaim Yosefy Faiez Zannad Robert Zhong Mahmoud Zureik Hiroshi Yoshida Wojciech Zareba Jianhui Zhu Jay L. Zweier Masayuki Yoshida Barry L. Zaret Xinsheng Zhu Noriko Yoshida Alan M. Zaslavsky Brenda K. Zierler Dimitri E. Zylberstein Cushman et al CRP and Coronary Disease in the Elderly 27

TABLE 1. Distribution of Cardiovascular Risk Factors by Baseline CRP Concentration

CRP, mg/L

Ͻ1 1–3 Ͼ3–10 Ͼ10 P for (nϭ1144) (nϭ1783) (nϭ811) (nϭ233) Trend Continuous variables Age, y 72.9 72.4 72.4 72.6 0.16 Body mass index, kg/m2 24.7 26.8 28.6 29.0 Ͻ0.001 Waist, cm 89.5 94.7 99.1 100.2 Ͻ0.001 LDL cholesterol, mmol/L (mg/dL) 3.29 (127) 3.47 (134) 3.44 (133) 3.23 (125) 0.21 HDL cholesterol, mmol/L (mg/dL) 1.50 (58) 1.40 (54) 1.35 (52) 1.32 (51) Ͻ0.001 Total cholesterol, mmol/L (mg/dL) 5.41 (209) 5.57 (215) 5.54 (214) 5.18 (200) 0.67 Pack-years (among ever smokers) 29.0 32.6 36.8 36.7 Ͻ0.001 Categorical variables, % Male sex 43.4 42.5 37.0 41.2 0.02 Black race 10.9 13.5 20.0 29.2 Ͻ0.001 Hypertension 46.7 56.5 64.6 66.5 Ͻ0.001 Diabetes IFG 11.2 14.0 16.9 19.4 Diabetes 8.2 13.7 21.8 25.0 Ͻ0.001 Smoking status Former 40.9 38.8 40.8 41.2 Current 8.5 11.9 16.6 20.2 Ͻ0.001 Regular aspirin use 19.2 19.1 22.5 20.2 0.17 Any subclinical cardiovascular disease 55.8 63.3 67.8 69.1 Ͻ0.001 Ankle-arm index Ͻ0.9 6.7 9.3 14.4 13.2 Ͻ0.001 Carotid intima-media thickness Ͼ80th percentile 23.1 30.3 35.1 36.1 Ͻ0.001 Major ECG abnormality 19.7 23.9 25.0 25.6 0.004 Carotid stenosis Ͼ25% 39.4 43.8 47.7 50.4 Ͻ0.001 Rose angina positive 3.1 2.4 2.6 4.3 0.78 Rose claudication positive 0.6 1.0 1.7 2.6 0.002 IFG indicates impaired fasting glucose. Values for continuous variables are means. increased risk for CRP Ͼ3 mg/L. There was no effect of Figure 4 shows the 10-year sex-specific incidence of CHD additional adjustment for baseline statin use. Further adjust- according to CRP concentration in categories of the Framing- ment for subclinical disease yielded little attenuation; a 37% ham Risk Score. In intermediate- and low-risk women, CRP increased risk of CHD for elevated CRP remained. When Ͼ3 mg/L added little to risk prediction, whereas in high-risk CRP was considered a continuous variable, with adjustment women, CRP provided additional risk information. Among for risk factors, the relative risk associated with a 1-ln-unit- women with a 10-year predicted risk Ͼ20%, for intermediate higher baseline CRP was 1.27 (95% CI, 1.12 to 1.44). There or elevated CRP, the observed incidences were 28%, and were no significant differences in associations by sex or race. 31%, respectively, compared with only 16% for those with The population-attributable risk percentage for elevated CRP low CRP. In men, CRP provided additional risk information was 11%. in intermediate- and high-Framingham-risk groups. Among Table 3 shows the relative risks of incident CHD for CRP men with a 10-year predicted risk 10% to 20%, those with Ͼ3 mg/L compared with CRP Ͻ1 mg/L in subgroups based CRP Ͼ3 mg/L had an observed risk of 32%. Among on the presence or absence of cardiovascular risk factors. high-Framingham-risk men, this observed risk was 41%. Baseline CRP was associated with CHD in all of these We investigated the utility of CRP Ͼ10 mg/L for deter- groups, including those without subclinical disease and those mining risk of CHD and the impact of hormone replacement at low risk by the Framingham Risk Score (P for interaction therapy among women. Of participants with elevated CRP, Ͼ0.05 for all). Although the relative risk did not differ by 22% were Ͼ10 mg/L. Of these 233 participants, 49 (21%) subclinical disease status, among men and women, the developed CHD during follow-up compared with 498 of 3738 presence of elevated CRP together with subclinical disease (13.3%) with lower CRP. The age-, sex-, and race-adjusted was associated with a higher incidence of CHD compared relative risk of CHD was 2.16 (95% CI, 1.55 to 3.00) for CRP with those with lower CRP and no subclinical disease Ͼ10 compared with Ͻ1 mg/L and 1.78 (95% CI, 1.26 to (Figure 3). 2.51) after adjustment for traditional risk factors. Among 326 28 Circulation July 5, 2005

Figure 2. Incidence rate per 1000 person-years of MI or CHD death by baseline CRP. Incidence rates were calculated within small intervals of CRP values and plotted with a scatterplot smoother. Association was well fit by a quadratic function of CRP, plotted with 95% confidence bands.

recent clinical guidelines were applied, intermediate CRP concentrations (1 to 3 mg/L) were weakly related to future CHD, and elevated CRP (Ͼ3.0 mg/L) was associated with a 1.45-fold increased risk of CHD, with adjustment for other vascular risk factors. There was little further confounding with adjustment for the presence of noninvasively assessed subclinical atherosclerosis. Elevated CRP was associated with CHD in all subgroups defined by conventional cardiac risk factors or subclinical disease. Among men with interme- Figure 1. Cumulative rate of MI or CHD death. Top, data for diate and high Framingham Risk Scores, CRP identified those men; bottom, data for women. Unadjusted hazard ratios and with higher-than-predicted risk. Among women, CRP dis- 95% CIs for each group compared with reference group (CRP Ͻ1 mg/L) are shown. Solid line indicates CRP Ͻ1 mg/L; dotted criminated risk best among those at high Framingham- line, CRP 1 to 3 mg/L; and dashed line, CRP Ͼ3 mg/L. predicted risk. The relative risk of CHD for elevated CRP observed here women excluded from analysis for hormone replacement was smaller than in most studies of middle-aged subjects and therapy use, the age-adjusted relative risk of CHD for CRP might seem modest at 1.45. However, event rates were high Ͼ3 mg/L was 1.35 (95% CI, 0.42 to 4.32). in this age group, so the attributable risk percent for elevated CRP was high at 11%, even given a modest relative risk.30 Discussion Thus, a much higher percentage of subjects with elevated In this 10-year prospective study in men and women Ն65 CRP subsequently had events in this study compared with years of age, CHD risk increased with increasing CRP. When studies of younger subjects.7,8 In a recent report by Danesh et

TABLE 2. Association of Baseline CRP With Incident MI or CHD Death Over 10 Years*

Relative Risk (95% CI) in CRP Categories

Ͻ1.0 mg/L 1.0–3.0 mg/L Ͼ3.0 mg/L (Nϭ1144, nϭ135) (Nϭ1783, nϭ230) (Nϭ1044, nϭ182) P Incidence rate, % (n of events) Men 17.1 (74) 20.6 (127) 33.3 (96) Women 10.4 (61) 11.0 (103) 15.5 (86) Model 1 1.0 (ref) 1.18 (0.96–1.46) 1.82 (1.46–2.28) Ͻ0.001 Model 2 1.0 (ref) 1.08 (0.86–1.35) 1.45 (1.14–1.86) Ͻ0.004 Model 3 1.0 (ref) 1.04 (0.82–1.31) 1.37 (1.06–1.78) 0.01 N is number at risk in given group; n, number of cases in given group; and ref, reference. *Model 1 is adjusted for age, race, and sex. Model 2 is adjusted for age, sex, race, field center, hypertension, diabetes, smoking status, log pack-years, body mass index, waist circumference, total cholesterol, HDL cholesterol, and regular aspirin use. Model 3 is adjusted for model 2 variables plus ankle-arm index Ͻ0.9, internal or common carotid intima-media thickness Ͼ80th percentile, positive responses to the Rose angina or claudication question- naires, major ECG abnormalities, and maximum stenosis of the carotid artery Ͼ25%. Cushman et al CRP and Coronary Disease in the Elderly 29

TABLE 3. Relative Risk of CHD for CRP >3 mg/L Compared With <1 mg/L by Categories of Baseline Risk Factors

Risk Factor Present Risk Factor Absent

Risk Factor n/N RR* (95% CI) n/N RR* (95% CI) Smoking (formerϩcurrent) 307/2076 1.84 (1.37–2.47) 240/1890 1.68 (1.18–2.40) Pack-years (Ͼmedian; ever smokers only) 171/967 2.13 (1.40–3.25) 120/1000 1.61 (1.02–2.53) Hypertension 365/2219 1.74 (1.32–2.29) 182/1748 1.55 (1.04–2.32) Diabetes or impaired fasting glucose 213/1130 1.49 (1.02–2.18) 334/2833 1.74 (1.30–2.32) Hyperlipidemia 139/966 1.99 (1.25–3.19) 404/2971 1.74 (1.34–2.26) Regular aspirin use 123/788 1.89 (1.17–3.05) 423/3176 1.79 (1.39–2.31) Estimated 10-year Framingham Risk Score Ͼ20% 181/772 2.00 (1.28–3.13) 357/3127 1.51 (1.15–1.98) Subclinical disease 417/2477 1.85 (1.42–2.42) 130/1494 1.42 (0.91–2.22) Carotid wall thickness Ͼ80th percentile 250/1167 1.56 (1.11–2.20) 293/2784 1.71 (1.26–2.32) Ankle-arm index Ͻ0.9 95/381 1.45 (0.82–2.59) 443/3514 1.78 (1.39–2.28) Carotid stenosis Ն25% 315/1726 1.68 (1.25–2.25) 230/2223 1.74 (1.22–2.47) Major ECG abnormality 177/886 1.73 (1.15–2.61) 355/2963 1.81 (1.38–2.38) n/N indicates number of events/number at risk in all 3 levels of CRP in specified category of each risk factor. *Adjusted for age, race, and sex. al,9 a similar adjusted relative risk was observed in a large of clinical disease, CRP was higher among those with any population, but in that case-control study, attributable risk single type of subclinical disease. Moreover, CHD incidence was not estimated. If elevated CRP represents a causal risk was higher among those with elevated CRP and subclinical factor as suggested by several experimental studies,31 our disease compared with groups with only 1 or neither of these estimate of attributable risk indicates a hypothesis that cor- risk factors. In this cohort, the 10-year stroke risk associated rection of elevated CRP could eliminate up to 11% of with elevated CRP was larger among those with higher incident CHD in this age group. compared with lower carotid intima-media thickness.39 In a It has been suggested that novel risk factors or atheroscle- short-term study, the risk of MI was higher among those with rosis imaging may identify those at intermediate CHD risk higher coronary artery calcium scores if CRP was also who might benefit from aggressive risk factor interventions.32 elevated.40 Taken together, findings from these few studies Along with findings in middle-aged populations,7,8 our data suggest possible roles for the assessment of both inflamma- provide evidence that CRP assessment can identify older tion and subclinical disease. It is also possible that CRP is a patients at higher or lower than their predicted risk of marker of subclinical disease, and if better measures of coronary events. Our findings with regard to women at low subclinical disease were available, adjustment for subclinical and intermediate risk differ from findings in middle-aged disease would further lessen the association of CRP with women in which CRP predicted cardiovascular events across CHD. the entire range of Framingham Risk Scores.33 Further work The CDC/AHA guideline for CRP testing suggest that is needed to validate our findings in this age group and to values Ͼ10 mg/L indicate acute inflammation and have determine appropriate values defining elevated CRP in vari- uncertain implications for vascular risk prediction.3 In this ous age and sex groups. older population, 6% of subjects had CRP Ͼ10 mg/L; when Other studies reported weak or no associations of CRP traditional risk factors were accounted for, these subjects had with subclinical disease measures.34–38 Here, in the absence a 1.8-fold increased risk of CHD, a higher risk estimate than

Figure 3. Incidence rates per 1000 person-years of first MI or CHD death by baseline CRP, stratified by sex and pres- ence of subclinical atherosclerosis. 30 Circulation July 5, 2005

measured in the whole cohort, incidence rates of CHD by baseline CRP were calculated, and subgroup analyses could be done. In conclusion, we extend previous reports on the associa- tion of CRP with CHD to men and women Ն65 years of age. CRP appears to be useful for risk assessment in this age group. Because event rates are high overall in older age, further study is required to determine optimal clinical roles of CRP measurement, especially as related to interventions for elevated CRP.

Acknowledgments This research was supported by contracts N01-HC-85079 through N01-HC-85086, N01-HC-35129, and N01-HC-15103 and grants HL-46696, HL-8329, and HL-03618 from the NIH National Heart, Lung and Blood Institute. A full list of participating CHS investiga- tors and institutions can be found at http://www.chs-nhlbi.org. References 1. Psaty BM, Furberg CD, Kuller LH, Bild DE, Rautaharju PM, Polak JF, Bovill E, Gottdiener JS. Traditional risk factors and subclinical disease measures as predictors of first myocardial infarction in older adults: the Cardiovascular Health Study. Arch Intern Med. 1999;159:1339–1347. 2. Psaty BM, Anderson M, Kronmal RA, Tracy RP, Orchard T, Fried LF, Lumley T, Robbins J, Burke G, Newman AB, Furberg CD. The asso- ciation between lipid levels and the risks of incident myocardial infarction, stroke, and total mortality: the Cardiovascular Health Study. Am J Geriatr Soc. 2004;52:1639–1647. 3. Pearson TA, Mensah GA, Alexander RW, Anderson JL, Cannon RO 3rd, Criqui M, Fadl YY, Fortmann SP, Hong Y, Myers GL, Rifai N, Smith SC Figure 4. Ten-year rate of CHD according to baseline CRP and Jr, Taubert K, Tracy RP, Vinicor F. Markers of inflammation and car- 10-year predicted risk from the Framingham Risk Score. Top, diovascular disease: application to clinical and Public Health practice: a data for women; bottom, data for men. Observed incidence statement for healthcare professionals from the Centers for Disease based on categories of CRP was determined within each cate- Control and Prevention and the American Heart Association. Circulation. gory of Framingham-predicted risk. For each category, numbers 2003;107:499–511. across top represent number of events per number at risk in 4. Ridker PM, Glynn RJ, Hennekens CH. C-reactive protein adds to the that group. predictive value of total and HDL cholesterol in determining risk of first myocardial infarction. Circulation. 1998;97:2007–2011. 5. Ridker PM, Hennekens CH, Buring JE, Rifai N. C-reactive protein and for CRP Ͼ3.0 mg/L. Our finding agrees with recently other markers of inflammation in the prediction of cardiovascular disease reported results in middle-aged women.33 Thus, CRP values in women. N Engl J Med. 2000;342:836–843. Ͼ10 mg/L appear to be important in CHD risk prediction. 6. Kervinen H, Palosuo T, Manninen V, Tenkanen L, Vaarala O, Manttari M. Joint effects of C-reactive protein and other risk factors on acute Limitations of this study merit consideration. The cohort, coronary events. Am Heart J. 2001;141:580–585. free-living elderly who were willing to enroll in the study, 7. Ridker PM, Rifai N, Rose L, Buring JE, Cook NR. Comparison of may not represent the general older population. The observa- C-reactive protein and low-density lipoprotein cholesterol levels in the prediction of first cardiovascular events. N Engl J Med. 2002;347: tional study design, even with extensive multivariate analysis, 1557–1565. cannot prove causal relationships. Competing risks may have 8. Koenig W, Lowel H, Baumert J, Meisinger C. C-reactive protein mod- diluted associations of CRP with CHD because CRP may be ulates risk prediction based on the Framingham score: implications for future risk assessment: results from a large cohort study in southern associated with other disease outcomes. In some cases, Germany. Circulation. 2004;109:1349–1353. analysis of subgroups was limited by small sample sizes. 9. Danesh J, Wheeler JG, Hirschfield GM, Eda S, Eiriksdottir G, Rumley A, Finally, CRP was measured only once at baseline, and it has Lowe GD, Pepys MB, Gudnason V. C-reactive protein and other circu- been suggested that repeated testing for confirmation be lating markers of inflammation in the prediction of coronary heart disease. N Engl J Med. 2004;350:1387–1397. 3 considered in those with high values. 10. Tracy RP, Lemaitre RN, Psaty BM, Ives DG, Evans RW, Cushman M, Strengths of this study include its large size, extensive Meilhan EN, Kuller LH. Relationship of C-reactive protein to risk of baseline data collection, and long-term event follow-up. cardiovascular disease in the elderly: results from the Cardiovascular Health Study and the Rural Health Promotion Project. Arterioscler Several new findings were observed on the basis of unique Thromb Vasc Biol. 1997;17:1121–1127. aspects of the study. First, we confirmed an association of 11. Harris TB, Ferruci L, Tracy RP, Corti MC, Wacholder S, Ettinger WH, elevated CRP with CHD incidence in an older age group; Heimovitz H, Cohen HJ, Wallace R. Associations of elevated most previous studies included younger subjects or clinical interleukin-6 and C-reactive protein levels with mortality in the elderly. Am J Med. 1999;106:506–512. trial participants. Second, independence of associations from 12. Tice JA, Browner W, Tracy RP, Cummings SR. The relation of noninvasively measured subclinical atherosclerosis was doc- C-reactive protein levels to total and cardiovascular mortality in older umented. Third, more complete adjustment for smoking U.S. women. Am J Med. 2003;114:199–205. 13. van der Meer IM, de Maat MP, Kiliaan AJ, van der Kuip DA, Hofman A, status was made by assessing pack-years, a major determinant Witteman JC. The value of C-reactive protein in cardiovascular risk of CRP concentration in smokers.34 Fourth, because CRP was prediction: the Rotterdam Study. Arch Intern Med. 2003;163:1323–1328. Cushman et al CRP and Coronary Disease in the Elderly 31

14. Cesari M, Penninx BW, Newman AB, Kritchevsky SB, Nicklas BJ, disease: prospective analysis from the Women’s Health Initiative obser- Sutton-Tyrrell K, Rubin SM, Ding J, Simonsick EM, Harris TB, Pahor M. vational study. JAMA. 2002;288:980–987. Inflammatory markers and onset of cardiovascular events: results from 28. Price TR, Psaty BM, O’Leary DH, Burke G, Gardin J, for the Cardio- the Health ABC Study. Circulation. 2003;108:2317–2322. vascular Health Study Research Group. Assessment of cerebrovascular 15. Jager A, van Hinsbergh VW, Kostense PJ, Emeis JJ, Yudkin JS, Nijpels disease in the Cardiovascular Health Study. Ann Epidemiol. 1993;3: G, Dekker JM, Heine RJ, Bouter LM, Stehouwer CD. von Willebrand 504–507. factor, C-reactive protein, and 5-year mortality in diabetic and nondi- 29. Ives DG, Fitzpatrick AL, Bild DE, Psaty BM, Kuller LH, Crowley PM, abetic subjects: the Hoorn Study. Arterioscler Thromb Vasc Biol. 1999; Cruise RG, Theroux S. Surveillance and ascertainment of cardiovascular 19:3071–3078. events: the Cardiovascular Health Study. Ann Epidemiol. 1995;5: 16. Mendall MA, Strachan DP, Butland BK, Ballam L, Morris J, Sweetnam 278–285. PM, Elwood PC. C-reactive protein: relation to total mortality, cardio- 30. Psaty BM, Koepsell TD, Manolio TA, Longstreth WT Jr, Wagner EH, vascular mortality and cardiovascular risk factors in men. Eur Heart J. Wahl PW, Kronmal RA. Risk ratios and risk differences in estimating the 2000;21:1584–1590. effect of risk factors for cardiovascular disease in the elderly. J Clin 17. Doggen CJM, Berckmans RJ, Sturk A, Manger Cats V, Rosendaal FR. Epidemiol. 1990;43:961–970. C-reactive protein, cardiovascular risk factors and the association with 31. Hirschfield GM, Pepys MB. C-reactive protein and cardiovascular myocardial infarction in men. J Intern Med. 2000;248:406–414. disease: new insights from an old molecule. QJM. 2003;96:793–807. 18. Gram J, Bladbjerg EM, Moller L, Sjol A, Jespersen J. Tissue-type plas- 32. Greenland P, Smith SC Jr, Grundy SM. Improving coronary heart disease minogen activator and C-reactive protein in acute coronary heart disease risk assessment in asymptomatic people: role of traditional risk factors a nested case-control study. J Intern Med. 2000;247:205–212. and noninvasive cardiovascular tests. Circulation. 2001;104:1863–1867. 19. Packard CJ, O’Reilly DS, Caslake MJ, McMahon AD, Ford I, Cooney J, 33. Ridker PM, Cook N. Clinical usefulness of very high and very low levels Macphee CH, Suckling KE, Krishna M, Wilkinson FE, Rumley A, Lowe of C-reactive protein across the full range of Framingham Risk Scores. GD. Lipoprotein-associated phospholipase A2 as an independent pre- Circulation. 2004;109:1955–1959. dictor of coronary heart disease: West of Scotland Coronary Prevention 34. Tracy RP, Psaty BM, Macy E, Bovill EG, Cushman M, Cornell ES, Study Group. N Engl J Med. 2000;343:1148–1155. Kuller LH. Lifetime smoking exposure affects the association of 20. Folsom AR, Aleksic N, Catellier D, Juneja HS, Wu KK. C-reactive C-reactive protein with cardiovascular disease risk factors and subclinical protein and incident coronary heart disease in the Atherosclerosis Risk in disease in healthy elderly subjects. Arterioscler Thromb Vasc Biol. 1997; Communities (ARIC) study. Am Heart J. 2002;144:233–238. 17:2167–2176. 21. Kuller LH, Shemanski L, Psaty BM, Borhani NO, Gardin J, Haan MN, 35. Redberg RF, Rifai N, Gee L, Ridker PM. Lack of association of O’Leary DH, Savage PJ, Tell GS, Tracy R. Subclinical disease as an C-reactive protein and coronary calcium by electron beam computed independent risk factor for cardiovascular disease. Circulation. 1995;92: tomography in postmenopausal women: implications for coronary artery 720–726. disease screening. J Am Coll Cardiol. 2000;36:39–43. 22. Fried LP, Borhani NO, Enright P, Furberg CD, Gardin JM, Kronmal RA, 36. Folsom AR, Pankow JS, Tracy RP, Arnett DK, Peacock JM, Hong Y, Kuller LH, Manolio TA, Mittelmark MB, Newman A, O’Leary DH, Psaty Djousse L, Eckfeldt JH. Association of C-reactive protein with markers of B, Rautaharju P, Tracy RP, Weiler PG, CHS Research Group. The prevalent atherosclerotic disease. Am J Cardiol. 2001;88:112–117. Cardiovascular Health Study: design and rationale. Ann Epidemiol. 1991; 37. Blackburn R, Giral P, Bruckert E, Andre JM, Gonbert S, Bernard M, 1:263–276. Chapman MJ, Turpin G. Elevated C-reactive protein constitutes an inde- 23. Macy EM, Hayes TE, Tracy RP. Variability in the measurement of pendent predictor of advanced carotid plaques in dyslipidemic subjects. C-reactive protein in healthy subjects: implications for reference intervals Arterioscler Thromb Vasc Biol. 2001;21:1962–1968. and epidemiological applications. Clin Chem. 1997;43:52–58. 38. Wang TJ, Nam BH, Wilson PW, Wolf PA, Levy D, Polak JF, D’Agostino 24. Rifai N, Tracy RP, Ridker PM. Clinical efficacy of an automated high- RB, O’Donnell CJ. Association of C-reactive protein with carotid ath- sensitivity C-reactive protein assay. Clin Chem. 1999;45:2136–2141. erosclerosis in men and women: the Framingham Heart Study. Arte- 25. Wilson PW, D’Agostino RB, Levy D, Belanger AM, Silbershatz H, rioscler Thromb Vasc Biol. 2002;22:1662–1667. Kannel WB. Prediction of coronary heart disease using risk factor cate- 39. Cao JJ, Thach C, Manolio TA, Psaty BM, Kuller LH, Chaves PH, Polak gories. Circulation. 1998;97:1837–1847. JF, Sutton-Tyrrell K, Herrington DM, Price TR, Cushman M. C-reactive 26. Cushman M, Legault C, Barrett-Connor E, Stefanick ML, Kessler C, Judd protein, carotid intima-media thickness, and incidence of ischemic stroke HL, Sakkinen PA, Tracy RP. Effect of postmenopausal hormones on in the elderly: the Cardiovascular Health Study. Circulation. 2003;108: inflammation-sensitive proteins: the Postmenopausal Estrogen/Progestin 166–170. Interventions (PEPI) study. Circulation. 1999;100:717–722. 40. Park R, Detrano R, Xiang M, Fu P, Ibrahim Y, LaBree L, Azen S. 27. Pradhan AD, Manson JE, Rossouw JE, Siscovick DS, Mouton CP, Rifai Combined use of computed tomography coronary calcium scores and N, Wallace RB, Jackson RD, Pettinger MB, Ridker PM. Inflammatory C-reactive protein levels in predicting cardiovascular events in nondi- biomarkers, hormone replacement therapy, and incident coronary heart abetic individuals. Circulation. 2002;106:2073–2077. Epidemiology

Obesity, Insulin Resistance, and the Metabolic Syndrome Determinants of Endothelial Dysfunction in Whites and Blacks

A.A. Lteif, MD; K. Han, MD; K.J. Mather, MD

Background—Insulin resistance is strongly associated with obesity and other components of the metabolic syndrome (MS). The relative importance of these components in the determination of endothelial function is unknown. Furthermore, there is conflicting evidence about whether ethnic differences exist in the relative importance of these components in regard to other cardiovascular outcomes. We evaluated the contributions of insulin resistance, obesity, and the other components of the MS to impaired endothelial function. Methods and Results—The relationships of the MS components (as defined according the National Cholesterol Education Program) and insulin resistance (estimated using the homeostasis model) with endothelium-dependent vasodilation were examined in 42 white and 55 black subjects. Endothelium-dependent vasodilation was assessed as the increment in leg blood flow (measured by thermodilution) after exposure to methacholine chloride. Waist circumference, glucose, blood pressure, and insulin resistance distributions did not differ between ethnic groups; blacks in our sample had higher HDL cholesterol (1.31 versus 1.09 mmol/L; PϽ0.001) and lower triglyceride levels (1.01 versus 1.37 mmol/L; Pϭ0.005) than white subjects. In the absence of the MS, black subjects exhibited reduced endothelium-dependent vasodilation compared with white subjects (Pϭ0.005), and both groups demonstrated significantly worse endothelial function when the MS was present (maximal increase in leg blood flow: blacks: 107Ϯ9% MS absent, 53Ϯ16% MS present; whites: 163Ϯ16% MS absent, 54Ϯ18% MS absent; Pϭ0.007, MS absent versus present; PϭNS for interaction of ethnicity and MS). Multivariable regression analysis examining relationships of endothelial function with the 5 MS components (analyzed as continuous variables) revealed independent relationships only with waist circumference (Pϭ0.01) and systolic blood pressure (Pϭ0.02). Waist circumference was no longer independently associated after adding insulin resistance to the modeling (Pϭ0.02 for log of homeostasis model index of insulin resistance, Pϭ0.02 for systolic blood pressure). Ethnicity still exerted an independent effect on endothelial function after accounting for the above components (Pϭ0.04 for an additional effect of ethnic status on endothelial function), with an ethnic difference in the effect of insulin resistance on endothelial function (Pϭ0.046 for interaction of ethnicity and log of homeostasis model index of insulin resistance). Conclusions—These findings suggest that insulin resistance and systolic blood pressure are the principal determinants of endothelial dysfunction in the MS and that there are ethnic differences in the relative importance of these factors. These differences may imply different benefits from treatments targeting blood pressure or insulin resistance in different ethnic groups. (Circulation. 2005;112:32-38.) Key Words: endothelium Ⅲ insulin resistance Ⅲ metabolic syndrome Ⅲ obesity

besity and the metabolic syndrome (MS) are associated Insulin resistance was found to improve the association Owith impaired endothelium-dependent vasodilation between MS and coronary artery calcification.6 In the Wom- (EDV).1,2 Insulin resistance, measured with formal euglyce- en’s Health Initiative, insulin resistance was a key component mic hyperinsulinemic clamp protocols, is closely related to of risk factor modeling to explain cardiovascular disease endothelial dysfunction.1,3 Insulin resistance is postulated to outcomes.7 The interrelationships of insulin resistance, the be the common underlying pathogenic link between the MS, and endothelial function have not been previously various components of the MS4 and may explain the presence reported. of the MS even in nonobese subjects.5 These interrelation- Relatively little information is available about ethnic dif- ships suggest that measures of insulin resistance should ferences in the relationship between insulin sensitivity and perhaps be included with other measures of obesity- and cardiovascular outcomes. In the Insulin Resistance Athero- MS-associated risks for cardiovascular outcomes. Compara- sclerosis Study (IRAS), the intrinsic associations of insulin tively few data are available on the value of this approach. resistance and other MS components did not differ across

Received November 5, 2004; revision received February 24, 2005; accepted March 4, 2005. From Indiana University, Indianapolis. Correspondence to Kieren Mather, 975 W Walnut St, IB424, Indianapolis, IN 46202. E-mail [email protected] © 2005 American Heart Association, Inc. Circulation is available at http://www.circulationaha.org DOI: 10.1161/CIRCULATIONAHA.104.520130 32 Lteif et al Metabolic Syndrome, Ethnicity, and the Endothelium 33 ethnicities.8,9 Although associations of subclinical atheroscle- The National Cholesterol Education Program (NCEP) Adult rosis with insulin resistance and other MS components were Treatment Panel III cutpoints for the presence or absence of the seen in this observational study,10 no ethnic differences have various components of the MS were used for analyses of the associations of these features, alone and in combination, with been reported. In other studies, the prevalence of the MS endothelial function. The MS was considered present if Ն3 of these among blacks appears to be similar to that in other ethnic features were present.23 groups.11–13 Despite this finding, endothelial function has been reported to be impaired in blacks compared with Assays weight-matched white subjects.14–16 Therefore, it is unclear Blood for serum glucose determinations was put in untreated whether the impact of the MS variables on various outcome polypropylene tubes and centrifuged with an Eppendorf microcen- trifuge (Brinkman). The glucose concentration of the supernatant parameters, including endothelial function, is uniform across was then measured by the glucose oxidase method with a glucose ethnicity. analyzer (model 2300, Yellow Springs Instruments). Blood for In light of these questions, we hypothesized that the determination of plasma insulin was collected in heparinized tubes, relationship of the MS components with endothelial function processed immediately, and frozen at Ϫ20°C. Insulin determinations would be dependent on insulin resistance and that this were made with a dual-site radioimmunoassay specific for human insulin and with cross-reactivity with proinsulin Ͻ0.2% (Linco). The relationship would be present in both black and white lower detection limit is 0.56 pmol/L, and in our laboratory, the subjects. We undertook a cross-sectional evaluation of the interassay and intra-assay coefficients of variation are 4.1% and relationships of insulin resistance and the components of the 2.6%, respectively. Standard methodologies for cholesterol and MS with direct in vivo measurements of endothelial function, triglyceride determinations were used for assays performed through specifically evaluating these relationships in black compared our the clinical laboratory of the local hospital. with white subjects. Endothelial Function All vascular studies were performed after an overnight fast, with Methods subjects in a supine position in a quiet, temperature-controlled room. Patients/Data Set A 6F sheath (Cordis Corp) was placed into the right femoral vein to allow the insertion of a custom-designed 5F double-lumen thermodi- Nondiabetic subjects entered into our database of vascular function lution catheter (Baxter Scientific, Edwards Division) to measure leg measurements from 1999 to 2004 were included in these analyses. blood flow (LBF). The right femoral artery was cannulated with a Subjects were community-dwelling volunteers who elected to par- 5.5F double-lumen catheter to allow simultaneous infusion of sub- ticipate in one of our ongoing protocols involving vascular function stances and invasive blood pressure monitoring via a vital signs measurements. These studies recruit subjects across the spectrum of monitor (Spacelabs). Baseline LBF and mean arterial pressure body habitus. We have not routinely studied subjects with frankly measurements were obtained after Ն30 minutes of rest after the elevated blood pressure or lipid levels; therefore, such subjects are insertion of the catheters. Femoral vein thermodilution curves were not overrepresented in our study population. Volunteers for our used to measure LBF rates, calculated by integration of the area studies of vascular function routinely undergo screening measure- under the curve, with a cardiac output computer (model 9520A, ment of fasting lipid profiles and oral glucose tolerance testing. American Edwards Laboratories). At baseline, 24 LBF measure- Diabetes mellitus, either previously diagnosed or discovered with Ϸ screening oral glucose tolerance testing, was an exclusion criterion ments were obtained at 30-second intervals. This was followed by for this analysis. Subjects were categorized prospectively as lean or measurements of the LBF response to graded intrafemoral arterial 2 infusions of the endothelium-dependent vasodilator methacholine obese according to body mass index (female cutpoint, 28 kg/m ; ␮ male cutpoint, 26 kg/m2). Ethnicity was self-reported by the partic- chloride (5, 10, and 15 g/min). Beginning 2 minutes after initiation of each infusion rate, 10 LBF measurements were obtained at ipants. All studies were approved by the local Institutional Review Ϸ Board, and all subjects gave written informed consent. All proce- 30-second intervals. The mean of these 10 measurements was dures were performed in accordance with institutional guidelines taken as the response at each step, and the percent increase in LBF Measurements included in the analysis included anthropomorphic relative to baseline was calculated for each subject. These measures measurements performed within 1 week of the vascular study, and were used as integrated measurements of endothelial function in the blood pressure and biochemical measurements were performed correlation analyses with the various baseline measures discussed above. These responses are specifically inhibited by coinfusion of immediately before the vascular study. Weight was measured on a G single stationary hospital-grade scale calibrated yearly. Height was inhibitors of nitric oxide synthase such as N -monomethyl-L- measured with a single wall-mounted stadiometer. Percentage body arginine, confirming their utility in assessing endothelium-dependent 24–26 fat was measured by dual energy x-ray absorptiometry or water vasodilator responses. This method has variability characteris- displacement (for subjects Ͼ122 kg, the upper limit for the dual tics superior to other commonly used measures of endothelial energy x-ray absorptiometry machine). Waist circumference was function, with within-subject coefficients of variation on the order of measured at the narrowest portion of the abdomen. Blood pressure 10%. measurements by wall-mounted sphygmomanometers were per- formed in duplicate with the subject supine at rest on the study day Statistical Analysis before initiation of study procedures. In some but not all cases, Statistical analyses were performed with JMP version 5.1 (SAS samples for a lipid profile were also obtained before the vascular Institute). The 97 nondiabetic subjects studied in this interval measurements. When this was not done, these values were excluded participated in 162 leg line studies; to take advantage of the from analysis if the screening lipid measurements and the vascular noise-averaging effect of including multiple measurements within study were separated by Ͼ4 weeks. individuals, random-effects modeling was used to account for these The data collected with these studies included duplicate fasting repeated measures. One-way ANOVA was used to compare the insulin and glucose measurements (separated by 10 minutes) before effect of categorical variables on endothelial function. Univariate initiation of the vascular study. These values were used to calculate analyses evaluated the isolated relationships of each of the continu- the homeostasis model index of insulin resistance (HOMA-IR).17 ous variables assessed with endothelial function. Simultaneous HOMA-IR has been found to closely relate to formal measurements multivariate analyses were used to explore the concurrent associa- of insulin sensitivity using hyperinsulinemic euglycemic clamp tions of these variables with endothelial function. Two-sided P procedures9,18 and has proved useful in epidemiological studies as an values of PϽ0.05 were taken as statistically significant. HOMA-IR index of insulin sensitivity.6,19–22 was normalized with a logarithmic transformation before inclusion in 34 Circulation July 5, 2005

TABLE 1. Characteristics of the Study Subjects TABLE 2. Prevalence of Positive MS Components and of the MS in the Study Population Using the Adult Treatment Whites Blacks Panel III Criteria

Parameter Mean SD Mean SD P Whites, % Blacks, % Age, y 36.4 6.9 34.5 7.4 NS Parameter (95% CI) (95% CI) Gender, M/F 31/11 35/20 NS Waist 13.6 (3.2–23.9) 13.6 (4.5–22.6) Patient category, lean/obese 25/17 30/25 NS Triglycerides 19.4 (7.4–31.4) 11.6 (3.1–20.1) BMI, kg/m2 26.1 5.5 29.2 6.1 NS HDL cholesterol 56.7 (44.7–68.7) 41.9 (28.9–54.9) Fat, % 26.6 11.9 27.1 11.9 NS Blood pressure 20.9 (8.6–33.2) 28.4 (16.5–40.3) Waist, cm 79.0 10.8 82.9 10.8 NS Glucose 4.5 (0–10.8) 2.1 (0–5.9) Glucose, mmol/L 5.2 0.5 5.0 0.4 NS MS 14.3 (9.6–19.0) 12.7 (11.2–14.2) Insulin, pmol/L 68.3 71.3 70.9 49.7 NS There were no differences across ethnic category in the prevalence of these HOMA-IR 2.7 2.5 2.8 1.8 NS components (PϭNS for all). SBP, mm Hg 116.7 11.7 122.7 12.8 NS 2ϭ DBP, mm Hg 71.5 9.0 75.9 10.0 NS analyses (model incorporating MS and ethnicity, r 0.64, PϽ0.001; MS ␤ coefficient, 38.8; Pϭ0.006; and ethnicity ␤ Cholesterol, mmol/L 4.26 0.67 4.36 0.79 NS coefficient, 28.4; Pϭ0.007). Adding logHOMA-IR to this LDL, mmol/L 2.50 0.49 2.58 0.71 NS model removed the independent effect of the MS, leaving insulin HDL, mmol/L 1.09 0.24 1.31 0.33 0.001 resistance and race as the independent variables (r2ϭ0.67, Triglycerides, mmol/L 1.37 0.81 1.01 0.39 0.005 PϽ0.001; logHOMA-IR ␤ coefficient, 93.6; Pϭ0.002; ethnicity BMI indicates body mass index; SBP, systolic blood pressure; and DBP, ␤ coefficient, 27.5; Pϭ0.009). diastolic blood pressure. We analyzed the contributions of the individual components of the MS to endothelial dysfunction using continuous variables parametric analyses,18 and values presented reflect the inverse rather than the simple categorical designations provided by transformation of the results from these analyses. Except when Adult Treatment Panel III cutpoints. All analyses were adjusted repeated-measures analysis for the dose-response to methacholine is as above. The MS components that were individually associated presented, the percent increase in LBF to the maximal dose of ϭϪ methacholine was used as the endothelial function end point. Age with impaired EDV were as follows: waist: r 0.33, and basal (unstimulated) blood flow were important confounders of PϽ0.0001; systolic blood pressure: rϭϪ0.29, Pϭ0.0002; and the relationships under study; therefore, these variables were in- triglycerides: rϭ0.19, Pϭ0.02. Insulin resistance (ie, cluded in all multivariable analyses as covariates, in addition to the logHOMA-IR) was also significantly associated by univariate random effect for repeated measures within individuals. analysis (rϭϪ0.34, PϽ0.0001). Results from multivariable linear regression analysis modeling of all of these features Results (adjusted as above) using all study subjects are presented in The characteristics of the study population are reported in Table 3. When only the defined MS variables (model 1) are Table 1. Notable differences between the white and black considered, waist and systolic blood pressure were indepen- subjects studied included higher HDL cholesterol levels and dently related to EDV. Adding insulin resistance to the model lower triglyceride levels among black subjects. These differ- (model 2) removed the contribution of waist circumference, ences were statistically significant but not large on the suggesting that waist circumference was serving as a surrogate clinical scale. Modest differences in body mass index and systolic blood pressure did not reach the significance thresh- old. Of note with regard to the present analysis, the 2 ethnic groups did not differ in measures of insulin, glucose, or HOMA-IR. The prevalence of the MS was similar in both ethnic groups (whites, 6 of 42, 14.3% [95% CI, 9.6 to 19.0]; blacks, 7 of 55, 12.7% [95% CI, 11.2 to 14.2]; PϭNS; Table 2). There were also no ethnic differences in the prevalence of each MS component (Table 2). EDV was impaired in black compared with white subjects (Figure; Pϭ0.03 by repeated-measures ANOVA including all subjects, Pϭ0.005 comparing only subjects without the MS). Both groups demonstrated significantly worse endothelial function when the MS was present (maximal increase in LBF: blacks: 107Ϯ9% MS absent, 53Ϯ16% MS present; whites: 163Ϯ16% MS absent, 54Ϯ18% MS absent; Pϭ0.007, MS absent versus present; PϭNS for interaction of ethnicity and MS). Impaired EDV with MS. Box and whisker plots indicating median value with 25th and 75th percentiles at the box boundaries and These relationships were essentially unchanged when ad- 5th and 95th percentiles at the whisker boundaries. MCh indi- justed for age and basal LBF in multivariable linear regression cates methacholine chloride infused at 5, 10, or 15 ␮g/min. Lteif et al Metabolic Syndrome, Ethnicity, and the Endothelium 35

TABLE 3. Multivariate Modeling for the Principal Determinants of EDV

Model 1 (R 2 adj, 0.67) Model 2 (R 2 adj, 0.69) Model 3 (R 2 adj, 0.68)

␤ F Ratio P ␤ F Ratio P ␤ F Ratio P Waist, cm Ϫ2.25 6.38 0.01 Ϫ1.13 1.27 NS Ϫ0.89 0.81 NS SPB, mm Hg Ϫ2.10 5.97 0.02 Ϫ2.10 5.35 0.02 Ϫ1.95 4.71 0.03 DBP, mm Hg 0.89 0.55 NS 1.06 0.77 NS 1.14 0.91 NS HDL, mmol/L Ϫ1.94 0.01 NS 3.15 0.01 NS 20.3 0.36 NS TG, mmol/L 15.03 0.52 NS 4.12 0.04 NS Ϫ5.55 0.07 NS Glucose, mmol/L 0.15 0.00 NS 11.18 0.35 NS 8.34 0.20 NS LogHOMA-IR, logU ⅐⅐⅐ ⅐⅐⅐ ⅐⅐⅐ Ϫ32.7 5.37 0.02 Ϫ33.7 5.87 0.02 Ethnicity ⅐⅐⅐ ⅐⅐⅐ ⅐⅐⅐ ⅐⅐⅐ ⅐⅐⅐ ⅐⅐⅐ 24.5 4.62 0.04 adj indicates adjusted; SPB, systolic blood pressure; DBP, diastolic blood pressure; and TG, triglycerides. ␤ Coefficients reflect the adjusted effect of a unit increment on percent vasodilation. All analyses are adjusted for age and baseline LBF and use random-effects modeling to account for repeated measures. measure of the effect of insulin resistance on EDV. Expanding sample size. Therefore, a retrospective power calculation the modeling to include ethnicity (model 3) revealed an addi- using the apparent effect sizes and observed variability of tional, independent effect of ethnicity on EDV. An interaction these variables was performed. With the present data, we had effect for ethnicity by logHOMA-IR was significant (␤ϭ58.1, an Ϸ7% chance of showing an independent relationship Pϭ0.046), but the interaction of ethnicity by systolic blood between logHOMA-IR and EDV in this subgroup; con- pressure was not. Including gender as a covariate did not alter versely, if such a relationship existed, because of the minimal the results of these analyses. apparent effect size, we would have required a sample size of Table 4 presents the results of analyzing model 2 within the Ϸ1300 subjects. Among whites, the retrospective power ethnicity subgroups. The white subjects exhibited largely the calculation suggested that we had an Ϸ12% chance of same relationship as seen for all subjects combined (Table 2), showing an independent relationship of systolic blood pres- namely a dominant contribution of logHOMA-IR, but the sure with EDV and would have required a sample size of effect of systolic blood pressure was not significant in these Ϸ320 to distinguish the apparent effect size given the subjects. In contrast, among black subjects, the dominant observed variability. In both cases, therefore, these are not correlate of impaired EDV was systolic blood pressure, with questions of a borderline loss of power resulting from sample no important independent effect of logHOMA-IR. As with size issues. the modeling for the entire study population, waist circum- ference did not contribute importantly to the determination of Discussion impaired EDV in either subgroup when the concurrent effects We have reported 2 novel findings in this analysis of the of these other variables were considered. cross-sectional determinants of impaired EDV. First, in anal- In light of the reduced sample sizes in these subgroup ysis of the complete study population, the addition of the analyses, we considered whether the loss of significance, HOMA index of insulin resistance (a simple combination of particularly for insulin resistance among the black subjects, fasting blood glucose and insulin levels) removed the inde- was an artifact of reduced statistical power caused by reduced pendent contribution of waist circumference in the determi- nation of endothelial dysfunction, suggesting that the effects TABLE 4. Ethnic Differences in the Principal Determinants of the central obesity are mediated by insulin resistance. of EDV Second, ethnicity was found to be an important independent determinant of endothelial function, with worse endothelial Whites Blacks function among black than white subjects and a reduced (R 2 adj, 0.73; Pϭ0.004) (R 2 adj, 0.52; Pϭ0.04) effect of insulin resistance to further worsen endothelial ␤ F Ratio P ␤ F Ratio P function among black subjects. Waist, cm Ϫ1.44 0.72 NS 0.02 0.00 NS Ϫ Ϫ Insulin Resistance, MS, and Impaired SPB, mm Hg 1.87 1.48 NS 2.52 5.35 0.03 Endothelial Function DBP, mm Hg 2.15 0.83 NS 1.29 0.91 NS Insulin resistance is thought to underlie the MS,27 and HDL, mmol/L Ϫ33.42 0.23 NS 47.02 2.04 NS epidemiological analyses suggest that measures of insulin TG, mmol/L Ϫ18.78 0.32 NS 15.41 0.32 NS resistance are an integral component of assessing the pres- Glucose, mmol/L 19.60 0.36 NS Ϫ12.68 0.35 NS ence of the MS.9,21,28,29 It is therefore logical to assume that LogHOMA-IR, logU Ϫ50.89 5.49 0.03 8.82 0.23 NS adding an assessment of insulin resistance to the NCEP- Abbreviations as in Table 3. ␤ Coefficients reflect the adjusted effect of a defined MS might alter the apparent contributions of the unit increment on percent vasodilation. All analyses are adjusted for age and syndrome components to a given vascular outcome. Insulin baseline LBF and use random-effects modeling to account for repeated resistance and the components of the MS have previously measures. been reported to associate individually with measurements of 36 Circulation July 5, 2005

EDV. The present study used multivariable analysis to subjects.14,16 Also, black subjects exhibited comparable cor- consider the simultaneous, mutually adjusted impact of these onary blood flow responses to endothelium-dependent vaso- factors. Insulin resistance was found to exert an important dilators but had an augmented corrective response to coinfu- effect on endothelial function and appeared to account for the sion with L-arginine.15 However, indexes of arterial elasticity contribution of central obesity to EDV. Conversely, waist were not found to differ between blacks and whites in circumference appeared to serve as a surrogate measure of the subjects across the range of hypertension severity.40 Very few effect of insulin resistance on EDV. molecular studies have been undertaken to explore these A similar effect of using the HOMA index of insulin differences. In one report, differences in the populations of resistance with the NCEP-defined MS has been found in receptors for endothelin (a proatherosclerotic vascular hor- evaluations of coronary calcification scores as the vascular mone) across ethnicity were seen.41 Overall, the present outcome.6 In studies of cardiovascular event rates and sur- finding is consistent with studies demonstrating ethnic differ- vival, the effect of insulin resistance measures is less appar- ences in other vascular end points, but the reasons for these ent,19,30–32 and it is not clear that there will be an incremental differences remain obscure. effect of insulin resistance beyond assessment of other MS In a combined analysis of the Cardiovascular Health Study components as data on “hard” cardiovascular outcomes be- and the Atherosclerosis Risk in Communities study, the comes available.33 Endothelial function is related to cardio- impact of traditional cardiovascular risk factors on cerebro- vascular outcomes,34–36 but it is not necessarily true that the vascular and cardiovascular atherosclerosis was found to determinants of vascular status in the short term (ie, physio- differ by age and gender.42 The present findings suggest that logical function) will carry the same weights over the long the metabolic determinants of endothelial function differ term. These differences in the additional contribution of between black and white subjects. This is a novel finding for insulin resistance to the MS components in predicting a which there are few confirmatory data. In the Women’s variety of vascular outcomes might therefore reflect true Health Initiative, the association of obesity with cardiovas- differences in vascular biology. It is equally possible, how- cular disease in postmenopausal women was lost among ever, that differing time courses of the effects of each variable blacks, whereas the relationship between blood pressure and or other unidentified modifying factors are obscuring the CVD was stronger among black than white women.7 This relationship of insulin resistance to vascular outcomes over important question has implications for the impact of thera- longer time frames. peutic interventions and warrants further study. There is general acceptance that the measurement of short-term physiological responses of the vasculature repre- Study Limitations sents a useful, accessible window into vascular biology. The Our study population consisted of volunteer subjects rather past decade has seen considerable effort expended on the use than a true population-based sample. The factors leading to of such proximate or alternative end points in assessing the participation are presumably equal across subgroups of the effects of a variety of interventions. In this setting, the present study population, so comparisons within the main study result implies that measurement of insulin resistance should population are likely valid. It is less clear, however, that these be included in the metabolic assessment of cardiovascular subjects are clearly representative of the general population; risk associated with obesity and the MS. thus, some caution must be used when our results are generalized. Impaired Vascular Responses in Blacks The observations made between ethnic subgroups of our Ethnic differences in the prevalence of the MS and of its population necessarily involved reductions in sample size, constituent components have been described.37 In the raising concerns about loss of statistical power. However, as NHANES III data set, no ethnic differences in the patterns of detailed in the Results section, when the apparent relation- principal component factor analysis associated with the ships of insulin resistance and systolic blood pressure with NCEP-defined MS were seen,12 suggesting that the interre- EDV segregated into white and black subjects, respectively, lationship of the NCEP-defined factors is comparable across the reason was not borderline power but rather weak effect ethnicities. A similar analysis in the IRAS cohort found that sizes. We are therefore confident that this observation is both direct and surrogate measures of insulin resistance were robust within our data set. integral to the “metabolic” factor in principal component The various protocols that contributed baseline data to this analysis,9 but again, no differences across ethnicity were seen analysis did not include measurements of any of the host of in these associations. These studies suggest that the biology novel inflammatory cardiovascular risk factors currently be- governing the interrelationships of the factors comprising the ing investigated as markers of cardiovascular disease. These MS is similar across ethnicities. questions are clearly relevant and will be addressed prospec- Differences in vascular biology across ethnicities have tively in further studies in our laboratory. long been recognized. Blacks, for example, have been found Conclusions to have an increased risk of macrovascular events at compa- The NCEP-defined MS is strongly associated with the pres- rable levels of blood pressure and lipid levels.38 Subclinical ence of endothelial dysfunction in black and white subjects. atherosclerosis is increased among blacks after adjustment for By multivariate analysis, the association of waist circumfer- major cardiovascular disease risk factors and insulin sensi- ence with endothelial dysfunction appeared to depend prin- tivity.39 Consistent with our findings, normotensive blacks cipally on insulin resistance, and the dominant determinants have been found to have worse EDV than normotensive white of endothelial function across our whole study population Lteif et al Metabolic Syndrome, Ethnicity, and the Endothelium 37 were insulin resistance and systolic blood pressure. The 14. Rosenbaum DA, Pretorius M, Gainer JV, Byrne D, Murphey LJ, Painter relative importance of these components differed by ethnic CA, Vaughan DE, Brown NJ. Ethnicity affects vasodilation, but not endothelial tissue plasminogen activator release, in response to bra- subgroup, with a difference specifically in the worsening of dykinin. Arterioscler Thromb Vasc Biol. 2002;22:1023–1028. endothelial function as a result of insulin resistance. These 15. Houghton JL, Philbin EF, Strogatz DS, Torosoff MT, Fein SA, Kuhner findings suggest that there are ethnic differences in vascular PA, Smith VE, Carr AA. The presence of African American race predicts biology as it relates to EDV. This possibility warrants further improvement in coronary endothelial function after supplementary L-arginine. J Am Coll Cardiol. 2002;39:1314–1322. study in specific ethnic subgroups because it may imply that 16. Gainer JV, Stein CM, Neal T, Vaughan DE, Brown NJ. Interactive effect different approaches to therapy are needed in black and white of ethnicity and ACE insertion/deletion polymorphism on vascular reac- patients. tivity. Hypertension. 2001;37:46–51. 17. Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, Turner RC. Homeostasis model assessment: insulin resistance and B cell function Acknowledgments from fasting plasma glucose and insulin concentrations in man. Diabe- This project was supported by the NIH (NIDDK 42469) and by a tologia. 1985;28:412–419. Junior Faculty Award from the American Diabetes Association to Dr 18. Mather KJ, Hunt AE, Steinberg HO, Paradisi G, Hook G, Katz A, Quon Mather. Dr Mather was also supported by the Sandra A. Daugherty MJ, Baron AD. Repeatability characteristics of simple indices of insulin Foundation. The Indiana University General Clinical Research resistance: implications for research applications. J Clin Endocrinol Center is supported by the NIH (M01 RR00750). We thank Dr Alain Metab. 2001;86:5457–5464. 19. Hanley AJ, Williams K, Stern MP, Haffner SM. Homeostasis model Baron for his guidance and mentorship on this and many other assessment of insulin resistance in relation to the incidence of cardiovas- projects. The expert nursing and technical support provided by Paula cular disease: the San Antonio Heart Study. Diabetes Care. 2002;25: Robinett, Donna Ford, Jan VanderLuitgaren, Chris Leffler, Kerry 1177–1184. Beverly, and Patti Merril is gratefully acknowledged. 20. Han TS, Sattar N, Williams K, Gonzalez-Villalpando C, Lean ME, Haffner SM. Prospective study of C-reactive protein in relation to the References development of diabetes and metabolic syndrome in the Mexico City 1. Steinberg HO, Chaker H, Leaming R, Johnson A, Brechtel G, Baron AD. Diabetes Study. Diabetes Care. 2002;25:2016–2021. Obesity/insulin resistance is associated with endothelial dysfunction: 21. Bonora E, Kiechl S, Willeit J, Oberhollenzer F, Egger G, Bonadonna RC, implications for the syndrome of insulin resistance. J Clin Invest. 1996; Muggeo M. Metabolic syndrome: epidemiology and more extensive phe- 97:2601–2610. notypic description: cross-sectional data from the Bruneck Study. Int J 2. Lavrencic A, Salobir BG, Keber I. Physical training improves flow- Obes Relat Metab Disord. 2003;27:1283–1289. mediated dilation in patients with the polymetabolic syndrome. Arte- 22. Resnick HE, Jones K, Ruotolo G, Jain AK, Henderson J, Lu W, Howard rioscler Thromb Vasc Biol. 2000;20:551–555. BV. Insulin resistance, the metabolic syndrome, and risk of incident 3. Petrie JR, Ueda S, Webb DJ, Elliott HL, Connell JM. Endothelial nitric cardiovascular disease in nondiabetic American Indians: the Strong Heart oxide production and insulin sensitivity: a physiological link with impli- Study. Diabetes Care. 2003;26:861–867. cations for pathogenesis of cardiovascular disease. Circulation. 1996;93: 23. Executive Summary of the Third Report of The National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and 1331–1333. Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel 4. Ferrannini E, Haffner SM, Mitchell BD, Stern MP. Hyperinsulinaemia: III). JAMA. 2001;285:2486–2497. the key feature of a cardiovascular and metabolic syndrome. Diabe- 24. Baron AD, Tarshoby M, Hook G, Lazaridis EN, Cronin J, Johnson A, tologia. 1991;34:416–422. Steinberg HO. Interaction between insulin sensitivity and muscle per- 5. Cheal KL, Abbasi F, Lamendola C, McLaughlin T, Reaven GM, Ford ES. fusion on glucose uptake in human skeletal muscle: evidence for capillary Relationship to insulin resistance of the Adult Treatment Panel III diag- recruitment. Diabetes. 2000;49:768–774. nostic criteria for identification of the metabolic syndrome. Diabetes. 25. Steinberg HO, Brechtel G, Johnson A, Fineberg N, Baron AD. Insulin- 2004;53:1195–1200. mediated skeletal muscle vasodilation is nitric oxide dependent: a novel 6. Reilly MP, Wolfe ML, Rhodes T, Girman C, Mehta N, Rader DJ. action of insulin to increase nitric oxide release. J Clin Invest. 1994;94: Measures of insulin resistance add incremental value to the clinical 1172–1179. diagnosis of metabolic syndrome in association with coronary atheroscle- 26. Mather KJ, Lteif A, Steinberg HO, Baron AD. Interactions between rosis. Circulation. 2004;110:803–809. endothelin and nitric oxide in the regulation of vascular tone in obesity 7. Howard BV, Criqui MH, Curb JD, Rodabough R, Safford MM, Santoro and diabetes. Diabetes. 2004;53:2060–2066. N, Wilson AC, Wylie-Rosett J. Risk factor clustering in the insulin 27. DeFronzo R, Ferrannini E. Insulin resistance: a multifaceted syndrome resistance syndrome and its relationship to cardiovascular disease in responsible for NIDDM, obesity, hypertension, dylipidemia, and athero- postmenopausal white, black, Hispanic, and Asian/Pacific Islander sclerotic cardiovascular disease. Diabetes Care. 1991;14:173–194. women. Metabolism. 2003;52:362–371. 28. Meigs JB, Wilson PW, Nathan DM, D’Agostino RB Sr, Williams K, 8. Festa A, D’Agostino R Jr, Howard G, Mykkanen L, Tracy RP, Haffner Haffner SM. Prevalence and characteristics of the metabolic syndrome in SM. Chronic subclinical inflammation as part of the insulin resistance the San Antonio Heart and Framingham Offspring Studies. Diabetes. syndrome: the Insulin Resistance Atherosclerosis Study (IRAS). Circu- 2003;52:2160–2167. lation. 2000;102:42–47. 29. Shen BJ, Todaro JF, Niaura R, McCaffery JM, Zhang J, Spiro A 3rd, 9. Hanley AJ, Karter AJ, Festa A, D’Agostino R Jr, Wagenknecht LE, Ward KD. Are metabolic risk factors one unified syndrome? Modeling Savage P, Tracy RP, Saad MF, Haffner S. Factor analysis of metabolic the structure of the metabolic syndrome X. Am J Epidemiol. 2003;157: syndrome using directly measured insulin sensitivity: the Insulin 701–711. Resistance Atherosclerosis Study. Diabetes. 2002;51:2642–2647. 30. Ruige JB, Assendelft WJJ, Dekker JM, Kostense PJ, Heine RJ, Bouter 10. Wagenknecht LE, Zaccaro D, Espeland MA, Karter AJ, O’Leary DH, LM. Insulin and risk of cardiovascular disease: a meta-analysis. Circu- Haffner SM. Diabetes and progression of carotid atherosclerosis: the lation. 1998;97:996–1001. Insulin Resistance Atherosclerosis Study. Arterioscler Thromb Vasc Biol. 31. Pyorala M, Miettinen H, Halonen P, Laakso M, Pyorala K. Insulin 2003;23:1035–1041. resistance syndrome predicts the risk of coronary heart disease and stroke 11. Okosun IS, Liao Y, Rotimi CN, Prewitt TE, Cooper RS. Abdominal in healthy middle-aged men: the 22-year follow-up results of the Helsinki adiposity and clustering of multiple metabolic syndrome in white, black Policemen Study. Arterioscler Thromb Vasc Biol. 2000;20:538–544. and Hispanic Americans. Ann Epidemiol. 2000;10:263–270. 32. Lakka HM, Laaksonen DE, Lakka TA, Niskanen LK, Kumpusalo E, 12. Ford ES. Factor analysis and defining the metabolic syndrome. Ethn Dis. Tuomilehto J, Salonen JT. The metabolic syndrome and total and cardio- 2003;13:429–437. vascular disease mortality in middle-aged men. JAMA. 2002;288: 13. Liese AD, Mayer-Davis EJ, Tyroler HA, Davis CE, Keil U, Duncan BB, 2709–2716. Heiss G. Development of the multiple metabolic syndrome in the ARIC 33. Isomaa B, Almgren P, Tuomi T, Forsen B, Lahti K, Nissen M, Taskinen cohort: joint contribution of insulin, BMI, and WHR: Atherosclerosis MR, Groop L. Cardiovascular morbidity and mortality associated with the Risk in Communities. Ann Epidemiol. 1997;7:407–416. metabolic syndrome. Diabetes Care. 2001;24:683–689. 38 Circulation July 5, 2005

34. Halcox JP, Schenke WH, Zalos G, Mincemoyer R, Prasad A, Waclawiw outcomes among black and white women with heart disease. Circulation. MA, Nour KR, Quyyumi AA. Prognostic value of coronary vascular 2003;108:1089–1094. endothelial dysfunction. Circulation. 2002;106:653–658. 39. D’Agostino RB Jr, Burke G, O’Leary D, Rewers M, Selby J, Savage PJ, 35. Neunteufl T, Heher S, Katzenschlager R, Wolfl G, Kostner K, Maurer G, Saad MF, Bergman RN, Howard G, Wagenknecht L, Haffner SM. Ethnic Weidinger F. Late prognostic value of flow-mediated dilation in the differences in carotid wall thickness: the Insulin Resistance Atheroscle- brachial artery of patients with chest pain. Am J Cardiol. 2000;86: rosis Study. Stroke. 1996;27:1744–1749. 207–210. 40. Prisant LM, Resnick LM, Hollenberg SM, Jupin D. Arterial elasticity 36. Schachinger V, Britten MB, Zeiher AM. Prognostic impact of coronary among normotensive subjects and treated and untreated hypertensive vasodilator dysfunction on adverse long-term outcome of coronary heart subjects: influence of race. Ethn Dis. 2002;12:63–68. disease. Circulation. 2000;101:1899–1906. 41. Ergul A, Tackett RL, Puett D. Distribution of endothelin receptors in 37. Park YW, Zhu S, Palaniappan L, Heshka S, Carnethon MR, Heymsfield saphenous veins of African Americans: implications of racial differences. SB. The metabolic syndrome: prevalence and associated risk factor J Cardiovasc Pharmacol. 1999;34:327–332. findings in the US population from the Third National Health and 42. Howard G, Manolio TA, Burke GL, Wolfson SK, O’Leary DH. Does the Nutrition Examination Survey, 1988–1994. Arch Intern Med. 2003;163: association of risk factors and atherosclerosis change with age? An 427–436. analysis of the combined ARIC and CHS cohorts: the Atherosclerosis 38. Jha AK, Varosy PD, Kanaya AM, Hunninghake DB, Hlatky MA, Waters Risk in Communities (ARIC) and Cardiovascular Health Study (CHS) DD, Furberg CD, Shlipak MG. Differences in medical care and disease Investigators. Stroke. 1997;28:1693–1701. Health Services and Outcomes Research

Adoption of Spironolactone Therapy for Older Patients With Heart Failure and Left Ventricular Systolic Dysfunction in the United States, 1998–2001

Frederick A. Masoudi, MD, MSPH; Cary P. Gross, MD; Yongfei Wang, MS; Saif S. Rathore, MPH; Edward P. Havranek, MD; JoAnne Micale Foody, MD; Harlan M. Krumholz, MD, SM

Background—Concerns have been raised about the appropriateness of spironolactone use in some patients with heart failure. We studied the adoption of spironolactone therapy after publication of the Randomized Aldactone Evaluation Study (RALES) in national cohorts of older patients hospitalized for heart failure. Methods and Results—This is a study of serial cross-sectional samples of Medicare beneficiaries Ն65 years old discharged after hospitalization for the primary diagnosis of heart failure and with left ventricular systolic dysfunction. The first sample was discharged before (April 1998 to March 1999, nϭ9758) and the second sample after (July 2000 to June 2001, nϭ9468) publication of RALES in September 1999. We assessed spironolactone prescriptions at hospital discharge in patient groups defined by enrollment criteria for the trial. Using multivariable logistic regression, we identified factors independently associated with prescriptions not meeting these criteria. Spironolactone use increased Ͼ7-fold (3.0% to 21.3% PϽ0.0001) after RALES. Of patients meeting enrollment criteria, 24.1% received spironolactone, as compared with 17.4% of those not meeting the criteria. Of all prescriptions after RALES, 30.9% were provided to patients not meeting enrollment criteria. Spironolactone was prescribed to 22.8% of patients with a serum potassium value Ն5.0 mmol/L, to 14.1% with a serum creatinine value Ն2.5 mg/dL, and to 17.3% with severe renal dysfunction (estimated glomerular filtration rate Ͻ30 mL · minϪ1 · 1.73 mϪ2). In multivariable analyses, factors associated with prescriptions not meeting enrollment criteria included advanced age, noncardiovascular comorbidities, discharge to skilled nursing facilities, and care provided by physicians without board certification. Conclusions—Spironolactone prescriptions increased markedly after the publication of RALES, and many treated patients were at risk for hyperkalemia. Simultaneously, many patients who might have benefited were not treated. These findings demonstrate the importance of balancing efforts to enhance use among appropriate patients and minimizing use in patients at risk for adverse events. (Circulation. 2005;112:39-47.) Key Words: heart failure Ⅲ aging Ⅲ aldosterone antagonists Ⅲ potassium

lthough the Randomized Aldactone Evaluation Study application of the trial results in case reports and series from A(RALES) demonstrated the benefits of aldosterone single centers.2–9 More recently, a study from Canada re- blockade with spironolactone in selected patients with se- ported simultaneous increases in spironolactone prescriptions verely symptomatic left ventricular systolic dysfunction and hospitalizations for hyperkalemia after the publication of (LVSD), this therapy has not been proven beneficial for all RALES.10 Although that study could not determine whether heart failure patients.1 Spironolactone may cause serious that observation resulted from spironolactone use in patients hyperkalemia, particularly in patients with marginal renal at high risk for hyperkalemia in clinical practice, it suggests function, type IV renal tubular acidosis, or relatively high the possibility of inappropriate use or inadequate monitoring baseline potassium levels. Because of the potential risks of of spironolactone in the community. Because aldosterone spironolactone therapy, concerns have been raised about the blockade may improve survival in some patients with heart

Received December 4, 2004; revision received March 1, 2005; accepted March 28, 2005. From the Division of Cardiology, Department of Medicine, Denver Health Medical Center (F.A.M., E.P.H.); and the Division of Cardiology, Department of Medicine (F.A.M., E.P.H.), and the Division of Geriatric Medicine, Department of Medicine (F.A.M.), University of Colorado Health Sciences Center, Denver, Colo; the Colorado Foundation for Medical Care (F.A.M., E.P.H., H.M.K.), Aurora, Colo; the Colorado Health Outcomes Program (F.A.M.), Aurora, Colo; the Section of General Internal Medicine, Department of Medicine (C.P.G.); the Section of Cardiovascular Medicine, Department of Medicine (S.S.R., Y.W., J.M.F., H.M.K.), and the Section of Health Policy and Administration, Department of Epidemiology and Public Health (H.M.K.), Yale University School of Medicine, New Haven, Conn; the Center for Outcomes Research and Evaluation (H.M.K., C.G.), Yale–New Haven Hospital, New Haven, Conn; and the Section of Cardiology, Department of Medicine, West Haven Veteran’s Administration Medical Center (J.M.F.), West Haven, Conn. Guest Editor for this article was Clyde W. Yancy, MD. Correspondence to Frederick A. Masoudi, MD, MSPH, Division of Cardiology, MC 0960 Denver Health Medical Center, 777 Bannock St, Denver, CO 80204. E-mail [email protected] © 2005 American Heart Association, Inc. Circulation is available at http://www.circulationaha.org DOI: 10.1161/CIRCULATIONAHA.104.527549 39 40 Circulation July 5, 2005

TABLE 1. Study Populations Before (1998–1999) and After TABLE 1. Continued (2000–2001) Publication of RALES Results

1998–1999 2000–2001 1998–1999 2000–2001 (nϭ9758) (nϭ9468) P (nϭ9758) (nϭ9468) P Meeting RALES enrollment criteria* 61.5 58.3 0.02 Other 2.5 3.7 Demographics Discharge medications Age, y 0.01 ACE inhibitor or ARB 64.2 60.9 Ͻ0.001 65–74 35.3 33.3 ARB 9.9 11.0 NS 75–84 45.3 45.0 Diuretic† 89.6 88.4 0.03 Ն85 19.4 21.6 ␤-Blocker 21.7 43.4 Ͻ0.001 Female 47.4 45.4 0.02 Potassium supplement 52.9 44.5 Ͻ0.001 Race NS Physician characteristics White 83.7 84.0 Specialty Ͻ0.001 Black 12.7 13.3 Cardiologist 29.4 27.6 Other 3.5 2.7 Internist with consult 22.0 25.7 Medical history Internist without consult 16.6 15.3 Hypertension 60.7 66.0 Ͻ0.001 Other with consult 19.9 21.1 Coronary artery disease 67.5 70.3 Ͻ0.001 Other without consult 12.1 10.3 Myocardial infarction 39.1 41.0 0.05 Physician board certification 71.8 75.5 Ͻ0.001 Coronary revascularization 35.8 39.8 Ͻ0.001 Hospital characteristics Cerebrovascular accident 16.5 19.1 Ͻ0.001 Teaching status 0.01 Chronic pulmonary disease 31.1 32.5 NS COTH member 16.4 17.0 Diabetes 40.4 41.9 NS Residency affiliated 24.1 21.1 Dementia 7.3 8.0 NS Nonteaching 59.5 61.9 Admission characteristics Cardiac care facilities Ͻ0.001 Systolic blood pressure NS Cardiac surgery capability 42.9 48.6 Ͻ100 6.0 5.4 Cardiac catheterization lab 21.0 21.6 100–140 45.0 46.9 No invasive facilities 36.1 29.8 Ͼ140 49.0 47.7 Ownership NS Spironolactone on admission 1.6 10.5 Ͻ0.001 Public 12.1 11.0 Laboratory/imaging data Not for profit 75.6 77.2 Atrial fibrillation 27.9 28.2 NS For profit 12.4 11.8 Pulmonary edema on chest radiograph 77.4 75.8 0.04 Census division 0.03 Potassium, mmol/dL NS Northeast 20.0 20.3 Ͻ3.5 5.7 5.1 Midwest 25.2 25.1 3.5–5.0 89.3 89.3 South 40.4 40.2 Ͼ5.0 5.1 5.7 West 14.4 14.4 Creatinine, mg/dL NS ARB indicates angiotensin receptor blocker; COTH, Council of Teaching Ͻ1.5 59.0 57.5 Hospitals. *RALES enrollment criteria applied included discharge prescription for ACE 1.5–2.4 32.7 33.1 inhibitor/ARB and diuretic, discharge serum creatinine Ͻ2.5 mg/dL, and Ͼ2.4 8.3 9.5 discharge serum potassium Ͻ5.0 mmol/L. Estimated GFR, mL ⅐ minϪ1 ⅐ 1.73 mϪ2 NS †Not including potassium-sparing diuretics. Ͼ89 5.7 4.9 60–89 26.6 25.7 failure but pose a serious risk in others, the simultaneous assessment of appropriate and potentially inappropriate use of 30–59 52.5 53.0 spironolactone is important. Ͻ30 15.1 16.4 Current clinical guidelines acknowledge this difficult bal- Ͻ Hematocrit 30% 8.2 8.2 NS ance, recommending that spironolactone be restricted to Severe LVSD 61.5 61.3 NS patients similar to those enrolled in RALES.11 Specifically, Discharge setting 0.001 the guidelines recommend spironolactone for patients with Home without home health care 56.0 56.5 LVSD; recent or current symptoms at rest despite the use of Home with home health care 23.4 22.1 digoxin, diuretics, an angiotensin-converting enzyme (ACE) ␤ Institutional setting 18.0 17.6 inhibitors, and (usually) a -blocker; a serum potassium value Ͻ5.0 mmol/L; and creatinine value Ͻ2.5 mg/dL before Masoudi et al Spironolactone Prescription in Heart Failure 41 initiation of treatment.11 Because many community-based (Pfizer), Aldactazide (Pfizer), hydrochlorothiazide/spironolactone, patients do not fit these criteria,12 widespread use of spirono- and spironolactone. lactone without consideration of these factors could expose Spironolactone treatment was also assessed in subgroups defined by the primary selection criteria used for RALES.1 Patients meeting many patients to adverse outcomes. all of the following criteria were designated as conforming to these The degree to which spironolactone is used for patients criteria: (1) discharge prescription for an ACE inhibitor or angioten- who might derive a benefit and is avoided among those at sin receptor blocker or a documented reason for nonuse of ACE high risk for adverse events is not known. To address this inhibitors; (2) discharge prescription for a diuretic (not including spironolactone); (3) a serum creatinine value closest to hospital issue, we assessed national patterns of spironolactone use discharge Ͻ2.5 mg/dL; and (4) a serum potassium value closest to before and after publication of the RALES results in Septem- hospital discharge of Ͻ5.0 mmol/L. All other patients were desig- ber 1999 in a community-based, nationally representative nated as not meeting RALES enrollment criteria. Because the serum sample of patients. This study was intended to assess the rate creatinine value may overestimate renal function, particularly in of adoption of spironolactone in the United States and the elderly patients, an estimated glomerular filtration rate (GFR) was also calculated for all patients according to the abbreviated Modifi- degree to which the selection of patients for treatment after cation of Diet in Renal Disease Study equation.15 publication of RALES conformed to study enrollment criteria. Additional Data Sources NHC data were linked with the American Medical Association Methods Physician Masterfile16,17 by using the unique physician identification number of the attending physician, defined as the clinician primarily National Heart Care Project responsible for the patient’s care during hospitalization.18 Hospital The National Heart Care (NHC) Project is an ongoing initiative characteristics were ascertained by linking with the 1998 and 2000 funded by the Centers for Medicare and Medicaid Services designed American Hospital Association Annual Surveys.19,20 to improve the quality of care for Medicare beneficiaries with heart failure.13 Fee-for-service Medicare beneficiaries hospitalized with a Statistical Analysis principal discharge diagnosis of heart failure (International Classifi- Patient, physician, and hospital characteristics, as well as documen- cation of Diseases, Ninth Revision, Clinical Modification codes tation of spironolactone prescription at discharge, were compared 402.01, 402.11, 402.91, 404.01, 404.91, or 428) between April 1998 between the 1998 to 1999 and 2000 to 2001 periods by ␹2 statistics. and March 1999 or July 2000 and June 2001, inclusive, were Correlates of spironolactone prescription were determined by char- identified. These periods straddled the publication of RALES. acteristics of the patients (demographics, medical history, comor- According to a sampling strategy described previously,13,14 dis- bidities, and admission variables), treating physician (board certifi- charges were grouped by state and sorted by age, sex, race, and cation, cardiology specialty), and hospital (teaching status, cardiac treating hospital, and up to 800 discharges were randomly selected care facilities, urban or rural location, ownership, and US census from each state. The selected records underwent detailed review by division). Differences in proportions were evaluated by ␹2 statistics. trained data abstractors in central data abstraction centers. Patients To assess changes in use between 1998 to 1999 and 2000 to 2001 with invalid social security numbers, those receiving long-term after adjustment for differences in patient, physician, and hospital hemodialysis, those transferred to another hospital, or those who left characteristics, logistic regression was performed, with spironolac- against medical advice based on data from the administrative records tone prescription at discharge as the dependent variable. Models or the chart abstraction were excluded. The NHC sample thus comprising statistically significant covariates (PϽ0.05) were con- consisted of 78 882 records, of which 39 477 were from 1998 to structed by backward selection. Subsequently, a variable indicating 1999 and 39 405 from 2000 to 2001. the sampling time frame was introduced into the model. The statistical significance of 2-way interactions between the time frame Study Sample variable and other covariates included in the final model was tested Because this was a study of the prescription of spironolactone at to identify heterogeneity in the rate of increase in spironolactone use. discharge, patients who died during hospitalization were excluded Finally, to identify characteristics associated with spironolactone (nϭ4126). Because RALES enrolled only patients with LVSD, we prescription that did not conform to RALES enrollment criteria, a further restricted the analysis to patients with an LV ejection fraction model in only those patients who were treated with spironolactone in Ͻ40% or a qualitative description of moderate or severe systolic 2000 to 2001 was constructed, with failure to conform to all RALES dysfunction, based on the evaluation closest to the discharge date enrollment criteria as the dependent variable. Because the character- (nϭ22 928). From this group, we excluded patients younger than 65 istics composing the criteria defined the dependent variable in this years because younger Medicare beneficiaries are not representative model, these variables were not candidate independent variables. All of the younger population as a whole. If a patient appeared more than analyses used probability weights to account for the NHC sampling once in the sample, 1 record was included by random selection. strategy, thus providing national level estimates. Statistical analyses Patients who were transferred to another acute care hospital, left were conducted with Stata 7.0 (Stata Corp) and SAS version 8.0 against medical advice, were receiving long-term renal dialysis, or (SAS Institute, Inc). with missing vital status were also excluded. A total of 3702 records of patients with LVSD were excluded by these selection criteria, Results resulting in a final study cohort of 19 226 patients (9758 in 1998 to Patient Population 1999 and 9468 in 2000 to 2001). In both samples, Ϸ20% were at least 85 years old, the Clinical Variables majority were male, and most had a history of hypertension The NHC database includes a wide range of variables, including and coronary artery disease (Table 1). Noncardiovascular demographic characteristics, medical history, characteristics on pre- comorbidities, including diabetes and chronic pulmonary sentation, diagnostic testing, laboratory data, and a comprehensive disease, were also common. An elevated potassium level at list of the names of medications prescribed at hospital discharge. For hospital discharge (Ͼ5.0 mmol/L) was present in Ͼ5% of the purposes of interpretability, drugs were assigned to classes. Ͼ Combination drugs were assigned to all classes contained in the patients. Although the prevalence of serum creatinine 2.5 combination. The following drugs were assessed to determine mg/dL was 8.3% in 1998 to 1999 and 9.5% in 2000 to 2001, spironolactone prescriptions (in alphabetical order): Aldactone the prevalence of a severely reduced estimated GFR (ie, Ͻ30 42 Circulation July 5, 2005

TABLE 2. Rates of Spironolactone Prescription Before TABLE 2. Continued (1998–1999) and After (2000–2001) Publication of RALES Results by Patient and Provider Characteristics

1998– 2000– Absolute 1998– 2000– Absolute 1999, 2001, Change, 1999, 2001, Change, % % %* P† % % %* P† Total 3.0 21.3 18.3 ⅐⅐⅐ Spironolactone on admission Ͻ0.001 Meeting RALES enrollment criteria‡ Ͻ0.001 No 1.9 15.3 13.3 No 3.2 17.4 14.2 Yes 65.1 72.5 7.4 Yes 2.9 24.1 21.2 Laboratory/imaging data Demographics Atrial fibrillation Ͻ0.001 Age, y 0.001 No 2.6 21.5 18.8 65–74 4.2 23.4 19.2 Yes 3.9 20.8 16.9 75–84 2.8 21.7 19.0 Pulmonary edema on chest radiograph 0.05 Ͼ84 1.3 17.0 15.7 No 4.0 23.0 19.0 Sex NS Yes 2.7 20.7 18.0 Female 2.6 21.0 18.4 Potassium, mmol/dL NS Male 3.3 21.5 18.2 Ͻ3.5 3.9 17.9 14.0 Race NS 3.5–5.0 2.9 21.4 18.4 White 3.0 21.5 18.6 Ͼ5.0 2.7 22.8 20.1 Black 2.7 20.8 18.1 Creatinine, mg/dL NS Other 3.8 15.1 11.3 Ͻ1.5 2.7 21.5 18.8 Medical history 1.5–2.4 3.8 23.0 19.2 Hypertension NS Ͼ2.4 1.8 14.1 12.3 Ϫ Ϫ No 3.5 23.6 20.1 Estimated GFR, mL ⅐ min 1 ⅐ 1.73 m 2 NS Yes 2.6 20.1 17.5 Ͼ89 2.8 17.9 15.1 Coronary artery disease 0.01 60–89 2.4 20.2 17.7 No 2.8 19.7 16.9 30–59 3.5 23.4 19.8 Yes 3.4 24.1 20.7 Ͻ30 2.2 17.3 15.1 Myocardial infarction 0.004 Hematocrit NS No 3.3 20.7 17.5 Ͻ30% 2.5 16.8 14.3 Yes 2.5 22.0 19.5 Ն30% 3.0 21.7 18.7 Coronary revascularization NS Severity of LVSD NS No 2.5 19.6 17.1 Moderate 2.4 18.4 16.1 Yes 3.8 23.7 19.9 Severe 3.4 23.1 19.7 Cerebrovascular accident NS Discharge characteristics No 3.0 21.4 18.4 Discharge setting NS Yes 3.0 21.0 18.0 Home without home health care 2.9 21.4 18.6 Chronic pulmonary disease NS Home with home health care 4.0 24.5 20.5 No 3.0 21.5 18.5 Institutional setting 1.8 17.7 15.9 Yes 3.0 20.9 17.8 Other 4.0 16.2 12.2 Ͻ Diabetes NS Discharge ACE-inhibitor prescription 0.001 No 2.7 21.1 18.4 No 4.1 17.2 13.1 Yes 3.4 21.5 18.1 Yes 2.4 23.8 21.4 Ͻ Dementia NS Discharge ARB prescription 0.001 No 3.1 21.9 18.8 No 2.9 20.6 17.7 Yes 1.4 14.0 12.6 Yes 4.4 26.5 22.1 Discharge ␤-blocker prescription NS Admission characteristics No 2.8 19.1 16.2 Systolic blood pressure Ͻ0.001 Yes 3.5 24.2 20.7 Ͻ100 4.5 27.5 23.0 100–140 4.2 23.8 19.5 Ͼ140 1.6 18.1 16.5 Masoudi et al Spironolactone Prescription in Heart Failure 43

TABLE 2. Continued Changes in Spironolactone Prescription Overall, spironolactone prescription rates in the cohort increased from 3.0% in 1998 to 1999 to 21.3% in 2000 to Ͻ 1998– 2000– Absolute 2001 (P 0.001, Table 2). Although the increase in pre- 1999, 2001, Change, scriptions between the 2 time periods was significant in all % % %* P† patient subgroups, the change was significantly greater in Discharge potassium supplement NS younger patients (P for interaction, 0.001); those with a Ͻ No 3.7 23.8 20.1 lower admission blood pressure (P 0.001); those with angina (PϽ0.001); and those not admitted who were Yes 2.3 17.1 15.8 already being treated with spironolactone (PϽ0.001). Physician characteristics Rates of increase did not differ significantly, however, by Specialty NS levels of discharge serum creatinine, estimated GFR, or Cardiologist 3.9 24.8 20.9 discharge potassium supplementation. Internist with consult 3.1 21.2 18.1 In the population of patients discharged in 2000 to 2001, Internist without consult 2.6 17.5 14.9 spironolactone was prescribed for 22.8% of patients with a Ͼ Other with consult 2.6 21.1 18.5 serum potassium value 5.0 mmol/L, for 14.1% of patients with a serum creatinine value Ͼ2.5 mg/dL, and for 17.3% of Other without consult 1.7 17.9 16.2 patients with severe renal dysfunction, as defined by an Physician board certification NS estimated GFR (Ͻ30 mL · minϪ1 · 1.73 mϪ2; the Figure). No 2.5 19.5 17.1 Among patients not receiving an ACE inhibitor prescription, Yes 3.2 21.8 18.7 17.2% received spironolactone, and among those receiving a Hospital characteristics prescription for potassium supplements, 17.1% were dis- Teaching status NS charged with a spironolactone prescription. Among patients COTH member 2.9 23.9 21.0 meeting study enrollment criteria after RALES, spironolac- tone was prescribed to 24.1% (a 21.2% absolute increase); Residency affiliated 3.6 23.1 19.6 among those not meeting these criteria, 17.4% received a Nonteaching 2.6 19.8 17.2 discharge prescription for spironolactone (14.2% absolute Cardiac care facilities NS increase). Of all spironolactone prescriptions in 2000 to 2001, Cardiac surgery capability 3.4 23.1 19.7 30.9% were provided to patients not meeting enrollment Cardiac catheterization lab 2.5 18.2 15.7 criteria. No invasive facilities 2.5 20.3 17.8 Correlates of Spironolactone Prescription Not Ownership 0.005 Meeting RALES Criteria Public 2.9 19.6 16.7 Among those patients receiving a prescription for spironolac- Not for profit 2.8 22.2 19.4 tone at discharge, prescriptions not conforming to RALES For-profit 3.7 16.2 12.6 criteria were more common among older patients (compared Census division NS with an age 65- to 74-year referent; age 75 to 84 years odds Northeast 2.1 19.4 17.4 ratio [OR], 1.40; 95% confidence interval [CI], 1.14 to 1.72; Ͻ ϩ ϭ Midwest 3.4 23.4 20.0 P 0.001; age 85 OR, 1.28; 95% CI, 0.96 to 1.70; P 0.09, Table 3) and were less common among women (OR, 0.80; South 3.1 20.5 17.4 95% CI, 0.66 to 0.97; Pϭ0.02). Prior history of heart failure, West 3.2 22.4 19.2 atrial fibrillation, and higher admission systolic blood pres- Abbreviations are as defined in text and in the footnote to Table 1. sures were associated with a lower OR of prescription not *P value comparing prescription rates between 1998–1999 and 2000–01 conforming to criteria, whereas noncardiovascular comor- Ͻ 0.001 for all patient groups. bidities, including chronic lung disease and anemia, were †P value for interaction between characteristic and time in multivariate model associated with higher ORs. Patients discharged to a skilled Ͼ ‡RALES enrollment criteria applied included discharge prescription for ACE nursing facility had 2-fold higher odds of receiving spi- inhibitor/ARB and diuretic, discharge serum creatinine Ͻ2.5 mg/dL, and ronolactone prescriptions not meeting criteria (OR, 2.31; 95% discharge serum potassium Ͻ5.0 mmol/L. CI, 1.71 to 3.11; PϽ0.001). Patients cared for by nongener- alist physicians without cardiology consultation were more mL · minϪ1 · 1.73 mϪ2) was higher (15.2% and 16.4%, respec- likely to conform to the RALES criteria (OR, 0.57; 95% CI, tively). Approximately 42% of patients in 2000 to 2001 did 0.39 to 0.84; Pϭ0.005) compared with those cared for by a not meet the appropriateness criteria for spironolactone cardiologist (1.00 referent), as were those cared for by a treatment. board-certified physician (OR, 0.70; 95% CI, 0.54 to 0.90; From 1998 to 1999 and 2000 to 2001, there was a decline Pϭ0.005) compared with those cared for by a physician in the proportion of patients treated at hospital discharge with without board certification. ACE inhibitors or angiotensin receptor blockers, diuretics (not including spironolactone), and potassium replacements. Discussion The proportion treated with ␤-blockers increased signifi- In this study of 2 cross-sectional cohorts of older patients cantly between periods. with heart failure and LVSD, prescription of spironolactone 44 Circulation July 5, 2005

Changes in proportions of patients receiving prescription for spironolactone at hospital discharge between 1998 to 1999 (before RALES) and 2000 to 2001 (after RALES) in all patients and patients stratified by serum potassium, creatinine, and estimated GFR.

increased Ͼ7-fold after publication of RALES. Of the dis- ulation predominantly treated with ACE inhibitors, the trial charge prescriptions written after RALES, almost one third did not enroll patients with severe renal insufficiency (de- were provided to patients not fitting the study enrollment fined as a serum creatinine value Ͻ2.5 mg/dL) or baseline criteria, many of whom were at high risk for hyperkalemia. hyperkalemia (potassium Ͼ5.0 mmol/L).1 Patients in the trial Slow adoption among patients who might benefit from were also assessed frequently after the initiation of therapy spironolactone simultaneously with rapidly increasing use for the development of hyperkalemia and worsening renal among those at higher risk for adverse consequences indi- insufficiency, which may not be the case in usual clinical cates that the integration of clinical trials results in practice practice.4 may not maximize either effectiveness or safety. These Since the publication of RALES, concerns have been findings demonstrate the complexity of adoption of informa- raised about the adoption of spironolactone outside the tion from new clinical trials, emphasizing the need for faster context of a carefully controlled clinical trial.2 Within a year adoption among appropriate patients and simultaneous efforts of the release of RALES, a single-center case series of 25 to minimize use in patients who might suffer severe adverse patients treated with spironolactone were admitted with events from inappropriate use. serious hyperkalemia. These patients were predominantly RALES established the efficacy of spironolactone in re- older, and none had a serum potassium value exceeding ducing mortality and hospitalization in selected patients with 4.8 mmol/L before admission.25 Several subsequent case heart failure and LVSD.1 In our study, a minority of patients series have described patients with hyperkalemia requiring who had serum potassium and creatinine levels within the hospitalization, some of whom died.5–9 Many patients in these ranges of RALES subjects and who were receiving other series were elderly, in whom the serum creatinine value often evidence-based therapy for heart failure were not treated with overestimates true renal function. The results of our study spironolactone at hospital discharge. This finding is concor- raise significant concerns about the safety of current patterns dant with other data demonstrating that standards of clinical of spironolactone prescription in community-based popula- care lag behind the evidence generated in clinical trials. The tions, particularly among older patients with noncardiovascu- development of mechanisms to facilitate rapid and appropri- lar comorbidities. ate diffusion of trial results into patient care is likely to A population-based study from Ontario, Canada, found improve important patient outcomes. that the use of spironolactone increased from 3.4% before the RALES, however, excluded patients who were not receiv- publication of RALES to 14.9% afterward.10 Simultaneously, ing evidence-based heart failure therapy and those at poten- hospitalization rates for hyperkalemia increased Ͼ4-fold, and tially high risk for hyperkalemia. Several factors predispose associated mortality increased by Ͼ6-fold. The authors of some patients with heart failure to the risk of potentially that study speculated that hyperkalemia resulted in part from life-threatening hyperkalemia with an agent that impairs the inappropriate use of spironolactone but were unable to aldosterone action. Both diabetes, which is associated with demonstrate whether the patterns they observed were due to type IV renal tubular acidosis,21 and renal insufficiency are misapplication of the RALES results. This is the first study in common in populations with heart failure.13,22–24 The addition a nationally representative sample demonstrating widespread of an aldosterone antagonist to the regimen of patients with use of spironolactone in patients for whom there is no good underlying abnormalities of potassium excretion already be- evidence for benefit, many of whom are at high risk for ing treated with ACE inhibitors, angiotensin receptor block- hyperkalemia. ers, ␤-blockers, or potassium supplements may increase the The finding of limited adoption of spironolactone use in risk of hyperkalemia to varying degrees. Although RALES patients like those enrolled in RALES with simultaneous demonstrated significant mortality benefits in a patient pop- widespread use in populations of patients at risk for hyper- Masoudi et al Spironolactone Prescription in Heart Failure 45

TABLE 3. Factors Independently Associated With Spironolactone Prescription Not Meeting RALES Enrollment Criteria (2000–2001*†)

Prescriptions Not Meeting Enrollment Unadjusted OR Adjusted OR Characteristic Criteria (95% CI) (95% CI) P‡ Age, y Ͻ75 29.9% Referent Referent ⅐⅐⅐ 75–84 36.8% 1.36 (1.12–1.66) 1.40 (1.14–1.72) 0.002 Ն85 35.6% 1.29 (1.00–1.67) 1.28 (0.96–1.70) NS Sex Male 35.7% Referent Referent ⅐⅐⅐ Female 32.0% 0.85 (0.71–1.02) 0.80 (0.66–0.97) 0.02 Race White 35.3% Referent Referent ⅐⅐⅐ Black 24.4% 0.59 (0.44–0.79) 0.69 (0.51–0.93) 0.02 Other 45.6% 1.53 (0.83–2.86) 1.78 (0.94–3.38) NS Admission source Noninstitutional 34.0% Referent Referent ⅐⅐⅐ Nursing facilities 34.6% 1.03 (0.68–1.55) 0.55 (0.34–0.89) 0.01 Other 34.9% 1.04 (0.67–1.64) 0.87 (0.54–1.41) NS Heart failure history No 37.1% Referent Referent ⅐⅐⅐ Yes 33.3% 0.85 (0.68–1.06) 0.76 (0.60–0.95) 0.01 Chronic lung disease 0.76 0.006 No 32.4% Referent Referent ⅐⅐⅐ Yes 37.6% 1.26 (1.05–1.52) 1.32 (1.08–1.61) 0.006 Atrial fibrillation No 35.0% Referent Referent ⅐⅐⅐ Yes 31.6% 0.86 (0.70–1.05) 0.78 (0.63–0.97) 0.02 Admission systolic blood pressure Ͻ100 37.2% 1.00 (0.70–1.42) 0.99 (0.69–1.42) NS 100–140 37.2% Referent Referent ⅐⅐⅐ Ͼ140 29.5% 0.71 (0.59–0.85) 0.71 (0.58–0.86) Ͻ0.001 Hematocrit 2.30 0.01 Ͻ30% 50.4% Referent Referent ⅐⅐⅐ Ն30% 32.9% 0.48 (0.34–0.68) 0.43 (0.30–0.62) Ͻ0.001 Attending physician specialty Cardiologist 35.1% Referent Referent ⅐⅐⅐ Internal medicine with cardiology consult 33.3% 0.93 (0.73–1.17) 0.93 (0.72–1.19) NS Internal medicine without consult 35.6% 1.02 (0.77–1.37) 0.95 (0.70–1.30) NS Other physician type with consultation 35.0% 1.00 (0.78–1.27) 0.84 (0.63–1.13) NS Other physician type without consult 28.0% 0.72 (0.51–1.02) 0.57 (0.39–0.84) 0.005 Attending physician with board certification No 37.3% Referent Referent ⅐⅐⅐ Yes 33.1% 0.83 (0.68–1.03) 0.70 (0.54–0.90) 0.005 Hospital ownership Public 29.2% 0.81 (0.59–1.10) 0.89 (0.64–1.23) NS Not for profit 34.1% Referent Referent ⅐⅐⅐ For profit 41.0% 1.35 (1.00–1.84) 1.39 (1.01–1.91) 0.64 Discharge setting Home 31.5% Referent Referent ⅐⅐⅐ Home with home health 31.8% 1.01 (0.82–1.26) 1.02 (0.82–1.28) NS Skilled nursing facility 45.9% 1.84 (1.44–2.37) 2.31 (1.71–3.11) Ͻ0.001 Other 43.0% 1.64 (0.98–2.74) 1.55 (0.90–2.67) NS Abbreviations are as defined in text and in the footnote to Table 1. *RALES enrollment criteria defined as all of the following: (1) treatment with ACE inhibitor at discharge or documentation for not prescribing an ACE inhibitor; (2) treatment with a diuretic at discharge; (3) serum creatinine Ͻ2.5 mg/dL at discharge; and (4) serum potassium Ͻ5.0 mmol/L at discharge. 46 Circulation July 5, 2005 kalemia has important implications for quality improvement the use of spironolactone on outcomes, there is evidence that and patient safety initiatives in heart failure. Typically, efforts use of spironolactone in elderly patients without consider- to improve patterns of pharmacological therapy in patients ation of serum potassium and creatinine values can result in with heart failure have focused exclusively on either underuse adverse outcomes, including death.7–9 (eg, ACE inhibitors or ␤-blockers) or overuse (eg, type I In conclusion, the prescription of spironolactone in older antiarrhythmic agents or nonsteroidal antiinflammatory patients with LSVD hospitalized with heart failure increased drugs).26–28 Given the patterns of care shown in this study, substantially in the United States after publication of RALES. any efforts to increase the use of spironolactone in patients More than one third of prescriptions were provided to patients with heart failure should also include mechanisms to ensure who did not meet enrollment criteria of the trial, many of that this drug is not used in populations in whom the risks whom had characteristics that placed them at high risk for may outweigh the benefits. hyperkalemia. The means of ensuring use only among those The demonstration that spironolactone is frequently pre- patients with an acceptably low risk of adverse events will be scribed to patients at high risk in a national sample of older important for optimizing the benefits and risks of aldosterone patients with heart failure in conjunction with evidence of antagonists in the clinical care of populations of patients with higher population rates of hyperkalemia associated with heart failure. increasing spironolactone use10 suggest the need for action to change current patterns of use of aldosterone-blocking drugs. Acknowledgments Although several mechanisms may be useful in reducing the Dr Masoudi is supported by NIH/NIA Research Career Award potential harm to patients, it is possible that changing the K08-AG01011. Dr Gross is supported by a Cancer Prevention, Control and Population Sciences Career Development Award labeling of aldosterone-blocking agents to reflect these con- (1K07CA-90402), the Claude D. Pepper Older Americans Indepen- cerns (eg, more stringent precautions) and efforts by industry dence Center at Yale (P30AG21342), and Paul Beeson career and professional associations to inform practitioners of these development award in Aging (K08 AG24842). Dr Foody is sup- hazards would have a positive influence. With the more ported by NIH/NIA Research Career Award K08-AG20623 and recent publication of EPHESUS, a study of the selective NIA/Hartford Foundation Fellowship in Geriatrics. Saif Rathore is supported by NIH/National Institute of General Medical Sciences aldosterone blocker eplerenone in patients with heart failure Medical Scientist Training Grant GM07205. after myocardial infarction,29 the use of aldosterone blockade is likely to proliferate further, increasing the importance of Disclosure ensuring that these agents are used only in those populations Dr Masoudi has received honoraria from Pfizer; Dr Foody has for whom the benefit outweighs the risk. received honoraria from Pfizer. The content of this publication does Certain issues should be considered in the interpretation of not necessarily reflect the views or policies of the Department of these results. First, it was not possible to determine the Health and Human Services, nor does mention of trade names, commercial products, or organizations imply endorsement by the US relation between the timing of initiation of spironolactone and government. The analyses on which this publication is based were the measurement of laboratory values. Thus, in some cases, performed under contract No. 500-02-CO01, entitled “Utilization serum potassium and creatinine levels may have reflected in and Quality Control Peer Review Organization for the State (com- part the results of treatment with spironolactone. Patients with monwealth) of Colorado,” sponsored by the Centers for Medicare and Medicaid Services (formerly Health Care Financing Adminis- elevated potassium levels and marked renal dysfunction, tration), Department of Health and Human Services. The authors however, are at relatively higher risk for hyperkalemia if assume full responsibility for the accuracy and completeness of the spironolactone is continued. Second, we used a relatively ideas presented herein. This article is a direct result of the Health conservative definition for RALES enrollment criteria. The Care Quality Improvement Program initiated by the Centers for data did not include symptom severity by New York Heart Medicare and Medicaid Services, which has encouraged identifica- tion of quality improvement projects derived from analysis of Association (NYHA) classification or the frequency of sub- patterns of care, and therefore required no special funding on the part sequent clinical assessments. Because RALES studied only of this contractor. Ideas and contributions to the author concerning those patients with advanced symptoms (NYHA class III or experience in engaging with issues presented are welcomed. IV) and provided very close follow-up for hyperkalemia, it is possible that our study overestimates the proportion of safe References prescriptions. Third, the doses of spironolactone prescribed at 1. Pitt B, Zannad F, Remme WJ, Cody R, Castaigne A, Perez A, Palensky J, Wittes J. The effect of spironolactone on morbidity and mortality in discharge were not available in our dataset. Because evidence patients with severe heart failure: Randomized Aldactone Evaluation from case series indicates that the risk of hyperkalemia Study Investigators. N Engl J Med. 1999;341:709–717. increases with higher spironolactone dose,6,25 such data might 2. Geraci JM, Knowlton AA. Spironolactone for heart failure: spiraling out identify a larger number of patients at risk for adverse of control. Chest. 2000;118:1522–1523. 3. Berry C, McMurray J. Serious adverse events experienced by patients outcomes. Third, we were unable to assess changes in with chronic heart failure taking spironolactone. Heart. 2001;85:e8–e9. spironolactone use after hospital discharge. Because of the 4. Bozkurt B, Agoston I, Knowlton AA. Complications of inappropriate use rapid adoption of spironolactone during the time period under of spironolactone in heart failure: when an old medicine spirals out of new guidelines. J Am Coll Cardiol. 2003;41:211–214. study, prescriptions at the time of hospital discharge may 5. Svensson M, Gustafsson F, Galatius S, Hildebrandt PR, Atar D. have underestimated use in the outpatient setting. This also Hyperkalaemia and impaired renal function in patients taking spirono- prohibits an accurate assessment of the relation between lactone for congestive heart failure: retrospective study. BMJ. 2003;327: spironolactone use and outcomes such as mortality or read- 1141–1142. 6. Wrenger E, Muller R, Moesenthin M, Welte T, Frolich JC, Neumann KH. mission because of the potential for significant misclassifica- Interaction of spironolactone with ACE inhibitors or angiotensin receptor tion of exposure. Although we could not assess the impact of blockers: analysis of 44 cases. BMJ. 2003;327:147–149. Masoudi et al Spironolactone Prescription in Heart Failure 47

7. Blaustein DA, Babu K, Reddy A, Schwenk MH, Avram MM. Estimation 17. Baldwin LM, Adamache W, Klabunde CN, Kenward K, Dahlman C, of glomerular filtration rate to prevent life-threatening hyperkalemia due Warren L. Linking physician characteristics and Medicare claims data: to combined therapy with spironolactone and angiotensin-converting issues in data availability, quality, and measurement. Med Care. 2002; enzyme inhibition or angiotensin receptor blockade. Am J Cardiol. 2002; 40(suppl):IV-82–IV-95. 90:662–663. 18. Iezzoni LI. Risk Adjustment for Measuring Health Care Outcomes, 2nd 8. Obialo CI, Ofili EO, Mirza T. Hyperkalemia in congestive heart failure ed. Chicago: Health Administration Press; 1997. patients aged 63 to 85 years with subclinical renal disease. Am J Cardiol. 19. American Hospital Association. The AHA Annual Survey Database: 2002;90:663–665. Fiscal Year 1998 Documentation. Chicago: Health Forum, AHA; 1999. 9. Vanpee D, Swine CH. Elderly heart failure patients with drug-induced 20. American Hospital Association. The AHA Annual Survey Database: serious hyperkalemia. Aging Clin Exp Res. 2000;12:315–319. Fiscal Year 2000 Documentation. Chicago: Health Forum, AHA; 2001. 10. Juurlink DN, Mamdani MM, Lee DS, Kopp A, Austin PC, Laupacis A, 21. Tan SY, Burton M. Hyporeninemic hypoaldosteronism: an overlooked Redelmeier DA. Rates of hyperkalemia after publication of the Ran- cause of hyperkalemia. Arch Intern Med. 1981;141:30–33. domized Aldactone Evaluation Study. N Engl J Med. 2004;351:543–551. 22. Nichols GA, Hillier TA, Erbey JR, Brown JB. Congestive heart failure in 11. Hunt SA, Baker DW, Chin MH, Cinquegrani MP, Feldman AM, Francis type 2 diabetes: prevalence, incidence, and risk factors. Diabetes Care. GS, Ganiats TG, Goldstein S, Gregoratos G, Jessup ML, Noble RJ, 2001;24:1614–1619. Packer M, Silver MA, Stevenson LW, Gibbons RJ, Antman EM, Alpert 23. Dries DL, Exner DV, Domanski MJ, Greenberg B, Stevenson LW. The JS, Faxon DP, Fuster V, Jacobs AK, Hiratzka LF, Russell RO, Smith SC prognostic implications of renal insufficiency in asymptomatic and symp- Jr. ACC/AHA guidelines for the evaluation and management of chronic tomatic patients with left ventricular systolic dysfunction. J Am Coll heart failure in the adult: executive summary a report of the American Cardiol. 2000;35:681–689. College of Cardiology/American Heart Association Task Force on 24. Hillege HL, Girbes AR, de Kam PJ, Boomsma F, de Zeeuw D, Practice Guidelines (Committee to Revise the 1995 Guidelines for the Charlesworth A, Hampton JR, van Veldhuisen DJ. Renal function, neu- Evaluation and Management of Heart Failure). Circulation. 2001;104: rohormonal activation, and survival in patients with chronic heart failure. 2996–3007. 12. Masoudi FA, Havranek EP, Wolfe P, Gross CP, Rathore SS, Steiner JF, Circulation. 2000;102:203–210. Ordin DL, Krumholz HM. Most hospitalized older persons do not meet 25. Schepkens H, Vanholder R, Billiouw JM, Lameire N. Life-threatening the enrollment criteria for clinical trials in heart failure. Am Heart J. hyperkalemia during combined therapy with angiotensin-converting 2003;146:250–257. enzyme inhibitors and spironolactone: an analysis of 25 cases. Am J Med. 13. Havranek EP, Masoudi FA, Westfall KA, Wolfe P, Ordin DL, Krumholz 2001;110:438–441. HM. Spectrum of heart failure in older patients: results from the National 26. Jencks SF, Huff ED, Cuerdon T. Change in the quality of care delivered Heart Failure project. Am Heart J. 2002;143:412–417. to Medicare beneficiaries, 1998–1999 to 2000–2001. JAMA. 2003;289: 14. Rathore SS, Foody JM, Wang Y, Smith GL, Herrin J, Masoudi FA, Wolfe 305–312. P, Havranek EP, Ordin DL, Krumholz HM. Race, quality of care, and 27. Wenger NS, Shekelle PG. Assessing care of vulnerable elders: ACOVE outcomes of elderly patients hospitalized with heart failure. JAMA. 2003; project overview. Ann Intern Med. 2001;135(pt 2):642–646. 289:2517–2524. 28. Shekelle PG, MacLean CH, Morton SC, Wenger NS. ACOVE quality 15. National Kidney Foundation. K/DOQI clinical practice guidelines for indicators. Ann Intern Med. 2001;135(pt 2):653–667. chronic kidney disease: evaluation, classification, and stratification. Am J 29. Pitt B, Remme W, Zannad F, Neaton J, Martinez F, Roniker B, Bittman Kidney Dis. 2002;39(suppl 1):S1–S266. R, Hurley S, Kleiman J, Gatlin M. Eplerenone, a selective aldosterone 16. Kenward K. The scope of the data available in the AMA’s Physician blocker, in patients with left ventricular dysfunction after myocardial Masterfile. Am J Public Health. 1996;86:1481–1482. infarction. N Engl J Med. 2003;348:1309–1321. Heart Failure

Effects of Candesartan on the Development of a New Diagnosis of Diabetes Mellitus in Patients With Heart Failure

Salim Yusuf, DPhil, FRCP; Jan B. Ostergren, MD, PhD; Hertzel C. Gerstein, MD, MSc; Marc A. Pfeffer, MD, PhD; Karl Swedberg, MD, PhD; Christopher B. Granger, MD; Bertil Olofsson, PhD; Jeffrey Probstfield, MD; John V. McMurray, MD; on behalf of the Candesartan in Heart Failure—Assessment of Reduction in Mortality and Morbidity Program (CHARM) Investigators

Background—Diabetes is a risk factor for heart failure, and both conditions are increasing. Identifying treatments that prevent both conditions will be clinically important. We previously reported that candesartan (an angiotensin receptor blocker) reduces cardiovascular mortality and heart failure hospitalizations in heart failure patients (CHARM: Candesartan in Heart Failure—Assessment of Reduction in Mortality and Morbidity Program). Methods and Results—We assessed the impact of candesartan versus placebo on the development of diabetes, a predefined secondary outcome in a randomized, controlled, double-blind study involving 5436 of the 7601 patients with heart failure, irrespective of ejection fraction, who did not have a diagnosis of diabetes at entry into the trial. Patients received candesartan (target of 32 mg once daily) or matching placebo for 2 to 4 years. One hundred sixty-three (6.0%) individuals in the candesartan group developed diabetes, as compared with 202 (7.4%) in the placebo group (hazard ratio [HR], 0.78 with a 95% confidence interval [CI] of 0.64 to 0.96; Pϭ0.020). The composite end point of death or diabetes occurred in 692 (25.2%) and 779 (28.6%), respectively, in the candesartan and placebo groups (HR, 0.86; 95% CI, 0.78 to 0.95; Pϭ0.004). The results were not statistically heterogeneous in the various subgroups examined, although the apparent magnitude of benefit appeared to be smaller among those treated concomitantly with angiotensin-converting enzyme inhibitors at trial entry (HR, 0.88; 95% CI, 0.65 to 1.20) compared with those not receiving these drugs (HR, 0.71; 95% CI, 0.53 to 0.93; P for heterogeneity, 0.28). Conclusions—The angiotensin receptor blocker candesartan appears to prevent diabetes in heart failure patients, suggesting that the renin-angiotensin axis is implicated in glucose regulation. (Circulation. 2005;112:48-53.) Key Words: renin Ⅲ diabetes mellitus Ⅲ prevention Ⅲ heart failure Ⅲ glucose

ype 2 diabetes mellitus (DM) is an important and renin-angiotensin-aldosterone system. Recent retrospective Tcommon risk factor for the development of heart failure analysis of the HOPE (Heart Outcomes Prevention Evalua- and for subsequent prognosis. This could be related to the tion)5 and ALLHAT (Antihypertensive and Lipid-Lowering established link between diabetes and either left ventricular treatment to prevent Heart Attack Trial)6 and data from a hypertrophy, coronary artery disease, or other risk factors (eg, single center in the SOLVD (Studies of Left Ventricular hypertension or obesity) with heart failure. Both DM and Dysfunction) trials7 suggest that angiotensin-converting en- heart failure are increasing1 and are associated with high rates zyme (ACE) inhibitors (ramipril and enalapril, respectively) of mortality and morbidity compared with unaffected indi- may reduce the risk of DM in individuals with atherosclerosis viduals. Both conditions lead to substantial economic costs to or heart failure. Data from the LIFE (Losartan Intervention for society. Preventing DM may prevent or delay the develop- End-point Reduction) study8 suggest that losartan, an angio- ment of many of its complications (eg, atherosclerotic vas- tensin receptor blocker (ARB), may reduce the risk of new cular disease or renal dysfunction). Recent evidence suggests DM compared with atenolol, a ␤-adrenergic blocker, in that lifestyle modification by weight loss2,3 and some patients with hypertension and left ventricular hypertrophy. glucose-lowering drugs (eg, metformin,2 acarbose4) can re- However, it is unclear whether this difference was caused by duce the risk of DM in high-risk individuals. A complemen- a lower rate of DM with losartan or an increased rate of DM tary approach may be to prevent DM by blocking the with atenolol. That angiotensin type 1 receptor blockade has

Received December 10, 2004; revision received March 10, 2005; accepted March 24, 2005. From the Population Health Research Institute (S.Y., H.C.G.), McMaster University, and Hamilton Health Sciences, Hamilton, Canada; Karolinska University Hospital (J.B.O.), Stockholm, Sweden; Brigham and Women’s Hospital (M.A.P.), Boston, Mass; Sahlgrenska University Hospital/Ostra (K.S.), Goteberg, Sweden; Duke University Medical Center (C.B.G.), Durham, NC; AstraZeneca (B.O.), Molndal, Sweden; University of Washington (J.P.), Seattle; and University of Glasgow (J.V.M.), Glasgow, Scotland. Correspondence to Salim Yusuf, DPhil, FRCP, Department of Medicine, McMaster Clinic, Room 252, 237 Barton St, East Hamilton, Ontario, Canada L8L 2X2. E-mail [email protected] © 2005 American Heart Association, Inc. Circulation is available at http://www.circulationaha.org DOI: 10.1161.CIRCULATIONAHA.104.528166 48 Yusuf et al Candesartan in Preventing Diabetes 49

Figure 1. Numbers of patients random- ized and final disposition.

a real effect in preventing DM is suggested by recent findings the occurrence of a new diagnosis of DM for all patients at the end from the Valsartan Antihypertensive Long-term Use Evalua- of the trial. If a diagnosis of DM was reported after randomization, tion (VALUE),9 in which valsartan significantly reduced the the date of diagnosis, details of the criteria used (fasting plasma glucose Ն7 mmol/L [126 mg/dL], fasting blood glucose new incidence of DM in comparison with amlodipine, a drug Ն6.1 mmol/L [110 mg/dL], 2-hour [oral glucose tolerance test] or a that is considered to be metabolically neutral. random glucose Ն11.1 mmol/L [200 mg/dL]), hypoglycemic med- In the CHARM (Candesartan in Heart Failure: Assessment ication prescribed, and lifestyle modifications prescribed were rec- of Reduction in Mortality and Morbidity) Program), we orded on the case record forms. The planned minimum follow-up prospectively specified that we would perform a secondary was 2 years with a maximum of 4 years, with the median being 3.1 years. At baseline and throughout the study, physicians were free to analysis of the effects of candesartan compared with placebo prescribe various treatments, including other cardiovascular drugs with respect to preventing the development of DM (and of (other than ARBs) or glucose-lowering drugs. DM plus all-cause mortality) in a broad range of patients with The prespecified outcomes for this report were the development of heart failure. DM alone or as a composite with all-cause mortality (to avoid the problem of competing risk). All analyses were based on an intention- to-treat approach with the Wald statistic in a Cox proportional- Methods hazards model and displayed by Kaplan-Meier plots according to Detailed descriptions of the methods and results of the CHARM treatment allocation. The hazard ratios (HRs) and 95% confidence program have been published previously.10–13 In brief, the CHARM intervals (CIs) comparing treatments, stratified by trial, were calcu- program consisted of 3 parallel trials involving complementary lated for the overall data. The results in a few key subgroups are also populations of patients with symptomatic heart failure. Those with a presented, along with the tests of heterogeneity based on the Cox Յ low ejection fraction ( 0.40) and with prior intolerance to an ACE regression model to evaluate whether the effects of candesartan inhibitor were included in the CHARM Alternative trial, and those varied in the subgroups. All patients provided written, informed receiving an ACE inhibitor were included in the CHARM Added consent, and the protocol was approved by the ethics committee at Ͼ trial. Those with an ejection fraction 0.40 and heart failure were each participating institution. enrolled into the CHARM Preserved trial (whether or not they were receiving an ACE inhibitor). Overall, 7601 patients were random- ized, of whom 2163 were known to have DM as reported by the Results investigators at randomization, and 5436 were not known to have The characteristics of the 5436 patients who were not known DM at baseline (Figure 1). This latter group of patients forms the to have DM at baseline are provided in Table 1 and were well basis for this report. Patients were randomized to receive either balanced between the groups allocated to candesartan or ϭ ϭ candesartan (n 2715) or placebo (n 2721). Patients received the placebo. Of note, 68% of patients had a body mass index Ͼ25 study drug in incremental doses (up to a maximum of 32 mg 2 ␤ candesartan once daily as tolerated) or matching placebo. Patients kg/m , 50% were hypertensive, 55% were taking -blockers, were followed up at 2, 4, and 6 weeks; 6 months; and then every 4 80% were taking diuretics, and 39% were receiving ACE months until the study end. The investigators were asked to report inhibitors. 50 Circulation July 5, 2005

TABLE 1. Baseline Characteristics

Candesartan Placebo (nϭ2715) (nϭ2721) Mean age, y 66Ϯ11 66Ϯ12 Age Ͼ75 years, % 24 24 Male, % 69 69 White, % 92 93 Current smokers, % 16 15 Previous smokers, % 49 47 Mean BMI, kg/m2 28 28 Ͻ25, % 32 32 Ն25 to Ͻ30, % 41 42 Ն30, % 27 27 Systolic blood pressure, mm Hg 130Ϯ19 131Ϯ19 NYHA class, % II 48 48 III/IV 52 52 Hypertension, % 50 50 Previous myocardial infarction, % 48 49 Baseline medication, % Any diuretic 80 81 Thiazide diuretic 12 12 ␤-Blocker 54 55 ACE inhibitor 39 38 BMI indicates body mass index. Proportions are in percent or meanϮSD. Figure 3. Results of candesartan in various subgroups. BMI indicates body mass index; MI, myocardial infarction; low., low- The proportion of patients receiving the study drug at 14 ering; and bl., blocker. months was 86.0% and 89.5% and at 26 months, was 82.4% and 86.7% in the active and placebo groups, respectively. Of those taking the study medication, the mean dose used at the DM was most marked in patients with preserved ejection last visit was 24.5 mg/d in the candesartan group and 28.0 mg fraction, with a favorable trend in the Alternative trial in the placebo group. (wherein candesartan was compared with placebo in the Overall, there were 163 (6.0%) individuals who reported a absence of an ACE inhibitor) and little apparent benefit in the new diagnosis of DM in the candesartan group compared Added trial (wherein candesartan or placebo was added to an with 202 (7.4%) in the placebo group (HR, 0.78; 95% CI, ACE inhibitor). However, the 95% CIs of these estimates 0.64 to 0.96; Pϭ0.020) (Figure 2). The composite of new DM overlap, and there was no statistically significant heterogene- or death occurred in 692 (25.2%) patients in the candesartan ity. Examining the results on the composite of death or group compared with 779 (28.6%) in the placebo group (HR, development of DM in various subgroups also provided no ϭ 0.86; 95% CI, 0.78 to 0.95; P 0.004). The difference in new evidence of statistical heterogeneity across the 3 component trials (Figures 3 and 4).

Figure 2. Effects of candesartan compared with placebo on incidence of DM. Figure 4. Results in each of 3 component trials. Yusuf et al Candesartan in Preventing Diabetes 51

TABLE 2. Previous Studies Evaluating ACE Inhibitors and Impact of Baseline Potassium and Changes in ARBs on Development of Diabetes Potassium Levels Relative Risk Potassium levels were available for 2743 patients involved in Study Comparison (95% CI) North America. The impact on reducing the rates of DM was ACE inhibitors similar in those with potassium values equal to or below and above the median (HR, 0.725; 95% CI, 0.47 to 1.11; HR, HOPE5 Ramipril vs placebo 0.66 (0.51–0.85) 0.81; 95% CI, 0.45 to 1.47, respectively; P for interaction, 26 ␤ CAPPP Captopril vs -blocker/diuretics† 0.79 (0.67–0.94) 0.81). There was a small decrease in potassium levels 7 SOLVD * Enalapril vs placebo 0.22 (0.10–0.46) (Ϫ0.028; SD, Ϫ0.497) in the placebo group and an increase ALLHAT6 Lisinopril vs amlodipine 0.83 (0.66–1.04) (ϩ0.144; SD, ϩ0.544) in the candesartan group (PϽ0.0001). Lisinopril vs diuretics† 0.70 (0.57–0.86) However, adjusting for this difference in potassium (with the ARBs use of time-dependent covariate analysis) between the 2 LIFE8 Losartan vs atenolol† 0.75 (0.63–0.88) groups did not alter the impact of candesartan on the SCOPE27 Candesartan vs placebo/other drugs 0.80 (0.62–1.03) development of DM. ALPINE28 Candesartan/felodipine vs 0.13 (0.02–0.99) ␤-blocker/diuretics† Impact of Postrandomization Differences VALUE9 Valsartan vs amlodipine 0.77 (0.69–0.86) in Hospitalization Because candesartan reduced the risk of hospitalizations *Based on data from only 1 center of 23 in the trial and thus may be open to biases. (during which patients are likely to be investigated more †Comparator could be diabetogenic. intensively), we examined whether the difference in new diagnoses of DM could be explained by differences in the Results by Diagnostic Criteria and rates of hospitalization. The rates of development of DM Hypoglycemic Treatment were similarly affected by candesartan in those with (58 DM was diagnosed by the patient’s physician on the basis of versus 74; HR, 0.76; 95% CI, 0.54 to 1.06) versus those a fasting glucose sample in 130 patients allocated to placebo without (105 versus 128; HR, 0.79; 95% CI, 0.61 to 1.02) an compared with 112 in the candesartan group and by an oral interim hospitalization. glucose tolerance test in 25 placebo patients compared with 21 candesartan patients. No criteria for the diagnosis of DM Discussion were specified on the case report forms provided for 47 Our study demonstrates that candesartan, an ARB, appears to placebo patients and 30 candesartan patients. DM was treated prevent the development of DM in patients with heart failure with insulin or an oral hypoglycemic drug in 151 placebo and no previous diagnosis of this condition. This effect was patients compared with 117 candesartan patients and with consistently observed in all subgroups examined, with no dietary modification alone in 51 placebo patients and 45 evidence of heterogeneity among the 3 component trials in candesartan patients. In 1 patient with candesartan treatment, this program. It appears that the magnitude of the effect may information on treatment for DM was missing. have been smaller in those receiving concomitant ACE inhibitors. Although the relative attenuation of the effects of Subgroups preventing DM among those receiving concomitant ACE The benefits of candesartan in preventing DM were seen in inhibitor is not conclusive (the test for interaction is not those with a high or low body mass index and in those significant), it is nevertheless plausible, as the mechanisms of receiving or not receiving ␤-blockers or diuretics (Figures 3 action of ARBs and ACE inhibitors have considerable over- and 4). Similar benefits were observed in those in New York lap. The reduction in DM was observed in those with varying Heart Association (NYHA) classes II and III. Although the severities of heart failure symptoms; those with varying magnitude of benefit appeared to be smaller in those receiv- levels of left ventricular ejection fraction; and those with ing a concomitant ACE inhibitor (HR, 0.88; 95% CI, 0.65 to different levels of body mass index, blood pressure, potas- 1.20) compared with those not on it (HR, 0.71; 95% CI, 0.53 sium, and concomitant drugs such as ␤-blockers or diuretics to 0.65), these differences were not heterogeneous from each (which may affect glycemic levels). The difference in DM is other (P for interaction, 0.28). The effect was most marked in not explained by differences in the rates of hospitalizations, patients with left ventricular ejection fractions Ͼ40% (the thereby excluding the possibility of detection biases. CHARM Preserved trial), but there was no significant het- Recent studies implicate angiotensin II (A-II) in the growth erogeneity in effect between the trials in the program (P for and development of adipose tissue.14 Angiotensinogen is heterogeneity, 0.14). induced early in the adipogenic differentiation of preadipo- cytes and is highly expressed in mature adipocytes. A-II Impact of Postrandomization Differences in Drugs inhibits the adipogenic differentiation of preadipocytes, and There was a higher proportion of patients in the placebo blockade of the angiotensin receptor enhances the response of group receiving a diuretic (76.5% placebo versus 72.6% preadipocytes to insulin. Increased expression of both A-II candesartan), a ␤-blocker (64.3% versus 60.0%), or an ACE and ACE has been demonstrated in subcutaneous, abdominal, inhibitor (38.1% versus 34.3%) by the end of the trial. adipose tissue of overweight and obese individuals.15 Reduc- Adjusting for these differences did not affect the overall ing the levels of A-II with either an ACE inhibitor or blunting impact of candesartan in preventing DM. the actions of A-II with ARBs could facilitate differentiation 52 Circulation July 5, 2005 of preadipocytes to mature adipocytes and subsequently jects With Cardiovascular Disease) study, which has increase lipid storage capacity in adipose tissue. Such an evaluated telmisartan in patients with impaired fasting glu- effect may reduce intramyocellular and hepatic fat and cose or impaired glucose tolerance and vascular disease.21 thereby improve insulin sensitivity. ARBs may also improve Furthermore, the ONTARGET (Ongoing Telmisartan Alone insulin sensitivity by raising adiponectin levels and by in- and in Combination With Ramipril Global End-point Trial) creasing the serine phosphorylation of insulin receptors, study21 is evaluating the relative impact of ramipril versus insulin receptor substrate-1, and phosphatidylinositol telmisartan versus their combination in Ͼ25 000 individuals 3-kinase.16 It is also possible that ACE inhibitors and ARBs (two thirds of whom do not have DM) in preventing DM, as increase blood flow to the pancreas and skeletal muscle or well as a range of vascular complications. improve insulin sensitivity or secretion by increasing potas- Whereas prevention of DM should be fundamentally ap- sium levels.17 Some ARBs have been demonstrated to have proached by reducing weight and increasing activity, the an agonist effect on the peroxisome proliferator–activated observations that ACE inhibitors and ARBs prevent the receptor-␥ enzyme, and this may also play an additional role development of DM has implications for certain populations in reducing glucose levels and the risk of DM.18 (such as heart failure, after myocardial infarction, vascular Our clinical observation of a reduction in the development disease, or hypertension) in whom such drugs have already of DM with candesartan, an ARB, is supported by several been shown to reduce major vascular events. In such popu- previous studies of ACE inhibitors or ARBs (Table 2). In the lations, preventing DM by blocking the renin-angiotensin-al- HOPE study, ramipril reduced the risk of new DM in those dosterone system is likely to confer additional clinical bene- with atherosclerosis.5 Similar observations have been made fits by reducing some of the risks associated with DM (such with enalapril in SOLVD in patients with low ejection as renal and vascular damage), especially during prolonged fractions.7 In the LIFE study, losartan reduced the develop- treatment (up to 10 years), which is well beyond the time ment of DM compared with a ␤-blocker; thus, that study is frame of current trials (usually 2 to 5 years). This suggests not able to differentiate between a protective effect of an that the clinical benefits observed during the planned ARB or an adverse effect of a ␤-blocker on the development follow-up of current trials of a few years may be an of DM.8 In the ALLHAT study of patients with hypertension, underestimate of the full benefits that may accrue from longer lisinopril reduced the rates of new DM compared with treatment. Supportive evidence for this hypothesis stems from amlodipine (which has a neutral effect) and thiazides (which the extended follow-up of SOLVD22 and the 7-year follow-up have an adverse effect on DM rates).6 Our observation in of the HOPE study.23 CHARM, wherein candesartan was compared with placebo, There are a few limitations of our study. First, we relied on indicates that the benefits are likely mediated through block- the clinical diagnosis of DM, rather that on serial testing of ing the effects of A-II. In the absence of concomitant therapy blood glucose levels or performing an oral glucose tolerance with an ACE inhibitor, there is an approximate one-third test. However, because this was a blinded study, no material relative risk reduction in DM with candesartan, which is biases in comparing candesartan versus placebo groups similar to the benefits of ramipril compared with placebo, would be expected to occur. However, we may have under- when used alone in the HOPE study. Therefore, although estimated the absolute rates of DM (and hence, the absolute CHARM is the only study that directly assessed the effects of benefits). Second, our population consisted of patients with an ARB against a placebo, the collective experience from heart failure, which is an intensely “diabetogenic” state. several trials with different comparator groups provides Therefore, further studies with ARBs to evaluate their impact persuasive and coherent evidence that ACE inhibitors and in preventing DM in other high-risk populations are required. ARBs prevent DM. Recently, our finding has also gained In conclusion, we have demonstrated that candesartan support from the results of the VALUE trial,9 in which reduces the risk of developing DM. This benefit occurs in valsartan prevented DM in hypertensives in comparison with addition to reductions in cardiovascular mortality and hospi- amlodipine, a drug that is considered to be metabolically talizations for heart failure,10–13, improvement in functional neutral. status (according to NYHA classification),24 and prevention CHARM is the only study to provide clear evidence of the of atrial fibrillation.25 These benefits of candesartan in pre- effects of an ARB in preventing DM in heart failure patients, venting multiple adverse outcomes in this high-risk popula- most of whom were receiving a diuretic. This suggests that tion provide persuasive evidence of the clinical benefits of blockade of the renin-angiotensin-aldosterone system to pre- ARBs in patients with heart failure. vent DM may be applicable to many different types of high-risk patients. Further data on this issue will be provided Acknowledgments by the DREAM (Diabetes Reduction Assessment With The CHARM program was funded by AstraZeneca, which was Ramipril and Rosiglitazone Medication) study,19 which has responsible for data collection and analysis. The Study Executive evaluated ramipril in Ϸ5200 individuals with impaired glu- Committee, consisting of all authors (except H. Gerstein and J. cose tolerance or impaired fasting glucose; the NAVIGATOR Probstfield), supervised the management of the study and were primarily responsible for the interpretation of the data, preparation, (Nateglinide and Valsartan in Impaired Glucose Tolerance review, and approval of the manuscript. Outcomes Research) study,20 which has evaluated valsartan in patients with impaired glucose intolerance and atherosclero- Disclosure sis or multiple risk factors; and the TRANSCEND (Telmis- Dr Yusuf has received research grants, has served on speakers’ artan Randomized Assessment Study in ACE Intolerant Sub- bureaus and/or received honoraria, and has served as a consultant to Yusuf et al Candesartan in Preventing Diabetes 53 several pharmaceutical companies, including AstraZeneca. Dr Pfef- reduced left ventricular systolic function treated with an ACE inhibitor: fer has received a research grant from, has served on the speakers’ the CHARM-Added trial. Lancet. 2003;362:767–771. bureau of and/or received honoraria from, and has consulted for 12. Granger CB, McMurray JJ, Yusuf S, Held P, Michelson EL, Olofsson B, AstraZeneca. Dr Olofsson is employed by AstraZeneca. Dr Swed- Ostergren J, Pfeffer MA, Swedberg K; CHARM Investigators and Com- berg has received research grants or other research support, has mittees. Effects of candesartan in patients with chronic heart failure and served on speakers’ bureaus and/or received honoraria, and has reduced left ventricular systolic function and intolerant to ACE inhibitors: served as a consultant. Dr Ostergren has received a research grant the CHARM-Alternative Trial. Lancet. 2003;362:772–776. from AstraZeneca; has served on the speakers’ bureaus of and/or 13. Yusuf S, Pfeffer MA, Swedberg K, Granger CB, Held P, McMurray JJ, received honoraria from AstraZeneca, Merck, Aventis, and Novartis; Michelson EL, Olofsson B, Ostergren J; CHARM Investigators and and has served as a consultant to AstraZeneca, Pfizer, Aventis, and Committees. Effects of candesartan in patients with chronic heart failure and preserved left ventricular systolic function: the CHARM-Preserved Novartis. Dr Gerstein is the principal investigator in a trial of ACEI trial. Lancet. 2003;362:777–781. to prevent diabetes and has received research support for CHARM 14. Sharma AM, Janke J, Gorzelniak K, Engeli S, Luft FC. Angiotensin CI, a substudy to prevent albuminuria. Dr Granger has received a blockade prevents type 2 diabetes by formation of fat cells. Hypertension. research grant from and has served as a consultant to AstraZeneca. 2002;40:609–611. Dr Probstfield has received research grants from King, Wyeth, and 15. Gorzelniak K, Engeli S, Janke J, Luft FC, Sharma AM. Hormonal regu- Boehringer; has served on the speakers’ bureaus of and/or received lation of the human adipose-tissue renin-angiotensin system: relationship honoraria from King, Wyeth, and Pfizer; and has served as a to obesity and hypertension. J Hypertens. 2002;20:965–973. consultant to King. Dr McMurray has received research grants or 16. Furuhashi M, Ura N, Higashiura K, Murakami H, Tanaka M, Moniwa N, other research support from, served on the speakers’ bureaus of Yoshida D, Shimamoto K. Blockade of the renin-angiotensin system and/or received honoraria from, and consulted for AstraZeneca and increases adiponectin concentrations in patients with essential hyper- Takeda. tension. Hypertension. 2003;42:76–81. 17. McGarry JD, Dobbins RL. Fatty acids, lipotoxicity and insulin secretion. References Diabetologia. 1999;42:128–138. 18. Schupp M, Janke J, Clasen R, Unger T, Kintscher U Angiotensin type 1 1. King H, Aubert RE, Herman WH. Global burden of diabetes, 1995–2025: ␥ prevalence, numerical estimates, and projections. Diabetes Care. 1998; receptor blockers induce peroxisome proliferator–activated receptor- 21:1414–1431. activity. Circulation. 2004;109:2054–2057. 2. Knowler WC, Barrett-Connor E, Fowler SE, Hamman RF, Lachin JM, 19. DREAM Trial Investigators. Rationale, design and recruitment charac- Walker EA, Nathan DM; Diabetes Prevention Program Research Group. teristics for a large simple international trial of diabetes prevention: the Reduction in the incidence of type 2 diabetes with lifestyle intervention or DREAM trial (Diabetes Reduction Assessment with Ramipril and Ros- metformin. N Engl J Med. 2002;346:393–403. iglitazone Medications). Diabetologia. 2004;47:1519–1527. 3. Tuomilehto J, Lindstrom J, Eriksson JG. Prevention of type 2 diabetes 20. Nateglinide and valsartan in impaired glucose tolerance outcomes mellitus by changes in lifestyle among subjects with impaired glucose research: rationale and design of the NAVIGATOR trial. Diabetes 2002; tolerance. N Engl J Med. 2001;344:1343–1350. 51(suppl 2):A116. Abstract. 4. Chiasson JL, Josse RG, Gomis R, Hanefeld M, Karasik A, Laakso M; 21. The ONTARGET/TRANSCEND Investigators. Rationale, design and STOP-NIDDM Trail Research Group. Acarbose for prevention of type 2 baseline characteristics of two large, simple randomized trials evaluating diabetes mellitus: the STOP-NIDDM randomised trial. Lancet. 2002;359: telmisartan, ramipril, and their combination in high-risk patients: the 2072–2077. ONTARGET/TRANSCEND trials. Am Heart J. 2004;148:52–61. 5. Yusuf S, Gerstein H, Hoogwerf B, Pogue J, Bosch J, Wolffenbuttel BH, 22. Jong P, Yusuf S, Rousseau MF, Ahn SA, Bangdiwala SI. Effect of Zinman B; HOPE Study Investigators. Ramipril and the development of enalapril on 12-year survival and life expectancy in patients with left diabetes. JAMA. 2001;286:1882–1885. ventricular systolic dysfunction: a follow-up study. Lancet. 2003;361: 6. ALLHAT Officers and Coordinators for the ALLHAT Collaborative 1843–1848. Research Group. Major outcomes in high-risk hypertensive patients ran- 23. HOPE and HOPE-TOO Investigators. Long-term effects of ramipril on domized to angiotensin-converting enzyme inhibitor or calcium channel cardiovascular events and diabetes: results of the HOPE Study Eval- blocker vs diuretic: the Antihypertensive and Lipid-Lowering Treatment uation. Circulation. In press. to Prevent Heart Attack Trial (ALLHAT). JAMA. 2002;288:2981–2997. 24. McMurray JJ, Ostergren J, Olofsson B, Granger CB, Michelson E, Young 7. Vermes E, Tardif JC, Bourassa MG, Racine N, Levesque S, White M, JB, Dunlap M, Yusuf S, Swedberg K, Pfeffer MA, for the CHARM Guerra PG, Ducharme A. Enalapril reduces the incidence of diabetes in Investigators. Candesartan improves functional class across a broad patients with chronic heart failure: insight from the Studies Of Left spectrum of patients with chronic heart failure: results of the candesartan Ventricular Dysfunction (SOLVD). Circulation. 2003;107:1291–1296. in heart failure—assessment of reduction in mortality and morbidity 8. Lindholm LH, Ibsen H, Dahlof B, Devereux RB, Beevers G, de Faire U, programme (CHARM). J Am Coll Cardiol. 2004;43(suppl A):206A. Fyhrquist F, Julius S, Kjeldsen SE, Kristiansson K, Lederballe-Pedersen Abstract. O, Nieminen MS, Omvik P, Oparil S, Wedel H, Aurup P, Edelman J, 25. Swedberg K, Cohen-Solal A, Granger C, McMurray J, Olsson L, Pfeffer Snapinn S; LIFE Study Group. Cardiovascular morbidity and mortality in M, Yusuf S, Ostergren J. Prevention of atrial fibrillation in symptomatic patients with diabetes in the Losartan Intervention For Endpoint reduction chronic heart failure by candesartan: results from CHARM. J Am Coll in hypertension study (LIFE): a randomised trial against atenolol. Lancet. Cardiol 2004;43(suppl A):222A. Abstract. 2002;359:1004–1010. 26. Effect of angiotensin-converting-enzyme inhibition compared with con- 9. Julius S, Kjeldsen SE, Weber M, Brunner HR, Ekman S, Hansson L, Hua ventional therapy on cardiovascular morbidity and mortality in hyper- T, Laragh J, McInnes GT, Mitchell L, Plat F, Schork A, Smith B, tension: the Captopril Prevention Project (CAPPP) randomised trial. Zanchetti A; VALUE trial group. Outcomes in hypertensive patients at Lancet. 1999;353:611–616. high cardiovascular risk treated with regimens based on valsartan or 27. Lithell H, Hansson L, Skoog I, Elmfeldt D, Hofman A, Olofsson B, amlodipine: the VALUE randomised trial. Lancet. 2004;363:2022–2031. Trenkwalder P, Zanchetti A; SCOPE Study Group. The Study on Cog- 10. Pfeffer MA, Swedberg K, Granger CB, Held P, McMurray JJ, Michelson nition and Prognosis in the Elderly (SCOPE): principal results of a EL, Olofsson B, Ostergren J, Yusuf S, Pocock S; CHARM Investigators randomized double-blind intervention trial. J Hypertens. 2003;21: and Committees. Effects of candesartan on mortality and morbidity in 875–886. patients with chronic heart failure: the CHARM-Overall Programme. 28. Lindholm LH, Persson M, Alaupovic P, Carlberg B, Svensson A, Lancet. 2003;362:759–766. Samuelsson O. Metabolic outcome during 1 year in newly detected 11. McMurray JJ, Ostergren J, Swedberg K, Granger CB, Held P, Michelson hypertensives: results of the Antihypertensive Treatment and Lipid EL, Olofsson B, Yusuf S, Pfeffer MA; CHARM Investigators and Com- Profile in a North of Sweden Efficacy Evaluation (ALPINE study). mittees. Effects of candesartan in patients with chronic heart failure and J Hypertens. 2003;21:1563–1574. ␣-Myosin Heavy Chain A Sarcomeric Gene Associated With Dilated and Hypertrophic Phenotypes of Cardiomyopathy

Elisa Carniel, MD; Matthew R.G. Taylor, MD, PhD; Gianfranco Sinagra, MD; Andrea Di Lenarda, MD; Lisa Ku, MS; Pamela R. Fain, PhD; Mark M. Boucek, MD; Jean Cavanaugh, MS; Snjezana Miocic, MD; Dobromir Slavov, PhD; Sharon L. Graw, PhD; Jennie Feiger, MS, MA; Xiao Zhong Zhu, BS; Dmi Dao, BA; Debra A. Ferguson, MS; Michael R. Bristow, MD, PhD; Luisa Mestroni, MD

Background—Mutations in the ␤-myosin heavy-chain (␤MyHC) gene cause hypertrophic (HCM) and dilated (DCM) forms of cardiomyopathy. In failing human hearts, downregulation of ␣MyHC mRNA or protein has been correlated with systolic dysfunction. We hypothesized that mutations in ␣MyHC could also lead to pleiotropic cardiac phenotypes, including HCM and DCM. Methods and Results—A cohort of 434 subjects, 374 (134 affected, 214 unaffected, 26 unknown) belonging to 69 DCM families and 60 (29 affected, 30 unaffected, 1 unknown) in 21 HCM families, was screened for ␣MyHC gene (MYH6) mutations. Three heterozygous MYH6 missense mutations were identified in DCM probands (P830L, A1004S, and E1457K; 4.3% of probands). A Q1065H mutation was detected in 1 of 21 HCM probands and was absent in 2 unaffected offspring. All MYH6 mutations were distributed in highly conserved residues, were predicted to change the structure or chemical bonds of ␣MyHC, and were absent in at least 300 control chromosomes from an ethnically similar population. The DCM carrier phenotype was characterized by late onset, whereas the HCM phenotype was characterized by progression toward dilation, left ventricular dysfunction, and refractory heart failure. Conclusions—This study suggests that mutations in MYH6 may cause a spectrum of phenotypes ranging from DCM to HCM. (Circulation. 2005;112:54-59.) Key Words: genetics Ⅲ myosin Ⅲ cardiomyopathy, hypertrophic Ⅲ cardiomyopathy, dilated

eventeen genes encoding cytoskeletal, sarcomeric, and case of elderly-onset sporadic HCM.8 On the basis of its Snuclear proteins, including ␤-myosin heavy chain behavior in human myocardial failure, we hypothesized that (␤MyHC), have been associated with dilated cardiomyopathy ␣MyHC may be relevant for myocardial function and that (DCM).1 Hypertrophic cardiomyopathy (HCM) is caused by mutations could cause a spectrum of cardiac phenotypes mutations in 9 genes encoding sarcomeric proteins; among ranging from HCM to DCM, as observed in the case of them, mutations in the ␤MyHC gene (MYH7) account for the ␤MyHC. majority of cases.2 Two cardiac MyHC isoforms3 have been identified in Methods humans, with the genes tandemly located on chromosome 14. Patient Population MYH6 encodes ␣MyHC and MYH7 encodes ␤MyHC.3 Ninety families, 69 with DCM (48 familial, 21 sporadic) and 21 with ␣MyHC and ␤MyHC are present in different amounts in HCM, for a total of 434 subjects, 163 of whom were affected, were mammalian hearts4; human hearts express predominantly studied in the Cardiology Divisions of the University of Colorado Hospital and the University Hospital of Trieste, Italy, and were ␤ 4–6 ␣ MyHC. In nonfailing human hearts, MyHC mRNA enrolled in the Familial Cardiomyopathy Registry.1 Informed con- represents 20% to 30% of the total myosin mRNA, whereas sent was obtained from all subjects enrolled in the study, according ␣MyHC protein represents Ϸ7% of the total MyHC. These to the institutional review committee. Accurate family history was are downregulated to 10% and Ͻ1%, respectively, in failing obtained from each individual, and family screening was performed. ␤ 5,6 All of the subjects underwent physical examination, ECG, and hearts, whereas MyHC is upregulated. laboratory analysis. Echocardiography was performed in 407 of 434 Few data exist on the role of MYH6 mutations in mam- individuals (echocardiograms were not obtained for 22 relatives mals.7 In humans, an MYH6 mutation has been found in one classified as healthy by history, physical examination, and ECG).

Received January 30, 2004; de novo received September 17, 2004; revision received February 24, 2005; accepted March 2, 2005. From the Familial Cardiomyopathy Registry Research Group. Reprint requests to Dr Luisa Mestroni, University of Colorado Cardiovascular Institute, Bioscience Park Center, 12635 E Montview Blvd, Suite 150, Aurora, CO 80010-7116. E-mail [email protected] © 2005 American Heart Association, Inc. Circulation is available at http://www.circulationaha.org DOI: 10.1161/CIRCULATIONAHA.104.507699 54 Carniel et al ␣MyHC Mutations in DCM and HCM 55

TABLE 1. Major and Minor Criteria for the Diagnosis of Familial DCM

Major Criteria Minor Criteria LV systolic dysfunction: EF Ͻ45% (2SD) and/or FS Ͻ25% (Ͼ2SD) Unexplained SV arrhythmias (AF or sustained arrhythmias) LV dilation: LVEDD Ͼ117% of predicted value Frequent (Ͼ1000/24 h) or repetitive (3 or more ectopic beats with a heart rate Ͼ120 bpm) Ventricular arrhythmias Ͻ50 years old LVEDD Ͼ112% of predicted value LVEF Ͻ50% or FS Ͻ28% Unexplained conduction abnormalities (grade II or III AV blocks, complete LBBB, or sinus nodal dysfunction) Unexplained SD or stroke Ͻ50 years old Segmental wall-motion abnormalities in Ͼ1 segment, or 1 if not previously present, in the absence of ischemic heart disease LV indicates left ventricular; FS, fractional shortening; EF, ejection fraction; LVEDD, LV end-diastolic diameter; SV, supraventricular; AF, atrial fibrillation; AV, atrioventricular; LBBB, left bundle-branch block; and SD, sudden death.

When clinically indicated, additional studies were performed, includ- graphic criteria, or 1 minor echocardiographic criterion and 1 ing right and left heart catheterization, ventriculography, coronary electrocardiographic criterion. angiography, endomyocardial biopsy, and neuromuscular evaluation. Molecular Genetic Screening Diagnostic Criteria of DCM and HCM Blood samples were collected from 427 of 434 subjects for DNA Criteria for the diagnosis of DCM were the presence of left analysis. MYH6 was screened for mutations by denaturing high- ventricular fractional shortening Ͻ25% (Ͼ2SD) and/or an ejection performance liquid chromatography and sequence analysis. fraction Ͻ45% (Ͼ2SD) and left ventricular end-diastolic diameter In the families in which we found either a putative disease-causing Ͼ117% of the predicted value by the Henry formula, corrected for mutation or a polymorphism, all available relatives were screened for age and body surface area.9 Exclusion criteria included any of the mutations. Criteria for classifying variants as putative disease-causing following conditions: blood pressure Ͼ160/110 mm Hg, obstruction mutations11 included changes in predicted amino acid sequences, Ͼ50% of a major coronary artery branch, alcohol intake Ͼ100 g/d, segregation within the family (when available), conservation across persistent high-rate supraventricular arrhythmia, systemic diseases, different species (http://www.ncbi.nlm.nih.gov/BLAST/), absence in a pericardial diseases, congenital heart diseases, cor pulmonale, and control population of at least 150 healthy ethnically similar subjects, and myocarditis. changes in protein secondary structure, with the use of GOR 412 and Familial DCM was defined by the presence of 2 or more affected NNPREDICT software.13 The control panels were screened by dena- subjects in the same family with DCM meeting the published turing high-performance liquid chromatography, and profiles different criteria.9 Family members were classified as affected, unaffected, or from the wild type were sequenced. unknown on the base of major and minor criteria (Table 1). The affected status was defined by the presence of 2 major criteria, 1 major criterion and 1 minor criterion, or 3 minor criteria. The Results unknown status was defined by the presence of 1 or 2 minor criteria Four putative disease-causing mutations were detected (Fig- and the unaffected status by the presence of a normal heart or the ures 1 and 2; Table 3) in 3 of 69 probands with DCM (4.3%) determination of other causes of myocardial dysfunction. and 1 of 21 probands with HCM (4.8%). HCM was diagnosed in the presence of unexplained left ventric- In exon 21, a P830L substitution was found in a sporadic ular hypertrophy,10 excluding secondary causes of cardiac hypertro- phy, such as hypertension and valvular disease. The diagnostic DCM case. The mutation affects a highly conserved residue criteria of HCM are summarized in Table 2. The affected status was of the globular head of ␣MyHC and is predicted to alter the defined by the presence of 1 major criterion, 2 minor echocardio- secondary structure of the light-chain binding domain.3,12–14

TABLE 2. Diagnostic Criteria of HCM

Major Criteria Minor Criteria Echocardiography LV wall thickness Ն13 mm in the anterior septum or posterior wall LV wall thickness of 12 mm in the anterior septum or posterior wall or 14 mm or Ն15 mm in the posterior septum or free wall. in the posterior septum or free wall Severe SAM (septal-leaflet contact) Moderate SAM (no leaflet-septal contact) Redundant MV leaflets Electrocardiography LVHϩrepolarization changes Complete BBB or (minor) interventricular conduction defect (in LV leads) TWI in leads I and aVL (Ն3mm) (with QRS–T-wave axis difference Minor repolarization changes in LV leads

Ն30°), V3–V6 (Ն3 mm) or II and III and aVF (Ն5 mm) Deep S wave in V2 (Ͼ25 mm) Abnormal Q wave (Ͼ40 ms or Ͼ25% R wave) in at least 2 leads

from II, III, aVF (in absence of LAFB, V1–V4; or I, aVL, V5–V6) Clinical features None Unexplained chest pain, dyspnea, or syncope SAM indicates systolic anterior motion of the mitral valve; LVH, left ventricular hypertrophy; TWI, T-wave inversion; LAFB, left anterior fascicular block; and BBB, bundle-branch block. See text and the footnote to Table 1 for explanation of other abbreviations. 56 Circulation July 5, 2005

Figure 1. MYH6 mutations detected in 4 families with DCM and HCM. Top, Pedi- grees of families. Filled symbols indicate affected subjects; open symbols with “N” (normal) indicate tested, unaffected sub- jects; open empty symbols indicate sub- jects without history of cardiomyopathy. In genotyped individuals, ϩ and Ϫ indi- cate presence of mutant allele and wild- type allele, respectively. Middle, Elution profiles for subjects carrying wild type (red) and mutation (blue). Bottom, Sequence analysis showing heterozy- gous nucleotide substitution.

In exon 23, an A1004S substitution was found in a different familial HCM. This mutation occurred in a highly conserved sporadic DCM case. The change leads to an alteration in residue of the rod domain and was absent in 2 unaffected polarity: An alanine (hydrophobic/nonpolar) is replaced by a relatives. The A1004S and Q1065H mutations occurred in the serine (hydrophilic/polar) in a highly conserved region of the fourth (g) and second (e) position, respectively, of the heptad rod domain. In exon 24, a glutamine (neutral) to histidine repeat motif of the ␣-helical coiled-coil15 (Figure 2B). In (basic) substitution (Q1065H) was found in a family with exon 31, a glutamic acid (acidic) to a lysine (basic) (E1457K)

Figure 2. A and B, Human ␣MyHC protein structure (Swiss-Prot P13533). Amino acid sequence, functional domains,3,14,15,24–28 and putative disease-causing mutations are shown. Gray shaded areas represent completely conserved sequences among human MyHC proteins.3 Mutated residues are highlighted in red. NBP indicates nucleotide-binding protein. B, S2 and LMM domains with heptad motif.15,25–28 ACD indicates assembly-competent domain. Carniel et al ␣MyHC Mutations in DCM and HCM 57

TABLE 3. Genotypes and Phenotypes of 4 Patients With a Putative Mutation of MYH6 and Their Relatives

Age at DNA Diagnosis, LVEDD, LVFS* LVEF, MWT, Arrhythmias Outcome Family Mutation Codon ID Phenotype Genotype y NYHA* mm* (%) %* mm* and Conduction (Age, y) TSSDC010 2489C3T P830L II:2 DCM ϩ/Ϫ 56 I 61 25 49 11 SVT, VF, LBBB, CHF (75) 1° AVB II:4 Unaffected NA 43† I NA NA NA NA ⅐⅐⅐ NS (53) III:2 Unaffected NA 17† I 43 42 NA 7 ⅐⅐⅐ NS (27) TSSDC021 3010G3T A1004S II:2 DCM ϩ/Ϫ 51 I 73 25 34 11 ⅐⅐⅐ CHF (59) III:1 Unaffected NA 16† I 47 36 63 8 ⅐⅐⅐ NS (24) II:1 Unaffected NA 47† I 50 36 67 10 ⅐⅐⅐ NS (55) TSSDC017 4369G3A E1457K II:2 DCM ϩ/Ϫ 44 II 73 18 31 12 LBBB, 1° AVB, CHF, heart PM transplant (57) III:2 Unaffected Ϫ/Ϫ 22† I 46 39 60 9 ⅐⅐⅐ NS (24) III:1 Unaffected NA 23† I 50 40 63 9 ⅐⅐⅐ NS (31) DNHOCM00 3195G3C Q1065H II:2 HCM ϩ/Ϫ 27 III 54 9 18 23 NSVT, RBBB Refractory CHF, death (45) III:1 Unaffected Ϫ/Ϫ 22† I 53 34 62 11 ⅐⅐⅐ NS (23) III:2 Unaffected Ϫ/Ϫ 15† I 47 38 81 9 ⅐⅐⅐ NS (17) ID indicates identification in pedigree; NYHA, New York Heart Association functional class; MWT, maximal wall thickness; ϩ/Ϫ, presence of the mutant allele; NA, not available; Ϫ/Ϫ, absence of the mutant allele; SVT, supraventricular tachycardia; VF, ventricular fibrillation; 1° AVB, first-degree atrioventricular block; PM, pacemaker; NSVT, nonsustained ventricular tachycardia; RBBB, right bundle-branch block; CHF, congestive heart failure; and NS, no symptoms. Other abbreviations are as defined in text and in the footnotes to Tables 1 and 2. *At enrollment. †Age at assessment. substitution was found in a sporadic DCM case. This muta- Disease Characteristics tion is predicted12 to alter the ␣-helix of the rod domain, Both patients and controls were white. DCM mutation carri- changing the conformation of a 4–amino acid region from an ers had a late onset of the disease (mean age, 50Ϯ6 years), organized ␣-helix to a random-coil pattern.12 All putative mild symptoms, and mild to moderate left ventricular dys- mutations were absent in at least 300 normal control chro- function (Table 3). All had slow progression of the disease mosomes (and absent in an overall number of Ͼ500 chromo- (follow-up, 8 to 19 years). The proband with HCM had an somes tested) and conserved across different species. early onset of the disease and evolution toward dilation and Mutations in lamin A/C, actin, ␤MyHC, troponin T, dysfunction, with death due to refractory heart failure while desmin, ␦-sarcoglycan, and lamina-associated polypeptide-2 awaiting heart transplantation. The family history was signif- genes were excluded in the DCM MYH6 mutation carriers. icant for sudden death at the age of 47 years in the proband’s Mutations in the ␤MyHC, troponin T, and myosin-binding affected mother. The proband’s offspring were clinically protein C genes were excluded in the HCM Q1065H carrier. unaffected and did not carry the mutation (Figure 1, Table 3). In addition to the putative disease-associated mutations, 7 Discussion nonsynonymous single-nucleotide polymorphisms were iden- This study provides genetic evidence of putative MYH6 tified: 6 new (G56R, I275N, A1130T, E1295Q, R1502Q, and mutations in patients that are associated with a spectrum of G1826N) and 1 (A1101V) already reported (http://www. phenotypes ranging from ventricular hypertrophy to dilation ϭ 16 ncbi.nlm.nih.gov/entrez/query.fcgi?db snp). Of these new (Figure 1). At enrollment, DCM mutation carriers (4.3% of variants, G56R was present in 8 different families in the DCM probands) had mild to moderate ventricular dysfunc- studied population, with no segregation with the disease tion at diagnosis and a slow progression of the disease. within the families, and was found in 10 of 150 healthy The Q1065H mutation found in familial HCM was asso- controls. A1130T and E1295Q were found in 3 and 2 ciated with a severe phenotype, characterized by early onset, different families, respectively, without cosegregation with severe hypertrophy, and evolution toward myocardial dila- the disease. The polymorphisms I275N, R1502Q, and tion, severe dysfunction and death in the fifth decade due to G1826S were present in the same Italian DCM family in refractory heart failure in the proband, or sudden death in the healthy relatives of affected subjects. The reported variants proband’s mother. Overall, the currently available data sug- did not meet the criteria for consideration as a disease- gest that ␣MyHC may represent a rare cause of HCM,8 and causing mutation and therefore, were classified as consequently, the small number of carriers limits genotype- polymorphisms. phenotype correlations. 58 Circulation July 5, 2005

The MYH6 mutation previously reported in 1 HCM case8 expression of the fast-contracting ␣MyHC can cause a loss of and the 4 novel putative mutations identified in our study are contractile function as observed in DCM. located in both the head and rod domains of ␣MyHC and In summary, we have provided genetic data suggesting that affect highly conserved residues of the protein (Figure 2).17–21 MYH6 mutations may lead to a spectrum of dilated and P830L occurs in a sharp bend (amino acids 829 to 832), hypertrophic phenotypes, including myocardial hypertrophy which connects a long ␣-helix region,3 the light-chain binding with evolution toward dilation and systolic dysfunction. site, with a short C-terminal ␣-helix.14 The proline-to-lysine Functional studies are currently in progress to clarify the role substitution could alter the binding of myosin light chain, of MYH6 mutations, to determine the mechanisms by which compromise the movement of the light chain on MyHC, and these mutations can lead to the development of a cardiomy- interfere with force generation. opathy phenotype, and to exclude whether any of the muta- The E1457K missense mutation is predicted to alter the tions reported here are rare polymorphisms. structure of the rod domain, its assembly, and its interactions with surrounding molecules. The A1004S and the Q1065H Familial Dilated Cardiomyopathy Registry mutations are expected to interrupt the heptad repeat motif of Research Group the ␣-helical coiled-coil and alter the hydrogen bonds that University of Colorado Cardiovascular Institute: Brian D. Lowes, stabilize the structure of the rod domain.15 MD; Human Medical Genetics Program: Katherine Gowan, MS; ␣ Hospital and University of Trieste, Italy: Mauro Driussi, MD, Giulio The evidence that MyHC, characterized by higher Scherl. ATPase activity and faster contraction,22 is downregulated in failing hearts5 supports the hypothesis that ␣MyHC is critical Acknowledgments for normal myocardial function. In human left ventricles, The authors are supported by grants from the NIH/NHLBI (1RO1 ␣MyHC mRNA represents 20% to 30% of the total MyHC HL69071-01, 5K23 HL67915-02), the Muscular Dystrophy Associ- RNA. However, the abundance of ␣MyHC protein is low, ation USA (PN0007056), and the American Heart Association Ϸ Ͻ (0250271N). We thank the family members for their participation in 7% of total MyHC in nonfailing hearts, decreasing to 1% this study and Stanislav Miertus, PhD, for his critical reading and in failing left ventricles.6 The relatively small amount of suggestions. ␣MyHC protein present in nonfailing left ventricles has called into question the physiological significance of MyHC References isoform changes in failing human ventricles.23 Interesting 1. Taylor MRG, Carniel E, Mestroni L. Familial dilated cardiomyopathy. observations come from studies of myocardial gene expres- Orphanet Databases, 2003. Available at: http://www.orpha.net/data/ patho/GB/uk-FDCardiomyopathy.pdf. Accessed June 8, 2005. 24–26 24 sion in patients with heart failure. Lowes et al studied 2. Richard P, Charron P, Carrier L, Ledeuil C, Cheav T, Pichereau C, 53 subjects (45 with DCM, 8 normal controls) assigned to Benaiche A, Isnard R, Dubourg O, Burban M, Gueffet JP, Millaire A, treatment with ␤-blockers or placebo. Before treatment, Desnos M, Schwartz K, Hainque B, Komajda M; EUROGENE Heart ␣ Failure Project. Hypertrophic cardiomyopathy: distribution of disease patients with DCM had downregulation of MyHC mRNA genes, spectrum of mutations, and implications for a molecular diagnosis and upregulation of ␤MyHC mRNA expression compared strategy. Circulation. 2003;107:2227–2232. with controls. Responders to ␤-blocker therapy had a signif- 3. Weiss A, Schiaffino S, Leinwand LA. Comparative sequence analysis of icant improvement in ejection fraction and functional capac- the complete human sarcomeric myosin heavy chain family: implications ␣ for functional diversity. J Mol Biol. 1999;290:61–75. ity, as well as a significant increase in the amount of MyHC 4. Nakao K, Minobe W, Roden R, Bristow MR, Leinwand LA. Myosin mRNA and a decrease in the level of ␤MyHC mRNA. The heavy chain gene expression in human heart failure. J Clin Invest. 1997; authors concluded that ␤-blocker therapy could reverse a 100:2362–2370. 5. Lowes BD, Minobe W, Abraham WT, Rizeq MN, Bohlmeyer TJ, Quaife pathological fetal gene program, leading to restoration of the RA, Roden RL, Dutcher DL, Robertson AD, Voelkel NF, Badesch DB, fast-contracting ␣MyHC fibers and to a consequent improve- Groves BM, Gilbert EM, Bristow MR. Changes in gene expression in the ment of myocardial function. Interestingly, similar changes intact human heart: down-regulation of ␣-myosin heavy chain in hyper- were also observed in placebo-treated patients who improved trophied, failing ventricular myocardium. J Clin Invest. 1997;100: 2315–2324. 26 spontaneously. Ladenson et al reported the case of a patient 6. Miyata S, Minobe W, Bristow MR, Leinwand LA. Myosin heavy chain with DCM and hypothyroidism: Treatment with levothyrox- isoform expression in the failing and nonfailing human heart. Circ Res. ine led to an improvement in the patient’s clinical and 2000;86:386–390. 7. Geisterfer-Lowrance AA, Christe M, Conner DA, Ingwall JS, Schoen FJ, echocardiographic parameters, associated with an 11-fold Seidman CE, Seidman JG. A mouse model of familial hypertrophic increase in ␣MyHC mRNA level. Similar data were obtained cardiomyopathy. Science. 1996;272:731–734. by Sabbah et al,27 who studied the effect of passive mechan- 8. Niimura H, Patton KK, McKenna WJ, Soults J, Maron BJ, Seidman JG, ical ventricular restraint with the Acorn cardiac support Seidman CE. Sarcomere protein gene mutations in hypertrophic cardio- myopathy of the elderly. Circulation. 2002;105:446–451. device in a dog model of chronic heart failure. At baseline, 9. Mestroni L, Maisch B, McKenna WJ, Schwartz K, Charron P, Rocco C, heart failure dogs had a decreased level of ␣MyHC mRNA Tesson F, Richter A, Wilke A, Komajda M. Guidelines for the study of and an increased amount of ␤MyHC mRNA compared with familial dilated cardiomyopathies. Eur Heart J. 1999;20:93–102. 10. Taylor MRG, Carniel E, Mestroni L. Familial hypertrophic cardiomyop- normal dogs. Therapy with the cardiac support device was athy: clinical features, melecular genetics and molecular genetic testing. associated with improved contractility and normalization of Exp Rev Mol Diagn. 2003;3:89–103. ␣MyHC mRNA levels. Finally, recent data from a study in 11. Strachan T, Read AP. Molecular pathology. In: Strachan T, Read AP. rat myocardium28 demonstrated that even a small amount of Human Molecular Genetics 2, 2nd ed. New York, New York: Wiley-Liss; ␣ 1999:377–399. MyHC may have physiological or biological significance. 12. Combet C, Blanchet C, Geourjon C. NPS@: Network Protein Sequence All of these studies support the hypothesis that decreased Analysis. Trends Biol Sci. 2000;25:147–150. Carniel et al ␣MyHC Mutations in DCM and HCM 59

13. Kneller DG, Cohen FE, Langridge R. Improvements in protein secondary 22. Swynghedauw B. Developmental and functional adaptation of contractile structure prediction by an enhanced neural network. J Mol Biol. 1990; proteins in cardiac and skeletal muscles. Physiol Rev. 1986;66:710–730. 214:171–182. 23. Reiser PJ, Portman MA, Ning XH, Schomisch Moravec C. Human 14. Rayment I, Rypniewski WR, Schmidt-Base K, Smith R, Tomchick DR, cardiac myosin heavy chain isoforms in fetal and failing adult atria and Benning MM, Winkelmann DA, Wesenberg G, Holden HM. Three- ventricles. Am J Physiol Heart Circ Physiol. 2001;280:H1814–H1820. dimensional structure of myosin subfragment-1: a molecular motor. 24. Lowes BD, Gilbert EM, Abraham WT, Minobe WA, Larrabee P, Science. 1993;261:50–58. Ferguson D, Wolfel EE, Lindenfeld J, Tsvetkova T, Robertson AD, 15. Blair E, Redwood C, de Jesus Oliveira M, Moolman-Smook JC, Brink P, Quaife RA, Bristow MR. Myocardial gene expression in dilated cardio- Corfield VA, Ostman-Smith I, Watkins H. Mutations of the light mero- myopathy treated with ␤-blocking agents. N Engl J Med. 2002;346: myosin domain of the ␤-myosin heavy chain rod in hypertrophic cardio- 1357–1365. myopathy. Circ Res. 2002;90:263–269. 25. Yasumura Y, Takemura K, Sakamoto A, Kitakaze M, Miyatake K. 16. National Center for Biotechnology Information. Submission of SNPs to Changes in myocardial gene expression associated with ␤-blocker dbSNP. Available at: http://www.ncbi.nlm.nih.gov/SNP/get_html.cgi? therapy in patients with chronic heart failure. J Card Fail. 2003;9: whichHtmlϭhow_to_submit. Accessed April 30, 2005. 469–474. 17. Weiss A, Leinwand LA. The mammalian myosin heavy chain gene 26. Ladenson PW, Sherman SI, Baughman KL, Ray PE, Feldman AM. family. Annu Rev Cell Dev Biol. 1996;12:417–439. Reversible alterations in myocardial gene expression in a young man with 18. McLachlan AD, Karn J. Periodic features in the amino acid sequence of dilated cardiomyopathy and hypothyroidism. Proc Natl Acad Sci U S A. nematode myosin rod. J Mol Biol. 1983;164:605–626. 1992;89:5251–5255. 19. Li Y, Brown JH, Reshetnikova L, Blazsek A, Farkas L, Nyitray L, Cohen 27. Sabbah HN, Sharov VG, Gupta RC, Mishra S, Rastogi S, Undrovinas AI, C. Visualization of an unstable coiled coil from the scallop myosin rod. Chaudhry PA, Todor A, Mishima T, Tanhehco EJ, Suzuki G. Reversal of Nature. 2003;424:341–345. chronic molecular and cellular abnormalities due to heart failure by 20. Sohn RL, Vikstrom KL, Strauss M, Cohen C, Szent-Gyorgyi AG, passive mechanical ventricular containment. Circ Res. 2003;93: Leinwand LA. A 29 residue region of the sarcomeric myosin rod is 1095–1101. necessary for filament formation. J Mol Biol. 1997;266:317–330. 28. Herron TJ, McDonald KS. Small amounts of ␣-myosin heavy chain 21. Cohen C, Parry DAD. A conserved C-terminal assembly region in isoform expression significantly increase power output of rat cardiac paramyosin and myosin rods. J Struct Biol. 1998;122:180–187. myocyte fragments. Circ Res. 2002;90:1150–1152. Hypertension

Elevated Blood Pressure Linked to Primary Hyperaldosteronism and Impaired Vasodilation in BK Channel–Deficient Mice

Matthias Sausbier, PhD*; Claudia Arntz, PhD*; Iancu Bucurenciu, MD; Hong Zhao, MD, PhD; Xiao-Bo Zhou, MD; Ulrike Sausbier, PhD; Susanne Feil, PhD; Simone Kamm; Kyrill Essin, PhD; Claudia A. Sailer, PhD; Usamah Abdullah; Peter Krippeit-Drews, PhD; Robert Feil, PhD; Franz Hofmann, MD; Hans-Günther Knaus, MD; Chris Kenyon, MD; Michael J. Shipston, PhD; Johan F. Storm, MD, PhD; Winfried Neuhuber, MD; Michael Korth, MD; Rudolf Schubert, MD; Maik Gollasch, MD; Peter Ruth, PhD

Background—Abnormally elevated blood pressure is the most prevalent risk factor for cardiovascular disease. The large-conductance, voltage- and Ca2ϩ-dependent Kϩ (BK) channel has been proposed as an important effector in the control of vascular tone by linking membrane depolarization and local increases in cytosolic Ca2ϩ to hyperpolarizing Kϩ outward currents. However, the BK channel may also affect blood pressure by regulating salt and fluid homeostasis, particularly by adjusting the renin-angiotensin-aldosterone system. Methods and Results—Here we report that deletion of the pore-forming BK channel ␣ subunit leads to a significant blood pressure elevation resulting from hyperaldosteronism accompanied by decreased serum Kϩ levels as well as increased vascular tone in small arteries. In smooth muscle from small arteries, deletion of the BK channel leads to a depolarized membrane potential, a complete lack of membrane hyperpolarizing spontaneous Kϩ outward currents, and an attenuated cGMP vasorelaxation associated with a reduced suppression of Ca2ϩ transients by cGMP. The high level of BK channel expression observed in wild-type adrenal glomerulosa cells, together with unaltered serum renin activities and corticotropin levels in mutant mice, suggests that the hyperaldosteronism results from abnormal adrenal cortical function in BKϪ/Ϫ mice. Conclusions—These results identify previously unknown roles of BK channels in blood pressure regulation and raise the possibility that BK channel dysfunction may underlie specific forms of hyperaldosteronism. (Circulation. 2005;112:60- 68.) Key Words: blood pressure Ⅲ ion channels Ⅲ vasoconstriction Ⅲ vasodilation Ⅲ hyperaldosteronism

ncreased arterial tone is a hallmark of elevated blood by pressure or vasoconstrictors.3 BK channels also mediate Ipressure. During hypertension, pressure-induced depolar- steady hyperpolarization and vasorelaxation as a result of ization of vascular smooth muscle cells contributes to in- transient outward currents carried by BK channels spontane- creased vascular tone by increasing Ca2ϩ influx through ously activated by local release of Ca2ϩ from intracellular voltage-dependent Ca2ϩ channels.1,2 In vitro evidence sug- stores via ryanodine receptor channels.4 The vascular BK gests that the unique large-conductance, voltage- and Ca2ϩ- channel consists of 4 ␣ subunits that form the ion-conducting activated Kϩ (BK) channel limits Ca2ϩ entry and thereby pore and 4 auxiliary ␤1 subunits. The ␤1 subunits, which are arterial contraction by repolarizing smooth muscle cells and restricted to smooth muscle, maintain the normal high voltage closing voltage-dependent Ca2ϩ channels previously opened and Ca2ϩ sensitivity of the pore-forming ␣ subunit.5 The role

Received August 4, 2004; revision received October 12, 2004; accepted October 25, 2004. From Pharmakologie und Toxikologie, Pharmazeutisches Institut der Universität Tübingen, Tübingen, Germany (M.S., C.A., I.B., H.Z., U.S., U.A., P.K.-D., P.R.); Institut für Pharmakologie für Pharmazeuten, Universitätsklinikum Hamburg-Eppendorf, Hamburg-Eppendorf, Germany (X.Z., M.K.); Institut für Pharmakologie und Toxikologie der Technischen Universität München, München, Germany (S.F., S.K., R.F., F.H.); Helios Franz-Volhard- Klinik, Med Klinik für Nephrologie und Intensivmedizin, MDC für Molekulare Medizin, Humboldt Universität Berlin, Berlin, Germany (K.E., M.G.); Institut für Biochemische Pharmakologie, Universität Innsbruck, Innsbruck, Austria (C.A.S., H.K.); Endocrinology Unit, School of Molecular and Clinical Medicine, Molecular Medicine Centre, Western General Hospital, Edinburgh, Scotland (C.K.); Membrane Biology Group, Division of Biomedical Science, University of Edinburgh, Edinburgh, Scotland (M.J.S.); Department of Physiology and Centre for Molecular Biology and Neuroscience, University of Oslo, Oslo, Norway (J.F.S.); Anatomisches Institut, Universität Erlangen-Nürnberg, Erlangen-Nürnberg, Germany (W.N.); and Institut für Physiologie der Universität Rostock, Rostock, Germany (R.S.). *Drs Sausbier and Arntz contributed equally to this work. The online-only Data Supplement can be found with this article at http://circ.ahajournals.org/cgi/content/full/01.CIR.0000156448.74296.FE/DC1. Correspondence to Dr Peter Ruth, Pharmakologie und Toxikologie, Pharmazeutisches Institut der Universität Tübingen, Tübingen, Germany. E-mail [email protected] © 2005 American Heart Association, Inc. Circulation is available at http://www.circulationaha.org DOI: 10.1161/01.CIR.0000156448.74296.FE 60 Sausbier et al BK Channel and Blood Pressure Regulation 61

of the BK channel auxiliary ␤1 subunit in blood pressure Electrophysiology of Tibial Artery and Aortic regulation has been tested previously by deleting its gene in Smooth Muscle Cells mice. The resulting loss of this subunit impaired the coupling For cell isolation, tibial artery, a fourth-order branch of the aorta, or 2ϩ between Ca2ϩ release and the activation of hyperpolarizing aorta was incubated in Ca -free physiological saline solution (PSS) 6,7 containing papain at 37°C for 30 minutes. Then the solution was BK currents, leading to systemic hypertension. Recent ϩ exchanged for PSS containing Ca2 , collagenase type H, and studies raise the possibility that changes in ␤1 subunit hyaluronidase, and digestion was continued for another 10 minutes at expression contribute to the development of hypertension in 37°C. For measuring of outward membrane currents (whole-cell ϩ rat8 and that gain of function mutation in the same subunit mode), the free Ca2 concentration in the pipette solution was 300 Ϫ Ϫ decreases the prevalence of diastolic hypertension in hu- nmol/L. The holding potential was 50 mV (arterial cells) and 20 mV (aortic cells), and test pulses of 300-ms duration were applied 9 ␤ mans. However, even in the absence of functional 1 every 5 seconds. For measuring membrane potentials (whole-cell subunits, the ␣ subunit can still form functional channels, perforated patch), the pipette solution contained nystatin. For record- which might be activated at physiological potentials if their ing of macroscopic Ca2ϩ channel currents, cells were voltage- voltage and Ca2ϩ sensitivity are increased by other factors clamped at a holding potential of Ϫ60 mV, and the potential was 10,11 12,13 stepped, for 300 ms every 5 seconds, in 10-mV increments up to 50 such as endothelial factors and/or phosphorylation. mV. The inward current was measured as peak inward current with Thus, functional BK channels may be operative in blood reference to zero current. vessels even when the ␤1 subunit is lacking. In addition, BK channels in tissues other than vasculature, such as the adrenal Luminal Diameter Analysis of Small Arteries gland,14 may also influence blood pressure regulation. There- by Videomicroscopy fore, we used mice lacking the BK channel ␣ subunit Tibial small arteries were equilibrated in buffer, and an intravascular (BKϪ/Ϫ15 to evaluate the global impact of BK channels on pressure of 80 mm Hg was applied under nonflow conditions. The chamber was continuously perfused at a rate of 2 mL/min with buffer blood pressure regulation. at 37.0Ϯ0.5°C. The small artery was allowed to equilibrate under videomicroscopic recording until a stable myogenic tone spontane- ously developed after 15 to 20 minutes. At the end of each Methods 2ϩ Details are given in the online-only Data Supplement. experiment, Ca -free buffer was applied to determine maximal vessel diameters. No significant difference in maximal vessel diam- Mice eter was detected between the 2 genotypes (WT, 69Ϯ3 ␮m; BKϪ/Ϫ, BKϪ/Ϫ mice were generated as described.15 Wild type (WT) and 63Ϯ2 ␮m). All compounds were administered to the adventitial side BKϪ/Ϫ mice with the hybrid SV129/C57BL6 background (always F2 of the pressurized small arteries. To exclude prostaglandin effects, generation) were used. Either litter- or age-matched animals were the buffer contained 1 ␮mol/L diclofenac. randomly assigned to the experimental procedures undertaken in 2؉ accordance with the German legislation on protection of animals. [Ca ]i Measurements in Aortic Smooth Muscle Cells Single aortic smooth muscle cells (prepared as above) were loaded ␮ 2ϩ Immunohistochemistry of Adrenal Gland with 2.5 mol/L fura 2-AM. For [Ca ]i measurement, cells were For immunofluorescence, on-slide 5-␮m cryostat slices from non- transferred to a glass coverslip coated with 0.01% poly-L-lysine Ϫ/Ϫ solution and continuously superfused with PSS at 36°C at a flow rate fixed WT and BK adrenal glands were incubated with anti- 2ϩ ␣ of 2 to 4 mL/min. [Ca ]i measurements were performed with the use BK (674–1115). BK expression was analyzed with a confocal laser- 2ϩ scanning microscope (Biorad MRC1000 attached to Nikon Diaphot of the dual-wavelength microfluorescence technique. Two [Ca ]i transients were elicited consecutively with a 15-minute interval in 300 and equipped with a krypton-argon laser). For peroxidase/DAB 2ϩ detection, 10-␮m cryosections of WT and BKϪ/Ϫ adrenal glands were between to allow refilling of intracellular Ca stores. perfused with 4% paraformaldehyde and incubated with anti- BK␣ . Data analysis was performed with a Zeiss Axioplan 2 Spontaneous Transient Outward Current and 2؉ (1115–674) microscope equipped with a Zeiss Axio Cam digital camera. Ca Spark Analysis Cerebral arteries were placed in Ca2ϩ-free Hanks’ solution supple- Determination of Renin Activity, Corticotropin, mented with papain for 15 minutes at 36°C. The segments were then and Cortisterone Levels From Serum placed in Hanks’ solution containing collagenase type F and H (ratio 30% and 70%, respectively) and 0.1 mmol/L CaCl2 for 6 minutes at Plasma renin concentration was measured as the generation of 36°C. After several washes in Ca2ϩ-free Hanks’ solution, single cells angiotensin I (ng/mL per hour) when plasma samples were incubated were isolated. Spontaneous transient outward currents (STOCs) were with excess renin substrate (plasma with no intrinsic renin activity measured in the perforated patch mode. Holding potential was set at from a binephrectomized rat). Angiotensin I was measured by Ϫ60 mV. The pipette solution contained 250 ␮g/mL amphotericin. radioimmunoassay as previously described.16 For the determination To measure Ca2ϩ sparks, cells were incubated with fluo 3-AM for 30 of corticotropin in plasma, we used a 2-site solid-phase immunora- minutes at room temperature in Ca2ϩ-free Hanks’ solution. Ca2ϩ diometric assay (IRMA) kit (Euria-acth IRMA kit) from Euro-Diag- sparks were measured as local fractional fluorescence increases nostica AB. This assay measures intact corticotropin (1–39); in the (F/F0) by confocal fluorescence microscopy at room temperature. assay used, the limit of detection was 5 pg/mL and the intra-assay The baseline fluorescence (F0) was determined by averaging line- variation Ͻ5%. Serum corticosterone was analyzed with an in-house scan images in the absence of Ca2ϩ sparks. specific radioimmunoassay, as described previously, and modified for microtiter plate scintillation proximity assay.17 Long-Term Radiotelemetric Blood Pressure Analysis Mean arterial blood pressure (MAP), heart rate (HR), and Serum Electrolytes and Serum Aldosterone Levels physical activity were analyzed in conscious male WT and BKϪ/Ϫ Serum was separated from nonheparinized blood collected by heart mice (nϭ7 for each genotype). Mice (aged 3 to 4 months), either puncture from WT and BKϪ/Ϫ mice euthanized by inhalation of litter- or age-matched, did not significantly differ in body weight carbon dioxide. The serum concentrations of Naϩ and Kϩ were (26.2Ϯ1.3 and 23.5Ϯ0.7 g, respectively). Mice were anesthetized measured by flame photometry. Serum concentrations of aldosterone with isoflurane. A ventral midline incision was performed before were measured by radioimmunoassay. careful isolation of the left common carotid artery. For ligation 62 Circulation July 5, 2005

Figure 1. Resting membrane potential and current density of L-type Ca2ϩ cur- rent in tibial artery smooth muscle cells. a, Current-voltage relationship of Kϩ outward currents in cells from 3 WT (blue circles) and 3 BKϪ/Ϫ (red squares) tibial arteries. Whole-cell currents were 2ϩ measured at 300 nmol/L [Ca ]i from a holding potential of Ϫ50mV without (blue circles, red squares) and with 300 nmol/L iberiotoxin (respective data points are covered by the red squares); nϭ11 (WT) and nϭ15 (BKϪ/Ϫ) cells. b, Membrane potential recordings from WT and BKϪ/Ϫ tibial artery cells with and without iberiotoxin and meanϮSEM membrane potential statis- tics (13 and 18 cells from WT and BKϪ/Ϫ mice, respectively); *PϽ0.05. c, Amplitudes of voltage-gated Ca2ϩ channel currents in WT and BKϪ/Ϫ cells do not differ. Currents were measured in the whole-cell configuration, and barium was used as charge carrier. Inward currents were evoked by step depolarizations (300-ms duration) of increasing amplitude from a holding potential of Ϫ60 mV up to ϩ50 mV in 10-mV increments. Current-voltage relationships of peak inward currents are shown (nϭ11 from 3 WT and nϭ7 from 3 BKϪ/Ϫ mice). MeanϮSEM current densities are plotted against the respective test potential. Inset, Averaged current trac- ings before (ctr) and after superfusion of a WT cell with 1 ␮mol/L nifedipine. Inward currents were activated by step depolariza- tions from Ϫ60 to ϩ10 mV and maintained for 300 ms. and retraction, 2 silk ligatures were passed under the vessel, one excluding a compensatory downregulation of this channel in Ϸ0.8 cm caudal to the bifurcation of the interior and exterior BKϪ/Ϫ arteries (Figure 1c). It is conceivable that the sustained Ϸ carotid artery and another 0.5 cm rostral from the caudal depolarization of BKϪ/Ϫ arterial cells compared with WT cells ligature. A tiny incision in the carotid artery was made for 2ϩ insertion of the catheter. The inserted catheter tip was advanced to could produce a sustained rise in cytosolic Ca because of a the thoracic aorta and fixed with suture. A subcutaneous pocket “window current” caused by incomplete inactivation of was performed along the right flank for placing the transmitter L-type Ca2ϩ channels.19 Interestingly, the steady state calcium body. After subcutaneous fixation of the transmitter, the incision window in smooth muscle cells was maximal at Ϫ30 mV,19 was closed with 6-0 silk. MAP, HR, and physical activity were which corresponds closely to the membrane potential in recorded at days 5 to 7 after surgery, when mice have regained Ϫ/Ϫ normal locomotor activity. Radiotelemetric signals were sampled BK cells. Ϫ/Ϫ for 1 minute at 5-minute intervals. Thus, the more depolarized membrane potential in BK cells may influence small-artery tone, which is obligatorily 2ϩ Results dependent on pressure-evoked depolarization triggering Ca entry through voltage-gated Ca2ϩ channels.20 In fact, at Increased Myogenic Tone and Attenuated physiological relevant pressures (80 and 120 mm Hg), tibial cGMP-Dependent Vasorelaxation in arteries from BKϪ/Ϫ mice showed increased myogenic tone ؊ ؊ BK / Small Arteries compared with arteries from WT mice (Figure 2a). However, In smooth muscle cells from the tibial artery of WT mice, the myogenic response, ie, the change in myogenic tone depolarization elicited large, iberiotoxin-sensitive outward induced by pressure steps, was not significantly different in ϩ K currents, but no such currents were detected in cells from BKϪ/Ϫ versus WT tibial arteries (Figure 2b), suggesting that Ϫ Ϫ Ϫ Ϫ BK / cells (Figure 1a). Furthermore, BK / cells did not BK channels do not determine the myogenic response per se. exhibit spontaneous hyperpolarizations and showed a more Apparently, membrane depolarization and increases in global Ϫ Ϯ 2ϩ 2ϩ depolarized mean membrane potential ( 31.0 2.3 mV) than [Ca ]i plus local Ca sparks in response to transmural WT cells (Ϫ41.0Ϯ3.9 mV). Depolarization was also ob- pressure are not sufficient to evoke a BK channel–mediated served in WT cells when BK channels were specifically negative feedback regulation of pressure-induced contraction. blocked by iberiotoxin (Figure 1b). The BK channel has Rather, BK channels seem to mediate essentially steady recently been found to be physically and functionally asso- dilatation over a large pressure range, inasmuch as their ciated with the L-type Ca2ϩ channel.18 In the arterial cells, deletion produced an almost parallel upward shift of the however, the expression of nifedipine-sensitive L-type Ca2ϩ pressure-tone relationship (Figure 2a). channels was not changed, as revealed by almost identical It is well established that the myogenic tone can be current densities in BKϪ/Ϫ and WT arterial cells, thereby decreased by activation of the NO/cGMP/protein kinase G Sausbier et al BK Channel and Blood Pressure Regulation 63

Figure 2. Increased myogenic tone and attenuated cGMP-dependent vasorelax- ation in BKϪ/Ϫ small arteries. a, Effect of transmural pressure on myogenic tone in tibial artery. Myogenic tone was calcu- lated as (1Ϫd/dmax), where d is the vessel diameter at a certain pressure and dmax the diameter in Ca2ϩ-free buffer at that pressure. b, Effect of pressure steps on myogenic tone change (⌬myogenic tone, ie, tone at end pressure minus tone at initial pressure of the step); nϭ4WTand 4BKϪ/Ϫ arteries. c, cGMP- and adeno- sine-induced relaxation of myogenic tone of tibial artery. Relaxation was calculated as drug-induced change of inner vessel diameter as a percentage of maximal diameter in Ca2ϩ-free buffer; nϭ6 arteries per genotype. All data are meanϮSEM; *PϽ0.05; **PϽ0.01.

(PKG) and cAMP/protein kinase A (PKA) pathways.21,22 transients was almost identical in BKϪ/Ϫ (nϭ76) and WT cells Application of 8-pCPT-cGMP, a cell-permeable cGMP ana- (nϭ64) (2.1Ϯ0.1 and 2.2Ϯ0.2 arbitrary units, respectively). logue that preferentially activates PKG, produced signifi- Additionally, the ratio between consecutively elicited Ca2ϩ cantly less relaxation of BKϪ/Ϫ arteries than WT arteries transients (AUC 2/AUC 1) was not significantly different in Ϫ/Ϫ (Figure 2c). In contrast, adenosine, which raises [cAMP]i via BK and WT cells (87Ϯ8% and 103Ϯ7%, respectively) 23 Ϫ/Ϫ A2 receptors, was equally potent in WT and BK arteries, (Figure 3). Preincubation with 8-Br-cGMP diminished the suggesting that the cGMP/PKG, but not the cAMP/PKA, second transients in WT cells to 38Ϯ7% of control, whereas pathway requires BK channel activation to promote arterial BKϪ/Ϫ cells were less affected (72Ϯ11%). The attenuated relaxation (Figure 2c). The partial relaxation produced by suppression in BKϪ/Ϫ cells suggests that BK channel activa- 8-pCPT-cGMP in BKϪ/Ϫ arteries may involve other PKG tion via cGMP/PKG is important for relaxation, probably substrates or cross-activation of PKA at high agonist involving hyperpolarization-induced inhibition of voltage- concentrations. gated Ca2ϩ channels. The small residual cGMP-induced Blood vessel relaxation by cGMP/PKG has been ascribed suppression of Ca2ϩ transients in BKϪ/Ϫ smooth muscle cells to suppression of Ca2ϩ transients.24 To test this hypothesis, may be caused by cGMP-mediated inhibition of Ca2ϩ release Ϫ/Ϫ 26 aortic smooth muscle cells from WT and BK mice, from IP3-sensitive stores. basically exhibiting the same differences in membrane poten- The amount of Ca2ϩ in intracellular stores of aortic cells tial as cells from tibial artery (Data Supplement Figure), were was apparently not changed by the absence of the BK 2ϩ stimulated with the ␣1-adrenergic receptor agonist phenyleph- channel. Caffeine at 10 mmol/L stimulated Ca release from rine, and the resulting Ca2ϩ transients (due to Ca2ϩ release and the sarcoplasmic reticulum to a similar extent in BKϪ/Ϫ influx via Ca2ϩ channels25) were measured in the absence of (3.6Ϯ0.2 arbitrary units of AUC; nϭ108) and WT cells 8-Br-cGMP. The area under the curve (AUC) of the Ca2ϩ (3.4Ϯ0.2; nϭ92).

Figure 3. Reduced suppression of Ca2ϩ transients by cGMP in BKϪ/Ϫ aortic smooth muscle cells. Left, Consecutive 2ϩ [Ca ]i transients (first and second tran- sients are shown), elicited by phenyleph- rine (10 ␮mol/L for 1 minute, arrows) at 15-minute intervals, in WT (blue) and BKϪ/Ϫ (red) aortic cells, after preincuba- tion with or without 8-Br-cGMP 2ϩ (1 mmol/L, 5 minutes). Right, [Ca ]i tran- sients were quantified as AUC (40 to 63 cells per genotype). All data are meanϮSEM; *PϽ0.05; **PϽ0.01. 64 Circulation July 5, 2005

Figure 4. Lack of STOCs but presence of normal Ca2ϩ sparks in BKϪ/Ϫ cerebral artery smooth muscle cells. a, STOC activity in WT and BKϪ/Ϫ cerebral arterial cells recorded at increasing membrane potentials and STOC frequency at Ϫ20 mV from 5 to 7 WT and BKϪ/Ϫ cells. b, Confocal line scans of fluo 3–loaded WT and BKϪ/Ϫ cells and time course of corresponding Ca2ϩ sparks. Spark amplitudes were measured as local fractional fluorescence increases (F/F0;F0 is baseline). Spark duration was measured at half-maximal ampli- tude; nϭ5 to 12 cells per genotype. All data are meanϮSEM; *PϽ0.05; **PϽ0.01.

Lack of STOCs in BK؊/؊ Arterial Cells Did Not (Figure 4a), implicating that STOCs are carried by BK Affect Ca2؉ Sparks channels. However, Ca2ϩ spark parameters were not affected Ϫ Ϫ Transient activation of outward hyperpolarizing currents in BK / cells, arguing against a feedback between STOCs 2ϩ 2ϩ (STOCs), carried by BK channels, has been proposed as a and Ca sparks via Ca entry (Figure 4b). In summary, 2 mechanism for the paradoxical Ca2ϩ spark–mediated relax- mechanisms important for determining vascular tone were Ϫ/Ϫ ation of small arteries.27 Furthermore, STOCs have been absent in BK small arteries: (1) BK channel activity as a major effector of cGMP/PKG-mediated relaxation and (2) the proposed to provide a negative feedback mechanism to 2ϩ ϩ steady hyperpolarizing effect in response to Ca sparks. control Ca2 spark dynamics through regulation of voltage- Hence, pathophysiological consequences for systemic blood dependent calcium entry and resultant changes in sarcoplas- Ϫ Ϫ pressure were expected in BK / mice. mic reticulum Ca2ϩ loading and ryanodine receptor sensitiza- ϩ Ϫ Ϫ tion. We examined STOCs and Ca2 spark activity in BK / Elevated Blood Pressure and Hyperaldosteronism Ϫ Ϫ cells. As expected, STOCs were completely absent in BK / in BK؊/؊ Mice cerebral arterial cells, even at depolarized membrane poten- MAP in the carotid artery, HR, and locomotor activity were tials that promote Ca2ϩ influx via voltage-gated Ca2ϩ channels measured by radiotelemetry in male mice. BKϪ/Ϫ mice were Sausbier et al BK Channel and Blood Pressure Regulation 65

locomotor activity because of their ataxia.15 In contrast to 6,7 Ϫ/Ϫ BK␤1 knockouts, the increase in blood pressure in BK mice was not accompanied by significant cardiac hypertrophy in 4- to 6-month-old animals, as indicated by heart/body weight ratio (WT, 4.6Ϯ0.2 mg/g; BKϪ/Ϫ, 4.8Ϯ0.1 mg/g) (Figure 5d). Nevertheless, the elevated MAP observed under resting conditions in BKϪ/Ϫ mice exhibiting HRs similar to those of WT mice supports the hypothesis that vascular BK channels may be important for the control of vascular tone and systemic blood pressure in vivo, mainly because of their effector role for both cGMP signaling and Ca2ϩ sparks. However, blood pressure regulation also involves essential endocrine mechanisms such as the renin-angiotensin-aldoste- rone system, which might also be affected by BK channels.28 Analysis of the serum electrolytes revealed a gender- ϩ ϩ independent decrease in the K concentration ([K ]serum)in mutants compared with WT (male WT 8.3Ϯ0.3 mmol/L versus male BKϪ/Ϫ 6.5Ϯ0.8 mmol/L; female WT 9.2Ϯ0.2 mmol/L versus female BKϪ/Ϫ 7.5Ϯ0.4 mmol/L). In ϩ general, [K ]serum values of WT mice were in agreement with ϩ previously published [K ]serum values of mice with a genetic Ϫ/Ϫ 29 ϩ background similar to the BK mice. However, [Na ]serum did not differ between the 2 genotypes (male WT 140Ϯ4 mmol/L versus male BKϪ/Ϫ 137Ϯ2 mmol/L; female WT 137Ϯ2 mmol/L versus female BKϪ/Ϫ 141Ϯ2 mmol/L) ϩ Ϫ/Ϫ (Figure 6a). The lower [K ]serum of BK mice was associated with markedly increased serum aldosterone levels in both genders (male WT 185Ϯ24 pg/mL versus male BKϪ/Ϫ 573Ϯ108 pg/mL; female WT 236Ϯ33 pg/mL versus female BKϪ/Ϫ 597Ϯ132 pg/mL) (Figure 6b). Consistent with this phenotype, we found high levels of BK channel ␣ subunit expression in the zona glomerulosa (Figure 6c). Here, these channels may control aldosterone production and/or release Figure 5. Elevated arterial blood pressure in BKϪ/Ϫ mice. a, by influencing the membrane potential and hence the open Locomotor activity recorded by radiotelemetry for 24 hours at probability of voltage-gated Ca2ϩ channels (L- and/or T-type) days 5, 6, and 7 after surgery. Bar indicates period of MAP and via the membrane potential.30,31 In contrast to glomerulosa HR measurement shown below; open bar indicates dark phase. b, Mean values of MAP, HR, and activity from telemetric blood cells, cortisol-synthesizing cells from zona fasciculata were pressure analysis are calculated hour by hour between 12 AM very weakly stained with the BK channel antibody (Figure Ϫ Ϫ and 6 PM, the period when WT and BK / mice exhibited com- 6c). In agreement herewith is the finding that corticosterone ϭ parable locomotor activity; n 7 per genotype. c, Correlations Ϫ/Ϫ between locomotor activity and MAP. Linear regression lines levels in the serum were not altered between WT and BK were yϭ4.56xϩ97.5 (R2ϭ0.85) (WT) and yϭ4.44xϩ103.6 mice (Figure 6d). A potential causative role of the renin/an- (R2ϭ0.69) (BKϪ/Ϫ). d, Absence of cardiac hypertrophy in BKϪ/Ϫ giotensin system and/or the pituitary for the observed in- mice. Statistics of heart weight (Hw)/body weight (Bw) (mg/g) crease in aldosterone was also considered. The determination from 10 to 11 mice per genotype are shown. All data are meanϮSEM; *PϽ0.05; **PϽ0.01. of serum corticotropin and serum renin activity, however, revealed no significant changes of these 2 parameters be- tween both genotypes, suggesting that the hyperaldosteron- significantly less active during the dark phase as a result of Ϫ/Ϫ 15 ism found in BK mice was of a primary nature. Thus, motor impairment (Figure 5a). Because physical activity Ϫ Ϫ elevation of blood pressure in BK / mice may be attributed affects MAP, we measured it during an interval (12 AM to 6 to both vascular and hormonal dysfunctions. PM) when the 2 genotypes showed similar activity (Figure 5a, Ϫ/Ϫ 5b). The BK mice showed a significantly (5.4 mm Hg) Discussion higher MAP than WT mice (Figure 5b). Additionally, dia- The deletion of the BK channel ␣ subunit permitted the stolic and systolic blood pressures were significantly elevated identification of physiological functions of this unique chan- in the mutants, whereas HR did not differ. This excludes the nel in regulating arterial blood pressure. BKϪ/Ϫ mice exhibit possibility that higher sympathetic tone and HR caused MAP a moderate increase in blood pressure that was traced back to Ϫ Ϫ elevation. Analysis of the MAP of BK / and WT mice as a vascular and endocrine abnormalities. Our data indicate that function of locomotor activity suggested that for any activity the vascular abnormalities arise from lack of spontaneous the MAP was shifted upward by 6.1 mm Hg in the mutants outward currents that contribute to the mean resting mem- (Figure 5c). However, BKϪ/Ϫ mice did not show high-range brane potentials in small and large vessels (Figure 1b; Data 66 Circulation July 5, 2005

Figure 6. Primary hyperaldosteronism in BKϪ/Ϫ mice and BK channel expression in adrenal glomerulosa cells. a, Statistics of serum electro- lyte concentrations from 6 to 9 male (M) and female (F) mice per genotype (WT, blue; BKϪ/Ϫ, red). b, Statistical analysis of serum aldosterone from 6 to 9 mice per genotype and gender. c, Left, Immunohistochemical detection of BK channels in the adrenal gland by peroxidase/DAB staining. Prominent staining is found in the zona glomerulosa (G), very weak staining in the zona fasciculata (F) and reticularis (R), and inter- mediate staining in the adrenal medulla (M). Sections from BKϪ/Ϫ adrenal were not stained under identical conditions. Right, Confocal single optical section demonstrates specific BK immunofluorescence on the surface (dashed line) of zona glomerulosa cells (arrows) of adrenal cor- tex (barϭ100 ␮m). BKϪ/Ϫ sections showed no staining (not shown). d, Statistical analysis of serum renin activity, corticosterone, and cortico- tropin (ACTH) levels from 8 to 16 mice per genotype. All data are meanϮSEM; *PϽ0.05; **PϽ0.01.

Supplement Figure, panel b). Thus, the lack of BK channels several mouse models with genetic ablations were used, have may account for the observed membrane depolarization shown that this pathway contributes to basal blood pressure (Figure 1b), which will tend to increase myogenic tone, ie, regulation.24,32,33 By contrast, cAMP-mediated relaxation of vessel tone in response to intravascular pressure (Bayliss small arteries involves effectors other than the BK channel, effect; Figure 2). The pathophysiology of an increased although cAMP kinase has also been shown to control BK vascular tone in BKϪ/Ϫ mice apparently includes several channel activity in vascular smooth muscle cells.34 mechanisms. First, we found that relaxation of myogenic tone Second, BK channel deficiency abolished the transient by cGMP is impaired in BKϪ/Ϫ vessels. This finding indicates outward Kϩ currents that are induced by local Ca2ϩ release that the vascular BK channel is an important effector of the from internal stores or, alternatively, by Ca2ϩ influx through cGMP/cGMP kinase pathway. Previous studies, in which T-type Ca2ϩ channels, a mechanism that recently was pro- Sausbier et al BK Channel and Blood Pressure Regulation 67 posed to operate in coronary vessels.35 Regardless of the Ca2ϩ tional pathways. Presumably, a selective deletion of only the source, the lack of transient Kϩ currents should increase the vascular BK␣ subunit would produce an even more severe 2ϩ Ϫ/Ϫ open probability of voltage-gated Ca channels, thus con- blood pressure elevation than BK␤1 deletion. However, in tributing to vasoconstriction of small arteries. In such arteries, terms of revealing the basis of genetically anchored diseases myogenic tone in response to intravascular pressure was and identifying susceptibility genes contributing to hyperten- absent when L-type Ca2ϩ channels of the Cav1.2 type were sion, the condition of a naturally occurring loss of function inactivated specifically in smooth muscle.36 This suggests mutation of the BK␣ gene is best reflected by the general BK that the increased myogenic tone in BKϪ/Ϫ arteries is medi- channel knockout as presented here. ated by enhanced Ca2ϩ influx through Cav1.2 channels, In conclusion, the results of this study indicate that the presumably because of the less negative membrane potential systemic blood pressure phenotype of complete BK channel of arterial muscle cells and an attendant increase of the deletion is relatively mild—milder than expected—at least Cav1.2 “window current.”19 The subsequent increase of under resting conditions. We hypothesize that the renovascu- 2ϩ global [Ca ]i, however, seems to be insufficient for stimu- lar role of the BK channel, which often acts as an “emergency lating the frequency or intensity of Ca2ϩ sparks evoked from brake”,41,42 may become more evident when studying micro- intracellular stores (Figure 4b). Apart from these vascular vasculature function and organ perfusions in BKϪ/Ϫ animals dysfunctions, we found a marked hyperaldosteronism that that are challenged with pathophysiological conditions such probably contributes to the blood pressure elevation in BKϪ/Ϫ as ischemia. mice. Aldosterone, via the mineralocorticoid receptor, is the major regulator of ENaC expression and activity in the Acknowledgments cortical collecting duct, thereby providing renal Naϩ reab- We thank Dan-Yang Huang and Dr Volker Vallon for assisting us in sorption, which also necessitates increased water reabsorption flame photometry analysis, Isolde Breuning and Janina Smykowsky to maintain Naϩ concentration at or near 140 mmol/L. Several for excellent technical assistance, and Deutsche Forschungsgemein- schaft, Fonds zur Förderung der Wissenschaftlichen Forschung, The rare forms of inherited hypertension are based on monoge- Wellcome Trust, Research Council of Norway, Thyssen-Stiftung and netic defects associated with increased mineralocorticoid Schilling Foundation for financial support. synthesis or dysfunctional aldosterone signaling (for a re- view, see Lifton et al37). We also found that BK channel References protein is localized in the cortices of adrenal glands and is 1. Harder DR, Smeda J, Lombard J. Enhanced myogenic depolarization in highest in zona glomerulosa cells therein. However, the hypertensive cerebral arterial muscle. Circ Res. 1985;57:319–322. 2. Wellman GC, Cartin L, Eckman DM, Stevenson AS, Saundry CM, physiological contribution of BK channels to aldosterone Lederer WJ, Nelson MT. Membrane depolarization, elevated Ca(2ϩ) secretion from this cell layer is controversial.28,38 Both T-type entry, and gene expression in cerebral arteries of hypertensive rats. and L-type voltage-dependent Ca2ϩ channels are expressed in Am J Physiol. 2001;281:H2559–H2567. 3. Brayden JE, Nelson MT. Regulation of arterial tone by activation of glomerulosa cells and have been implicated in steroidogene- calcium-dependent potassium channels. Science. 1992;256:532–535. sis and aldosterone secretion.30,31,39,40 Interestingly, both 4. Nelson MT, Cheng H, Rubart M, Santana LF, Bonev AD, Knot HJ, types have been described to be functionally associated with Lederer WJ. Relaxation of arterial smooth muscle by calcium sparks. BK channels in other cell types.18,35 Thus, loss of control of Science. 1995;270:633–637. 2ϩ 2ϩ 5. Wallner M, Meera P, Ottolia M, Kaczorowski GJ, Latorre R, Garcia ML, Ca influx through Ca channels in glomerulosa cells may Stefani E, Toro L. Characterization of and modulation by a beta-subunit be a plausible mechanism for the observed hyperaldosteron- of a human maxi KCa channel cloned from myometrium. Receptors ism in BKϪ/Ϫ mice. In support of this idea is the finding that Channels. 1995;3:185–199. neither the renin/angiotensin pathway nor pituitary hormones 6. Brenner R, Perez GJ, Bonev AD, Eckman DM, Kosek JC, Wiler SW, Patterson AJ, Nelson MT, Aldrich RW. Vasoregulation by the beta1 seem to evoke the hyperaldosteronism. subunit of the calcium-activated potassium channel. Nature. 2000;407: Ϫ Ϫ Considering that the BK / mice showed several synergis- 870–876. tic mechanisms that all tend to increase blood pressure, it 7. Plüger S, Faulhaber J, Fürstenau M, Löhn M, Waldschütz R, Gollasch M, Haller H, Luft FC, Ehmke H, Pongs O. Mice with disrupted BK channel seems surprising that their blood pressure elevation was not beta1 subunit gene feature abnormal Ca(2ϩ) spark/STOC coupling and more severe. The relative mildness of their blood pressure elevated blood pressure. Circ Res. 2000;87:E53–E60. elevation was confirmed by the absence of significant cardiac 8. Amberg GC, Santana LF. Downregulation of the BK channel beta1 hypertrophy in 4- to 6-month-old mutants (Figure 5d). We subunit in genetic hypertension. Circ Res. 2003;93:965–971. 9. Fernández-Fernández JM, Tomás M, Vazquez E, Orio P, Latorre R, Sentí hypothesize that other peripheral or central regulators that are M, Marrugat J, Valverde MA. Gain-of-function mutation in the Ϫ Ϫ able to dampen blood pressure may be more active in BK / KCNMB1 potassium channel subunit is associated with low prevalence than in WT mice. This idea is supported by the finding that of diastolic hypertension. J Clin Invest. 2004;113:1032–1039. deletion of the less widespread regulatory BK channel ␤1 10. Fisslthaler B, Popp R, Kiss L, Potente M, Harder DR, Fleming I, Busse R. Cytochrome P450 2C is an EDHF synthase in coronary arteries. subunit caused a more severe blood pressure elevation6,7 (and Nature. 1999;401:493–497. consequently also significant cardiac hypertrophy in mutant 11. Archer SL, Gragasin FS, Wu X, Wang S, McMurtry S, Kim DH, Platonov mice6) than the deletion of the BK␣ subunit in the present M, Koshal A, Hashimoto K, Campbell WB, Falck JR, Michelakis ED. ␤ Endothelium-derived hyperpolarizing factor in human internal mammary study. The predominant expression of the BK 1 subunit in artery is 11,12-epoxyeicosatrienoic acid and causes relaxation by acti- 5 Ϫ/Ϫ smooth muscle suggests that BK␤1 mice rather than vating smooth muscle BK(Ca) channels. Circulation. 2003;107:769–776. BK␣Ϫ/Ϫ mice represent the more selective “vascular” BK 12. Han G, Kryman JP, McMillin PJ, White RE, Carrier GO. A novel transduction mechanism mediating dopamine-induced vascular relax- channel deletion, although lack of the BK␤1 subunit leaves ␣ ation: opening of BKCa channels by cyclic AMP-induced stimulation of intact subunits that can be still activated at physiological the cyclic GMP-dependent protein kinase. J Cardiovasc Pharmacol. ϩ voltages by high local Ca2 , cGMP/cGMP kinase, and addi- 1999;34:619–627. 68 Circulation July 5, 2005

13. Zhou XB, Arntz C, Kamm S, Motejlek K, Sausbier U, Wang GX, Ruth 28. Ganz MB, Nee JJ, Isales CM, Barrett PQ. Atrial natriuretic peptide P, Korth M. A molecular switch for specific stimulation of the BKCa enhances activity of potassium conductance in adrenal glomerulosa cells. channel by cGMP and cAMP kinase. J Biol Chem. 2001;276: Am J Physiol. 1994;266:C1357–C1365. 43239–43245. 29. McDonald FJ, Yang B, Hrstka RF, Drummond HA, Tarr DE, McCray PB 14. Payet MD, Bilodeau L, Drolet P, Ibarrondo J, Guillon G, Gallo-Payet N. Jr, Stokes JB, Welsh MJ, Williamson RA. Disruption of the beta subunit Modulation of a Ca(2ϩ)-activated Kϩ channel by angiotensin II in rat of the epithelial Naϩ channel in mice: hyperkalemia and neonatal death adrenal glomerulosa cells: involvement of a G protein. Mol Endocrinol. associated with a pseudohypoaldosteronism phenotype. Proc Natl Acad 1995;9:935–947. Sci U S A. 1999;96:1727–1731. 15. Sausbier M, Hu H, Arntz C, Feil S, Kamm S, Adelsberger H, Sausbier U, 30. Hausdorff WP, Catt KJ. Activation of dihydropyridine-sensitive calcium Sailer CA, Feil R, Hofmann F, Korth M, Shipston MJ, Knaus HG, Wolfer channels and biphasic cytosolic calcium responses by angiotensin II in rat DP, Pedroarena CM, Storm JF, Ruth P. Cerebellar ataxia and Purkinje adrenal glomerulosa cells. Endocrinology. 1988;123:2818–2826. cell dysfunction caused by Ca2ϩ-activated Kϩ channel deficiency. Proc 31. Lotshaw DP. Role of membrane depolarization and T-type Ca2ϩ Natl Acad Sci U S A. 2004;101:9474–9478. channels in angiotensin II and Kϩ stimulated aldosterone secretion. Mol 16. Millar JA, Leckie BJ, Morton JJ, Jordan J, Tree M. A microassay for Cell Endocrinol. 2001;175:157–171. active and total renin concentration in human plasma based on antibody 32. Huang PL, Huang Z, Mashimo H, Bloch KD, Moskowitz MA, Bevan JA, trapping. Clin Chim Acta. 1980;101:5–15. Fishman MC. Hypertension in mice lacking the gene for endothelial nitric 17. MacPhee IA, Antoni FA, Mason DW. Spontaneous recovery of rats from oxide synthase. Nature. 1995;377:239–242. experimental allergic encephalomyelitis is dependent on regulation of the 33. Lopez MJ, Wong SK, Kishimoto I, Dubois S, Mach V, Friesen J, Garbers immune system by endogenous adrenal corticosteroids. J Exp Med. 1989; DL, Beuve A. Salt-resistant hypertension in mice lacking the guanylyl 169:431–445. cyclase-A receptor for atrial natriuretic peptide. Nature. 1995;378:65–68. 18. Liu G, Shi J, Yang L, Cao L, Park SM, Cui J, Marx SO. Assembly of a 34. Barman SA, Zhu S, White RE. Protein kinase C inhibits BKCa channel Ca(2ϩ)-dependent BK channel signaling complex by binding to beta2 activity in pulmonary arterial smooth muscle. Am J Physiol. 2004;286: adrenergic receptor. EMBO J. 2004;23:2196–2205. L149–L155. 19. Fleischmann BK, Murray RK, Kotlikoff MI. Voltage window for sus- 35. Chen CC, Lamping KG, Nuno DW, Barresi R, Prouty SJ, Lavoie JL, tained elevation of cytosolic calcium in smooth muscle cells. Proc Natl Cribbs LL, England SK, Sigmund CD, Weiss RM, Williamson RA, Hill Acad Sci U S A. 1994;91:11914–11918. JA, Campbell KP. Abnormal coronary function in mice deficient in 20. Schubert R, Mulvany MJ. The myogenic response: established facts and alpha1H T-type Ca2ϩ channels. Science. 2003;302:1416–1418. attractive hypotheses. Clin Sci (Lond). 1999;96:313–326. 36. Moosmang S, Schulla V, Welling A, Feil R, Feil S, Wegener JW, 21. Schubert R, Serebryakov VN, Mewes H, Hopp HH. Iloprost dilates rat Hofmann F, Klugbauer N. Dominant role of smooth muscle L-type small arteries: role of K(ATP)- and K(Ca)-channel activation by cAMP- dependent protein kinase Am J Physiol. 1997;272:H1147–H1156. calcium channel Cav1.2 for blood pressure regulation. EMBO J. 2003; 22. Sausbier M, Schubert R, Voigt V, Hirneiss C, Pfeifer A, Korth M, 22:6027–6034. Kleppisch T, Ruth P, Hofmann F. Mechanisms of NO/cGMP-dependent 37. Lifton RP, Gharavi AG, Geller DS. Molecular mechanisms of human vasorelaxation. Circ Res. 2000;87:825–830. hypertension. Cell. 2001;104:545–556. ϩ ϩ 23. Kleppisch T, Nelson MT. Adenosine activates ATP-sensitive potassium 38. Lotshaw DP. Effects of K channel blockers on K channels, membrane channels in arterial myocytes via A2 receptors and cAMP-dependent potential, and aldosterone secretion in rat adrenal zona glomerulosa cells. protein kinase. Proc Natl Acad Sci U S A. 1995;92:12441–12445. Endocrinology. 1997;138:4167–4175. 24. Pfeifer A, Klatt P, Massberg S, Ny L, Sausbier M, Hirneiss C, Wang GX, 39. Rossier MF, Python CP, Capponi AM, Schlegel W, Kwan CY, Vallotton Korth M, Aszodi A, Andersson KE, Krombach F, Mayerhofer A, Ruth P, MB. Blocking T-type calcium channels with tetrandrine inhibits steroi- Fassler R, Hofmann F. Defective smooth muscle regulation in cGMP dogenesis in bovine adrenal glomerulosa cells. Endocrinology. 1993;132: kinase I-deficient mice. EMBO J. 1998;17:3045–3051. 1035–1043. 25. Furutani H, Zhang XF, Iwamuro Y, Lee K, Okamoto Y, Takikawa O, 40. Spat A, Hunyady L. Control of aldosterone secretion: a model for con- Fukao M, Masaki T, Miwa S. Ca2ϩ entry channels involved in con- vergence in cellular signaling pathways. Physiol Rev. 2004;84:489–539. tractions of rat aorta induced by endothelin-1, noradrenaline, and vaso- 41. Gribkoff VK, Starrett JE, Dworetzky SI. Maxi-K potassium channels: pressin. J Cardiovasc Pharmacol. 2002;40:265–276. form, function, and modulation of a class of endogenous regulators of 26. Schlossmann J, Ammendola A, Ashman K, Zong X, Huber A, Neubauer intracellular calcium. Neuroscientist. 2001;7:166–177. G, Wang GX, Allescher HD, Korth M, Wilm M, Hofmann F, Ruth P. 42. Hu H, Shao LR, Gu N, Chavoshy S, Tieb M, Behrens R, Laake P, Pongs Regulation of intracellular calcium by a signalling complex of IRAG, IP3 O, Knaus HG, Ottersen OP, Storm JF. Presynaptic Ca2ϩ-activated Kϩ receptor and cGMP kinase Ibeta. Nature. 2000;404:197–201. channels in glutamatergic hippocampal terminals and their role in spike 27. Jaggar JH, Porter VA, Lederer WJ, Nelson MT. Calcium sparks in repolarization and regulation of transmitter release. J Neurosci. 2001;21: smooth muscle. Am J Physiol. 2000;278:C235–C256. 9585–9597. Imaging

Risk of Embolism and Death in Infective Endocarditis: Prognostic Value of Echocardiography A Prospective Multicenter Study

Franck Thuny, MD; Giovanni Disalvo, MD; Olivier Belliard, MD; Jean-François Avierinos, MD; Valeria Pergola, MD; Valerie Rosenberg, MD; Jean-Paul Casalta, MD; Joanny Gouvernet, MD, PhD; Geneviève Derumeaux, MD; Diana Iarussi, MD; Pierre Ambrosi, MD; Raffaello Calabro, MD; Alberto Riberi, MD; Frédéric Collart, MD; Dominique Metras, MD; Hubert Lepidi, MD; Didier Raoult, MD, PhD; Jean-Robert Harle, MD; Pierre-Jean Weiller, MD; Ariel Cohen, MD; Gilbert Habib, MD

Background—The incidence of embolic events (EE) and death is still high in patients with infective endocarditis (IE), and data about predictors of these 2 major complications are conflicting. Moreover, the exact role of echocardiography in risk stratification is not well defined. Methods and Results—In a multicenter prospective European study, including 384 consecutive patients (aged 57Ϯ17 years) with definite IE according to Duke University criteria, we tested clinical, microbiological, and echocardiographic data as potential predictors of EE and 1-year mortality. Transesophageal echocardiography was performed in all patients. Embolism occurred before or after IE diagnosis (total-EE) in 131 patients (34.1%) and after initiation of antibiotic therapy (new-EE) in 28 patients (7.3%). Staphylococcus aureus and Streptococcus bovis were independently associated with total-EE, whereas vegetation length Ͼ10 mm and severe vegetation mobility were predictors of new-EE, even after adjustment for S aureus and S bovis. One-year mortality was 20.6%. In multivariable analysis, independently of the other predictors of death (age, female sex, creatinine serum Ͼ2 mg/L, moderate or severe congestive heart failure, and S aureus) and comorbidity, vegetation length Ͼ15 mm was a predictor of 1-year mortality (adjusted relative riskϭ1.8; 95% CI, 1.10 to 2.82; Pϭ0.02). Conclusions—In IE, vegetation length is a strong predictor of new-EE and mortality. In combination with clinical and microbiological findings, echocardiography may identify high-risk patients who will need a more aggressive therapeutic strategy. (Circulation. 2005;112:69-75.) Key Words: echocardiography Ⅲ embolism Ⅲ endocardium Ⅲ prognosis

espite recent improvement in diagnostic1 and therapeu- some authors reported an increased risk of EE and/or mor- Dtic strategies,2 infective endocarditis (IE) is still associ- tality in patients with large and mobile vegetation,9,10,13,17–19 ated with high in-hospital mortality, ranging from 16% to others did not find such a correlation.8,11,14–16,20 Several 25%,3–5 and a high incidence of embolic events (EE), ranging reasons may explain such conflicting results, including the from 13% to 49%.6 These wide ranges of complications changing pattern of the disease over time,4,21 the retrospective underscore the heterogeneity of the disease and the critical design of some studies,5,7–9,11–16,21 the small number of need for baseline risk stratification in order to focus aggres- patients,8–11 the heterogeneous definition of IE,8–12,18 the sive management toward high-risk subsets of patients. How- underutilization of transesophageal echocardiography (TEE),8,9,12,16 ever, previous studies that attempted to identify baseline which enhances sensitivity of vegetation detection,2 and the predictors of mortality5,7–14 and embolism8,10,13,15–19 led to inclusion of EE occurring before echocardiography instead of conflicting results. In particular, despite the key role played postdiagnosis events.10,15,19 Subsequently, no general agree- by echocardiography in IE diagnosis,1 its prognostic value ment emerges from currently available guidelines,22,23 which has been questioned, specifically that of vegetation charac- give discordant recommendations with regard to surgical teristics, namely, vegetation length and mobility. Whereas indications on the basis of vegetation length.

Received February 14, 2004; de novo received July 18, 2004; revision received February 12, 2005; accepted March 11, 2005. From the Departments of Cardiology of La Timone Hospital, Marseille, France (F.T., J.A., J.C., J.G., P.A., A.R., F.C., D.M., H.L., D.R., J.H., P.W., G.H.); Saint-Antoine Hospital, Paris, France (O.B., V.R., A.C.); Charles Nicolle Hospital, Rouen, France (G. Derumeaux); and Second University, Naples, Italy (G. Disalvo, V.P., D.I., R.C.). Correspondence to Dr Gilbert Habib, Département de Cardiologie, Hôpital de la Timone, Boulevard Jean Moulin, 13005, Marseille, France. E-mail [email protected], [email protected] © 2005 American Heart Association, Inc. Circulation is available at http://www.circulationaha.org DOI: 10.1161/CIRCULATIONAHA.104.493155 69 70 Circulation July 5, 2005

To resolve these issues, we undertook a large multicenter, neurologist during the clinical course, who was unaware of the prospective study of patients with a definite diagnosis of IE microbiological and echocardiographic findings. Cutaneous manifes- by current diagnostic criteria in the contemporary era, with tations and EE occurring after surgery were not included. The outcome at 1 year was obtained by contacting the patients’ systematic use of TEE, to assess the predictive value of physicians. clinical and echocardiographic parameters on the subsequent risk of embolism and death. We hypothesized that in addition Statistical Analysis to clinical and microbiological variables, echocardiography For discrete variables, the relation between a variable and an event ␹2 provides accurate baseline risk stratification of patients diag- was studied by test or Fisher exact test (2 tailed) if the expected count in any cell was Ͻ5. Mann-Whitney test were used for nosed with IE. continuous variables. For the end point of total-EE, logistic regression analysis was Methods performed with the use of clinical and microbiological variables as previously defined; among echo variables, vegetation length and Patient Sample vegetation mobility were not included in this analysis because these From January 1993 to March 2003, all consecutive patients admitted 2 parameters defined at index echo and potentially evolving with in 4 referral European centers with a suspected diagnosis of IE were time could not be used as predictors of past events. Variables eligible for study entry (nϭ613). The only exclusion criterion was significantly associated with this end point in single-variable analysis pacemaker IE (nϭ34). Written informed consent was obtained from (PϽ0.05) were included as candidate predictors in an ascending all participating patients, as required by the institutional review stepwise logistic regression analysis. board under an approved protocol. Monthly screening of echocardi- For the second end point of new-EE, only the echo variables ography and microbiology databases of all patients hospitalized for vegetation length and vegetation mobility were tested as potential suspected IE was performed by each center to ensure that a predictors, first in a single-variable analysis and then after adjust- consecutive sample of all definite diagnoses was obtained. Blood ment for predictors of total-EE. cultures, serological assessment, transthoracic echocardiography One-year survival was estimated by the Kaplan-Meier method. (TTE), TEE, and cerebral and thoracoabdominal CT scans were Baseline clinical, microbiological, and echocardiographic variables systematically performed within 48 hours of admission in all except were tested as potential predictors of 1-year mortality with Cox 7 patients, who underwent abdominal echography and no CT scan proportional hazards modeling. Variables with PϽ0.10 were in- because of severe renal insufficiency. Patients with definite diagno- cluded into the multivariable model. sis of IE according to Duke University criteria1 formed the final Receiver operating characteristic (ROC) curve analysis was per- prospective cohort (nϭ384). Antibiotic therapy was started immedi- formed to determine the optimal cutoff value of vegetation length ately after diagnosis. that best predicted end points. PϽ0.05 was considered significant. All analyses were performed with SPSS for Windows, release Clinical Data 10.0.1999.Chicago (SPSS Inc). The following clinical and biological parameters were prospectively Finally, interobserver variability was good for both vegetation collected at diagnosis and during hospitalization: age, sex, fever length (␬ϭ0.8) and mobility (␬ϭ0.75). (temperature Ͼ38°C), previous heart disease, intravenous drug abuse, HIV infection, diabetes, history of cancer, comorbidity,24 Results moderate or severe congestive heart failure (CHF) diagnosed as previously described,25 and serum creatinine Ͼ2 mg/dL. Early Patient Characteristics on Admission surgery was defined as valve replacement or repair performed during Among the 384 patients with a definite diagnosis of IE, 294 the course of antibiotic therapy. had 2 major clinical Duke University criteria, 89 had 1 major and 3 minor criteria, and only 1 patient had 5 minor criteria. Echocardiography Baseline patient features are reported in Table 1, and micro- TTE and TEE were performed in all cases, as previously reported,19 biological data are reported in Table 2. and were reviewed by 2 experienced echocardiographers, blinded to MeanϮSD age was 57Ϯ17 years (range, 16 to 94 years), and patients’ clinical status. Echocardiographic data included the pres- ence, maximal length, and mobility of vegetation.19 Measurements 26% of patients were older than 70 years. Ninety-eight patients of vegetation length were performed in various planes, and maximal (25%) presented with moderate or severe CHF on admission, length was used. In the presence of multiple vegetations, the largest and 103 patients (26.8%) had already had an EE, including length was used for analysis. Mobility was evaluated with the use of stroke in 46 patients (12%). A vegetation was identified by TEE 9 a 4-point scale, as follows: absent, fixed vegetation with no in 320 patients (83%) but in only 192 patients (50%) by TTE. An detectable independent motion; low, vegetation with a fixed base but with a mobile free edge; moderate, pedunculated vegetation that abscess was identified by TEE in 94 patients (24%), and a new remains within the same chamber throughout the cardiac cycle; and moderate to severe regurgitation was identified in 209 (54%). severe, prolapsing vegetation that crosses the coaptation plane of the leaflets during the cardiac cycle. An abscess was defined as a Indications for Surgery thickened area or mass with a heterogeneous echogenic or echolu- Early surgery was performed in 201 patients (52.3%) at a 26 cent appearance. Valvular regurgitations were assessed semiquan- median time of 12 days (range, 0 to 50) after institution of titatively.27,28 Echocardiographic data were stored electronically and used unaltered for subsequent analysis. antibiotic treatment. One hundred nine patients (28.4%) were operated on in an urgent setting, Ͻ15 days after diagnosis and End Points the beginning of antibiotic therapy, 60 (15.6%) between 15 End points were embolic events that occurred before or after and 30 days, and 32 (8.3%) after 30 days. The indications for initiation of antibiotic therapy (total-EE), embolic events that oc- surgery included moderate or severe CHF in 72 cases, curred after initiation of antibiotic therapy (new-EE), and 1-year persistent vegetation after systemic embolization in 56 cases, mortality. Diagnosis of EE was based on clinical or CT scans data or both. CT scans, systematically performed at study entry, were abscess formation in 77 cases, acute severe aortic or mitral subsequently repeated if clinically indicated. Specific diagnosis of regurgitation without CHF in 29 cases, and early prosthetic cerebral embolism was eventually confirmed by an experienced valve IE in 9 cases. Forty-three patients with moderate or Thuny et al Prognosis of Infective Endocarditis 71

TABLE 1. Clinical and Laboratory Findings in 384 Cases of IE With and Without Embolic Events Occurring Before or After Initiation of Antibiotic Therapy (Total-EE)

All Patients Total-EE Without Total-EE (nϭ384) (nϭ131) (nϭ253) P* Age, meanϮSD, y 57Ϯ17 56Ϯ17 57Ϯ17 0.53 Male 274 (71) 92 (70) 182 (72) 0.72 Valve localization Mitral 191 (50) 70 (53) 121 (48) 0.33 Aortic 214 (53) 67 (51) 147 (58) 0.20 Prosthetic valves 91 (24) 28 (21) 63 (25) 0.70 Multivalvular† 60 (16) 24 (18) 36 (14) 0.30 Right-heart IE‡ 34 (9) 18 (14) 16 (6) 0.02 Previous heart disease§ 180 (47) 54 (41) 126 (50) 0.13 IVDA 24 (6) 13 (10) 11 (4) 0.04 HIV infection 13 (3) 4 (3) 9 (4) 1.0 History of cancer 23 (6) 8 (6) 15 (6) 1.0 Diabetes 27 (7) 12 (9) 15 (6) 0.29 Comorbidity index Ͼ2ʈ 69 (18) 26 (20) 43 (17) 0.49 Moderate or severe CHF 98 (25) 37 (28) 61 (24) 0.38 Serum creatinine Ͼ2 mg/L 70 (18) 28 (21) 42 (17) 0.27 Values are number (%). IVDA indicates intravenous drug abuse. *Comparison between total-EE group and without total-EE group. Bold values are significant. †At least 2 locations. ‡Only right heart localization. §Including 91 patients with prosthetic valve, 33 congenital heart disease (27 bicuspid aortic valves, 3 interventricular septal defects, 2 Fallot, and 1 quadricuspid aortic valve), 27 mitral valve diseases, 26 aortic valve diseases, 3 obstructive cardiomyopathies. ʈCharlson comorbidity scale. severe CHF or abscess or both were not operated on because (49 cases), kidney (22 cases), lungs (16 cases), peripheral of severe comorbidity in 41 and refusal in 2. The median arteries (10 cases), mesentery (3 cases), coronary circulation duration of hospital stay was 43 days (range, 0 to 167). (2 cases), and eye (1 case). Thirty-three patients (8.6%) presented with Ͼ1 total-EE. Embolism was silent in 19 Embolic Risk patients (4.9%). There was no statistical difference between Total Embolic Events patients with and without total-EE with regard to age, sex, Among 384 patients, 131 (34.1%) had Ն1 total-EE. Sites of presence of previous heart disease or prosthetic valve, local- embolization were central nervous system (62 cases), spleen ization of infection (mitral or aortic), and predisposing factors

TABLE 2. Microbiological Findings in 384 Cases of IE With and Without Embolic Events Occurring Before or After Initiation of Antibiotic Therapy (Total-EE)

All Patients Total-EE Without Total-EE (nϭ384) (nϭ131) (nϭ253) P* S bovis 63 (16) 32 (24) 31 (12) 0.003 Enterococci 28 (7) 14 (11) 14 (5) 0.07 Oral streptococci 95 (25) 24 (18) 71 (28) 0.045 S aureus 82 (21) 37 (28) 45 (18) 0.03 Coagulase-negative staphylococci 17 (4) 4 (3) 13 (5) 0.44 Others† 40 (10) 15 (19) 25 (10) 1 Negative blood cultures‡ 76 (20) 15 (11) 61 (24) 0.003 Values are number (%). *Comparison between total-EE group and without total-EE group. Bold values are significant. †Including Q fever (nϭ9), HACEK group (Haemophilus, Actinobacillus, Cardiobacterium, Eikenella, Kingella ͓nϭ6͔), Candida species (nϭ6), Escherichia coli (nϭ4), Enterobacter cloacae (nϭ2), Gemella morbillorum (nϭ3), Corynebacterium (nϭ3), Bartonella quintana (nϭ2), Bartonella henselae (nϭ2), Streptococcus agalactiae (nϭ1), Mycoplasma hominis (nϭ1), Propionibacterium acnes (nϭ1). ‡Definite IE diagnosis using clinical Duke criteria: 1 major and 3 minor criteria (nϭ75), 5 minor criteria (nϭ1). 72 Circulation July 5, 2005

TABLE 3. Predictors of Embolic Events in Multivariate Analysis TABLE 5. Predictors of 1-Year Mortality (Cox Multivariable Analysis) Adjusted P Odds Ratio 95% CI Adjusted RR 95% CI P Total-EE Age 1.02 1.01–1.04 0.007 S bovis Ͻ0.001 3.9 1.86–8.21 Female sex 1.6 1.01–2.58 0.048 S aureus 0.002 2.4 1.15–4.83 Comorbidity index Ͼ2 1.6 0.92–2.64 0.1 New-EE Serum creatinine Ͼ2 mg/L 1.9 1.16–3.23 0.01 Vegetation length Ͼ10 mm 0.004 9 1.98–40.8 Prosthetic valve 1.6 0.99–2.68 0.053 Severe vegetation mobility 0.04 2.4 1.02–5.42 S aureus IE 2 1.19–3.24 0.001 S bovis 0.19 1.9 0.73–4.74 Moderate or severe CHF 1.6 1.02–1.54 0.04 S aureus 0.12 2 0.84–4.76 Vegetation length Ͼ15 mm 1.8 1.10–2.82 0.02 and comorbidity (Table 1). However, total-EE were more among 117 patients with severe vegetation mobility (16.2% ϭ frequent in patients with intravenous drug abuse (P 0.04), [19/117] versus 3.4% [9/267]; PϽ0.001). Conversely, ϭ ϭ right-side IE (P 0.02), positive blood cultures (P 0.003), new-EE occurred in only 2 patients with both vegetation ϭ Streptococcus bovis (P 0.003), and Staphylococcus aureus length Ͻ10 mm and no severe mobility. After adjustment for ϭ IE (P 0.03). Total-EE were less frequently observed in oral the 2 multivariate predictors of total-EE, ie, S aureus and S ϭ streptococcal IE (P 0.04) (Table 2). By multivariable anal- bovis, vegetation length Ͼ10 mm and severe vegetation ysis, S bovis and S aureus remained the only predictors of mobility remained the only predictors of new-EE (Table 3). total-EE (Table 3). In the subgroup of 14 patients with new cerebral embolism, Ͼ Embolic Events During Antibiotic Therapy vegetation length was 10 mm in all 14 patients. In those New-EE occurred in 28 patients (7.3%). Sites of embolization patients, vegetation mobility was severe in 12 patients. were central nervous system (14 cases), spleen (9 cases), kidney (5 cases), peripheral arteries (5 cases), eye (2 cases), Mortality coronary circulation (2 cases), and pulmonary circulation (2 Incidence and Causes of Death cases). These events occurred at a median time of 7 days One-year mortality was 20.6%. Thirty-seven patients (9.6%) (range, 1 to 38) after institution of adequate antibiotic therapy died during their hospital stay at a median time of 16 days (20 [71.4%] in the first 15 days). (range, 0 to 73) after institution of antibiotic therapy. The By single-variable analysis, vegetation length was predic- causes for death were severe CHF (10 cases), multiorgan tive of new-EE (PϽ0.001). Vegetation length (median) was failure (9 cases), cerebral embolism (9 cases), septic shock (6 larger in patients with new-EE than in those without cases), cerebral hemorrhage (3 cases), atrioventricular block (15.5 mm [range, 0 to 40] versus 9 mm [range, 0 to 50], (1 case), and myocardial infarction (1 case). Forty-two respectively; PϽ0.001). A vegetation length threshold of patients died after dismissal. The cause for late death was a 10 mm was identified as having the highest predictive value direct consequence of the cardiac lesions induced by IE in 26 for new-EE by ROC curve analysis. New-EE occurred more patients (severe valve regurgitation in 21 and postoperative frequently in patients with vegetation length Ͼ10 mm than in left ventricular dysfunction in 5). In the remaining 16 pa- those with vegetation length Յ10 mm (13.7% [26/190] versus tients, the cause for late death was not directly related to 1% [2/194]; PϽ0.001). New-EE were also more frequent cardiac lesions, including stroke in 2 patients, myocardial infarction in 1, IE recurrence in 1, noncardiac cause in 8, and TABLE 4. Predictors of 1-Year Mortality unknown in 4. (Cox Single-Variable Analysis) Factors associated with 1-year mortality are summarized in RR 95% CI P Table 4. Vegetation length was predictive of 1-year mortality (relative risk [RR]ϭ1.03 per millimeter; 95% CI, 1.01 to Age 1.03 1.01–1.04 0.0003 1.06; Pϭ0.01), and ROC curve demonstrated vegetation Female sex 1.8 1.16–2.86 0.009 length Ͼ15 mm to have the best predictive value (RRϭ2.1; Ͼ Comorbidity index 2 1.8 1.10–2.90 0.03 95% CI, 1.34 to 3.26; Pϭ0.001). The Figure shows 1-year Serum creatinine Ͼ2 mg/L 2.9 1.80–4.53 Ͻ0.0001 survival curves according to vegetation length. Prosthetic valve 1.6 0.99–2.58 0.05 By multivariable analysis, baseline predictors of 1-year S aureus IE 2.1 1.35–3.39 0.002 mortality were vegetation length Ͼ15 mm (adjusted RRϭ1.7; Cerebral embolism 1.4 0.83–2.48 0.2 95% CI, 1.10 to 2.64; Pϭ0.03), age (adjusted RRϭ1.02; 95% Moderate or severe CHF 1.9 1.21–3.01 0.005 CI, 1.01 to 1.04; Pϭ0.009), female sex (adjusted RRϭ1.6; ϭ Ͼ Abscess 1.2 0.72–1.92 0.53 95% CI, 1.01 to 2.57; P 0.04), serum creatinine 2 mg/L ϭ ϭ Moderate or severe regurgitation 1.1 0.72–1.75 0.61 (adjusted RR 2.1; 95% CI, 1.29 to 3.46; P 0.003), S aureus (adjusted RRϭ1.9; 95% CI, 1.16 to 3.14; Pϭ0.01), and Vegetation length, mm 1.03 1.01–1.06 0.01 moderate or severe CHF (adjusted RRϭ1.6; 95% CI, 1.02 to Ͼ Vegetation length 15 mm 2.1 1.34–3.26 0.001 2.54; Pϭ0.04). Vegetation length Ͼ15 mm remained a Thuny et al Prognosis of Infective Endocarditis 73

Embolic Risk in IE Embolism represents one of the most frequent and severe complications of IE and has been reported to occur in 13% to 49% with IE.6 If the total risk of embolism associated with IE is very high, the risk of new embolism occurring after initiation of therapy is much lower, from 6% to 21% in past series10,17,19,29 and 7.3% in the present series. The risk of embolism seems particularly high during the first 2 weeks after diagnosis,8 and this point was confirmed by the present study because 71.4% of new-EE occurred during the first 15 days after diagnosis. The exact role of echocardiography in predicting embolism has been largely debated,8,10,13,15–19 and past studies gave conflicting results. The causes for these discrepancies are well known. Limitations of past studies include small sample size,8–11 use of TTE alone,8,9,12,16 inclusion of EE occurring before echocardiography,10,15,19 and poor standardization of diagnostic criteria. The present study overcomes all these limitations because it included prospectively a large cohort of patients with definite IE according to Duke University crite- ria, with special attention to new-EE. In addition, this is the Ϯ One-year survival ( SE) according to vegetation length (L). largest study in which vegetation characteristics were pro- spectively collected and TEE was systematically performed. predictor even when comorbidity index was included in the The main result is that the echocardiographic characteristics model (Table 5). of vegetation are clearly associated with the embolic risk. In Among the 114 patients with vegetation length Ͼ15 mm, a previous study from our center,19 178 patients with IE were 75 had at least 1 standard clinical indication for surgery included, and a significant correlation was found between EE (CHF, nϭ35; acute severe aortic regurgitation without CHF, and vegetation size and mobility. However, a limitation of nϭ24; abscess, nϭ28; recurrent embolism despite appropri- this study was the inclusion of patients with previous EE and ate antibiotic therapy, nϭ18; prosthetic valve IE with para- the relative small number of new-EE. The present large valvular leak, nϭ9). multicenter study was initiated to overcome these limitations and allowed us both to include a larger number of patients and Risk of Embolism and Death in Prosthetic Versus to analyze a more significant number of new-EE. The results Native Valve Endocarditis of the present study confirm that large vegetations Ͼ10 mm Embolic Risk or severe vegetation mobility or both are associated with an New-EE were observed in 22 (7.5%) among the 293 patients increased embolic risk. Conversely, new-EE were infrequent with native valve IE and in 6 (6.5%) among the 91 patients in a low-risk subgroup of patients with both vegetation length with prosthetic valve IE. Vegetation length and mobility Ͻ10 mm and no severe mobility. remained predictors of new-EE in native valve IE subgroup Predictors other than vegetation characteristics were iden- (vegetation length Ͼ10 mm [odds ratioϭ5.9; Pϭ0.006] and tified by several studies. For example, antiphospholipid severe vegetation mobility [odds ratioϭ3.5; Pϭ0.03]) but not antibodies, coagulation parameters, and endothelial cell acti- in prosthetic valve IE subgroup. vation have been associated with an increased embolic risk.30 Bacteriologic factors and localization of IE have also been Mortality previously reported to influence the incidence of EE. For One-year mortality was 18.4% and 27.5% in the subgroups of example, S aureus31 and S bovis32 have been associated with patients with native and prosthetic valve IE, respectively. an increased embolic risk. In our study, S aureus and S bovis Vegetation length Ͼ15 mm remained associated with death infection were associated with an increased risk of total-EE. (RRϭ2.2; 95% CI, 1.27 to 3.71; Pϭ0.004) in the native valve However, in regard to the occurrence of new-EE, vegetation IE subgroup but not in the prosthetic valve IE subgroup, in length and mobility remained the only predictors after adjust- which only S aureus was predictive of death (RRϭ2.9; 95% ment for these microbiological variables. Finally, our study CI, 1.30 to 6.67; Pϭ0.009). did not confirm the previously reported higher incidence of EE in mitral valve IE.2,10 Discussion The present study shows that echocardiography, performed Mortality early in the course of IE, has a strong prognostic value. Along Mortality is still high in IE, although it has declined in recent with baseline clinical and microbiological features, the as- years.11,13,21 The low mortality rate observed in our study may sessment of vegetation characteristics (length and mobility) be related to a more aggressive surgical approach33 (28.4% of allows identification of patients who are at highest risk for all patients underwent surgery before the 15th day of antibi- new-EE and death. otic therapy) and a lower incidence of S aureus IE compared 74 Circulation July 5, 2005 with the most recent American studies.20,21 In addition, a high Study Limitations incidence of S bovis IE was observed in our study, and IE This study has several limitations. First, it was subject to a caused by this microorganism has been associated with a referral bias because it was performed in referral centers. The good prognosis.4 early surgery policy of these centers could have reduced the Mortality in IE may be related to factors related to the incidence of new-EE. Moreover, a repeated CT scan was not patient or to factors related to the disease, the former being systematically performed after antibiotic therapy in all pa- potentially preventable. Thus, the identification of factors tients, and therefore the exact incidence of new silent EE may associated with increased mortality is a crucial challenge have been underestimated. In addition, the incidence of EE because it will allow the identification of high-risk patients in was low in the subgroup of patients with prosthetic valve IE, whom an aggressive strategy will be potentially useful. and therefore no definite conclusion can be drawn concerning Several markers have previously been identified in past the value of TEE in predicting EE in this particular subgroup. studies, including age,34 occurrence of complications, staph- The rate of negative blood cultures was relatively high in ylococcal infection, and prosthetic valve IE.2 Other studies our study. The causes for negative blood cultures may include found different results; for example, Netzer et al5 found that prior antibiotic therapy and a relatively high incidence of Q only neurological symptoms, arthralgia, and weight loss were fever endocarditis in our countries, as previously reported.35 independent predictors of mortality, and Wallace et al7 found It is our policy to perform a systematic serological assessment that clinical indices such as abnormal white cell count, serum of multiple microorganisms in case of suspected IE, allowing albumin concentration, serum creatinine concentration, or identification of a high rate of “atypical” microorganisms visible vegetation were the best predictors of bad prognosis. (Coxiella burnetii, Bartonella species, Mycoplasma species) The most comprehensive study has recently been published consistent with IE and not identified by blood cultures. by Hasbun et al.14 They studied 6-month mortality in a series Conclusion of 513 patients with complicated IE. They found that comor- Echocardiography has a strong predictive value in IE. Inde- bidity, abnormal mental status, moderate to severe CHF, pendently of other baseline characteristics, vegetation length staphylococcal infection, and medical therapy were indepen- has major prognostic implications by predicting both EE dent predictors of mortality. In these series, the presence of under antibiotic therapy and mortality. Thus, the measure- vegetation was not associated with increased mortality, but ment of vegetation length at the time of diagnosis of IE is the vegetation size and mobility were not specifically ana- strongly recommended as part of the initial risk stratification. lyzed. In addition, the presence of vegetation was considered Patients with the largest vegetations should be considered at a criterion for “complicated IE” in this study, and only high risk for subsequent serious complications. Whether a patients with complicated IE were included; thus, some more aggressive therapeutic strategy (ie, early surgery) is patients without vegetation were probably not included, and required in those patients requires further prospective studies. this may explain why the presence of vegetation did not predict mortality in this study. Similarly, in the recent study References of Chu et al,20 early echocardiographic findings were not 1. Durack DT, Lukes AS, Bright DK. New criteria for diagnosis of infective endocarditis: utilization of specific echocardiographic findings: Duke predictive of death, but in this report, which included possible Endocarditis Service. Am J Med. 1994;96:200–209. and definite diagnosis of IE, TEE was performed in only 66% 2. Bayer AS, Bolger AF, Taubert KA, Wilson W, Steckelberg J, Karchmer of patients, and data on vegetation size were not prospectively AW, Levison M, Chambers HF, Dajani AS, Gewitz MH, Newburger JW, collected. Gerber MA, Shulman ST, Pallasch TJ, Gage TW, Ferrieri P. Diagnosis and management of infective endocarditis and its complications. Circu- In our study several factors were associated with a poor lation. 1998;98:2936–2948. prognosis, including age, female sex, serum creatinine, S 3. Mylonakis E, Calderwood SB. Infective endocarditis in adults. N Engl aureus, and moderate or severe CHF. More importantly, we J Med. 2001;345:1318–1330. Ͼ 4. Hoen B, Alla F, Selton-Suty C, Béguinot I, Bouvet A, Briançon S, Casalta also found vegetation length 15 mm to be a predictor of JP, Danchin N, Delahaye F, Etienne J, Le Moing V, Leport C, Mainardi 1-year mortality, even after adjustment for the other predic- JL, Ruimy R, Vandenesch F. Changing profile of infective endocarditis: tors and comorbidity. Older series failed to give a prognostic results of a 1-year survey in France. JAMA. 2002;288:75–81. value to the presence or size of vegetation, probably because 5. Netzer RO, Zollinger E, Seiler C, Cerny A. Infective endocarditis: clinical spectrum, presentation and outcome: an analysis of 212 cases 1980–1995. of the small number of patients studied or use of TTE Heart. 2000;84:25–30. alone.8,11,16 However, our results are in agreement with the 6. Habib G. Embolic risk in subacute bacterial endocarditis: role of trans- recent series of Cabell et al,13 who found a direct relationship esophageal echocardiography. Curr Cardiol Rep. 2003;5:129–136. 7. Wallace SM, Walton BI, Kharbanda RK, Hardy R, Wilson AP, Swanton between vegetation size and mortality at 30 days and 1 year. RH. Mortality from infective endocarditis: clinical predictors of outcome. Finally, the fact that very large vegetations were indepen- Heart. 2002;88:53–60. dently associated with a worse prognosis is not surprising 8. Steckelberg JM, Murphy JG, Ballard D, Bailey K, Tajik AJ, Taliercio CP, because this feature is frequently associated with severe valve Guiliani ER, Wilson WR. Emboli in infective endocarditis: the prognostic value of echocardiography. Ann Intern Med. 1991;114:635–640. destruction and a high embolic risk. Interestingly, among the 9. Sanfilippo AJ, Picard MH, Newell JB, Rosas E, Davidoff R, Thomas JD, patients with vegetation length Ͼ15 mm, 34% had no other Weyman AE. Echocardiographic assessment of patients with infectious indication for surgery and would have been identified as endocarditis: prediction of risk for complications. J Am Coll Cardiol. 1991;18:1191–1199. high-risk patients by echocardiography before standard indi- 10. Mugge A, Daniel WG, Frank G, Lichtlen PR. Echocardiography in cations for surgery were met. infective endocarditis: reassessment of prognostic implications of vege- Thuny et al Prognosis of Infective Endocarditis 75

tation size determined by the transthoracic and the transesophageal diagnosis and treatment of infective endocarditis: full guidelines: the Task approach. J Am Coll Cardiol. 1989;14:631–638. Force on Infective Endocarditis executive summary; the Task Force on 11. Jaffe WM, Morgan DE, Pearlman AS, Otto CM. Infective endocarditis, Infective Endocarditis of the European Society of Cardiology. Eur 1983–1988: echocardiographic findings and factors influencing morbidity Heart J. 2004;25:267–276. and mortality. J Am Coll Cardiol. 1990;15:1227–1233. 24. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of 12. Malquarti V, Saradarian W, Etienne J, Milon H, Delahaye JP. Prognosis classifying prognostic comorbidity in longitudinal studies: development of native valve infective endocarditis: a review of 253 cases. Eur Heart J. and validation. J Chronic Dis. 1987;40:373–383. 1984;5(suppl C):11–20. 25. Richardson JV, Karp RB, Kirklin JW, Dismukes WE. Treatment of 13. Cabell CH, Pond KK, Peterson GE, Durack DT, Corey GR, Anderson DJ, infective endocarditis: a 10-year comparative analysis. Circulation. 1978; Ryan T, Lukes AS, Sexton DJ. The risk of stroke and death in patients 58:589–597. with aortic and mitral valve endocarditis. Am Heart J. 2001;142:75–80. 26. Daniel WG, Mugge A, Martin RP, Lindert O, Hausmann D, Nonnast- 14. Hasbun R, Vikram HR, Barakat LA, Buenconsejo J, Quagliarello Daniel B, Laas J, Lichtlen PR. Improvement in the diagnosis of abscess VJ. Complicated left-sided native valve endocarditis in adults: risk clas- associated with endocarditis by transesophageal echocardiography. sification for mortality. JAMA. 2003;289:1933–1940. N Engl J Med. 1991;324:795–800. 15. De Castro S, Magni G, Beni S, Cartoni D, Fiorelli M, Venditti M, 27. Helmcke F, Nanda NC, Hsiung MC, Soto B, Adey CK, Goyal RG, Schwartz SL, Fedele F, Pandian NG. Role of transthoracic and trans- Gatewood RP. Color Doppler assessment of mitral regurgitation with esophageal echocardiography in predicting embolic events in patients orthogonal planes. Circulation. 1987;75:175–183. with active infective endocarditis involving native cardiac valves. 28. Perry GJ, Helmcke F, Nanda NC, Byard C, Soto B. Evaluation of aortic Am J Cardiol. 1997;80:1030–1034. insufficiency by Doppler color flow mapping. J Am Coll Cardiol. 1987; 9:952–959. 16. Heinle S, Wilderman N, Harrison JK, Waugh R, Bashore T, Nicely LM, 29. Rohmann S, Erbel R, Gorge G, Makowski T, Mohr-Kahaly S, Nixdorff Durack D, Kisslo J. Value of transthoracic echocardiography in pre- U, Drexler M, Meyer J. Clinical relevance of vegetation localization by dicting embolic events in active infective endocarditis: Duke Endocarditis transoesophageal echocardiography in infective endocarditis. Eur Service. Am J Cardiol. 1994;74:799–801. Heart J. 1992;13:446–452. 17. Vilacosta I, Graupner C, San Roman JA, Sarria C, Ronderos R, Fernandez 30. Kupferwasser LI, Hafner G, Mohr-Kahaly S, Erbel R, Meyer J, Darius H. C, Mancini L, Sanz O, Sanmartin JV, Stoermann W. Risk of embolization The presence of infection-related antiphospholipid antibodies in infective after institution of antibiotic therapy for infective endocarditis. J Am Coll endocarditis determines a major risk factor for embolic events. J Am Coll Cardiol. 2002;39:1489–1495. Cardiol. 1999;33:1365–1371. 18. Tischler MD, Vaitkus PT. The ability of vegetation size on echocardiog- 31. Erbel R, Liu F, Ge J, Rohmann S, Kupferwasser I. Identification of raphy to predict clinical complications: a meta-analysis. J Am Soc Echo- high-risk sub-groups in infective endocarditis and the role of echocardi- cardiogr. 1997;10:562–568. ography. Eur Heart J. 1995;16:588–602. 19. Di Salvo G, Habib G, Pergola V, Avierinos JF, Philip E, Casalta JP, 32. Pergola V, Di Salvo G, Habib G, Avierinos JF, Philip E, Vailloud JM, Vailloud JM, Derumeaux G, Gouvernet J, Ambrosi P, Lambert M, Thuny F, Casalta JP, Ambrosi P, Lambert M, Riberi A, Ferracci A, Ferracci A, Raoult D, Luccioni R. Echocardiography predicts embolic Mesana T, Metras D, Harle JR, Weiller PJ, Raoult D, Luccioni R. events in infective endocarditis. J Am Coll Cardiol. 2001;37:1069–1076. Comparison of clinical and echocardiographic characteristics of Strepto- 20. Chu VH, Cabell CH, Benjamin DK, Kuniholm EF, Fowler VG, coccus bovis endocarditis with that caused by other pathogens. Engemann J, Sexton DJ, Corey GR, Wang A. Early predictors of Am J Cardiol. 2001;88:871–875. in-hospital death in infective endocarditis. Circulation. 2004;109: 33. Alexiou C, Langley SM, Stafford H, Lowes JA, Livesey SA, Monro JL. 1745–1749. Surgery for active culture-positive endocarditis: determinants of early and 21. Cabell CH, Jollis JG, Peterson GE, Corey GR, Anderson DJ, Sexton DJ, late outcome. Ann Thorac Surg. 2000;69:1448–1454. Woods CW, Reller LB, Ryan T, Fowler VG. Changing patient charac- 34. Di Salvo G, Thuny F, Rosenberg V, Pergola V, Belliard O, Derumeaux teristics and the effect on mortality in endocarditis. Arch Intern Med. G, Cohen A, Iarrussi D, Giorgi R, Casalta JP, Caso P, Habib G. Endo- 2002;162:90–94. carditis in the elderly: clinical, echocardiographic, and prognosis features. 22. Bonow RO, Carabello B, de Leon AC, Edmunds LH, Fedderly BJ, Freed Eur Heart J. 2003;24:1575–1582. MD, Gaasch WH, McKay CR, Nishimura RA, O’Gara PT, O’Rourke RA, 35. Habib G, Derumeaux G, Avierinos JF, Casalta JP, Jamal F, Volot F, Rahimtoola SH. ACC/AHA guidelines for the management of valvular Garcia M, Lefevre J, Biou F, Maximovitch-Rodaminoff A, Fournier PE, heart disease. Circulation. 1998;98:1949–1984. Ambrosi P, Velut JG, Cribier A, Harle JR, Weiller PJ, Raoult D, Luccioni 23. Horstkotte D, Follath F, Gutschik E, Lengyel M, Oto A, Pavie A, R. Value and limitations of the Duke criteria for the diagnosis of infective Soler-Soler J, Thiene G, von Graevenitz A. Guidelines on prevention, endocarditis. J Am Coll Cardiol. 1999;33:2023–2029. Molecular Cardiology

Bcl-xL Gene Transfer Inhibits Bax Translocation and Prolongs Cardiac Cold Preservation Time in Rats

Jianhua Huang, MD; Kiminori Nakamura, DDS, PhD; Yoshinori Ito, MD; Takeshi Uzuka, MD; Masayuki Morikawa, MD, PhD; Sachie Hirai, BS; Kei Tomihara, DDS; Toshihiro Tanaka, MD, PhD; Yukari Masuta, BS; Keiji Ishii, MD; Kazunori Kato, PhD; Hirofumi Hamada, MD, PhD

Background—Apoptosis is an important cause of early graft loss after heart transplantation. Bcl-xL was reported to protect the heart against normothermic ischemia and reperfusion injury. In this study, we determined whether overexpression of Bcl-xL could inhibit tissue injury resulting from prolonged cold preservation followed by warm reperfusion of heart transplants. Methods and Results—Lewis rat hearts were transduced with an adenovirus vector harboring Bcl-xL cDNA (AxCAhB- clxL) 4 days before collection of tissue. After preservation in University of Wisconsin solution at 4°C for 24 hours, the heart was either perfused with a Langendorff device ex vivo or used for heterotopic heart transplantation in vivo. Bcl-xL gene transfer significantly reduced the infarct size (23.0Ϯ2.6% versus 47.7Ϯ7.0% in saline control and 48.6Ϯ6.1% in vector control, PϽ0.01) after 2-hour reperfusion at 37°C with the Langendorff device and significantly decreased creatine kinase release (0.82Ϯ0.27 IU, versus 1.57Ϯ0.33 and 1.50Ϯ0.37 IU in saline and vector controls, respectively; PϽ0.05). In heart transplantation, overexpresson of Bcl-xL inhibited Bax translocation from the cytosol to the mitochondria, resulting in decreased cytochrome c release from the mitochondria; it also significantly decreased cardiac cell apoptosis and improved graft survival rate after long cold preservation, followed by warm reperfusion. Conclusions—Bcl-xL gene transfer inhibited the translocation of Bax and prolonged the cold preservation time of cardiac transplants. This may be a potential therapeutic method in clinical practice. (Circulation. 2005;112:76-83.) Key Words: gene therapy Ⅲ ischemia Ⅲ reperfusion Ⅲ transplantation Ⅲ apoptosis

ith the development of new technologies and improve- reperfusion injury7,8 or cold ischemia and reperfusion Wments in immunotherapy methods, there has been a injury2 in the heart. Several studies proved that cardiac worldwide increase in the number of cases of cardiac trans- dysfunction could be prevented by the antiapoptotic Bcl-2 plantation in recent decades. One of the limitations of heart gene in rats9 or transgenic mice.10 On the other hand, transplantation is the lack of donor hearts, attributable at least cobalt protoporphyrin (heme oxygenase-1 inducer) was in part to the current duration of heart preservation time, reported to prolong graft survival after 24-hour preserva- which is limited to 4 to 6 hours in clinical practice.1 Although tion at 4°C in a cardiac transplantation model in which using cardiac preservation liquids such as the University of Bcl-2 expression was elevated.11 Thus, the strategy of a Wisconsin solution (UWS) prolongs the preservation time of combination of antiapoptotic gene transfer and cold pres- heart grafts, 18 hours is the maximum preservation time for ervation seems to be reasonable for success of heart functional recovery of the heart in different experimental transplants. protocols.2 Bcl-xL belongs to the same family as Bcl-2, which was shown to play an important role in cardiac protection by Seep6 hepatocyte growth factor12 and insulin-like growth factor- Long cold preservation and warm reperfusion are con- 1.13 Bcl-xL was also reported to inhibit Bax translocation sidered important causes of early cardiac graft loss. Al- to the mitochondria, decrease the release of cytochrome c though calcium flux,3 acidosis,4 ATP depletion,5 and (cyto-c) from the mitochondria to the cytosol, and protect reactive oxygen species6 have been defined as putative retinal cell apoptosis in transgenic mice.14 Because cold causative factors, there may be other mediator factors that ischemia and reperfusion cause mitochondrial dysfunction remain to be identified. Indeed, besides necrosis, apoptosis and release of cyto-c from the mitochondria to the cy- is one of the early events of either warm ischemia and tosol,15 we hypothesized that Bcl-xL would be useful in

Received June 27, 2004; de novo received January 12, 2005; revision received February 25, 2005; accepted March 8, 2005. From the Department of Molecular Medicine (J.H., K.N., Y.I., S.H., K.T., T.T., Y.M., K.I., K.K., H.H.), the Division of Gene Therapy (Y.I., H.H.), and the Department of Thoracic and Cardiovascular Surgery (T.U., M.M.), Sapporo Medical University, Sapporo, Japan. Correspondence to Hirofumi Hamada, MD, PhD, Department of Molecular Medicine, Sapporo Medical University, South 1, West 17, Chuo-ku, Sapporo, 060-8556, Japan. E-mail [email protected] © 2005 American Heart Association, Inc. Circulation is available at http://www.circulationaha.org DOI: 10.1161/CIRCULATIONAHA.105.535740 76 Huang et al Bcl-xL Prolongs Cardiac Cold Preservation 77 preventing cold ischemia and warm reperfusion injury of PharMingen) was applied, and the Vector ABC kit (Vector Labora- the heart. tories) was used as described by the manufacturer. We previously demonstrated that adenovirus-mediated Langendorff Perfusion and CK Measurement Bcl-xL gene transduction to the heart improved cardiac To determine the effect of long cold preservation and warm function and decreased the infarct size of the heart after warm reperfusion on the heart, we first performed the Langendorff perfu- ischemia and reperfusion.16 In the present study, we used sion of the heart ex vivo. After 24 hours of cold preservation, the adenovirus-mediated Bcl-xL gene transfer to the heart to heart was transferred to the Langendorff perfusion system and ascertain whether it could inhibit cold ischemia and warm perfused with a modified Krebs-Henseleit solution gassed with 95% O2 and 5% CO2 through the aorta at 85 mm Hg of perfusion reperfusion injury and prolong the cold preservation time of pressure.16 The heart was perfused at 37°C for 2 hours, the coronary heart grafts in a rat heterotopic transplantation model. effluent flow for the first 15 minutes after the start of reperfusion was collected for creatine kinase (CK) measurement,21 and CK activity Methods was measured by an enzymatic assay. Adenoviral Vectors TTC Staining The adenoviral vector (Adv) encoding human Bcl-xL17 or Esche- Triphenyltetrazolium chloride (TTC) staining has been a standard for richia coli ␤-galactosidase (LacZ)18,19 was used in the experiment. the measurement of infarct size and has been used previously for Adv propagation and purification were described previously.18 Be- assessment of infarct size resulting from apoptosis and necrosis.22 fore use, the viral titer (particle units/mL [PU/mL]) and the contam- Thus, we performed TTC staining to detect the infarct size of the ination of replication-competent Adv in the viral stock were evalu- heart after 2 hours of Langendorff perfusion. The heart was cut ated as previously reported.19 evenly into 4 slices transversely from the apex to the base and immersed into 1% TTC (Sigma-Aldrich) in PBS at 37°C for 30 Animal Model minutes, and then the reaction was stopped by addition of 10% Syngeneic Lewis rats (male, 250 to 350 g) were purchased from neutralized formalin. The heart sections were photographed and Sankyo Labo Service (Tokyo, Japan). The rats were anesthetized by captured as a digital image. Infarct size was analyzed with NIH intramuscular injection with ketamine (40 mg/kg) and xylazine (4 software, and infarct size percentage was calculated as infarct area (white area)/total area of the left ventricle (brick red areaϩwhite mg/kg). The donor rat was intubated with a 17-gauge needle tube and area).16 supported by a rodent respirator. A median thoracotomy was made and the aorta was exposed. AxCAhBclxL (1010 PU in 100 ␮Lof 0.9% saline) was injected into the heart at the root of the aorta. The Histological Examination aorta and pulmonary artery were clamped for 30 seconds after Adv After 2-hour perfusion in the Langendorff device or 24 hours of injection and then released (nϭ10). For control, 1010 PU AxCAZ3 in heterotopic heart transplantation, tissue was excised from the middle 100 ␮L of 0.9% saline or 100 ␮L of 0.9% saline was injected into rat of the heart between the apex and base, fixed with 10% neutralized hearts in the same way (nϭ10 each). formalin, dehydrated, and embedded in paraffin. Five-micron sec- After 4 days, the thoracotomy was re-opened, and the heart was tions were cut, and randomly selected sections were stained with arrested by perfusion with UWS at 4°C, collected, and stored in the hematoxylin and eosin. The sections were observed under the light microscope. UWS at 4°C for 24 hours. The heterotopic heart transplantation was performed as described previously.20 In brief, a laparotomy was made in the recipient rat. The infrarenal aorta and inferior vena cava TUNEL Staining were exposed. The heart transplantation was performed by an After 24 hours of cold preservation and 15 minutes after heart end-to-side anastomosis of the donor’s aorta to the recipient’s transplantation, tissue was excised from the middle of the heart infrarenal aorta and the donor’s pulmonary artery to the recipient’s between the apex and base, and cryosections were cut and randomly inferior vena cava. Graft survival was determined by daily palpation. selected for terminal dUTP nick end-labeling (TUNEL) staining as 16 The study was performed in accordance with institutional guidelines previously described. Nuclei were counterstained with methyl for animal experiments. green. The total numbers of nuclei and of TUNEL-positive nuclei were counted in 5 randomly chosen fields of view per tissue section in a blinded manner, and results were expressed as the number of Detection of Bcl-xL Expression in the Heart After TUNEL-positive nuclei divided by the number of counterstained Adv Transduction of the Heart nuclei. To determine the effectiveness of Bcl-xL gene transfer in the heart via aortic root injection, both X-gal staining of hearts injected with Immunochemical Analysis of Bcl-xL, Bax, and 1010 PU AxCAZ3 and Bcl-xL immunostaining of hearts injected with 10 Cyto-c in the Heart 10 PU AxCAhBclxL were performed. X-gal staining was per- After 24 hours of cold preservation and 15 minutes after transplan- 16 formed as described previously. For Bcl-xL immnofluorescence tation, the heart was collected. Bcl-xL expression was determined by detection, cryosections of heart were fixed with 4% paraformalde- Western blotting as previously described.16 hyde, permeabilized with 0.1% Triton X-100 in phosphate-buffered To determine Bax translocation and cyto-c release after 15 saline (PBS), and blocked with 3% bovine serum albumin in PBS. minutes of heart transplantation, the heart was immersed in 4 mL The sections were incubated with anti-human Bcl-xL antibody lysis buffer containing 250 mmol/L sucrose, 20 mmol/L HEPES, (65186E, BD PharMingen) and anti–␣-sarcomeric actinin monoclo- 10 mmol/L KCl, 1 mmol/L MgCl2, 1 mmol/L EDTA, 1 mmol/L nal antibody (EA-53, Sigma-Aldrich) for 30 minutes, washed with EGTA, 1 mmol/L dithiothreitol, and 1 mmol/L phenylmethylsulfo- PBS, and then incubated with goat anti-rabbit IgG and goat anti- nyl fluoride, pH 7.5, and incubated for 5 minutes on ice.23 The heart mouse IgG secondary antibody coupled to Alexa 488 or Alexa 594 was homogenized with a mixing homogenizer (Kinematica AG), and (Molecular Probes, Inc). All images were captured with a confocal the suspension was centrifuged at 750g for 10 minutes at 4°C to microscope (Zeiss/Pascal). sediment the nuclear fraction. The supernatant was collected and To determine the duration of Bcl-xL expression after Bcl-xL gene centrifuged at 12 000g for 10 minutes at 4°C to sediment the transfer, the time course of Bcl-xL expression in the heart was mitochondrial fraction. The resultant supernatant was further centri- determined by Western blotting and immunostaining. Western blot- fuged at 14 000g for 10 minutes at 4°C and then filtered through a ting of Bcl-xL was performed as previously described.16 For immu- 0.22-␮m ultrafilter (Millipore) to generate a purified cytosolic nostaining of Bcl-xL, an anti-human Bcl-xL antibody (65186E, BD fraction.24 78 Circulation July 5, 2005

Immunoblotting was performed according to standard protocols. The protein concentration was measured by the bicinchoninic acid method (Pierce). Aliquots of 40 ␮g of each sample were loaded on 15% sodium dodecyl sulfate polyacrylamide gels, subjected to electrophoresis, and then transferred to nitrocellulose membranes. The membranes were blocked with 5% milk in Tris-buffered saline containing 0.1% Tween 20 at 4°C overnight and then probed with rabbit antibody to Bax (sc-526, Santa, Cruz Biotechnology) and mouse antibody to cyto-c (7H8.2C12, PharMingen), followed ac- cordingly by horseradish peroxidase–conjugated anti-rabbit IgG (Zymed) or horseradish peroxidase–conjugated anti-mouse IgG (Amersham Biosciences) at room temperature for 1 hour. The horseradish peroxidase was detected with a chemoluminescence ECL-Plus kit (Amersham Biosciences UK). Cyto-c oxidase IV was detected with mouse antibody to cyto-c oxidase IV (COX, A-6431, Molecular Probes) as the mitochondrial marker, and ␣-tubulin was detected with mouse antibody to ␣-tubulin (B-5-1-2, Sigma-Aldrich) as an internal protein control.

Statistical Analysis Data were expressed as meanϮSD. Statistical comparisons were performed with ANOVA followed by Bonferroni/Dunn testing. A probability value Ͻ0.05 was considered statistically significant.

Results Efficiency of Bcl-xL Gene Transfer and Duration of Expression of Bcl-xL in the Heart X-gal staining showed that the Adv expression was distrib- uted in both the left and right ventricles after aortic root injection of Adv vector (Figure 1A) and that the transductive efficiency for cardiomyocytes was Ϸ45% to 50% (data not shown). Bcl-xL expression was also clearly proved by im- munostaining methods (Figure 1B). Both immunostaining and Western blotting showed that after Adv gene transduc- tion, the expression of Bcl-xL was strong for 2 to 7 days after gene transfer, then decreased, and almost reached normal levels after 21 days (Figure 1C and 1D). Figure 1. A and B, LacZ and Bcl-xL expression in rat hearts after Adv gene transduction via aortic root injection. Rat Bcl-xL Gene Transfer Decreased Infarct Size and hearts were harvested after 4 days of Adv gene transduction, Inhibited CK Release in the Heart and LacZ and Bcl-xL expressions were detected by X-gal staining and separate immunostaining . A, LacZ detection by The effects of Adv-mediated Bcl-xL gene transfer on infarct X-gal staining in whole hearts. B, Bcl-xL detection by immu- size and CK release were assessed in isolated rat hearts nostaining. Left, normal heart; right, Bcl-xL gene–transduced perfused with the Langendorff device. As shown in Figure 2A heart. Cardiac myocytes appear red, Bcl-xL–positive cardiac myocytes appear green, Barϭ50 ␮m. C and D, Time course and 2B, after 24-hour preservation in UWS at 4°C and 2-hour of Bcl-xL expression after Adv Bcl-xL gene transduction via reperfusion at 37°C, infarct size was 23.0Ϯ2.6% in the aortic root injection. C, Bcl-xL expression detected by immu- AxCAhBclxL-treated group, 47.7Ϯ7.0% in the saline group, nostaining. D, Bcl-xL expression detected by Western blot- Ϯ ϭ ting. ␤-Tubulin was used as internal control. 1, Normal heart and 48.6 6.1% in the AxCAZ3 group (n 4 in each group). and after AxCAhBclxL transfection: 2, 2 days; 3, 4 days; 4, 7 Bcl-xL gene transfer significantly reduced the infarct size days; 5, 14 days; 6, and 21 days. Magnification ϫ400. compared with saline or AxCAZ3 (PϽ0.01). Histological examination showed that in control hearts, a large proportion of cardiomyocytes underwent cell membrane rupture, Bcl-xL Gene Transfer Prolonged Survival of whereas AxCAhBclxL-treated heart did not show such Cardiac Isografts After Cold Ischemia and Warm changes (data not shown). The coronary effluent was col- Reperfusion Insult lected during the first 15 minutes of reperfusion, and total CK To determine whether Bcl-xL prolonged the preservation level was measured as an index of myocardial damage. As time, rat hearts that had been transduced with Bcl-xL gene shown in Figure 2C, the CK level in hearts treated with were harvested and preserved in UWS at 4°C for 24 hours; AxCAhBclxL was 0.82Ϯ0.27 IU; in contrast, the CK level subsequently, heterotopic heart transplantation was was 1.57Ϯ0.33 IU in saline-treated hearts and 1.50Ϯ0.37 IU performed. in AxCAZ3-treated hearts (nϭ4 each). Bcl-xL gene transfer As shown in Figure 3A and 3B, 70% (7 of 10) of the significantly decreased the CK level in the perfusate control hearts that underwent 24 hours of cold ischemia (PϽ0.05). ceased to beat as early as Ͻ15 minutes after transplantation Huang et al Bcl-xL Prolongs Cardiac Cold Preservation 79

Figure 2. Effect of Bcl-xL gene transfer on infarct size and CK release. Rat hearts were stored at 4°C for 24 hours, perfused in Langendorff device at 37°C for 2 hours, and then cut evenly into 4 slices trans- versely from apex to base. A, Representa- tive TTC-stained sections from perfused heart. Top, saline-treated heart; middle, AxCAZ3 (LacZ)-treated heart; bottom, AxCAhBclxL (Bcl-xL)-treated heart. B, Infarct size determined by TTC staining. Data are meanϮSD, nϭ4, **PϽ0.01. C, CK activity in coronary effluent measured for first 15 minutes of reperfusion. Value rep- resent meanϮSD, nϭ4, *PϽ0.05.

into syngeneic Lewis rats. Only 1 graft survived Ͼ14 days in clear leukocytes in the interstitium of the myocardium. In the control group. In contrast, 88% (8 of 9 rats [1 rat died contrast, in the AxCAhBclxL-treated hearts, myocardial because of complications of surgery]) of cardiac isografts structure was well preserved, and only a few infiltrating with Bcl-xL gene transfer survived Ͼ14 days. leukocytes were observed in the myocardium. Bcl-xL Gene Transfer Decreased Cardiac Cell Bcl-xL Expression After 24 Hours of Cold Death and Inflammatory Cell Infiltration After Preservation and 15 Minutes After Heterotopic Heart Transplantation Cardiac cell apoptosis was detected after 24 hours of cold Heart Transplantation preservation and 15 minutes after heart transplantation. The We previously reported that preadministration of AxCAhB- 16 average number of TUNEL-positive nuclei was 6.5Ϯ1.9% in clxL resulted in a robust Bcl-xL expression in the rat heart. AxCAhBclxL-treated hearts compared with 18.9Ϯ6.1% in To determine whether cold preservation and warm reperfu- saline-treated hearts (Figure 4). Thus, Bcl-xL significantly sion had some effect on the expression of Bcl-xL, immuno- decreased cardiac cell apoptosis (PϽ0.01). blot analysis of the heart after 24 hours of cold preservation Hematoxylin-eosin staining was performed after 24 hours with or without heart transplantation was performed. of heart transplantation. As shown in Figure 5, there were As shown in Figure 6, expression of Bcl-xL decreased after large numbers of cardiomyocytes that underwent necrosis in 24 hours of cold preservation and 15 minutes after heart the control saline and AxCAZ3-treated rat hearts, as evi- transplantation in control hearts (lanes 6, 9). In contrast, even denced by infiltration of polymorphonuclear and mononu- after 24 hours of cold preservation with or without heart

Figure 3. Graft survival after 24 hours of cold preservation in rat heterotopic heart transplantation model. A, Survival curve after heterotopic heart transplantation in hearts treated with AxCAhBclxL, saline, and AxCAZ3 (nϭ10 in each group). B, Survival time of individual hearts in each group after heterotopic heart transplantation. 80 Circulation July 5, 2005

Figure 4. Detection of apoptosis in hearts after 24 hours of cold preservation and 15 minutes after heart transplanta- tion by TUNEL staining. a, normal heart; b, heart treated with saline; c, heart treated with AxCAhBclxL. Normal nuclei appear light blue; TUNEL-positive nuclei appear dark brown. Magnification ϫ400. Value represent meanϮSD, nϭ4, *PϽ0.01.

transplantation, Bcl-xL still showed high expression in and reperfusion caused Bax translocation from the cytosol to AxCAhBclxL-transfected hearts (lanes 2, 3). the mitochondria, resulting in increased release of cyto-c from the mitochondria to the cytosol. Overexpression of Overexpression of Bcl-xL Inhibited Bax Translocation Bcl-xL inhibited Bax translocation and cyto-c release. Cold ischemia and reperfusion cause dysfunction of mito- chondria and cyto-c release.15 To examine the effect of Discussion Bcl-xL overexpression on Bax and cyto-c after 24 hours of Miniati et al25 found that Adv up-regulation of Bcl-2 inhibited cold preservation and 15 minutes after heart transplantation, oxidative stress and graft coronary artery disease in rat heart we determined the effect of Bcl-xL overexpression on the transplants. The authors performed the syngeneic-allogeneic subcellular localization of Bax and cyto-c by immunoblot retransplantation technique to allow for maximal translation analysis. of the Bcl-2 gene product at the time of allogeneic transplant As shown in Figure 7, 24-hour cold preservation and reperfusion. In the present study, we preadministered Adv 15-minute reperfusion after heart transplantation resulted in vector to the donor heart 4 days before transplantation and Bax loss in the cytosol and release of cyto-c from the found that Bcl-xL was efficiently expressed in the heart after mitochondria to the cytosol in controls (lanes 2, 3). In Adv vector transduction via the root of aorta. Overexpression contrast, overexpression of Bcl-xL by Adv transduction of the of Bcl-xL in the heart inhibited the cold ischemia and warm heart prevented Bax loss from the cytosol and decreased the reperfusion injury by decreasing cardiac cell apoptosis and cyto-c release (lane 4). The data show that cold preservation prolonged the preservation time of heart grafts. Rat hearts

Figure 5. Histological examination of hearts after 24 hours of cold preservation and 24 hours after heart transplantation. a, e, normal heart, no cold preservation; b, f, heart treated with saline; c, g, heart treated with AxCAZ3; d, h, heart treated with AxCAhBclxL. Magni- fications are ϫ100 (upper panel) and ϫ400 (low panel). In hearts treated with AxCAhBclxL, inflammatory and polymorphonuclear cell infiltration was obviously less than in hearts treated with saline or AxCAZ3. Huang et al Bcl-xL Prolongs Cardiac Cold Preservation 81

Figure 6. A, Representative Western blot analysis of Bcl-xL after 24 hours of cold preservation and 15 minutes after heart transplanta- tion. Lanes 1, 4, and 7, normal control heart; lanes 2 and 3, heart treated with AxCAhBclxL; lanes 5 and 6, heart treated with saline; lanes 8 and 9, heart treated with AxCAZ3; lanes 2, 5, and 8, heart after 24 hour of cold preservation; lanes 3, 6, and 9, heart after 24 hours of cold preservation and 15 minutes after heart transplantation. ␤-Tubulin was used as internal control. B, Densitomery of Bcl-xL expression in heart after 24 hours of cold preservation and 15 minutes after heart transplantation. Value represent meanϮSD, nϭ3, *PϽ0.05. exposed to a prolonged period of cold ischemia (24 hours, In the field of heart transplantation, various attempts 4°C) failed to function after transplantation into syngeneic have been made to prolong cardiac preservation time. Most recipients.26 Our study suggests that Bcl-xL can serve as a of them have been designed to limit energy loss, maintain powerful cardiac preservation agent in heart transplantation. intracellular ionic composition, and prevent hypothermic myocyte swelling; however, even with optimal preserva- tion after a long period of cold preservation, reperfusion of the heart causes dysfunction and apoptosis and/or necrosis of cardiac myocytes. There have been reports that Adv- mediated Bcl-xL16 or TAT-BH427 inhibited apoptosis of cardiac myocytes caused by warm ischemia and reperfu- sion injury. In the present study, overexpression of Bcl-xL was shown to decrease cardiac injury in a stringent and clinically relevant model of 24-hour cold ischemia and warm reperfusion followed by syngeneic transplantation. Ex vivo experiments also showed that after 24-hour cold preservation and 2-hour reperfusion at 37°C, infarct size and CK levels were significantly reduced by Bcl-xL gene transfer. Mitochondria play a critical role in cell apoptosis.28,29 In the heart, prolonged cold ischemia and warm reperfusion Figure 7. A, Representative Western blot analysis of Bax and jeopardize the myocardial capacity to regenerate energy by cyto-c in mitochondrial and cytosolic fractions after 24 hours of 30 cold preservation and 15 minutes after heart transplantation. mitochondrial oxidative phosphorylation and are corre- Left, mitochondrial fraction; right, cytosolic fraction. 1, Normal lated with a progressive loss of mitochondrial function.31 control heart; 2, heart treated with saline; 3, heart treated with Cyto-c release from mitochondria occurs in both normo- AxCAZ3; 4, heart treated with AxCAhBclxL. COX-V was used as thermic32 and cold ischemia and reperfusion,15 and cyto-c mitochondrial marker; ␣-tubulin was used as internal control of cytosolic fraction. B, Densitometry of cytosolic Bax. Results are release activates the caspase pathway, with consequent representative of 3 independent experiments. induction of apoptosis.33 Furthermore, in pathophysiolog- 82 Circulation July 5, 2005 ical settings, apoptosis frequently induces inflammation 4. Thatte HS, Rhee JH, Zagarins SE, Treanor PR, Birjiniuk V, Crittenden because of the onset of secondary necrosis and stimulation MD, Khuri SF. Acidosis-induced apoptosis in human and porcine heart. Ann Thorac Surg. 2004;77:1376–1383. of cytokine and chemokine expression, which inevitably 5. Harrison GJ, Willis RJ, Headrick JP. Extracellular adenosine levels and result in the demise of adjacent cells that are not directly cellular energy metabolism in ischemically preconditioned rat heart. Car- damaged by the original insult. Necrosis is another form of diovasc Res. 1998;40:74–87. cardiac cell death after cold preservation and reperfusion. 6. Flitter WD. Free radicals and myocardial reperfusion injury. Br Med Bull. 1993;49:545–555. The cardiac necrosis that was evidenced by infarct size 7. Fliss H, Gattinger D. Apoptosis in ischemic and reperfused rat myocar- after ex vivo perfusion (Figure 2A) and inflammatory dium. Circ Res. 1996;79:949–956. response after heart transplantation (Figure 5d and 5h) 8. Zhao ZQ, Nakamura M, Wang NP, Wilcox JN, Shearer S, Ronson RS, Guyton RA, Vinten-Johansen J. Reperfusion induces myocardial apoptot- were decreased in the AxCAhBclxL-treated group, which ic cell death. Cardiovasc Res. 2000;45:651–660. indicated that Bcl-xL may play a role in antinecrosis. 9. Chatterjee S, Stewart AS, Bish LT, Jayasankar V, Kim EM, Pirolli T, Recently, it has been suggested that cyto-c release poten- Burdick J, Woo YJ, Gardner TJ, Sweeney HL. Viral gene transfer of the tially induces necrosis by depleting cellular ATP levels15; antiapoptotic factor Bcl-2 protects against chronic postischemic heart failure. Circulation. 2002;106(suppl I):I-212–I-217. thus, the pathway leading to apoptosis or necrosis can be 10. Brocheriou V, Hagege AA, Oubenaissa A, Lambert M, Mallet VO, shared.34 The heterogeneity and extent of cyto-c release are Duriez M, Wassef M, Kahn A, Menasche P, Gilgenkrantz H. Cardiac critical for regulating the switch between alternative de- functional improvement by a human Bcl-2 transgene in a mouse model of 15 ischemia/reperfusion injury. J Gene Med. 2000;2:326–333. velopment of apoptosis or necrosis. 11. Katori M, Buelow R, Ke B, Ma J, Coito AJ, Iyer S, Southard D, Busuttil Thus, before irreversible changes occur in the mitochon- RW, Kupiec-Weglinski JW. Heme oxygenase-1 overexpression protects dria, inhibition of cyto-c release and protection of mitochon- rat hearts from cold ischemia/reperfusion injury via an antiapoptotic drial integrity would be important in preventing heart injury pathway. Transplantation. 2002;73:287–292. 12. Nakamura T, Mizuno S, Matsumoto K, Sawa Y, Matsuda H, Nakamura caused by prolonged cold preservation and warm reperfusion. T. Myocardial protection from ischemia/reperfusion injury by endoge- Cyclosporin A, an inhibitor of mitochondrial permeability nous and exogenous HGF. J Clin Invest. 2000;106:1511–1519. transition, was proven to inhibit the apoptosis of cultured 13. Yamamura T, Otani H, Nakao Y, Hattori R, Osako M, Imamura. IGF-I differentially regulates Bcl-xL and Bax and confers myocardial protection neonate cardiac myocytes by decreasing the release of cyto- in the rat heart. Am J Physiol Heart Circ Physiol. 2001;280: c.23 In the present study, we showed that prolonged cold H1191–H1200. expression and warm reperfusion decreased Bcl-xL level of 14. He L, Perkins GA, Poblenz AT, Harris JB, Hung M, Ellisman MH, Fox the heart, but in Bcl-xL gene–transduced hearts, even after 24 DA. Bcl-xL overexpression blocks bax-mediated mitochondrial contact site formation and apoptosis in rod photoreceptors of lead-exposed mice. hours of cold preservation followed by 15 minutes of heart Proc Natl Acad Sci U S A. 2003;100:1022–1027. transplantation, Bcl-xL still had strong expression. TUNEL 15. Kuznetsov AV, Schneeberger S, Seiler R, Brandacher G, Mark W, staining showed that after a long period of cold preservation, Steurer W, Saks V, Usson Y, Margreiter R, Gnaiger E. Mitochondrial cardiac cell apoptosis during the early stage of reperfusion in defects and heterogeneous cytochrome c release after cardiac cold ische- mia and reperfusion. Am J Physiol Heart Circ Physiol. 2004;286: the control group was significantly higher than in the H1633–H1641. AxCAhBclxL-treated group. We found that overexpression 16. Huang J, Ito Y, Morikawa M, Uchida H, Kobune M, Sasaki K, Abe T, of Bcl-xL inhibited the translocation of Bax to the mitochon- Hamada H. Bcl-xL gene transfer protects the heart against ischemia/ reperfusion injury. Biochem Biophys Res Commun. 2003;311:64–70. drial and decreased the release of cyto-c from the mitochon- 17. Shinoura N, Yoshida Y, Asai A, Kirino T, Hamada H. Relative level of dria to the cytosol. This is in accordance with results in a expression of Bax and Bcl-XL determines the cellular fate of apoptosis/ transgenic mice model in which overexpression of Bcl-xL necrosis induced by the overexpression of Bax. Oncogene. 1999;18: prevented retinal cell apoptosis by inhibiting the translocation 5703–5713. 18. Dehari H, Ito Y, Nakamura T, Kobune M, Sasaki K, Yonekura N, 14 of Bax. Therefore, inhibition of Bax translocation may be Kohama G, Hamada H. Enhanced antitumor effect of RGD fiber- one of the mechanisms through which Bcl-xL protects the modified adenovirus for gene therapy of oral cancer. Cancer Gene Ther. heart against prolonged cold preservation and warm reperfu- 2003;10:75–85. 19. Takahashi K, Ito Y, Morikawa M, Kobune M, Huang J, Tsukamoto M, sion injury. Sasaki K, Nakamura K, Dehari H, Ikeda K, Uchida H, Hirai S, Abe T, Our finding validated the feasibility of use of the Bcl-xL Hamada H. Adenoviral-delivered angiopoietin-1 reduces the infarction gene in heart transplantation. This may contribute to the and attenuates the progression of cardiac dysfunction in the rat model of development of a novel method aimed at prolongation of acute myocardial infarction. Mol Ther. 2003;8:584–592. 20. Ono K, Lindsey ES. Improved technique of heart transplantation in rats. cardiac cold preservation time. J Thorac Cardiovasc Surg. 1969;57:225–229. 21. Granville DJ, TASHAkkor B, Takeuchi C, Gustafsson AB, Huang C, Acknowledgments Sayen MR, Wentworth P Jr, Yeager M, Gottlieb RA. Reduction of This work was supported in part by a grant to Drs Hamada and Ito ischemia and reperfusion-induced myocardial damage by cytochrome from the Ministry of Education, Science, Japan. P450 inhibitors. Proc Natl Acad Sci U S A. 2004;101:1321–1326. 22. McCully JD, Wakiyama H, Hsieh YJ, Jones M, Levitsky S. Differential contribution of necrosis and apoptosis in myocardial ischemia- References reperfusion injury. Am J Physiol Heart Circ Physiol. 2004;286: 1. Stringham JC, Southard JH, Hegge J, Triemstra L, Fields BL, Belzer FO. H1923–H1935. Limitations of heart preservation by cold storage. Transplantation. 1992; 23. Shiraishi J, Tatsumi T, Keira N, Akashi K, Mano A, Yamanaka S, Matoba 53:287–294. S, Asayama J, Yaoi T, Fushiki S, Fliss H, Nakagawa M. Important role 2. Masters TN, Fokin AA, Schaper J, Pool L, Gong G, Robicsek F. Changes of energy-dependent mitochondrial pathways in cultured rat cardiac in the preserved heart that limit the length of preservation. J Heart Lung myocyte apoptosis. Am J Physiol Heart Circ Physiol. 2001;281: Transplant. 2002;21:590–599. H1637–H1647. 3. Siegmund B, Zude R, Piper HM. Recovery of anoxic-reoxygenated car- 24. Deng Y, Wu X. Peg3/Pw1 promotes p53-mediated apoptosis by inducing diomyocytes from severe Ca2ϩ overload. Am J Physiol. 1992;263(pt Bax translocation from cytosol to mitochondria. Proc Natl Acad Sci 2):H1262–1269. USA. 2000;97:12050–12055. Huang et al Bcl-xL Prolongs Cardiac Cold Preservation 83

25. Miniati DN, Lijkwan MA, Murata S, Martens J, Coleman CT, Hoyt EG, 30. Kay L, Daneshrad Z, Saks VA, Rossi A. Alteration in the control of Robbins RC. Effects of adenoviral upregulation of bcl-2 on oxidative mitochondrial respiration by outer mitochondrial membrane and creatine stress and graft coronary artery disease in rat heart transplants Transplan- during heart preservation. Cardiovasc Res. 1997;34:547–556. tation. 2003;76:382–386. 31. Kuwabara M, Takenaka H, Maruyama H, Onitsuka T, Hamada M. Effect 26. Akamatsu Y, Haga M, Tyagi S, Yamashita K, Graca-Souza AV, Ollinger of prolonged hypothermic ischemia and reperfusion on oxygen con- R, Czismadia E, May GA, Ifedigbo E, Otterbein LE, Bach FH, Soares sumption and total mechanical energy in rat myocardium: participation of MP. Heme oxygenase-1–derived carbon monoxide protects hearts from mitochondrial oxidative phosphorylation. Transplantation. 1997;64: transplant associated ischemia reperfusion injury. FASEB J. 2004;18: 577–583. 32. Weiss JN, Korge P, Honda HM, Ping P. Role of the mitochondrial 771–772. permeability transition in myocardial disease. Circ Res. 2003;93: 27. Sugioka R, Shimizu S, Funatsu T, Tamagawa H, Sawa Y, Kawakami T, 292–301. Tsujimoto Y. BH4-domain peptide from Bcl-xL exerts anti-apoptotic 33. Gottlieb RA, Burleson KO, Kloner RA, Babior BM, Engler RL. Reper- activity in vivo. Oncogene. 2003;22:8432–8440. fusion injury induces apoptosis in rabbit cardiomyocytes. J Clin Invest. 28. Kluck RM, Bossy-Wetzel E, Green DR, Newmeyer DD. The release of 1994;94:1621–1628. cytochrome c from mitochondria: a primary site for Bcl-2 regulation of 34. Tatsumi T, Shiraishi J, Keira N, Akashi K, Mano A, Yamanaka S, Matoba apoptosis. Science. 1997;275:1132–1136. S, Fushiki S, Fliss H, Nakagawa M. Intracellular ATP is required for 29. Kroemer G, Reed JC. Mitochondrial control of cell death. Nat Med. mitochondrial apoptotic pathways in isolated hypoxic rat cardiac 2000;6:513–519. myocytes. Cardiovasc Res. 2003;59:428–440. Augmented Cardiac Hypertrophy in Response to Pressure Overload in Mice Lacking the Prostaglandin I2 Receptor Akiyoshi Hara, PhD; Koh-ichi Yuhki, PhD; Takayuki Fujino, MD; Takehiro Yamada, PhD; Koji Takayama, MD; Shuhko Kuriyama, MD; Osamu Takahata, MD; Hideji Karibe, PhD; Yuji Okada, MD; Chun-Yang Xiao, PhD; Hong Ma, MD; Shuh Narumiya, MD; Fumitaka Ushikubi, MD

Background—In the heart, the expressions of several types of prostanoid receptors have been reported. However, their roles in cardiac hypertrophy in vivo remain unknown. We intended to clarify the roles of these receptors in pressure overload–induced cardiac hypertrophy using mice lacking each of their receptors. Methods and Results—We used a model of pressure overload–induced cardiac hypertrophy produced by banding of the transverse aorta in female mice. In wild-type mice subjected to the banding, cardiac hypertrophy developed during the Ϫ/Ϫ observation period of 8 weeks. In mice lacking the prostaglandin (PG) I2 receptor (IP ), however, cardiac hypertrophy and cardiomyocyte hypertrophy were significantly greater than in wild-type mice at 2 and 4 weeks but not at 8 weeks, whereas there was no such augmentation in mice lacking the prostanoid receptors other than IP. In addition, cardiac fibrosis observed in wild-type hearts was augmented in IPϪ/Ϫ hearts, which persisted for up to 8 weeks. In IPϪ/Ϫ hearts, the expression level of mRNA for atrial natriuretic peptide, a representative marker of cardiac hypertrophy, was significantly higher than in wild-type hearts. In vitro, cicaprost, an IP agonist, reduced platelet-derived growth factor–induced proliferation of wild-type noncardiomyocytes, although it could not inhibit cardiotrophin-1–induced hypertrophy of cardiomyocytes. Accordingly, cicaprost increased cAMP concentration efficiently in noncardiomyocytes. Conclusions—IP plays a suppressive role in the development of pressure overload–induced cardiac hypertrophy via the inhibition of both cardiomyocyte hypertrophy and cardiac fibrosis. Both effects have been suggested as originating from the action on noncardiomyocytes rather than cardiomyocytes. (Circulation. 2005;112:84-92.) Key Words: hypertrophy Ⅲ myocardium Ⅲ pressure Ⅲ prostaglandins Ⅲ thromboxanes

6–8 ardiac hypertrophy in response to pressure overload is a system. PGI2 is known as a vasodilator and an inhibitor of Ccompensatory mechanism to maintain circulatory ho- platelet activation. In contrast, thromboxane A2 is a potent meostasis.1 When the mechanical overload on the heart is vasoconstrictor and a platelet stimulator. In the heart, several severe and prolonged, however, it is an important risk factor prostanoids, such as PGE2 and PGI2, have been reported to in cardiac morbidity and mortality.2 In addition, many pa- show cardioprotective actions against ischemia/reperfusion tients suffer from this condition because hypertension is a injury.9–11 In regard to cardiac hypertrophy, several reports common cardiovascular disease that over time leads to suggest the participation of prostanoids in its development. cardiac hypertrophy. Therefore, the mechanisms underlying These include inhibitory effects of PGI2 and its analogue on the development of pressure overload–induced cardiac hyper- angiotensin II–induced hypertrophy in cultured cardiomyo- trophy have been studied extensively. Until now, several cytes12 and their inhibitory effects on proliferation and extracellular signaling molecules, such as endothelin, angio- collagen synthesis in cultured cardiac fibroblasts.13 In addi- tensin II, and cardiotrophin-1, have been proposed as medi- tion, synthesis of PGE2 and PGI2 has been reported to be ators promoting pressure overload–induced cardiac hypertro- elevated in the hypertrophied and failing heart14,15 along with phy.3–5 In contrast, little is known about these signaling an upregulation of cyclooxygenase (COX)-2.16 The hypertro- molecules that ameliorate pressure overload–induced cardiac phic effect of exogenously administered PGF2␣ on cardio- hypertrophy. myocytes has also been reported.17,18 Furthermore, several Prostanoids, consisting of prostaglandins (PGs) and types and subtypes of prostanoid receptors have been re- thromboxane, exert various actions in the cardiovascular ported to be expressed in the heart.6,7 These results suggest

Received June 28, 2004; de novo received December 1, 2004; revision received February 28, 2005; accepted March 9, 2005. From the Department of Pharmacology, Asahikawa Medical College, Asahikawa (A.H., K.Y., T.F., T.Y., K.T., S.K., O.T., H.K., Y.O., C.-Y.X., H.M., F.U.), and the Department of Pharmacology, Kyoto University Faculty of Medicine, Kyoto (S.N.), Japan. The online-only Data Supplement can be found with this article at http://circ.ahajournals.org/cgi/content/full/CIRCULATIONAHA.104.527077/DC1. Correspondence to Fumitaka Ushikubi, MD, Department of Pharmacology, Asahikawa Medical College, Midorigaoka-Higashi 2-1-1-1, Asahikawa 078-8510, Japan. E-mail [email protected] © 2005 American Heart Association, Inc. Circulation is available at http://www.circulationaha.org DOI: 10.1161/CIRCULATIONAHA.104.527077 84 Hara et al Role of PGI2 Receptor in Cardiac Hypertrophy 85 some roles played by these receptors in the development of TABLE 1. Effects of Aortic Banding on Heart Rate and Blood pressure overload–induced cardiac hypertrophy; their exact Pressure in Mice Lacking Prostanoid Receptors roles in vivo, however, remain unknown. n HR, bpm SBP, mm Hg DBP, mm Hg DP, EP, FP, IP, and TP are the receptors for PGD , PGE , 2 2 WT 14 631Ϯ18 100Ϯ372Ϯ2 PGF2␣, PGI2, and thromboxane A2, respectively. In addition, Ϫ/Ϫ Ϯ Ϯ Ϯ 6,7 EP2 5 649 11 90 2652 there are 4 subtypes of EP: EP1,EP2,EP3, and EP4. In the Ϫ/Ϫ Ϯ Ϯ Ϯ present study we intended to clarify the roles of prostanoid EP3 5 598 19 106 2762 receptors in pressure overload–induced cardiac hypertrophy FPϪ/Ϫ 4 643Ϯ21 105Ϯ368Ϯ5 using mice each lacking one of these receptors. IPϪ/Ϫ 8 636Ϯ21 101Ϯ576Ϯ4 TPϪ/Ϫ 4 637Ϯ17 97Ϯ470Ϯ5 Methods F2 WT 7 646Ϯ12 100Ϯ372Ϯ5 Ϫ/Ϫ Mice EP4 7 607Ϯ30 110Ϯ481Ϯ4 The generation and maintenance of mice lacking EP2,EP3,EP4, FP, Heart rate (HR), systolic blood pressure (SBP), and diastolic blood pressure Ϫ/Ϫ Ϫ/Ϫ Ϫ/Ϫ Ϫ/Ϫ Ϫ/Ϫ Ϫ/Ϫ IP, or TP (EP2 ,EP3 ,EP4 ,FP ,IP , and TP mice, Ϫ/Ϫ Ϫ/Ϫ (DBP) were measured by the tail-cuff method in wild-type (WT), EP2 ,EP3 , 19–24 respectively) have been reported. All mice, including the wild- Ϫ/Ϫ Ϫ/Ϫ Ϫ/Ϫ Ϫ/Ϫ Ϫ/Ϫ FP ,IP ,TP , F2 wild-type (F2 WT), and EP4 mice at 4 weeks of aortic type control mice, but with the exception of EP4 mice, have a Ϯ Ϫ/Ϫ banding. Each value represents mean SEM. genetic background of C57BL/6 mice. EP4 mice have a mixed genetic background of 129sv/ola and C57BL/6 mice.21 For the Ϫ/Ϫ experiments in which EP4 mice were used, F2 wild-type mice with were prepared from the middle segment of the heart at intervals of Ϫ/Ϫ a mixed genetic background similar to that of EP4 mice were used 0.3 mm. The sections were stained with hematoxylin and eosin for as a control. All experiments, which were approved by the Asa- examination of their gross appearance and were stained by the van hikawa Medical College Committee on Animal Research, were Gieson method for measurements of cardiomyocyte hypertrophy and performed with the use of 12- to 15-week-old female mice or 18- to cardiac fibrosis. Cardiomyocyte hypertrophy was assessed by mea- 20-day fetuses. suring cross-sectional area of 100 cardiomyocytes having nearly circular capillary profiles and nuclei in the left ventricle near the Expression of mRNAs for Prostanoid Receptors endocardial region. Cardiac fibrosis was assessed separately as To examine the expression of mRNAs for prostanoid receptors in the interstitial and perivascular fibrosis by calculating the ratio of van heart, we prepared total RNA from the heart using Isogen (Nippon Gieson–stained area of interstitial or perivascular fibrosis to total area of cardiac tissue in each section. These analyses were performed Gene) and performed reverse transcription–polymerase chain reac- by digital planimetry with the use of NIH Image computer tion (RT-PCR) as reported.11,25 A similar procedure was used in the software.27 examination of IP mRNA expression in cardiomyocytes and noncardiomyocytes. Expression of mRNAs for COXs and Atrial A Model of Pressure Overload–Induced Natriuretic Peptide Cardiac Hypertrophy We examined whether the cardiac expression of mRNAs for COX-1, a constitutive isoform, and COX-2, an inducible isoform, changes on Pressure overload was produced by banding of the transverse aorta pressure overload. We also examined cardiac expression of mRNA (transverse aortic constriction procedure) according to the reported for atrial natriuretic peptide (ANP), a representative marker of method26 with minor modifications. Briefly, mice anesthetized with cardiac hypertrophy. Total RNA was prepared from the left ventri- ketamine (100 mg/kg IP) and xylazine (5 mg/kg IP) were maintained cles at indicated times,11 and the expression levels of the mRNAs under a respirator (model 480-7, Shinano Industry). After thoracot- were determined by RT-PCR with the use of the expression level of omy, the transverse aorta was exposed between the carotid arteries 18S ribosomal RNA (rRNA) as an internal control. The sequences and was constricted by ligation with 7-0 nylon string along with a of the primers used were as follows: COX-1 (sense), 5ЈCTGCT- blunted 24-gauge needle, which needle was then pulled out. As GAGAAGGGAGTTCATT3Ј; COX-1 (antisense), 5ЈGTCGCA- controls, sham-operated mice were produced; they received essen- CACGCGGTTATGTT3Ј; COX-2 (sense), 5ЈACACTCTATCACTG- tially the same operation except for the aortic ligation. At indicated GCACCC3Ј; COX-2 (antisense), 5ЈGGACGAGGTTTTTCCA- times after the banding, wet weight of the heart (HW) and body CCAG3Ј; ANP (sense), 5ЈATGGGCTCCTTCTCCATCACC3Ј; ANP weight (BW) were measured, and their ratios (HW/BW ratio) were (antisense), 5ЈTCCACTCTGGGCTCCAATCCTGT3Ј; 18S rRNA used as indices of cardiac hypertrophy. (sense), 5ЈATCCTGCCAGTAGCATATGC3Ј; and 18S rRNA (anti- To assess the degree of pressure overload, a polyethylene cannula, sense), 5ЈCCGAGGTTATCTAGAGTCAC3Ј. connected to a pressure transducer, was inserted into the right and Ϫ Ϫ left carotid arteries in some of the wild-type and IP / mice. Arterial ␣ pressure was measured before and at a steady state within 10 minutes Measurements of 6-Keto-PGF1 Contents after the banding. in the Heart Blood pressure and heart rate of conscious mice were measured by To determine whether pressure overload increases PGI2 production ␣ the tail-cuff method (BP-98A, Softron) between 9 AM and noon at in the heart, we measured the tissue contents of 6-keto-PGF1 ,a indicated times after the banding, as reported.10 The values were stable metabolite of PGI2. Tissue samples were prepared from the left obtained by averaging at least 5 measurements. The results showed ventricle of wild-type mice at indicated times after the aortic ␣ that blood pressure and heart rate were not significantly different banding, and the contents of 6-keto-PGF1 were measured with an between wild-type and IPϪ/Ϫ mice at 1, 2, 4, or 8 weeks of the EIA kit (Cayman Chemical). banding (data not shown). Similarly, blood pressure and heart rate at 4 weeks of the banding were not significantly different among In Vitro Examination of Cardiomyocyte Ϫ/Ϫ Ϫ/Ϫ Ϫ/Ϫ Ϫ/Ϫ wild-type, EP2 ,EP3 ,FP , and TP mice or between F2 Hypertrophy and Noncardiomyocyte Proliferation Ϫ/Ϫ wild-type and EP4 mice (Table 1). Cultures of cardiomyocytes and noncardiomyocytes were performed as reported.11 In short, cardiac ventricles of fetal mice were minced Histological Analysis of the Heart and then incubated with a buffer containing 0.1% collagenase At indicated times, hearts were fixed with 10% formalin and (Wako) for 60 minutes at 37°C. The cells were filtrated through a embedded in paraffin. Five transverse sections (5-␮m thickness) nylon mesh, suspended in a culture medium (DMEM/F-12 supple- 86 Circulation July 5, 2005 mented with 100 U/mL penicillin and 100 ␮g/mL streptomycin) TABLE 2. Effects of Aortic Banding on BW, HW, and BW/HW containing 2.5% fetal calf serum, and then preplated onto a dish to Ratio in Mice Lacking Prostanoid Receptors separate cardiomyocytes from noncardiomyocytes. After incubation for 30 minutes, unattached cardiomyocytes were harvested and used HW/BW, for an assay. Attached noncardiomyocytes were grown to near n BW, g HW, mg mg/g confluence and then used for an assay. Sham-operated mice Hypertrophy of cardiomyocytes was estimated by [14C]leucine incorporation, and proliferation of noncardiomyocytes was by WT 14 22.5Ϯ0.6 104.3Ϯ2.6 4.7Ϯ0.1 3 Ϫ/Ϫ [ H]thymidine incorporation and cell number. Cardiomyocytes were EP2 6 21.6Ϯ0.8 99.4Ϯ3.9 4.6Ϯ0.1 5 plated into 24-well culture plates at 10 cells per well. Noncardio- EP Ϫ/Ϫ 5 22.8Ϯ0.2 106.0Ϯ2.1 4.6Ϯ0.1 myocytes were plated into 24-well culture plates at 104 cells per well 3 Ϫ Ϫ and into 6-well culture plates at 5ϫ104 cells per well for examination FP / 5 21.5Ϯ0.5 103.9Ϯ2.2 4.8Ϯ0.1 of [3H]thymidine incorporation and cell number, respectively. After IPϪ/Ϫ 8 23.7Ϯ0.8 113.1Ϯ5.7 4.6Ϯ0.1 48 hours of culture, culture medium was changed to a fresh one Ϫ/Ϫ Ϫ TP 5 22.0Ϯ0.7 105.9Ϯ2.9 4.8Ϯ0.1 containing vehicle or cicaprost (10 5 mol/L, Schering), and indo- methacin (10Ϫ5 mol/L, Sigma). In cardiomyocytes, cardiotrophin-1 F2 WT 7 21.6Ϯ0.5 99.9Ϯ3.3 4.7Ϯ0.2 Ϫ (10 10 mol/L, Genzyme/Techne) and [14C]leucine (0.1 mCi/mL, Ϫ/Ϫ EP4 7 22.9Ϯ1.5 110.7Ϯ5.0 4.9Ϯ0.1 Amersham) were added to the culture medium, and the cells were cultured for 48 hours. In noncardiomyocytes, platelet-derived growth Aortic-banded mice factor (PDGF) (Peprotec) or fetal calf serum was added at 5 ng/mL WT 8 22.7Ϯ0.5 151.6Ϯ9.1* 6.7Ϯ0.5* or 0.5%, respectively, to the culture medium 30 minutes after the Ϫ/Ϫ EP2 5 23.1Ϯ0.5 147.5Ϯ9.1* 6.4Ϯ0.3* medium change. For examination of [3H]thymidine incorporation, Ϫ/Ϫ Ϯ Ϯ Ϯ the cells were cultured for 24 hours, and then [3H]thymidine (2 EP3 7 25.1 0.6 158.1 9.6* 6.3 0.3* mCi/mL, Amersham) was added, and the cells were cultured for an FPϪ/Ϫ 5 21.7Ϯ1.0 152.7Ϯ11.7* 7.1Ϯ0.8* additional 6 hours. For cell number count, the cells were cultured for IPϪ/Ϫ 6 24.2Ϯ0.4 195.5Ϯ14.1*† 8.2Ϯ0.7*† 48 hours, harvested by trypsin-EDTA treatment, and counted with a Ϫ/Ϫ hemocytometer. Amounts of [14C]leucine and [3H]thymidine incor- TP 5 24.4Ϯ1.0 160.9Ϯ8.1* 6.6Ϯ0.4* porated into cardiomyocytes and noncardiomyocytes, respectively, F2 WT 6 23.8Ϯ0.7 150.9Ϯ12.4* 6.3Ϯ0.4* were quantified by liquid scintillation counter. Ϫ/Ϫ EP4 5 24.3Ϯ0.9 162.2Ϯ12.0* 6.7Ϯ0.5* Ϫ/Ϫ Ϫ/Ϫ Ϫ/Ϫ Ϫ/Ϫ Measurements of cAMP Accumulation BW and HW were measured in wild-type (WT), EP2 ,EP3 ,FP ,IP , Ϫ/Ϫ Ϫ/Ϫ The cardiomyocytes and noncardiomyocytes were preincubated for TP , F2 wild-type (F2 WT), and EP4 mice at 4 weeks of aortic banding or 30 minutes at 37°C in the culture medium containing 10Ϫ5 mol/L sham operation. Each value represents meanϮSEM. indomethacin and 1 mmol/L 3-isobutyl-1-methylxanthine (Sigma- *PϽ0.05 vs corresponding sham-operated mice; †PϽ0.05 vs aortic-banded Aldrich). Then the cardiomyocytes were stimulated with vehicle, wild-type mice. cicaprost (10Ϫ6 or 10Ϫ5 mol/L), or isoproterenol (10Ϫ7 mol/L, Nacalai Tesque), and the noncardiomyocytes were stimulated with vehicle or Ϫ/Ϫ Ϫ6 Ϫ5 ing that the procedure induced cardiac hypertrophy. In IP cicaprost (10 or 10 mol/L). After stimulation for 30 minutes at mice, however, HW and HW/BW ratio were significantly 37°C, the levels of intracellular cAMP ([cAMP]i) were measured as reported.25 We determined the protein contents of the cells by use of greater than those in wild-type mice, suggesting that IP a BCA protein assay kit (Pierce Chemical). mediated an antihypertrophic effect (Table 2 and Figures 1B and 1C). In contrast, there were no such differences among Ϫ/Ϫ Ϫ/Ϫ Ϫ/Ϫ Ϫ/Ϫ Statistical Analysis wild-type, EP2 ,EP3 ,FP , and TP mice or between All values are expressed as meanϮSEM. Statistical analysis was Ϫ/Ϫ F2 wild-type and EP4 mice (Table 2). In sham-operated performed with 1-way (Figure 5 and Tables 1 and 2) or 2-way groups, there was no significant difference in HW, BW, and (Figures 1C, 2B, 3B, and 4 and Table 3) ANOVA followed by Ϫ/Ϫ Ϫ/Ϫ Ϫ/Ϫ Duncan’s test for multiple comparisons. A difference was considered HW/BW ratio among wild-type, EP2 ,EP3 ,FP , and Ϫ/Ϫ Ϫ/Ϫ statistically significant at PϽ0.05 from the 2-tailed test. All data TP mice or between F2 wild-type and EP4 mice (Table were analyzed with the software program Super ANOVA, version 2). These results suggest that IP is the only prostanoid 1.11. receptor able to affect the development of pressure overload– induced cardiac hypertrophy, at least in the present model. Results When the time course of the hypertrophy was examined, it Expression of mRNAs for Prostanoid Receptors in was already apparent at 1 week after the banding, reached a the Heart maximum level at 2 weeks, and continued at a similar level We first examined which types and subtypes of prostanoid thereafter (Figure 1C and Table 3). The HW or HW/BW ratio receptors are expressed in the heart using the RT-PCR at 1 week was not significantly different between wild-type Ϫ/Ϫ method. We found the expression of mRNAs for EP2,EP3, and IP mice. At 2 and 4 weeks, however, both HW and Ϫ/Ϫ EP4, FP, IP, and TP but not for EP1 and DP (Figure 1A). HW/BW ratio in IP mice were significantly greater than those in wild-type mice (Figure 1C and Table 3). At 8 weeks In Vivo Model of Pressure Overload–Induced after the banding, however, there was no significant differ- Cardiac Hypertrophy ence in the HW/BW ratio between wild-type and IPϪ/Ϫ mice, On the basis of the results of RT-PCR analysis, we next whereas HW was still significantly greater in IPϪ/Ϫ mice than examined cardiac hypertrophy using a model of pressure in wild-type mice. There was no significant difference in BW Ϫ/Ϫ Ϫ/Ϫ Ϫ/Ϫ overload–induced cardiac hypertrophy in EP2 ,EP3 , after the banding between wild-type and IP mice through- Ϫ/Ϫ Ϫ/Ϫ Ϫ/Ϫ Ϫ/Ϫ EP4 ,FP ,IP , and TP mice. In wild-type mice, HW out the experiment (Table 3). These results clearly indicate and HW/BW ratio at 4 weeks after aortic banding were that IP participates in the suppression of pressure overload– significantly higher than that in sham-operated mice, indicat- induced cardiac hypertrophy. Hara et al Role of PGI2 Receptor in Cardiac Hypertrophy 87

Figure 2. Cardiomyocyte hypertrophy induced by pressure overload in wild-type and IPϪ/Ϫ mice. A, Histological manifesta- Figure 1. Expression of mRNAs for prostanoid receptors in the tion of cardiomyocytes at 2 weeks after aortic banding. Magnifi- ϫ heart and degree of cardiac hypertrophy induced by pressure cation 300. B, Cross-sectional area of cardiomyocytes was ϭ overload in mice lacking prostanoid receptors individually. A, measured at 1, 2, 4, and 8 weeks after aortic banding. n 4to Ͻ Ͻ Expressions of mRNAs for prostanoid receptors were deter- 9. *P 0.05 vs corresponding sham-operated mice; †P 0.05 vs mined in cardiac ventricles of wild-type mice by RT-PCR. B, aortic-banded wild-type mice. Representative photographs of whole heart isolated from wild- Ϫ Ϫ type and IP / mice at 2 weeks of aortic banding and corre- At 1 week after the aortic banding, the cross-sectional area sponding midtransverse sections stained with hematoxylin and eosin. C, HW/BW ratios at 1, 2, 4, and 8 weeks after aortic of cardiomyocytes was not significantly different between Ϫ/Ϫ banding in wild-type and IPϪ/Ϫ mice. nϭ4to7.*PϽ0.05 vs cor- wild-type and IP hearts (Figure 2B). At 2 and 4 weeks after responding sham-operated mice; †PϽ0.05 vs aortic-banded the banding, however, it was significantly more augmented in wild-type mice. IPϪ/Ϫ hearts than in wild-type hearts, indicating that IP- mediated signal suppressed the development of cardiomyo- Before the banding, the systolic pressures in carotid arter- cyte hypertrophy. At 8 weeks after the banding, the cross- Ϯ ϭ Ϯ ϭ Ϫ Ϫ ies were 71.0 8.8 (n 4) and 70.0 6.4 (n 3) mm Hg in sectional area of cardiomyocytes in IP / hearts was not Ϫ/Ϫ wild-type and IP mice, respectively. After the banding, the significantly different from that in wild-type hearts, suggest- systolic pressures in right carotid artery increased to ing that mechanisms leading to cardiomyocyte hypertrophy Ϯ Ϯ 88.5 8.5 and 87.3 7.5 mm Hg, respectively, and those in caught up with the suppressive effect of IP. left carotid artery decreased to 65.5Ϯ6.9 and In wild-type hearts, the increase in the area of perivascular 65.0Ϯ4.9 mm Hg, respectively, resulting in a pressure gradi- fibrosis was already apparent at 1 week after the banding, and ent of 23.0Ϯ1.6 and 22.3Ϯ2.7 mm Hg, respectively. This it further increased gradually thereafter (Figure 3B, top). In suggests that a similar degree of pressure overload was given Ϫ Ϫ contrast, the increase in the area of interstitial fibrosis was not between wild-type and IP / hearts by the procedure. apparent at 1 week after the banding, reached a maximum Cardiomyocyte Hypertrophy and Cardiac Fibrosis level at 2 weeks, and then declined gradually (Figure 3B, Cardiac hypertrophy is associated frequently with both car- bottom). The difference in the appearance times of perivas- diomyocyte hypertrophy and cardiac fibrosis, with the latter cular and interstitial fibrosis may be derived from the differ- consisting of both interstitial and perivascular fibrosis. To ent timing of inflammatory cell infiltration into these 2 determine whether the antihypertrophic role of IP is due to areas,28 and the late-phase decline of interstitial fibrosis may inhibition of cardiomyocyte hypertrophy or cardiac fibrosis, suggest decreased contents of type III collagen during pro- we measured the cross-sectional area of cardiomyocytes gression of cardiac fibrosis.29 At 1 week after the banding, (Figure 2) and the area of cardiac fibrosis (Figure 3). In both fibrotic areas were not significantly different between Ϫ Ϫ wild-type mice, the cross-sectional area of cardiomyocytes wild-type and IP / hearts (Figure 3B). At 2 and 4 weeks, and the area of cardiac fibrosis increased significantly after however, these areas were significantly enlarged in IPϪ/Ϫ the banding compared with those in sham-operated mice, hearts compared with those in wild-type hearts. In contrast to indicating the development of both cardiomyocyte hypertro- cardiomyocyte hypertrophy, these fibrotic areas at 8 weeks phy and cardiac fibrosis. were still significantly larger in IPϪ/Ϫ hearts than in wild-type 88 Circulation July 5, 2005

There was no significant difference in the cross-sectional area of cardiomyocytes and the area of interstitial and Ϫ/Ϫ Ϫ/Ϫ Ϫ/Ϫ perivascular fibrosis among wild-type, EP2 ,EP3 ,FP , Ϫ/Ϫ Ϫ/Ϫ and TP mice or between F2 wild-type and EP4 mice at 4 weeks after the banding (see online-only Data Supplement).

Expression of ANP mRNA in the Heart We next examined whether IP deficiency affects the cardiac expression of hypertrophy-related genes, in which we used ANP mRNA as a marker. In wild-type mice, the expression level of ANP mRNA increased significantly at 2 weeks of the banding compared with that in sham-operated mice (Figure 4), indicating that pressure overload induced gene expression along with the development of cardiac hypertrophy. In IPϪ/Ϫ hearts, however, expression levels were significantly higher than those in wild-type hearts throughout the experimental period. Moreover, in IPϪ/Ϫ hearts, its significant increase was apparent as early as 1 week after the banding. Interestingly, the expression level in sham-operated IPϪ/Ϫ mice was slightly but significantly higher than that in sham-operated wild-type mice, suggesting an inhibitory effect of basally produced

PGI2 on ANP mRNA expression. These results suggest that stimulation of IP is able to modulate the expression of hypertrophy-related genes in the heart.

Expression of COX mRNAs and Production of PGI2 in the Heart To determine whether pressure overload affects the expres-

sion of COXs and production of PGI2, we measured the ␣ Figure 3. Cardiac fibrosis induced by pressure overload in wild- expression levels of COX mRNAs and 6-keto-PGF1 con- type and IPϪ/Ϫ mice. A, Histological manifestation of van Gie- tents in the left ventricle of wild-type mice. The expression son–stained interstitial and perivascular fibrosis in sections of level of mRNA for COX-1 or COX-2 did not change cardiac ventricles at 2 weeks after aortic banding. Arrows indi- cate fibrotic area. Magnification ϫ20. B, Ratios of area of significantly after the banding throughout the experimental perivascular (top) and of interstitial (bottom) fibrosis to total car- period compared with that before the banding (data not diac area at 1, 2, 4, and 8 weeks after aortic banding were pre- shown). In addition, the expression of COX-2 mRNA was sented. nϭ4 to 10. *PϽ0.05 vs corresponding sham-operated mice; †PϽ0.05 vs aortic-banded wild-type mice. barely detectable by the present method, and its level was significantly lower than that of COX-1 mRNA compared with the use of an 18S rRNA as an internal control. Accord- hearts, suggesting a potent antifibrotic role of IP. These ingly, there was no significant change in the cardiac level of results indicate that IP participates in the suppression of both 6-keto-PGF1␣ throughout the experimental period; these cardiomyocyte hypertrophy and cardiac fibrosis. values were 10.4Ϯ2.1 (nϭ4) and 9.3Ϯ2.3 (nϭ5) pg/mg wet weight in aortic-banded and sham-operated groups, respec- tively, at 1 week of the banding. These results indicate that the pressure overload did not affect COX mRNA expression

or PGI2 production in the heart, whereas significant amounts

of PGI2 were being produced in the heart irrespective of the presence or absence of pressure overload.

Effects of Cicaprost on Noncardiomyocyte Proliferation and Cardiomyocyte Hypertrophy To determine whether the antihypertrophic effect of IP is derived from its action on noncardiomyocytes or cardiomyo- cytes, we examined the effects of cicaprost, an IP agonist, on noncardiomyocyte proliferation and cardiomyocyte hypertro- Figure 4. Changes in expression levels of ANP mRNA induced phy. In wild-type noncardiomyocytes, PDGF increased by pressure overload in wild-type and IPϪ/Ϫ mice. Expression [3H]thymidine incorporation 113% above control, indicating levels of ANP mRNA in left ventricle were determined by that PDGF could stimulate DNA synthesis of noncardiomyo- RT-PCR at 1, 2, 4, and 8 weeks after aortic banding. nϭ3to5. *PϽ0.05 vs corresponding sham-operated mice; †PϽ0.05 vs cytes (Figure 5A). Cicaprost significantly inhibited the corresponding wild-type mice. PDGF-induced increase in [3H]thymidine incorporation in Hara et al Role of PGI2 Receptor in Cardiac Hypertrophy 89

TABLE 3. Effects of Aortic Banding on BW and HW in Wild-Type and IP؊/؊ Mice

Weeks After Aortic Banding

12 4 8 BW, g WT (SO) 20.8Ϯ0.6 (4) 22.7Ϯ1.4 (4) 22.5Ϯ1.2 (4) 24.1Ϯ1.0 (4) WT (AB) 20.7Ϯ0.6 (10) 21.7Ϯ0.4 (9) 22.7Ϯ0.4 (9) 23.5Ϯ0.7 (6) IPϪ/Ϫ (SO) 21.8Ϯ1.6 (4) 22.8Ϯ1.3 (4) 22.2Ϯ0.8 (4) 24.1Ϯ0.8 (4) IPϪ/Ϫ (AB) 20.7Ϯ0.3 (7) 22.7Ϯ0.3 (8) 23.6Ϯ0.7 (7) 26.1Ϯ0.9 (9) HW, mg WT (SO) 98.0Ϯ1.6 (4) 104.6Ϯ5.5 (4) 104.6Ϯ5.0 (4) 115.5Ϯ3.9 (4) WT (AB) 113.9Ϯ4.6 (10) 144.2Ϯ3.7 (9)* 143.8Ϯ7.6 (9)* 151.1Ϯ9.3 (6)* IPϪ/Ϫ (SO) 107.7Ϯ5.4 (4) 108.9Ϯ5.5 (4) 99.0Ϯ3.4 (4) 109.7Ϯ7.0 (4) IPϪ/Ϫ (AB) 122.7Ϯ6.6 (7) 172.1Ϯ6.4 (8)*† 177.2Ϯ11.6 (7)*† 181.5Ϯ10.4 (9)*† BW and HW were measured in wild-type (WT) and IPϪ/Ϫ mice at 1, 2, 4, and 8 weeks of aortic banding (AB) or sham operation (SO). Each value represents meanϮSEM. Numbers in parentheses represent number of animals. *PϽ0.05 vs corresponding sham-operated mice; †PϽ0.05 vs aortic-banded wild-type mice. wild-type noncardiomyocytes, an effect that disappeared in Discussion Ϫ Ϫ IP / noncardiomyocytes, indicating that the effect was me- In the present study pressure overload produced by the aortic diated by IP. Interestingly, the slight inhibitory effect of banding caused marked cardiac hypertrophy, consisting of cicaprost on [3H]thymidine incorporation was also observed cardiomyocyte hypertrophy and cardiac fibrosis. The degree in PDGF-unstimulated wild-type noncardiomyocytes. Ac- of cardiac hypertrophy after the banding was significantly cordingly, cicaprost significantly inhibited PDGF-induced greater in IPϪ/Ϫ mice than in wild-type mice and was increase in cell number only in wild-type noncardiomyocytes accompanied by augmentation of both cardiomyocyte hyper- (Figure 5B). These results indicate that stimulation of IP trophy and cardiac fibrosis. In addition, the increase in the could inhibit proliferation of noncardiomyocytes induced by expression level of ANP mRNA after the banding was PDGF. In contrast, cicaprost failed to inhibit serum-induced significantly augmented in IPϪ/Ϫ hearts compared with in proliferation of wild-type noncardiomyocytes (data not wild-type hearts. These findings provide direct evidence that shown), suggesting a difference in signaling mechanisms IP-mediated signaling plays a suppressive role in the devel- between PDGF and serum. In wild-type cardiomyocytes, 14 opment of pressure overload–induced cardiac hypertrophy cardiotrophin-1 increased [ C]leucine incorporation 43% and fibrosis. above control, indicating that cardiotrophin-1 could induce Synthesis of PGI has been reported to be elevated in the cardiomyocyte hypertrophy (Figure 5C). However, cicaprost 2 hypertrophied and failing heart15 along with an upregulation could not modulate the cardiotrophin-1–induced increase in of COX-2.16 In the present study, however, we found no [14C]leucine incorporation, presenting the possibility that IP significant increase in COX mRNA expression and PGI signaling is defective in cardiomyocytes. 2 production after the aortic banding. This discrepancy may be To further evaluate the IP signaling in cardiomyocytes and derived from the differences in experimental conditions used noncardiomyocytes, we examined IP expression and its or diseases studied. Accordingly, COX-2 upregulation in the second messenger system in these cell types. When IP mRNA expression was examined by the RT-PCR method, we found heart has been reported in patients having congestive heart 16 that both cell types from wild-type mice expressed IP mRNA failure, and increased cardiac PGI2 production was transient 15 but those from IPϪ/Ϫ mice did not, and we found that its in dogs subjected to aortic banding. Recently, involvement expression level was apparently much higher in noncardio- of proinflammatory cytokines in the pathogenesis of heart 30 myocytes than in cardiomyocytes when GAPDH mRNA failure due to a variety of causes has been suggested. Our expression was used as a reference (Figure 5D). We next procedure of aortic banding was relatively mild in degree and examined functionally whether stimulation of IP could acti- did not induce apparent cardiac failure, suggesting little vate the second messenger system in these cell types by participation of inflammatory cytokines, a common inducer of COX-2, in the pathogenesis of the present cardiac hyper- measuring [cAMP]i because IP belongs to a Gs-coupled ␣ receptor family. In noncardiomyocytes, cicaprost increased trophy. Nevertheless, cardiac 6-keto-PGF1 concentrations

[cAMP]i prominently: the increase was 260% above control before and after the aortic banding were grossly estimated to Ϫ5 at a concentration of 10 mol/L (Figure 5E). In contrast, the be 30 to 50 nmol/L, a well-working range of PGI2. These increase in cardiomyocytes was only 40%, whereas the results suggest that basally produced PGI2 could exhibit an ␤-adrenergic agonist isoproterenol could increase it exten- antihypertrophic effect on the heart on a pressure overload sively in this cell type. These results indicate that IP-mediated and that increased production of PGI2 via COX-2 induction in signaling works more efficiently in noncardiomyocytes than failing hearts would further contribute to the suppression of in cardiomyocytes. cardiac hypertrophy. 90 Circulation July 5, 2005

noid produced by cultured cardiac fibroblasts and that an IP agonist inhibited fibroblast proliferation and expression of collagen type I and III mRNAs.13 In addition, antifibrotic

effects of PGI2 have also been reported in extracardiac tissues, such as blood vessels and the kidney.32,33 These results indicate that stimulation of IP reduces the proliferation of noncardiomyocytes and suggest that stimulation of IP on fibroblasts accounts for its suppressive effect on pressure overload–induced cardiac fibrosis. In wild-type cardiomyocytes, cicaprost failed to reduce the cardiotrophin-1–induced hypertrophy, agreeing with a previ-

ous report that PGI2 did not show a direct antihypertrophic action on cultured rat cardiomyocytes.12,17 It is noteworthy that angiotensin II–induced cardiomyocyte hypertrophy was totally dependent on several factors secreted from noncardio- myocytes, such as endothelin and cardiotrophin-1.34,35 Ac- cordingly, the action of IP on noncardiomyocytes may be involved in the underlying mechanism of its suppressive effects on cardiomyocyte hypertrophy in vivo. In support of

this idea, PGI2 was reported to be capable of attenuating hypertrophy of cardiomyocytes when they were cocultured with noncardiomyocytes, although it did not have a direct antihypertrophic effect on cardiomyocytes.12 These results

suggest that PGI2 acts on noncardiomyocytes and inhibits their release of hypertrophy-inducible factors, leading to the suppression of cardiomyocyte hypertrophy. In the present study the expression of IP mRNA and the

IP-mediated increase in [cAMP]i were apparently much greater in noncardiomyocytes than in cardiomyocytes, sug- Figure 5. Effects of cicaprost on noncardiomyocyte proliferation gesting further that activation of IP on noncardiomyocytes and cardiomyocyte hypertrophy. A, Incorporation of [3H]thymi- would be more important in the suppression of cardiac dine in PDGF-stimulated or -unstimulated noncardiomyocytes. hypertrophy. However, we could not exclude a role of IP on Values are percentage of [3H]thymidine incorporation in PDGF- untreated controls, the mean values of which were 2319 and cardiomyocytes in the suppression of cardiac hypertrophy 2496 cpm per well in wild-type and IPϪ/Ϫ noncardiomyocytes, because IP mRNA was expressed in cardiomyocytes and respectively. nϭ5 to 12. B, Cell number in PDGF-stimulated or because its stimulation induced a small but significant in- -unstimulated noncardiomyocytes. Values are percentage of cell crease in [cAMP] . Further studies would be required to number in PDGF-untreated controls. nϭ4. C, Incorporation of i [14C]leucine in the cardiotrophin (CT)-1–stimulated or clarify the detailed mechanisms of antihypertrophic role of IP -unstimulated cardiomyocytes from wild-type mice. Values are in pressure overload–induced cardiac hypertrophy. percentage of [14C]leucine incorporation in cardiotrophin-1–un- We found the expression of mRNAs for EP2,EP3,EP4, FP, treated controls, the mean value of which was 2527 cpm per well. nϭ4 to 12. Concentrations of cicaprost, PDGF, and and TP in addition to that for IP in the heart. In a model of cardiotrophin-1 were 10Ϫ5 mol/L, 5 ng/mL, and 10Ϫ10 mol/L, pressure overload–induced cardiac hypertrophy in vivo, how- Ͻ Ϫ/Ϫ Ϫ/Ϫ respectively. *P 0.05 vs PDGF- or cardiotrophin-1–untreated ever, the degree of cardiac hypertrophy in EP2 ,EP3 , Ͻ Ϫ Ϫ Ϫ Ϫ Ϫ Ϫ control group; †P 0.05 vs PDGF-treated control group. D, EP / ,FP / ,orTP / mice was not significantly different Expression of IP mRNA in noncardiomyocytes (NCM) and car- 4 diomyocytes (CM) from wild-type and IPϪ/Ϫ mice. GAPDH compared with that in their respective wild-type mice. This mRNA was used as loading control. E, Effects of cicaprost on result suggests that neither EP subtype, FP, or TP plays a [cAMP]i in noncardiomyocytes (NCM) and cardiomyocytes (CM) major role in the present model of pressure overload–induced from wild-type and IPϪ/Ϫ mice. These cells were incubated with vehicle or cicaprost (10Ϫ6 or 10Ϫ5 mol/L). Isoproterenol (10Ϫ7 cardiac hypertrophy. In contrast, there are several reports mol/L) increased [cAMP]i to 1115.0Ϯ37.9 pmol/mg protein in suggesting hypertrophic effects of exogenously administered ϭ Ͻ 17,18 cardiomyocytes. n 4or5.*P 0.05 vs control group. PGF2␣ on rat cardiomyocyte both in vivo and in vitro, although there has been no report showing the hypertrophic

Because pressure overload–induced cardiac hypertrophy effect of endogenous PGF2␣. The apparent discrepancy in the was accompanied by both cardiomyocyte hypertrophy and effect of PGF2␣ may be derived from a species difference, as cardiac fibrosis, we next investigated whether the antihyper- reported; the hypertrophic effect of PGF2␣ on cultured trophic role of IP depends on its effect on cardiomyocytes or cardiomyocytes found in rats was not observed in mice.36 noncardiomyocytes. In wild-type noncardiomyocytes, cica- Alternatively, amounts of endogenously produced PGF2␣ prost significantly suppressed PDGF-induced proliferation; may be insufficient to affect pressure overload–induced this growth factor was reportedly implicated in the develop- cardiac hypertrophy in vivo. 31 ment of cardiac hypertrophy and fibrosis. This result is in PGI2 is a major prostanoid in the cardiovascular system and good agreement with a report that PGI2 was a major prosta- exerts a variety of actions in the system. It exhibits a Hara et al Role of PGI2 Receptor in Cardiac Hypertrophy 91 cardioprotective effect in ischemia/reperfusion injury10 and late phase of ischemic preconditioning in conscious rabbits. Proc Natl plays a part in the late phase of ischemic preconditioning.9 A Acad Sci U S A. 2000;97:10197–10202. 10. Xiao CY, Hara A, Yuhki K, Fujino T, Ma H, Okada Y, Takahata O, recent study has shown that neointimal formation in the Yamada T, Murata T, Narumiya S, Ushikubi F. Roles of prostaglandin I2 carotid artery after endothelial injury is markedly enhanced in and thromboxane A2 in cardiac ischemia-reperfusion injury: a study using Ϫ/Ϫ mice lacking their respective receptors. Circulation. 2001;104: IP mice, suggesting an antiproliferative effect of PGI2 on 37 2210–2215. vascular smooth muscle cells. In agreement with this effect, 11. Xiao CY, Yuhki K, Hara A, Fujino T, Kuriyama S, Yamada T, Takayama PGI2 analogues have long been used for the treatment of K, Takahata O, Karibe H, Taniguchi T, Narumiya S, Ushikubi F. Pros- 38 primary pulmonary hypertension. Recently, a gene therapy taglandin E2 protects the heart from ischemia-reperfusion injury via its using PGI synthase gene transfection has been tested in rat receptor subtype EP4. Circulation. 2004;109:2462–2468. 2 12. Ritchie RH, Schiebinger RJ, LaPointe MC, Marsh JD. Angiotensin models of cardiovascular diseases, such as pulmonary hyper- II–induced hypertrophy of adult rat cardiomyocytes is blocked by nitric tension,39 vascular remodeling,40 and peripheral vascular oxide. Am J Physiol. 1998;275:H1370–H1374. occlusion,41 and has been shown to be promising. Further- 13. Yu H, Gallagher AM, Garfin PM, Printz MP. Prostacyclin release by rat cardiac fibroblasts: inhibition of collagen expression. Hypertension. more, the present study showed a novel role of IP in pressure 1997;30:1047–1053. overload–induced cardiac hypertrophy, emphasizing, along 14. Zamorano B, Carmona MT. Prostaglandin-E2 and cyclic adenosine 3Ј-5Ј monophosphate levels in hypertrophied rat heart. Biol Res. 1992;25: with previous reports, a potent therapeutic potential of PGI2 for cardiovascular diseases. It should be noted, however, that 85–89. 15. Newman WH, Frankis MB, Halushka PV. Increased myocardial release there may be a species difference in the effect of IP, as shown of prostacyclin in dogs with heart failure. J Cardiovasc Pharmacol. in the present study for the effect of FP on cardiomyocyte 1983;5:194–201. hypertrophy, indicating that the usefulness of PGI for the 16. Wong SC, Fukuchi M, Melnyk P, Rodger I, Giaid A. Induction of 2 cyclooxygenase-2 and activation of nuclear factor-␬B in myocardium of treatment of cardiac hypertrophy and heart failure remains to patients with congestive heart failure. Circulation. 1998;98:100–103. be determined in future studies. 17. Adams JW, Migita DS, Yu MK, Young R, Hellickson MS, Castro-Vargas ␣ In conclusion, the present study clearly showed that IP- FE, Domingo JD, Lee PH, Bui JS, Henderson SA. Prostaglandin F2 stimulates hypertrophic growth of cultured neonatal rat ventricular mediated signaling suppresses the development of pressure myocytes. J Biol Chem. 1996;271:1179–1186. overload–induced cardiac hypertrophy via its inhibition of 18. Lai J, Jin H, Yang R, Winer J, Li W, Yen R, King KL, Zeigler F, Ko A, ␣ both cardiomyocyte hypertrophy and cardiac fibrosis. The Cheng J, Bunting S, Paoni NF. Prostaglandin F2 induces cardiac finding should contribute to better understanding of the myocyte hypertrophy in vitro and cardiac growth in vivo. Am J Physiol. 1996;271:H2197–H2208. mechanism underlying cardiac hypertrophy. 19. Hizaki H, Segi E, Sugimoto Y, Hirose M, Saji T, Ushikubi F, Matsuoka T, Noda Y, Tanaka T, Yoshida N, Narumiya S, Ichikawa A. Abortive Acknowledgments expansion of the cumulus and impaired fertility in mice lacking the prostaglandin E receptor subtype EP . Proc Natl Acad Sci U S A. 1999; This work was supported by a grant-in-aid for scientific research 2 96:10501–10506. from the Ministry of Education, Science, Sports, and Culture of 20. Ushikubi F, Segi E, Sugimoto Y, Murata T, Matsuoka T, Kobayashi T, Japan and by the research grant for cardiovascular disease (14A-1) Hizaki H, Tuboi K, Katsuyama M, Ichikawa A, Tanaka T, Yoshida N, from the Ministry of Health and Welfare of Japan. This work was Narumiya S. Impaired febrile response in mice lacking the prostaglandin also supported by grants from Ono Pharmaceutical Co, Smoking E receptor subtype EP3. Nature. 1998;395:281–284. Research Foundation, Akiyama Foundation, and Hokkaido Heart 21. Segi E, Sugimoto Y, Yamasaki A, Aze Y, Oida H, Nishimura T, Murata Association. T, Matsuoka T, Ushikubi F, Hirose M, Tanaka T, Yoshida N, Narumiya S, Ichikawa A. Patent ductus arteriosus and neonatal death in prostaglan- References din receptor EP4-deficient mice. Biochem Biophys Res Commun. 1998; 246:7–12. 1. Scheuer J, Buttrick P. The cardiac hypertrophic responses to pathologic 22. Sugimoto Y, Yamasaki A, Segi E, Tsuboi K, Aze Y, Nishimura T, Oida and physiologic loads. Circulation. 1987;75(suppl I):63–68. H, Yoshida N, Tanaka T, Katsuyama M, Hasumoto K, Murata T, Hirata 2. Levy D, Garrison RJ, Savage DD, Kannel WB, Castelli WP. Prognostic M, Ushikubi F, Negishi M, Ichikawa A, Narumiya S. Failure of partu- implications of echocardiographically determined left ventricular mass in rition in mice lacking the prostaglandin F receptor. Science. 1997;277: the Framingham Heart Study. N Engl J Med. 1990;322:1561–1566. 681–683. 3. Ito H, Hiroe M, Hirata Y, Fujisaki H, Adachi S, Akimoto H, Ohta Y, 23. Murata T, Ushikubi F, Matsuoka T, Hirata M, Yamasaki A, Sugimoto Y, Marumo F. Endothelin ETA receptor antagonist blocks cardiac hypertro- Ichikawa A, Aze Y, Tanaka T, Yoshida N, Ueno A, Oh-ishi S, Narumiya phy provoked by hemodynamic overload. Circulation. 1994;89: S. Altered pain perception and inflammatory response in mice lacking 2198–2203. prostacyclin receptor. Nature. 1997;388:678–682. 4. Senbonmatsu T, Ichihara S, Price E Jr, Gaffney FA, Inagami T. Evidence 24. Kabashima K, Murata T, Tanaka H, Matsuoka T, Sakata D, Yoshida N, for angiotensin II type 2 receptor–mediated cardiac myocyte enlargement Katagiri K, Kinashi T, Tanaka T, Miyasaka M, Nagai H, Ushikubi F, during in vivo pressure overload. J Clin Invest. 2000;106:R25–R29. Narumiya S. Thromboxane A2 modulates interaction of dendritic cells 5. Takimoto Y, Aoyama T, Iwanaga Y, Izumi T, Kihara Y, Pennica D, and T cells and regulates acquired immunity. Nat Immunol. 2003;4: Sasayama S. Increased expression of cardiotrophin-1 during ventricular 694–670. remodeling in hypertensive rats. Am J Physiol. 2002;282:H896–H901. 25. Ma H, Hara A, Xiao CY, Okada Y, Takahata O, Nakaya K, Sugimoto Y, 6. Coleman RA, Smith WL, Narumiya S. International Union of Pharma- Ichikawa A, Narumiya S, Ushikubi F. Increased bleeding tendency and cology classification of prostanoid receptors: properties, distribution, and decreased susceptibility to thromboembolism in mice lacking the prosta-

structure of the receptors and their subtypes. Pharmacol Rev. 1994;46: glandin E receptor subtype EP3. Circulation. 2001;104:1176–1180. 205–229. 26. Rockman HA, Ross RS, Harris AN, Knowlton KU, Steinhelper ME, Field 7. Narumiya S, Sugimoto Y, Ushikubi F. Prostanoid receptors: structures, LJ, Ross J Jr, Chien KR. Segregation of atrial-specific and inducible properties, and functions. Physiol Rev. 1999;79:1193–1226. expression of an atrial natriuretic factor transgene in an in vivo murine 8. Ushikubi F, Hirata M, Naruniya S. Platelet prostaglandin receptors. In: model of cardiac hypertrophy. Proc Natl Acad Sci U S A. 1991;88: von Bruchhausen F, Walter U, eds. Hand Book of Experimental Phar- 8277–8281. macology, Volume 126: Platelets and Their Factors. Berlin, Germany: 27. Akishita M, Iwai M, Wu L, Zhang L, Ouchi Y, Dzau VJ, Horiuchi M. Springer-Verlag; 1997:135–154. Inhibitory effect of angiotensin II type 2 receptor on coronary arterial 9. Shinmura K, Tang XL, Wang Y, Xuan YT, Liu SQ, Takano H, Bhatnagar remodeling after aortic banding in mice. Circulation. 2000;102: A, Bolli R. Cyclooxygenase-2 mediates the cardioprotective effects of the 1684–1689. 92 Circulation July 5, 2005

28. Nicoletti A, Michel JB. Cardiac fibrosis and inflammation: interaction 35. Kuwahara K, Saito Y, Harada M, Ishikawa M, Ogawa E, Miyamoto Y, with hemodynamic and hormonal factors. Cardiovasc Res. 1999;41: Hamanaka I, Kamitani S, Kajiyama N, Takahashi N, Nakagawa O, 532–543. Masuda I, Nakao K. Involvement of cardiotrophin-1 in cardiac myocyte- 29. Weber KT, Janicki JS, Shroff SG, Pick R, Chen RM, Bashey RI. Collagen nonmyocyte interactions during hypertrophy of rat cardiac myocytes in remodeling of the pressure-overloaded, hypertrophied nonhuman primate vitro. Circulation. 1999;100:1116–1124. myocardium. Circ Res. 1988;62:757–765. 36. Deng XF, Rokosh DG, Simpson PC. Autonomous and growth fac- tor–induced hypertrophy in cultured neonatal mouse cardiac myocytes: 30. Murray DR, Freeman GL. Proinflammatory cytokines: predictors of a comparison with rat. Circ Res. 2000;87:781–788. failing heart? Circulation. 2003;107:1460–1462. 37. Cheng Y, Austin SC, Rocca B, Koller BH, Coffman TM, Grosser T, 31. Suzuki J, Baba S, Ohno I, Endoh M, Nawata J, Miura S, Yamamoto Y, Lawson JA, FitzGerald GA. Role of prostacyclin in the cardiovascular Sekiguchi Y, Takita T, Ogata M, Tamaki K, Ikeda J, Shirato K. Immu- response to thromboxane A2. Science. 2002;296:539–541. nohistochemical analysis of platelet-derived growth factor-B expression 38. Archer S, Rich S. Primary pulmonary hypertension: a vascular biology in myocardial tissues in hypertrophic cardiomyopathy. Cardiovasc and translational research “work in progress.” Circulation. 2000;102: Pathol. 1999;8:223–231. 2781–2791.

32. Kouchi Y, Esato K, O-Hara M, Zempo N. Effect of prostaglandin I2 39. Nagaya N, Yokoyama C, Kyotani S, Shimonishi M, Morishita R, analogue TRK-100 on the suppression of intimal fibrous proliferation. J Uematsu M, Nishikimi T, Nakanishi N, Ogihara T, Yamagishi M, Vasc Surg. 1992;16:232–238. Miyatake K, Kaneda Y, Tanabe T. Gene transfer of human prostacyclin 33. Yokoyama C, Yabuki T, Shimonishi M, Wada M, Hatae T, Ohkawara S, synthase ameliorates monocrotaline-induced pulmonary hypertension in Takeda J, Kinoshita T, Okabe M, Tanabe T. Prostacyclin-deficient mice rats. Circulation. 2000;102:2005–2010. 40. Todaka T, Yokoyama C, Yanamoto H, Hashimoto N, Nagata I, Tsukahara develop ischemic renal disorders, including nephrosclerosis and renal T, Hara S, Hatae T, Morishita R, Aoki M, Ogihara T, Kaneda Y, Tanabe infarction. Circulation. 2002;106:2397–2403. T. Gene transfer of human prostacyclin synthase prevents neointimal 34. Harada M, Itoh H, Nakagawa O, Ogawa Y, Miyamoto Y, Kuwahara K, formation after carotid balloon injury in rats. Stroke. 1999;30:419–426. Ogawa E, Igaki T, Yamashita J, Masuda I, Yoshimasa T, Tanaka I, Saito 41. Hiraoka K, Koike H, Yamamoto S, Tomita N, Yokoyama C, Tanabe T, Y, Nakao K. Significance of ventricular myocytes and nonmyocytes Aikou T, Ogihara T, Kaneda Y, Morishita R. Enhanced therapeutic interaction during cardiocyte hypertrophy: evidence for endothelin-1 as a angiogenesis by cotransfection of prostacyclin synthase gene or optimi- paracrine hypertrophic factor from cardiac nonmyocytes. Circulation. zation of intramuscular injection of naked plasmid DNA. Circulation. 1997;96:3737–3744. 2003;108:2689–2696. Neuronal Nitric Oxide Synthase Mediates Statin-Induced Restoration of Vasa Nervorum and Reversal of Diabetic Neuropathy

Masaaki Ii, MD, PhD; Hiromi Nishimura, MD, PhD; Kengo F. Kusano, MD, PhD; Gangjian Qin, MD; Young-sup Yoon, MD, PhD; Andrea Wecker, BS; Takayuki Asahara, MD, PhD; Douglas W. Losordo, MD

Background—Peripheral neuropathy is a frequent and major complication of diabetes. Methods and Results—Severe peripheral neuropathy developed in type II diabetic mice, characterized by significant slowing of motor and sensory nerve conduction velocities. Rosuvastatin restored nerve vascularity, including vessel size, and nerve function also recovered to the levels of nondiabetic mice. Neuronal nitric oxide synthase expression in sciatic nerves was reduced in diabetic mice but was preserved by rosuvastatin. Coadministration of a nitric oxide synthase inhibitor with rosuvastatin attenuated the beneficial effects of rosuvastatin on nerve function and limited the recovery of vasa nervorum and nerve function. In vitro, rosuvastatin inhibited downregulation of neuronal nitric oxide synthase expression induced by high-glucose conditions in cultured Schwann cells. Furthermore, Akt phosphorylation in Schwann cells, downregulated by high-glucose conditions, was also restored by rosuvastatin, consistent with the change of neuronal nitric oxide synthase expression. Akt inhibition independently reduced neuronal nitric oxide synthase expression in Schwann cells in low-glucose cultures. Conclusions—These data indicate that the HMG-CoA reductase inhibitor rosuvastatin has a favorable effect on diabetic neuropathy independent of its cholesterol-lowering effect. Our data provide evidence that this effect may be mediated in part via neuronal nitric oxide synthase/nitric oxide and phosphatidylinositol 3-kinase/Akt-signaling pathways and also suggest that restoration or preservation of the microcirculation of the sciatic nerve may be involved. (Circulation. 2005; 112:93-102.) Key Words: diabetes mellitus Ⅲ nervous system Ⅲ nitric oxide synthase

eripheral neuropathy is one of the major complications of ripheral nervous system associated with destruction of vasa Pdiabetes that can lead to significant morbidity. Sensory nervorum. These prior studies have also documented that abnormalities predominate,1 leading to a failure to detect vascular endothelial growth factor (VEGF) gene transfer minor trauma of the lower extremities, ultimately contribut- successfully modifies the neuropathy by restoring microcir- ing to skin ulcerations. Indeed, some reports indicate that culation in vasa nervorum.12,13 In addition, prior studies from 20% of all hospital admissions among diabetic patients in the our laboratory and others have revealed that HMG-CoA United States are for foot problems.2 Even with intensive reductase inhibitors, or “statins,” can exert an angiogenic therapy, including insulin and more effective oral agents, the effect in ischemic tissues.14–16 Moreover, recent data have incidence of neuropathy has increased to as much as 16%3 of suggested a potential beneficial effect of statins that extends the 17 million diabetic patients in the United States alone. beyond the traditional indications of hyperlipidemia and that The pathogenic mechanisms of diabetic neuropathy that have may be especially prominent in people with diabetes.17–22 been considered include the following: degeneration of pro- Accordingly, we investigated the hypothesis that diabetic teins critical to neural function by nonenzymatic glycosyla- neuropathy can be reversed by administration of statins. In tion,4 altered neural polyol metabolism,5,6 reduction of neu- the present report we document that the disordered peripheral rotrophic factors,1,7 and microvascular disease with impaired nerve physiology resulting from experimental diabetes is blood flow.6,8,9 Ischemia in diabetic nerves10,11 has also been associated with marked destruction of the vasa nervorum of considered in the pathogenesis of diabetic complications. the sciatic nerve and that neuronal function may be success- Indeed, we have shown previously that experimentally in- fully restored by administration of rosuvastatin, a new HMG- duced diabetes causes physiological dysfunction in the pe- CoA reductase inhibitor. Nerve recovery induced by rosuv-

Received October 7, 2004; revision received March 8, 2005; accepted March 11, 2005. From the Division of Cardiovascular Research, Caritas St Elizabeth’s Medical Center, Tufts University School of Medicine, Boston, Mass (M.I., K.F.K., G.Q., Y.Y., A.W., T.A., D.W.L.), and Regenerative Medicine, Institute for Biomedical Research and Innovation, Kobe, Japan (H.N., T.A.). Correspondence to Douglas W. Losordo, MD, Division of Cardiovascular Research, Caritas St Elizabeth’s Medical Center, Tufts University School of Medicine, 736 Cambridge St, Boston, MA 02135. E-mail [email protected] © 2005 American Heart Association, Inc. Circulation is available at http://www.circulationaha.org DOI: 10.1161/CIRCULATIONAHA.104.511964 93 94 Circulation July 5, 2005 astatin is accompanied by replenishment of the vasa fluorescent microscope (ϫ20), and vessel cross-sectional areas were nervorum; however, our results also suggest that this statin measured by NIH Image (version 1.62) in the same sections. Because may have neurogenic effects that augment the demonstrated BS1-lectin binds to vascular endothelium, the fluorescent labeling that occurs by intracardiac injection of the FITC-conjugated BS1- vascular effects. Together, these findings provide additional lectin is restricted to vascular structures. Vessels are identified as evidence of a vasculogenic etiology of diabetic neuropathy, small circular/punctate areas of fluorescence in the cross-sectional documenting statin-induced angiogenesis and vasodilatation images and were counted in 10 randomly selected cross sections ϫ of the vasa nervorum, and suggest potential direct neurotro- under the fluorescent microscope ( 20). Vessel cross-sectional area, which reflects vessel size, was measured with the use of NIH Image phic effects of statins. (version 1.62) in the same nerve cross sections. Methods Double-Fluorescent Immunohistochemistry Reagent Sciatic nerves were isolated from mice at the time of euthanasia and prepared for frozen sections. Frozen tissue sections (4 ␮min Rosuvastatin was supplied by AstraZeneca UK. Rosuvastatin is a thickness) were air dried and fixed with 100% acetone for 5 minutes hydrophilic inhibitor of HMG-CoA reductase.23,24 at Ϫ20°C. Nonspecific protein binding was blocked with 10% normal horse serum. Sections were incubated overnight at 4°C with Animals a rabbit polyclonal antibody against endothelial nitric oxide synthase All protocols were approved by St Elizabeth’s Institutional Animal (eNOS) (1:250), neuronal nitric oxide synthase (nNOS) (1:200), and Care and Use Committee. In all experiments, investigators perform- VEGF (1:500) (Santa Cruz Biotechnology, Santa Cruz, Calif) diluted ing the follow-up examinations were blinded to the treatment in 2% goat serum/PBS, followed by 30 minutes of incubation with a ϩ/ϩ db administered. Male C57BLKS/J-m Lepr homozygous (db/db) secondary antibody, Cy3-conjugated anti-rabbit IgG (1:500) (Jack- mice (Jackson Laboratories, Bar Harbor, Maine) aged 8 to 12 weeks son ImmunoResearch). Normal rabbit IgG was used as a negative were used. Age-matched heterozygotes (db/m), a nonpenetrant control. After they were washed with PBS, sections were incubated genotype (Jackson Laboratories), were used as the control animals. with second primary antibodies biotinylated isolectin B4 (ILB4) Mice were treated with daily subcutaneous injection of rosuvastatin (1:100, Vector Laboratories) for detection of endothelial cells or goat (1 mg/kg) or saline as a vehicle. Serum glucose and total cholesterol polyclonal anti-S100 (1:100, Santa Cruz) for detection of Schwann levels were measured with the use of an instant check meter (Roche) cells at 4°C overnight, followed by FITC-conjugated streptavidin at days 0 and 28 after each treatment. (1:500) or Cy2-conjugated anti-goat IgG (1:500) (Jackson Immu- noResearch) as a secondary antibody, respectively. Sections were Neurophysiological Measurements counterstained with DAPI (1:5000) and mounted in aqueous mount- Sciatic nerve conduction velocity was measured with standard ing medium. orthodromic surface recording techniques and a Teca TD-10 (Oxford Instruments) portable recording system in all mice at baseline (before Immunocytochemistry for BrdU treatment) and then at 2 and 4 weeks after treatment, as described Cells were fixed with 4% paraformaldehyde for 5 minutes in 4-well previously.25 Briefly, motor nerve conduction velocity (MCV) was chamber slides (Nalge Nunc). After they were washed with PBS, calculated by dividing the distance between stimulating electrodes by slides were incubated with 2N HCl for 20 minutes at 37°C followed the average latency difference between the peaks of the compound by incubation with 100 mmol/L sodium borate buffer for 20 minutes, muscle action potentials evoked from 2 sites (sciatic notch and and nonspecific protein binding was blocked with 10% normal goat ankle). Sensory nerve conduction velocity (SCV) was calculated by serum. Sections were incubated with biotinylated sheep anti-BrdU dividing the distance between stimulating and recording electrodes antibody (1:100, BIODESIGN) at 4°C overnight, followed by by the latency of the signal from the stimulation artifact to the onset rhodamine-conjugated streptavidin (1:500, Jackson ImmunoRe- of the peak signal. For each nerve, maximal velocities were deter- search) as a secondary antibody. Nuclei were stained with DAPI mined bilaterally. All measured data from both sides were averaged. (1:5000), and slides were mounted in aqueous mounting medium. Proliferation activity was evaluated under fluorescent microscopy as Tail-Flick Testing an average percentage of BrdU-positive cells in 5 randomly selected This behavioral test examined the response of each mouse to a high-powered fields per well. thermal noxious stimulus administered to its tail with a variation of the tail immersion test. This measure was chosen because of the role nNOS Inhibition Study that small fiber dorsal root ganglia sensory neurons play in pain Mice were treated with a daily subcutaneous injection of rosuvastatin transmission. The animals were loosely restrained with their tails (1 mg/kg) or rosuvastatin (1 mg/kg) plus selective nNOS inhibitor immersed in a beaker of water to a depth of Ϸ2 cm. The beaker was 1,2-trifluoromethylphenyl imidazole (TRIM) (50 mg/kg) (Sigma). uniformly heated at a rate of Ϸ2°C per minute beginning at 38°C. A MCV and SCV were measured at 2 and 4 weeks after the treatment. magnetic stirring bar was used to maintain a uniform temperature. The assessment of vascularity in sciatic nerve with FITC-conjugated The temperature at which each animal flicked its tail out of the water BS1-lectin was performed at the time of euthanasia, as described was recorded to the nearest 0.5°C. above. Assessment of Vascularity in Sciatic Nerve Cell Culture Vascularity of sciatic nerves from both nondiabetic and diabetic mice Primary mouse Schwann cells were obtained from sciatic nerve by was assessed by in situ fluorescent staining with the use of the explant method as described previously.26 Cells were seeded on endothelial cell–specific marker FITC-conjugated BS1-lectin (Vec- 10-cm plastic culture dishes coated with rat type I collagen (300 tor Laboratories). After anesthesia, BS1-lectin (0.1 mg per mouse) ␮g/mL) (Sigma) and cultured in DMEM F-12 (GIBCO) supple- was injected systemically by direct cardiac puncture. Ten minutes mented with 20% fetal bovine serum (FBS), 10 ng/mL progesterone, later, the animals were euthanized, and sciatic nerves were harvested 5 ␮g/mL transferrin, 5 ␮g/mL insulin, 10 ␮g/mL putrescine (Sigma), and fixed in 2% paraformaldehyde for 2 hours. After fixation, nerves B27 supplement, 25 mmol/L HEPES, 100 U/mL penicillin, and 100 were either whole mounted for longitudinal analysis or embedded in ␮g/mL streptomycin (GIBCO). Cultured Schwann cells were iden- OCT compound for frozen cross section. Samples were analyzed tified on the basis of cell soma and nuclear morphology, and the with the use of a computer-assisted Nikon fluorescence microscope purity was also confirmed by immunocytochemical labeling for with a digital camera (ECLIPS TE200, Nikon Inc). The number of cytoplasmic S100 protein and glial fibrillary acidic protein to be vessels was counted in 10 randomly selected cross sections under Ͼ95% (data not shown). Mouse Schwann cell line MSC80 was a Ii et al Statins Reverse Diabetic Neuropathy 95

kind gift from Dr Jean-Jaques Hauw (Raymond Escourolle Neuro- with 10% nonfat dry milk and 0.2% Tween-20 in PBS, pH 7.4, then pathology Laboratory, Paris, France). MSC80 was cultured and immunoblotted overnight at 4°C with rabbit polyclonal antibodies maintained in DMEM (GIBCO) supplemented with 4 mmol/L against mouse nNOS (1:1000; BD Pharmingen), phospho-Akt/Akt glutamine, 100 U/mL penicillin, 100 ␮g/mL streptomycin, and 10% (1:1000; Cell Signaling), and a goat polyclonal antibody against FCS (GIBCO). ␣-actin (1:1000, Santa Cruz). Blots were washed with 0.2% Primary human adult dermal microvascular endothelial cells Tween-20 in PBS and incubated for 45 minutes with horseradish (HMVECs), obtained from BioWhittaker, were cultured in endothe- peroxidase–linked goat antibody against rabbit IgG (1:5000; Santa lial basal medium supplemented with 5% FBS, 0.1 mg/mL bovine Cruz). Immunoreactive bands were visualized with ECL reagent brain extract, 10 ng/mL epidermal growth factor, 0.5 mg/mL (Amersham). Densitometric analyses for the blots were performed hydrocortisone, and 100 ␮g/mL gentamicin (Cambrex). All the cells with the use of NIH Image software.

were maintained at 37C° and 5% CO2. Measurement for NO Production in Schwann Cells Proliferation Assay MSC80s were cultured on 24-well plates at 70% to 80% confluence The number of viable Schwann cells was determined with a validated under low-glucose (5 mmol/L) and high-glucose (25 mmol/L) nonradioactive cell proliferation assay kit (CellTiter 96; Promega). conditions with or without rosuvastatin (1 and 10 ␮mol/L) for 48 Cells were seeded in 96-well plates at 80% confluence (1ϫ104 cells hours. Culture medium was collected for the measurement of per well) and cultured in 0.5% FBS medium for 48 hours. Then nitrate/nitrite. NO production was evaluated by its final metabolite rosuvastatin was added at concentrations from 0.0001 to 10 ␮mol/L nitrate/nitrite. The concentration of nitrite was measured as an OD in 20% FBS medium. Twenty percent FBS medium alone was used value at 540 nm with the use of a Nitrate/Nitrite Colorimetric Assay as a control. After 48 hours, 15 ␮L of dye solution was added per Kit (Cayman) according to the manufacturer’s instructions. well, and cells were incubated for 4 hours before measurement of absorbance at 570 nm. The proliferation activity of HMVECs was Statistical Analysis evaluated by BrdU incorporation as described previously.27 Cells All results are presented as meanϮSEM. Statistical comparisons ␮ were seeded in human fibronectin-coated (50 g/mL at 37°C for 1 between 2 groups were performed by Student t test, and ANOVA hour) 4-well Laboratory-Tek chamber slides (Nalge Nunc) at a was performed for serial analysis. Probability values Ͻ0.05 were ϫ 5 density of 1 10 cells per well and cultured in 0.5% FBS medium considered statistically significant. All in vitro experiments were for 24 hours. Then rosuvastatin was added at concentrations from repeated at least in triplicate and analyzed. 0.01 to 10 ␮mol/L in 5% FBS medium. Five percent FBS medium was used as a control. After 18 hours of exposure to each treatment, BrdU (Sigma) was added to each well at a concentration of 10 Results ␮mol/L, and its incorporation into the cells was determined after 6 Nerve Conduction Velocities and Characterization hours. Immunocytochemistry for BrdU was performed as described of Mice above. BrdU-positive and total cells in each well were counted in 10 high-powered fields (ϫ40), and the percentage of BrdU-positive db/db mice develop a severe peripheral neuropathy at 8 cells was calculated. weeks, as described previously.28 Electrophysiological re- cordings in db/db mice documented significant slowing of Migration Assay MCV and SCV (MCVϭ29.1Ϯ1.2 m/s and SCVϭ24.6Ϯ1.4 Schwann cell and HMVEC migrations were evaluated with a m/s) compared with those of control mice (MCVϭ46.7Ϯ2.4 modified Boyden’s chamber assay as described previously.25 Briefly, m/s and SCVϭ52.5Ϯ1.5 m/s; PϽ0.01 for both MCV and the polycarbonate filter (8-␮m pore size) (Poretics) was placed between upper and lower chambers. Cells were preincubated in SCV). Placebo-injected db/db mice showed persistent, stable DMEM F-12 or endothelial basal medium for 30 minutes with neuropathy 4 weeks after initiation of treatment rosuvastatin at concentrations from 0.0001 to 10 ␮mol/L. Cell (MCVϭ27.0Ϯ0.5 m/s and SCVϭ29.6Ϯ1.0 m/s), whereas suspensions (1ϫ104 cells per well) were placed in the upper rosuvastatin-injected db/db mice demonstrated nearly com- chamber, and the lower chamber was filled with medium containing plete recovery of MCV and SCV back to levels of nondia- human recombinant nerve growth factor (100 ng/mL) or VEGF (50 ng/mL) (R&D Systems). The chamber was incubated for 4 hours at betic mice. Specifically, MCV and SCV in rosuvastatin- Ϯ Ϯ 37°C and 5% CO2. Migration activity was evaluated as the mean injected db/db mice were 47.2 1.6 and 44.7 1.8 m/s, and number of migrated cells in 5 high-powered fields (ϫ40) per those in nondiabetic mice were 49.1Ϯ2.1 and 47.3Ϯ2.3 m/s, chamber. respectively (nondiabetic mice versus rosuvastatin-injected diabetic mice; PϭNS) (Figure 1a and 1b). Tail-flick testing, Dominant Negative Akt and Myristoylated Forms a measure of the function of small fiber dorsal root ganglia of Akt/Adenovirus Infection sensory neurons, was also performed 4 weeks after treatment. Cultured MSC80s were transduced with an adenoviral construct encoding dominant negative Akt (dn-Akt) or myristoylated Akt In saline-injected db/db mice, tail-flick temperatures were (myr-Akt) at a multiplicity of infection of 500 for 3 hours in DMEM significantly increased to 46.1Ϯ0.2°C (PϽ0.001 versus non- with supplement. On the following day, cells were used for Western diabetic mice). In contrast, tail-flick temperatures recorded in blot analysis. rosuvastatin-injected db/db mice did not differ significantly from those of age-matched nondiabetic control animals Western Blot Analysis (44.6Ϯ0.3°C in rosuvastatin-injected mice versus MSC80s were cultured with the following conditions for 48 hours: Ϯ ϭ mannitol (25 mmol/L), glucose (5 and 25 mmol/L), high glucose 44.9 0.3°C in control mice; P NS) (Figure 1c). Blood (25 mmol/L) in the presence of rosuvastatin ranging from 0.001 to 10 glucose and serum total cholesterol levels before and after ␮mol/L, and high glucose (25 mmol/L) with rosuvastatin (1 ␮mol/L) treatment in db/db mice were not significantly different in in the presence of the phosphatidylinositol 3-kinase (PI3K)/Akt rosuvastatin-injected versus saline-injected mice (Table). inhibitor LY294002 (10 ␮mol/L). Cells were lysed in 100 ␮L SDS sample buffer (20% SDS, 5% ␤-mercaptoethanol, 10% glycerol in 50 mmol/L Tris-HCl, pH 7.4) per 35-mm culture dish. Protein Vascularity of Vasa Nervorum in Sciatic Nerve extracts were separated by 7.5% SDS-PAGE and transferred to a To investigate the potential role of microvascular pathology 0.2-␮m PVDF membrane (Bio-Rad). The membranes were blocked in the development of diabetic neuropathy, we performed in 96 Circulation July 5, 2005

(Figure 2a and 2d) (24.0Ϯ1.2 versus 24.0Ϯ1.6 vessels per cross section [PϭNS] and 78.4Ϯ5.8 versus 79.0Ϯ4.7 ␮m2 per vessel [PϭNS], respectively; nϭ8) (Figure 2g and 2h).

nNOS, eNOS, and VEGF Expression in Sciatic Nerve Prior investigations have established the roles of eNOS and VEGF signaling in the angiogenic response to ischemia. To investigate the contribution of these molecules in the recov- ery of vasa nervorum, we performed fluorescent immunohis- tochemistry for not only eNOS and VEGF but also nNOS expression in vasa nervorum in diabetic mice. There were marked differences of nNOS (Figure 3a), eNOS (Figure 3b), and VEGF (Figure 3c) expression between saline-injected db/db mice and rosuvastatin-injected db/db mice. The expres- sion of nNOS was most notably decreased in saline-injected db/db mice in comparison to rosuvastatin-injected db/db mice. In contrast, the expression of nNOS in rosuvastatin- injected db/db mice was similar to that seen in nondiabetic mice (data not shown).

TRIM Partly Reversed the Effect of Rosuvastatin on Nerve Dysfunction Figure 1. Rosuvastatin (ROS) treatment reverses nerve conduc- To further investigate whether NO production is essential for tion slowing and sensory deficit in diabetic mice. Nondiabetic mice were used as controls (control; nϭ10), and db/db mice rosuvastatin-mediated improvement in diabetic neuropathy, were randomly assigned to saline injection (saline; nϭ10) or we inhibited nNOS by coadministration of a selective nNOS rosuvastatin injection (ROS; nϭ10). Data are expressed as inhibitor, TRIM, with rosuvastatin. In rosuvastatin-injected Ϯ mean SEM. Electrophysiological parameters MCV (a) and SCV db/db mice, significant restoration of MCV and SCV was (b) were performed before and 2 and 4 weeks after the initiation of treatment. PϭNS, *PϽ0.01, **PϽ0.0001, saline vs rosuvasta- observed at 2 and 4 weeks after treatment (MCVϭ37.3Ϯ0.9 tin. Tail-flick testing, to assess sensory function, was performed and 47.2Ϯ1.6 m/s for 2 and 4 weeks versus 26.4Ϯ1.3 m/s at 4 weeks after initiation of treatment (c). PϭNS, *PϽ0.01 com- week 0; SCVϭ38.2Ϯ2.1 and 44.7Ϯ1.8 m/s for 2 and 4 weeks pared with nondiabetic control. versus 22.7Ϯ1.3 at week 0, respectively; PϽ0.01 for both MCV and SCV; nϭ10 in each group). In contrast, situ fluorescent imaging of the vasa nervorum by in vivo rosuvastatin-injected mice receiving TRIM did not demon- perfusion of FITC-labeled BS1-lectin. In the nerves of strate the full extent of neurological recovery seen in saline-injected db/db mice, the vascularity of vasa nervorum rosuvastatin-injected mice in which nNOS function was was markedly reduced (Figure 2b and 2e). Cross sections unimpaired (Figure 4a and 4b). Nerve conduction velocity from these nerves disclosed a significant reduction not only in measurements in rosuvastatin- versus rosuvastatinϩTRIM- the number (capillary density) but also in the size (cross- treated diabetic mice revealed significant reduction in the sectional area) of vasa nervorum compared with that of beneficial effect of rosuvastatin on nerve function resulting nerves from control mice (16.4Ϯ0.8 versus 24.0Ϯ1.6 vessels from inhibition of nNOS (2 weeks [SCV, 38.2Ϯ2.1 for per cross section [PϽ0.01] and 53.0Ϯ2.8 versus 79.0Ϯ4.7 rosuvastatin versus 32.7Ϯ1.3 m/s for rosuvastatinϩTRIM; ␮m2 per vessel [PϽ0.01], respectively; nϭ8) (Figure 2g and PϽ0.01] [Figure 4b] and 4 weeks [MCV, 47.2Ϯ1.6 for 2h). In contrast, in the nerves of rosuvastatin-injected db/db rosuvastatin versus 35.8Ϯ1.3 m/s for rosuvastatinϩTRIM; mice, the vascularity of vasa nervorum was well preserved SCV, 44.7Ϯ1.8 for rosuvastatin versus 34.9Ϯ0.9 m/s for (Figure 2c and 2f), and cross-sectional images from BS1- rosuvastatinϩTRIM; PϽ0.01] [Figure 4a and 4b]) Inhibition lectin–perfused db/db mice revealed vessel numbers and of nNOS also prevented the normalization of tail-lick tempera- cross-sectional areas similar to those of nondiabetic mice ture by rosuvastatin (rosuvastatin 44.6Ϯ0.3°C versus

Blood Glucose and Serum Total Cholesterol Levels

Glucose, mg/dL Cholesterol, mg/dL

0 Weeks 4 Weeks 0 Weeks 4 Weeks db/m mice (control; nϭ5) 168Ϯ23.5 168.3Ϯ28.7 155.0Ϯ3.0 156.9Ϯ2.5 db/db mice (saline; nϭ10) 530Ϯ14.8 539Ϯ18.0 155.9Ϯ1.0 157.8Ϯ1.7 db/db mice (ROS; nϭ10) 538Ϯ11.6 546Ϯ14.7 154.3Ϯ1.1 153.3Ϯ1.2 Statistical significance NS NS Values are meanϮSEM. Statistical comparisons between 0 and 4 weeks were made by Student t test. Ii et al Statins Reverse Diabetic Neuropathy 97

Effect of Rosuvastatin on HMVEC Proliferation and Migration To determine whether the direct action of rosuvastatin on the vasa nervorum includes direct modulation of endothelial cell phenotype, we investigated the effect of rosuvastatin on HMVEC proliferation and migration at concentrations from 0.001 to 10 ␮mol/L. Rosuvastatin significantly promoted proliferative activity of HMVECs from 60% to 28% at concentrations from 0.001 to 0.1 ␮mol/L (Figure 5a). In contrast, rosuvastatin significantly promoted VEGF-induced HMVEC migration by 30% at lower concentrations (0.001 ␮mol/L) and inhibited migration by 40% and 50% at higher concentrations (1 and 10 ␮mol/L, respectively) (Figure 5b).

Rosuvastatin Directly Modulates Schwann Cell Proliferation and Migration To determine whether the effect of rosuvastatin on the nerve recovery might also be the result of a direct neurotrophic effect, we investigated the effect of rosuvastatin on Schwann cell proliferation and migration at concentrations from 0.001 to 10 ␮mol/L. Rosuvastatin significantly promoted prolifer- ative activity of Schwann cells by up to 86% at concentra- tions of 0.1 ␮mol/L (Figure 5c). Rosuvastatin also enhanced nerve growth factor–induced Schwann cell migration signif- icantly by 25% to 62% at concentrations from 0.01 to 10 ␮mol/L in a dose-dependent manner (Figure 5d).

Rosuvastatin Preserves NO Production Figure 2. Rosuvastatin (ROS) treatment restores vasa nervorum Downregulated by High Glucose in Schwann Cells in diabetic mice. Representative fluorescence photomicrographs To confirm that rosuvastatin upregulates not only nNOS show longitudinal views of whole-mounted mouse sciatic nerves expression under high-glucose conditions but also promotes (a, b, and c) and their respective cross sections (d, e, and f) 4 weeks after treatment. The network of vasa nervorum is mark- NO production in Schwann cells, we evaluated the actual NO edly reduced in the diabetic saline injection group (b and e). The production by measuring its final metabolite nitrate/nitrite in vascularity of vasa nervorum in the rosuvastatin group (c and f) culture medium. Rosuvastatin significantly preserved NO appears well preserved, and the number and size of visible ves- production by Schwann cells despite high-glucose conditions sels in the cross sections appear similar to those of nondiabetic controls (a and d). Barϭ200 ␮m. Vascularity was quantified in (Figure 5e). tissue cross sections of mouse sciatic nerve. Before euthanasia at 4 weeks after treatment, mice were perfused with FITC- Effect of Rosuvastatin on Downregulated nNOS conjugated BS1-lectin to visualize vasa nervorum. Ten cross Expression in Schwann Cells by High Glucose sections per frozen sample were randomly selected from each specimen, and the number (capillary density) (g) and size (cross- The expression of nNOS in MSC80 was reduced by 60% in sectional vessel area) (h) of vessels per cross section were high-glucose conditions. In contrast, the same concentration quantified and averaged. Data are expressed as meanϮSEM of osmotic substance, mannitol, did not affect nNOS expres- (nϭ5 per study group). PϭNS, *PϽ0.01 compared with control. sion (Figure 6a). Rosuvastatin significantly preserved nNOS expression at doses ranging from 0.1 to 10 ␮mol/L despite rosuvastatinϩTRIM 45.7Ϯ0.2°C, PϽ0.01; saline 46.1Ϯ0.2°C high-glucose conditions. Phosphorylated Akt expression was ϩ Ϯ Pϭ versus rosuvastatin TRIM 45.7 0.2°C, NS) (Figure 4c). also evaluated and coincided with nNOS expression, showing The vascularity of vasa nervorum documented by BS1-lectin in a decrease in high-glucose conditions with restoration of ϩ rosuvastatin TRIM-treated db/db mice was also less developed phospho-Akt expression by increasing concentrations of ro- (Figure 4f and 4i) as well as saline-injected db/db mice (Figure suvastatin (Figure 6a). To provide further evidence of the 4d and 4g) compared with that in the rosuvastatin-injected group association between Akt and nNOS, we performed Western (Figure 4e and 4h). These findings indicate that nNOS- blot for nNOS using a PI3K inhibitor and dn-Akt adenoviral dependent NO signaling is required for a portion of the recovery vector. Rosuvastatin restored nNOS expression in high- in nerve function induced by rosuvastatin. Quantitative analyses glucose conditions. However, both a PI3K inhibitor and also show significant reversal of the improvement in nerve dn-Akt adenoviral transduction reversed the effect of rosuv- vascularity in both the number (capillary density) and the size astatin on nNOS expression. On the other hand, overexpres- (cross-sectional vessel area) in the rosuvastatinϩTRIM group sion of myr-Akt by adenoviral transduction resulted in compared with the rosuvastatin group (15.7Ϯ1.4 versus preservation of nNOS expression despite high-glucose con- 24.0Ϯ1.6 vessels per cross section [PϽ0.01] and 50.4Ϯ2.9 ditions. The expression of phosphorylated Akt was also versus 79.0Ϯ4.7 ␮m2 per vessel [PϽ0.01], respectively). consistent with that of nNOS. These data indicate that 98 Circulation July 5, 2005

Figure 3. Representative photomicrographs of flu- orescent immunohistochemistry for nNOS (a), eNOS (b), and VEGF (c) in sciatic nerves in nondia- betic and diabetic mice 4 weeks after treatment. Reduced expression of nNOS, eNOS, and VEGF, identified by red fluorescence, was observed in saline injection group (saline) compared with the rosuvastatin injection group (ROS) and nondiabetic control. Schwann cells and capillaries were identi- fied by specific markers S100 and ILB4 (green), respectively. Merged images indicate colocalization of S100 and nNOS, ILB4 and eNOS, and ILB4 and VEGF (orange). Barϭ100 ␮m. Ii et al Statins Reverse Diabetic Neuropathy 99

rosuvastatin acts via PI3K to phosphorylate Akt, which in turn mediates the effect of rosuvastatin on nNOS (Figure 6b).

Discussion This is the first study to investigate the effect of HMG-CoA reductase inhibitors on diabetic peripheral neuropathy. The data demonstrate lipid-independent effects of the HMG-CoA reductase inhibitor rosuvastatin on the recovery of vasa nervorum and nerve function in diabetic neuropathy. More- over, these data indicate that HMG-CoA reductase inhibitors may have direct neurotrophic effects. In our in vivo studies, the development of diabetic neurop- athy is related to loss of vasa nervorum responsible for perfusion of peripheral nerves. Nerve function correlated with morphological observations of the vasa nervorum in the affected nerves of the diabetic mice. The overall restoration of the number and cross-sectional area of vasa nervorum in db/db mice to a pattern similar to that of nondiabetic mice was observed after rosuvastatin administration. The coinci- dence of restoration of vasa nervorum accompanied by functional nerve recovery has now been documented in diabetic animal models with the use of 3 distinct angiogenic agents, VEGF,12 sonic hedgehog,13 and now rosuvastatin. In the present study the recovery of vasa nervorum was mainly associated with recovery of nNOS/NO production to nondi- abetic levels. Moreover, we found that coadministration of rosuvastatin and an nNOS specific inhibitor, TRIM, partially reversed the effect of rosuvastatin. These results suggest that NO production through nNOS might play an important role in the regeneration of vasa nervorum. Multiple prior reports have provided evidence of a link between statins, eNOS, and angiogenesis14,16,29,30 and that rosuvastatin enhances release of NO from the rat aortic vascular endothelium. Weis et al31 have reported that cerivastatin increases endothelial VEGF release and modulates VEGF receptor-2 expression in endo- thelial cells. As indicated in these studies, the evidence that statins regulate eNOS and VEGF in endothelial cells has already been shown. However, no reports have shown that statin also regulate nNOS, which is mainly expressed in neuronal tissue and has the potential to produce more NO than eNOS. Our in vitro data suggest that rosuvastatin directly upregulates nNOS/NO in Schwann cells via the PI3K/Akt signaling pathway. Together with these prior find- ings, our data support a vascular mechanism for at least part

meanϮSEM. a and b, PϭNS, *#PϽ0.01, **PϽ0.0001 compared with saline. c, PϭNS, *PϽ0.01 compared with nondiabetic con- trol. PϭNS, saline vs rosuvastatinϩTRIM. d to i, Fluorescence photomicrographs of representative longitudinal views of whole- mounted mouse sciatic nerves (d to f) and their respective cross sections (g to i) 4 weeks after treatment. Total network of vasa Figure 4. Rosuvastatin (ROS)-induced restoration of nerve dys- nervorum in rosuvastatin injection group was well developed function is nNOS dependent. To determine whether and preserved (e and h). In contrast, vascularity of vasa nervo- rosuvastatin-induced recovery of nerve function required nNOS, rum in the rosuvastatinϩTRIM group (f and i) was reduced and we used the nNOS inhibiter TRIM and assessed MCV, SCV, and appears similar to that in saline injection group (d and g). tail-flick temperature in diabetic mice. db/db mice were ran- Barϭ200 ␮m. j and k, Quantitative analyses of nerve vascularity domly assigned to saline control, rosuvastatin injection (nϭ5), or (j, capillary density; k, cross-sectional vessel area) in the rosuvastatin injectionϩTRIM administration (nϭ5). Sciatic nerve rosuvastatinϩTRIM group compared with the rosuvastatin group conduction measurements were performed at the time of treat- in 10 randomly selected cross sections per nerve. Data are ment (0 week) and then at 2 and 4 weeks. Tail-flick testing was expressed as meanϮSEM (nϭ5 per study group). PϭNS, performed at 4 weeks after treatment. Data are expressed as *PϽ0.01 compared with control. 100 Circulation July 5, 2005

Figure 6. Rosuvastatin (ROS) restores downregulated Akt phos- phorylation and nNOS expression in high glucose–exposed Schwann cells. a, MSC80 cells were cultured in low glucose (5 mmol/L [L]), high glucose (25 mmol/L [H]), mannitol (25 mmol/L [M]), and high glucose with rosuvastatin in concen- trations ranging from 0.001 to 10 ␮mol/L (HϩROS). After 48 hours in culture, cells were harvested and processed for West- ern analysis. Immunoblots were quantified and expressed as relative values compared with low glucose. Immunoblots of ␣-actin and t-Akt were used as loading controls for nNOS and p-Akt, respectively. Data are expressed as meanϮSEM. a, High glucose reduces nNOS and p-Akt in Schwann cells, and rosuv- astatin restores both nNOS and p-Akt in a dose-dependent manner. b, MSC80s were cultured in low (5 mmol/L) or high (25 mmol/L) glucose in the presence of rosuvastatin (1 ␮mol/L), dn-Akt, LY294002 (10 ␮mol/L), or myr-Akt for 48 hours. As shown in b, dn-Akt and LY294002 both prevent rosuvastatin- mediated restoration of nNOS and p-Akt in high glucose–ex- Figure 5. Effects of rosuvastatin (ROS) on HMVEC or Schwann cell posed Schwann cells. PϭNS, *PϽ0.01 compared with high- (ScC) proliferation and migration. For proliferation assay, HMVECs glucose condition alone. were incubated with rosuvastatin for 24 hours after 24-hour serum deprivation. BrdU (10 ␮mol/L) was added 6 hours before evaluation. Proliferation activity was assessed by quantifying BrdU-positive cells of the documented statin-mediated restoration of nerve as a percentage of total cells (a). *PϽ0.05, **PϽ0.001 compared with non-rosuvastatin control. Schwann cells were incubated with rosuvas- function. tatin for 48 hours after 30-hour serum deprivation. Proliferation activity Prior reports about the effect of statins on angiogenesis was assessed by absorbance at OD 570 nm (c). For migration assay, have been conflicting between a stance of promotion14,15,32 cells were incubated with rosuvastatin for 4 hours in a modified Boy- 33,34 den chamber. Migrated cells were counted in 3 randomly selected and inhibition. These discrepant data may result from the high-powered fields and averaged in each well. *PϽ0.001 compared different cell types used or may be attributed to the different with non-rosuvastatin control (HMVEC, b). *PϽ0.05, **PϽ0.001 com- statin concentrations. Indeed, recent data indicate that statins pared with non-rosuvastatin control (ScC, d). e, Effect of rosuvastatin have biphasic effects on angiogenesis.30,31 In our in vitro on NO production in Schwann cells. MSC80s were cultured in low glucose (5 mmol/L [L]), high glucose (25 mmol/L [H]), and high glu- studies, we used a wide range of concentrations of rosuvas- cose with rosuvastatin (1 ␮mol/L [H-ROS1] and 10 ␮mol/L tatin to permit identification of a dose-response effect.30 [H-ROS10]). After 48-hour incubation, nitrite concentrations in the cul- Rosuvastatin promoted HMVEC proliferation at low concen- ture medium from 6 wells were measured by colorimetric method as ␮ NO productivity and averaged. PϭNS, *PϽ0.0001 compared with low trations (0.0001 to 0.1 mol/L). However, a dose-dependent glucose. All data are expressed as meanϮSEM. inhibitory effect of rosuvastatin on HMVEC migration was Ii et al Statins Reverse Diabetic Neuropathy 101 observed at concentrations from 0.1 to 10 ␮mol/L. On the In summary, rosuvastatin, a new HMG-CoA reductase other hand, rosuvastatin dose-dependently promoted both inhibitor, has a favorable effect on diabetic neuropathy that is Schwann cell proliferation and migration at concentrations independent of its cholesterol-lowering effect and that is from 0.0001 to 10 ␮mol/L. The interpretation of these in vitro associated with restoration of vasa nervorum. The effect of findings, specifically attempting to use these data to explain rosuvastatin on vasa nervorum appears to be mediated via an the in vivo observations, is complicated by several factors. NO-dependent pathway. Moreover, we found that rosuvasta- First, the cell culture results are observations made in mo- tin restituted downregulated nNOS expression in Schwann noculture and therefore lack the dynamic environment of the cells under high-glucose conditions at least in part via a intact organism. Second, the concentrations used, which PI3K/Akt signaling pathway. Rosuvastatin also has a direct bracket the serum concentrations in patients, may not per- neurotrophic effect, promoting proliferation and migration fectly reflect the microenvironment of the nerve. With these activity of Schwann cells. caveats, however, the in vitro data demonstrate direct effects In the clinical setting, statins generally have been shown to of statins on neural elements. Indeed, in vivo proliferation in reduce cardiovascular events in association with reducing the nerve assessed by the BrdU labeling technique indicated lipid levels. In addition, however, recent data have revealed that statin slightly promoted Schwann cell proliferative ac- that cardioprotective effects are extended to populations of tivity for at least 4 weeks after initiation of treatment (data not patients without significant lipid abnormalities. Our observa- shown). These results may explain why TRIM incompletely tions suggest an additional lipid-independent activity of reversed electrophysiological recovery induced by rosuvasta- statins that may be of therapeutic relevance. tin. The restoration of vasa nervorum resulting from rosuv- astatin treatment was almost completely reversed by admin- Acknowledgments istration of TRIM, as demonstrated by in situ fluorescent This study was supported in part by National Institutes of Health imaging of whole-mounted explants of sciatic nerve (Figure grants (HL-53354, HL-60911, HL-63414, HL-63695, HL-66957) and a grant from AstraZeneca US. We thank M. Neely and Deirdre 4). These data suggest that rosuvastatin may have not only a Costello for secretarial assistance. vasodilating effect on the vasa nervorum but also direct neurotrophic effects on the sciatic nerve itself. Administration Disclosure of TRIM to diabetic mice, without rosuvastatin, was not Drs Ii and Losordo have received a research grant from AstraZeneca. performed in this study because our goal was to evaluate the effect of nNOS antagonism on the salutary effects of rosuv- References astatin rather than in the native diabetic state. Nevertheless, 1. Tomlinson DR, Fernyhough P, Diemel LT. Role of neurotrophins in one could expect further worsening of nerve function by diabetic neuropathy and treatment with nerve growth factors. Diabetes. 1997;46(suppl 2):S43–S49. inhibiting nNOS expression, which is already reduced in 2. Reiber GE, Smith DG, Carter J, Fotieo G, Deery HG II, Sangeorzan JA, diabetic mice, because nNOS has been shown to play a role Lavery L, Pugh J, Peter-Riesch B, Assal JP, del Aguila M, Diehr P, as a neuroprotective factor35,36 as well as a NO producer. The Patrick DL, Boyko EJ. A comparison of diabetic foot ulcer patients data also suggest that statins could act by upregulating nNOS managed in VHA and non-VHA settings. J Rehabil Res Dev. 2001;38: 309–317. expression via Akt activation in Schwann cells. The in vivo 3. The Diabetes Control and Complications Trial Research Group. The data show that rosuvastatin resulted in recovery of both vessel effect of intensive treatment of diabetes on the development and pro- number and size in diabetic mice up to the nondiabetic level. gression of long-term complications in insulin-dependent diabetes mellitus. N Engl J Med. 1993;329:977–986. It is tempting to conclude, on the basis of this association, that 4. Cullum NA, Mahon J, Stringer K, McLean WG. Glycation of rat sciatic statin-induced angiogenesis led to nerve recovery; however, nerve tubulin in experimental diabetes mellitus. Diabetologia. 1991;34: further study will be needed to provide definitive evidence of 387–389. the sequence of events induced by statin administration that 5. Greene DA, Lattimer SA, Sima AA. Sorbitol, phosphoinositides, and sodium-potassium-ATPase in the pathogenesis of diabetic complications. led to restoration of nerve function. N Engl J Med. 1987;316:599–606. Certain pieces of prior data are consistent with our present 6. Cameron NE, Cotter MA, Low PA. Nerve blood flow in early experi- findings. For example, Cameron et al37 showed improvement mental diabetes in rats: relation to conduction deficits. Am J Physiol. in nerve function in a streptozotocin-induced model of 1991;261:E1–E8. 7. Apfel SC, Arezzo JC, Brownlee M, Federoff H, Kessler JA. Nerve growth diabetes after administration of rosuvastatin. The same au- factor administration protects against experimental diabetic sensory neu- thors documented the impact of rosuvastatin on NO- ropathy. Brain Res. 1994;634:7–12. dependent function in aorta and corpus cavernosum of dia- 8. Tuck RR, Schmelzer JD, Low PA. Endoneurial blood flow and oxygen 38 tension in the sciatic nerves of rats with experimental diabetic neuropathy. betic mice. Finally, and perhaps most importantly, Fried et Brain. 1984;107:935–950. al39 studied patients with diabetes and showed a potential 9. Tesfaye S, Harris N, Jakubowski JJ, Mody C, Wilson RM, Rennie IG, attenuating effect of statin therapy on the advent of neurop- Ward JD. Impaired blood flow and arterio-venous shunting in human athy in diabetic patients. In contrast, there is a significant diabetic neuropathy: a novel technique of nerve photography and fluo- rescein angiography. Diabetologia. 1993;36:1266–1274. body of literature suggesting a positive relationship between 10. Dyck PJ. Hypoxic neuropathy: does hypoxia play a role in diabetic statin use and the onset of neuropathy.40 Together these prior neuropathy. Neurology. 1989;39:111–118. studies and the present data suggest that additional study is 11. Stevens EJ, Carrington AL, Tomlinson DR. Nerve ischaemia in diabetic required to achieve comprehensive understanding of both the rats: time-course of development, effect of insulin treatment plus com- parison of streptozotocin and BB models. Diabetologia. 1994;37:43–48. mechanisms of diabetic neuropathy and the potential thera- 12. Schratzberger P, Walter DH, Rittig K, Bahlmann FH, Pola R, Curry C, peutic impact of statins. Silver M, Krainin JG, Weinberg DH, Ropper AH, Isner JM. Reversal of 102 Circulation July 5, 2005

experimental diabetic neuropathy by VEGF gene transfer. J Clin Invest. Favorable effect of VEGF gene transfer on ischemic peripheral neurop- 2001;107:1083–1092. athy. Nat Med. 2000;6:405–413. 13. Kusano K, Allendoerfer KL, Munger W, Pola R, Bosch-Marce M, 26. Chi H, Horie H, Hikawa N, Takenaka T. Isolation and age-related char- Kirchmair R, Yoon YS, Curry C, Silver M, Kearney M, Asahara T, acterization of mouse Schwann cells from dorsal root ganglion explants in Losordo DW. Sonic hedgehog induces arteriogenesis in diabetic vasa type I collagen gels. J Neurosci Res. 1993;35:183–187. nervorum and restores function in diabetic neuropathy. Arterioscler 27. Ii M, Hoshiga M, Fukui R, Negoro N, Nakakoji T, Nishiguchi F, Kohbayashi E, Thromb Vasc Biol. 2004;24:2102–2107. Ishihara T, Hanafusa T. Beraprost sodium regulates cell cycle in vascular smooth 14. Kureishi Y, Luo Z, Shiojima I. The HMG-CoA reductase inhibitor sim- muscle cells through cAMP signaling by preventing down-regulation of vastatin activates the protein kinase Akt and promotes angiogenesis in p27(Kip1). Cardiovasc Res. 2001;52:500–508. normocholesterolemic animals. Nat Med. 2000;6:1004–1010. 28. Robertson DM, Sima AA. Diabetic neuropathy in the mutant mouse 15. Llevadot J, Murasawa S, Kureishi Y, Uchida S, Masuda H, Kawamoto A, [C57BL/ks(db/db)]: a morphometric study. Diabetes. 1980;29:60–67. Walsh K, Isner JM, Asahara T. HMG-CoA reductase inhibitor mobilizes 29. Stalker TJ, Lefer AM, Scalia R. A new HMG-CoA reductase inhibitor, bone marrow–derived endothelial progenitor cells. J Clin Invest. 2001;108: rosuvastatin, exerts anti-inflammatory effects on the microvascular endo- 399–405. thelium: the role of mevalonic acid. Br J Pharmacol. 2001;133:406–412. 16. Dimmeler S, Aicher A, Vasa M, Mildner-Rihm C, Adler K, Tiemann M, 30. Urbich C, Dernbach E, Zeiher AM, Dimmeler S. Double-edged role of Rutten H, Aeiher AM. HMG-CoA-reductase inhibitors (statins) increase statins in angiogenesis signaling. Circ Res. 2002;90:737–744. endothelial progenitor cells via the P13 kinase/Akt pathway. J Clin Invest. 31. Weis M, Heeschen C, Glassford AJ, Cooke JP. Statins have biphasic 2001;108:391–397. effects on angiogenesis. Circulation. 2002;105:739–745. 17. Lacoste L, Lam JY, Hung J, Letchacovski G, Solymoss CB, Waters D. 32. Pourati I, Kimmelstiel C, Rand W, Karas RH. Statin use is associated Hyperlipidemia and coronary disease: correction of the increased throm- with enhanced collateralization of severely diseased coronary arteries. Am bogenic potential with cholesterol reduction. Circulation. 1995;92: Heart J. 2003;146:876–881. 3172–3177. 33. Sato I, Ma L, Ikeda M, Morita I, Murota S. Simvastatin, a potent 18. Bustos C, Hernandez-Presa MA, Ortego M, Tunon J, Ortega L, Perez F, HMG-CoA reductase inhibitor, inhibits the proliferation of human and Diaz C, Hernandez G, Egido J. HMG-CoA reductase inhibition by ator- bovine endothelial cells in vitro. J Atheroscler Thromb. 1998;4:102–106. vastatin reduces neointimal inflammation in a rabbit model of atheroscle- 34. Vincent L, Chen W, Hong L, Mirshahi F, Mishal Z, Mirshahi-Khorassani rosis. J Am Coll Cardiol. 1998;32:2057–2064. T, Vannier JP, Soria J, Soria C. Inhibition of endothelial cell migration by 19. Laufs U, Liao JK. Post-transcriptional regulation of endothelial nitric cerivastatin, an HMG-CoA reductase inhibitor: contribution to its anti- oxide synthase mRNA stability by RHo GTPase. J Biol Chem. 1998;273: angiogenic effect. FEBS Lett. 2001;495:159–166. 24266–24271. 35. Thippeswamy T, Jain RK, Mumtaz N, Morris R. Inhibition of neuronal 20. West of Scotland Coronary Prevention Study Group. Influence of pravastatin nitric oxide synthase results in neurodegenerative changes in the axo- and plasma lipids on clinical events in the West of Scotland Coronary tomised dorsal root ganglion neurons: evidence for a neuroprotective role Prevention Study (WOSCOPS). Circulation. 1998;97:1440–1445. of nitric oxide in vivo. Neurosci Res. 2001;40:37–44. 21. Heart Protection Study Collaborative Group. MRC/BHF Heart Protection 36. Keilhoff G, Fansa H, Wolf G. Nitric oxide synthase, an essential factor in Study of cholesterol lowering with simvastatin in 20,536 high-risk indi- peripheral nerve regeneration. Cell Mol Biol (Noisy-le-grand). 2003;49: viduals: a randomised placebo-controlled trial. Lancet. 2002;360:7–22. 885–897. 22. Collins R, Armitage J, Parish S, Sleigh P, Peto R. MRC/BHF Heart 37. Cameron N, Cotter M, Inkster M, Nangle M. Looking to the future: Protection Study of cholesterol-lowering with simvastatin in 5963 people diabetic neuropathy and effects of rosuvastatin on neurovascular function with diabetes: a randomised placebo-controlled trial. Lancet. 2003;361: in diabetes models. Diabetes Res Clin Pract. 2003;61(suppl 1):S35–S39. 2005–2016. 38. Nangle MR, Cotter MA, Cameron NE. Effects of rosuvastatin on nitric 23. McTaggart F, Buckett L, Davidson R, Holdgate G, McCormick A, oxide–dependent function in aorta and corpus cavernosum of diabetic Schneck D, Smith G, Warwick M. Preclinical and clinical pharmacology mice: relationship to cholesterol biosynthesis pathway inhibition and lipid of rosuvastatin, a new 3-hydroxy-3-methylglutaryl coenzyme A reductase lowering. Diabetes. 2003;52:2396–2402. inhibitor. Am J Cardiol. 2001;87:28B–32B. 39. Fried LF, Forrest KY, Ellis D, Chang Y, Silvers N, Orchard TJ. Lipid 24. Olsson AG. Statin therapy and reductions in low-density lipoprotein modulation in insulin-dependent diabetes mellitus: effect on micro- cholesterol: initial clinical data on the potent new statin rosuvastatin. vascular outcomes. J Diabetes Complications. 2001;15:113–119. Am J Cardiol. 2001;87:33B–36B. 40. Chong PH, Boskovich A, Stevkovic N, Bartt RE. Statin-associated pe- 25. Schratzberger P, Schratzberger G, Silver M, Curry C, Kearney M, ripheral neuropathy: review of the literature. Pharmacotherapy. 2004;24: Magner M, Alroy J, Adelman LS, Weinberg DH, Ropper AH, Isner JM. 1194–1203.

CLINICAL PERSPECTIVE

Peripheral polyneuropathy occurs in 50% of patients with long-standing diabetes. The sensory deficits that characterize this condition are a major factor in the development of skin ulcerations that contribute to a major source of morbidity in diabetic patients. Despite the frequency of diabetic neuropathy, its pathophysiology remains incompletely characterized, and, as a result, no uniformly effective therapy has been developed for preventing or reversing this condition. Among the prevailing hypotheses for the pathophysiology of diabetic neuropathy, attrition of the vasa nervorum recently has gained momentum as a result of clinical and experimental studies. Tesfaye et al (N Engl J Med. 2005;352:341–350) recently reported the strong association between vascular risk factors and the advent of neuropathy in people with diabetes, who also exhibit reduced circulating levels of endothelial progenitor cells, and our laboratory has shown in multiple models that nerve dysfunction in diabetes is preceded by loss of the vasa and that the process can be prevented or reversed by angiogenic factors. In diabetic patients treated with vascular endothelial growth factor gene therapy, improvements in nerve function have been observed. The statins have been shown to act as angiogenic factors by multiple potential mechanisms. In the present study, we show that statins may reverse or prevent the onset of diabetic neuropathy by preserving the integrity of the vasa nervorum. The data suggest that this effect is mediated in part by a previously unrecognized salutary effect of statins on neuronal nitric oxide synthase expression by neural elements. These findings may add to the growing list of indications for statin administration early in the course of diabetes, but of course this recommendation awaits confirmation in clinical trials. Pediatric Cardiology

Early Structural and Functional Changes of the Vasculature in HIV-Infected Children Impact of Disease and Antiretroviral Therapy

Marietta Charakida, MD; Ann E. Donald, AVS; Hannah Green, MSc; Clare Storry, BSc, AVS; Margaret Clapson, RGN, RSCN, RHV, BSc; Muriel Caslake, PhD; David T. Dunn, PhD; Julian P. Halcox, MD, MA, MRCP; Diana M. Gibb, MD, MSc; Nigel J. Klein, PhD, FRCPCH; John E. Deanfield, BA, BChir, MB, FRCP

Background—Premature cardiovascular disease is increasingly recognized in HIV-infected patients, but the mechanisms involved are unclear. The purpose of this study was to determine the impact of HIV infection and antiretroviral therapy (ART) on markers of early vascular disease in children. Methods and Results—We studied 83 HIV-infected children (56 had taken ART, of whom 31 received a regimen containing protease inhibitors [PIs]; 27 were never treated) and a control group of 59 healthy children. Carotid intima-media thickness (IMT) and brachial artery flow-mediated dilatation (FMD) were measured. IMT was significantly greater in HIV-infected children compared with the control subjects (PϽ0.001). Among the HIV-infected children, age and treatment were significantly associated with increased IMT. Children exposed to PIs had greater IMT compared with both non–PI-treated children and untreated children (Pϭ0.02). FMD was also significantly reduced in the HIV-infected children compared with control subjects (Pϭ0.02). Pairwise comparisons of different treatment exposure groups revealed that FMD was impaired by a mean of 3.6% (95% CI, 1.8 to 5.3; PϽ0.001) for children exposed to PIs compared with untreated children and by a mean of 1.8% (95% CI, 0.01 to 3.5; Pϭ0.05) compared with non–PI-treated children. HIV-infected children had lipid abnormalities, but they did not account for the observed differences in either FMD or IMT. Conclusions—HIV infection in childhood is associated with adverse structural and functional vascular changes that are most pronounced in children exposed to PI therapy. Longitudinal studies are required to differentiate the relative impact of HIV disease and ART and to assess the potential for prevention. (Circulation. 2005;112:103-109.) Key Words: endothelium Ⅲ HIV Ⅲ protease inhibitors

uman immunodeficiency virus (HIV) infection is a major of atherogenesis begins earlier.7 Numerous cardiovascular Hcause of morbidity and mortality worldwide. In developed risk factors have been shown to affect both endothelial countries, life expectancy has increased considerably as a result vascular function and early structural arterial wall changes of antiretroviral therapy (ART), and cardiovascular disease has from the first decade of life.8,9 emerged as an important late concern.1,2 Recent studies have We therefore studied children who acquired HIV through shown a 26% increase in myocardial infarction in HIV-infected mother-to-child transmission to examine the potential effect of adults per year exposure to ART and that metabolic abnormal- chronic HIV infection and ART on measures of vascular 3–5 ities related to the use of combination therapy may contribute. structure and function. This population provides the opportunity Protease inhibitors (PIs), in particular, are associated with not only to investigate the impact of HIV infection and ART on dyslipidemia, insulin resistance, and lipodystrophic phenotype, the vasculature but to do so without the confounding effect of the which may be due to hepatic overproduction of VLDL and, to a cumulative risk factor burden present by adulthood. lesser extent, impaired clearance.5,6 It has been difficult, how- ever, to separate the effects of HIV disease itself, drug treatment, metabolic consequences, or interactions with the exposure to Methods classic risk factors for atherosclerosis. Study Population and Design Although clinical manifestations of atherosclerosis do not We studied HIV-infected children attending Great Ormond Street typically present until middle and late adulthood, the process Hospital (GOSH) NHS Trust (London, UK). Children with current

Received October 26, 2004; revision received February 7, 2005; accepted March 8, 2005. From the Vascular Physiology Unit (M. Charakida, A.E.D., C.S., J.P.H., J.E.D.) and Infectious Diseases and Microbiology Unit (N.J.K.), Institute of Child Health, London (N.J.K.); Medical Research Council Clinical Trials Unit, London (H.G., D.T.D., D.M.G.); Infectious Diseases Unit, Great Ormond Street Hospital for Children, NHS TRUST, London (M. Clapson); Department of Vascular Biochemistry, Division of Cardiovascular and Medical Sciences, University of Glasgow, Glasgow (M. Caslake), UK. Correspondence to Marietta Charakida, Vascular Physiology Unit, 30 Guilford St, London, WC1N 1EH, UK. E-mail [email protected] © 2005 American Heart Association, Inc. Circulation is available at http://www.circulationaha.org DOI: 10.1161/CIRCULATIONAHA.104.517144 103 104 Circulation July 5, 2005 opportunistic infections or cardiovascular risk factors, including obtained. The optimal longitudinal image was acquired on the R hypertension, diabetes, and renal failure, were excluded. Of the 96 wave of the ECG and continuously recorded on videotape for 5 HIV-infected children identified as eligible for the study, 83 agreed seconds. Measurements of the posterior wall of the artery were made to participate. Fifty-nine healthy volunteer children were recruited as from stored images with electronic calipers. IMT was calculated as control subjects from healthy siblings of the HIV-infected children the distance between the first bright line (lumen-intima interface) and and children of staff working at the hospital. No formal matching the leading edge of the second bright line (media-adventitia inter- was used in selecting the control children (apart from siblings). None face). Six measurements, the 3 maximum measurements of the right had a current or recent infectious illness, nor were they receiving any common carotid artery in 3 different frames and the 3 maximum medication or vitamin supplementation. measurements of the left common carotid artery in 3 different All children had a clinical examination, including blood pressure frames, were averaged. measurements, blood sampling, anthropometry, and vascular mea- surements. Details of past clinical and ART history were extracted FMD Study from the UK Collaborative HIV Pediatric Study (CHIPS) database.10 Each child underwent measurement of endothelium-dependent vas- The HIV clinic at GOSH does not use rigid criteria for initiating cular responses of the right brachial artery by high-resolution ART. However, ART is usually started for persistently low or ultrasound imaging with an Acuson 5- to 10-MHz linear probe as declining CD4% counts and/or clinical deterioration. The nature of previously reported.8 Brachial artery diameter was measured offline the ART regime was not based on measures of disease severity. by an automatic edge detection system (Brachial Tools) and ex- Ethnicity data were obtained from the National Study of HIV in pressed as a percentage change from baseline diameter.14 Doppler- Pregnancy and Childhood (NSHPC). derived flow measurements (using a pulsed-wave Doppler signal at Institutional Review Board approval was received, and all parents a 70° angle) were also obtained continuously. The increase in blood or caregivers and children, when appropriate, gave written informed flow after the release of the cuff was expressed as a percentage consent to participate in the study at the time of the visit. change from the baseline flow. Endothelium-independent response to glyceryl trinitrate (GTN) was also calculated as a percentage Anthropometric Measurements change from the baseline diameter after sublingual administration of Weight and height were recorded, and body mass index (BMI; 25 ␮g GTN in the HIV-infected children only. kg/m2) was calculated. Values were converted to age- and sex- adjusted z scores with the use of UK reference curves.11 Blood Statistical Analysis pressure was measured as the average of 3 seated readings with an In descriptive analyses, parametric summary statistics and signifi- automated oscillometric device (Omron). A pediatric cuff was used cance tests were used when the data were approximately normally when the arm circumference was Ͻ25 cm and an adult cuff when the distributed; otherwise, nonparametric methods were used. All FMD arm circumference was Ͼ25 cm. and GTN analyses were adjusted for baseline diameter; GTN analyses were adjusted for weight. Multivariate regression analysis Blood Sample Analysis was used to examine relationships between vascular measurements Nonfasting blood was taken from all children in heparinized tubes, and HIV infection status and ART exposure with adjustment for and 1 mL of plasma was stored at Ϫ80°C. Lipid levels and potential confounders. Confounders included age, sex, ethnicity, inflammatory markers (see below) were measured in all subjects, and Centers for Disease Control (CDC) stage, hsCRP, glucose, and markers of disease activity were assessed in the HIV-infected selected lipid parameters. Variables that were not normally distrib- children. uted were log10 transformed to reduce the influence of outlying observations. Markers of HIV Severity and Inflammation HIV RNA viral load was measured by branched DNA assay (Chiron Results Diagnostics) with a lower limit of detection of 50 copies per 1 mL. Study Population CD4% was measured by flow cytometry. For viral load and CD4% The 83 HIV-infected children had a mean age of 11.0 years measurements, the closest measurements (within 3 months before the study visit) were used if available. High-sensitivity C-reactive (range, 5.4 to 17.7 years); 70% were black; and all had protein (hsCRP) was analyzed with an in-house ELISA method.12 acquired HIV from mother-to-child transmission. The control children were 1 year older than the HIV-infected children; Lipoprotein Analysis 80% were white and had similar age-adjusted BMI scores Total cholesterol was measured by enzymatic colorimetry; HDL was (Table 1). measured directly in plasma after precipitation with polyethylene Twenty-seven HIV-infected children had never received glycol (Roche Diagnostics). Apolipoprotein B (apoB) and lipopro- tein (a) [Lp(a)] were measured by immunoturbimetry kits from ART. The remaining 56 children had taken ART for a median WAKO and Diasys Diagnostic Systems, respectively. LDL peak and of 5 years (range, 0.2 to 14 years). Of these 56, 31 had mean particle diameters were determined by PAGE, as described in received regimens containing a PI (Table 2). At the time of detail elsewhere.13 the study, 48 children were taking ART (23 including a PI), and 35 children were not receiving any ART. Of the 31 Vascular Measurements PI-exposed children, 8 were on combination therapy that did All vascular measurements were undertaken by 2 experienced investigators. Measurement on a random sample of 10 HIV-infected not include a PI at the time of the study visit (median duration and 10 control children performed by both investigators showed no since a PI was taken, 1.3 years; range, 0.5 to 2.8 years). evidence of systematic observer bias (mean difference in intima- Twenty HIV-infected children (24%) had previously experi- media thickness [IMT], 0.003 [Pϭ0.80, paired t test]; mean differ- enced a CDC stage C disease15; all 20 had been treated with ϭ ence in flow-mediated dilation [FMD], 0.25 [P 0.40, paired t test]). ART (15 exposed to PI therapy) (Table 2). IMT Study Lipoprotein Parameters The right and left common carotid arteries were scanned with a 5- to 10-MHz linear-array transducer (Acuson). The carotid bulb was Triglycerides, non-HDL cholesterol, apoB, and Lp(a) were identified, and longitudinal 2D ultrasonographic images of the significantly higher in HIV-infected children compared with common carotid artery 1 to 2 cm proximal to the carotid bulb were control subjects. In addition, HIV-infected children had Charakida et al Vascular Consequences of HIV and Its Treatment 105

TABLE 1. Demographic Data, Clinical Parameters, Lipid Results, and Vascular Measurements in HIV-Infected Children and Control Subjects HIV-Infected Children Control Subjects (nϭ83) (nϭ59) P* Demographic data Age, y 11.0 (3.1) 12.2 (2.8) 0.014 Male, % 57 47 0.31 Height z score (IQR) Ϫ0.31 (Ϫ1.26–0.27) 0.30 (Ϫ0.15–0.94) 0.0002 BMI z score (IQR) 0.04 (Ϫ0.78–0.87) 0.20 (Ϫ0.46–1.04) 0.39 Clinical parameters Systolic blood pressure, mm Hg 104 (10) 110 (12) 0.0009 Diastolic blood pressure, mm Hg 61 (9) 63 (8) 0.09 Heart rate, bpm 80 (13) 72 (12) 0.0003 Glucose, mmol/L 4.4 (0.7) 5.1 (0.3) Ͻ0.0001 hsCRP, median (IQR), mg/L 0.6 (0.2–2.3) 0.2 (0.1–0.6) Ͻ0.0001 Lipid/blood results Total cholesterol, mmol/L 4.1 (0.9) 3.8 (0.7) 0.06 HDL cholesterol, mmol/L 1.2 (0.5) 1.4 (0.3) 0.02 Non-HDL cholesterol, mmol/L 2.9 (0.8) 2.4 (0.6) 0.0007 Mean particle diameter, nm 28.2 (1.0) 29.0 (1.1) Ͻ0.0001 Peak particle diameter, nm 27.8 (1.2) 28.6 (1.4) 0.0009 ApoB, mg/dL 75 (17) 64 (11) 0.0001 Lp(a), median (IQR), mg/dL 42 (23–69) 23 (14–51) 0.02 Triglycerides, median (IQR), mmol/L 1.0 (0.7–1.3) 0.6 (0.5–0.8) Ͻ0.001 Vascular measurements IMT, mm 0.6 (0.07) 0.47 (0.06) Ͻ0.0001 FMD,† % 7.9 (4.9) 9.4 (5.8) 0.02 GNT-mediated dilation,‡ % 15.1 (5.3) IQR indicates interquartile range. Height, BMI, blood pressure, heart rate, and lipid measurements only were available for 56, 56, 55, 55, and 43 control children, respectively. Values are given as mean (SD) unless otherwise stated. *Based on the 2-sample t test or Wilcoxon 2-sample test when appropriate. †Adjusted for baseline diameter; estimates given for mean baseline diameter of 2.86 mm. ‡Adjusted for baseline diameter and weight; estimates given for mean baseline diameter of 2.81 mm and mean weight of 36.9 kg. significantly lower HDL cholesterol levels and mean and Vascular Measurements peak LDL particle sizes (Table 1). IMT Assessment Within the HIV cohort, differences in lipid parameters IMT measurements were available for 77 of the 83 HIV- were observed between those receiving different treatment infected children and 45 of the 53 control subjects. IMT was regimens (Table 2). Total cholesterol was higher in children higher in the HIV-infected children compared with the receiving ART (both with and without PI) compared with control subjects (PϽ0.0001; Table 1) in both unadjusted and untreated HIV-infected children. Lp(a), apoB, and triglycer- adjusted analyses. There was a significant association be- ide levels were higher in PI-treated children compared with tween IMT and age in the HIV-infected children that was not non–PI-treated and untreated HIV-infected children. HDL observed in the control subjects (Figure 1A); for every year cholesterol levels were lower in the untreated HIV-infected increase in age, IMT was increased on average by 0.005 mm children compared with treated children, and the latter had (95% CI, 0.0003 to 0.01). There was no association between levels similar to those in the control children. ethnicity and IMT in either HIV-infected or control children. There was evidence of a treatment effect when the HIV- CRP Measurement infected children were categorized as ever exposed to PIs, hsCRP was significantly higher in the HIV-infected children non–PI-treated, and untreated (Pϭ0.03; Table 2). Exposure to compared with control subjects (Table 1). Eight HIV-infected PIs was associated with a higher mean IMT compared with children had markedly increased hsCRP values (CRP Ͼ10 both those untreated (Pϭ0.04) and the non–PI-treated group mg/L); although they did not have clinical evidence of acute (Pϭ0.01; Figure 1B). Similar results were found when the infection, these CRP levels are compatible with acute inflam- children were categorized according to their treatment regi- mation. However, the difference in CRP values between men at the time of the test (data not shown). There was no control subjects and HIV-infected children remained when association between IMT and CD4% or viral load at the time the analysis was performed with these 8 children omitted. of the study, but more advanced CDC stage was associated 106 Circulation July 5, 2005

TABLE 2. Comparison of ART Exposure in HIV-Infected Children PI Treated Non–PI Treated Untreated (nϭ31) (nϭ25) (nϭ27) P* Demographic data Age, y 11.1 (3.5) 11.5 (3.0) 10.3 (2.5) 0.39 Male, % 62 44 63 0.72 Height z score, median (IQR) Ϫ0.59 (Ϫ1.35–0.18) Ϫ0.40 (Ϫ0.94–0.06) Ϫ0.24 (Ϫ1.32–0.38) 0.75 BMI z score, median (IQR) Ϫ0.20 (Ϫ0.81–0.75) 0.04 (Ϫ0.93–0.56) 0.62 (Ϫ0.25–1.19) 0.13 Clinical parameters Systolic blood pressure, mm Hg 103 (10) 103 (8) 106 (10) 0.37 Diastolic blood pressure, mm Hg 61 (12) 60 (7) 61 (6) 0.90 Heart rate, bpm 79 (11) 79 (12) 83 (15) 0.37 Glucose, mmol/L 4.4 (0.8) 4.3 (0.6) 4.4 (0.7) 0.79 hsCRP, median (IQR), mg/L 1.1 (0.2–2.0) 1.2 (0.4–6.7) 0.4 (0.4–2.0) 0.14 Clinical parameters related to HIV and ART CDC stage C, % 48 20 0 Ͻ0.0001 CD4%, median (IQR) 26 (21–34) 26 (14–29) 22 (19–24) 0.18

HIV RNA, log10 copies/mL 2.30 (0.98) 2.68 (1.11) 4.08 (0.70) Ͻ0.0001 Exposure to ART, median (IQR), y 5.3 (4.7–6.8) 3.8 (2.3–6.0) Exposure to PIs, median (IQR), y 4.2 (2.7–5.1) Lipid/blood results Total cholesterol, mmol/L 4.6 (0.9) 4.0 (0.8) 3.6 (0.8) 0.002 HDL cholesterol, mmol/L 1.4 (0.5) 1.3 (0.5) 1.0 (0.3) 0.0007 Non-HDL cholesterol, mmol/L 3.2 (1.0) 2.7 (0.7) 2.7 (0.7) 0.02 ApoB, mg/dL 80 (20) 73 (15) 72 (15) 0.13 Mean particle diameter, nm 28.0 (1.1) 28.3 (0.7) 28.3 (1.2) 0.52 Peak particle diameter, nm 27.7 (1.2) 27.0 (1.3) 27.8 (1.3) 0.74 Lp(a), median (IQR), mg/dL 50 (38–86) 33 (17–45) 34 (21–63) 0.01 Triglycerides, median (IQR), mmol/L 1.1 (0.8–1.4) 0.9 (0.8–1.2) 0.9 (0.6–1.4) 0.31 Vascular measurements IMT, mm 0.62 (0.07) 0.58 (0.06) 0.58 (0.06) 0.03 FMD,† % 6.3 (5.4) 8.1 (5.9) 9.9 (5.7) 0.0005 GTN-mediated dilation,‡ % 14.5 (8.4) 14.0 (9.5) 17.2 (8.9) 0.08 IQR indicates interquartile range. Eight of the PI-treated children were on combination therapy that did not include a PI at the time of the study visit. Eight children in the non-PI group were off all treatment at the time of the study visit. CD4% and HIV RNA: closest measurement within 3 months before the time of the study visit (nϭ78 for CD4%, nϭ77 for HIV RNA). There are 4 CDC stages that reflect the extent of HIV disease progression: N signifies no signs or symptoms; A, B, and C, increasingly severe disease; and C, AIDS). Values are given as mean (SD) unless otherwise stated. *Global probability values are based on ANOVA or the Kruskal-Wallis test when appropriate. †Adjusted for baseline diameter; estimates given for mean baseline diameter of 2.80 mm. ‡Adjusted for baseline diameter and weight; estimates given for mean baseline diameter of 2.80 mm and mean weight of 36.9 kg. with greater IMT. However, in multivariate analyses, the Among the HIV-infected children, there were no differ- higher IMT in children with exposure to PIs could not be ences in baseline vessel diameter, baseline flow, or reactive accounted for by CDC stage or other potential confounders. hyperemia in the different treatment groups. FMD was Notably, there was no association between IMT and duration significantly lower in those exposed to PIs compared with of PI therapy. non–PI-treated (Pϭ0.05) and untreated (PϽ0.001) children. Similar results were obtained when the analysis was per- FMD Assessment FMD measurements were available for 82 of the 83 HIV- formed by current therapy (Figure 2B). The response to GTN infected children and 57 of the 59 control children. Resting was not associated with exposure to ART (Table 2). As for vessel size, blood flow, and reactive hyperemia were similar IMT, CDC stage C was associated with lower FMD. FMD in both groups. FMD was lower in the HIV-infected children was also inversely related to hsCRP, and this relationship than in the control subjects after adjustment for baseline diam- remained after the 8 HIV-infected children whose hsCRP was eter (Pϭ0.02; Table 1 and Figure 2A). The relationship re- Ͼ10 mg/L were excluded. However, the relationship between mained after adjustment for age, sex, and other potential con- FMD and ART exposure remained after adjustment for CDC founders. There was no evidence of an association between age, stage, hsCRP, and other potential confounders in multivariate sex, ethnicity, lipids, and hsCRP in the whole cohort. analysis. Charakida et al Vascular Consequences of HIV and Its Treatment 107

Figure 2. Endothelial function in HIV-infected children and con- trol subjects. A, FMD was significantly reduced in HIV-infected Figure 1. IMT in HIV-infected children and control subjects. A, children. *Pϭ0.02. B, FMD according to ART received at time of IMT and age. IMT was increased on average by 0.005 mm/y study visit. FMD was reduced in children receiving PIs. (95% CI, 0.0003 to 0.01; Pϭ0.04) (solid line) in HIV-infected chil- *Pϭ0.006, **PϽ0.0001. dren, whereas no association with age was found in control subjects (dotted line). B, IMT according to exposure to treat- ment. IMT was significantly higher in children who had received developed countries. Thus, HIV-infected children have the PIs. *Pϭ0.04, **Pϭ0.01. potential to survive even to the third or fourth decade of life. Because cardiovascular disease is emerging as an important Discussion health concern at these ages, it is critically important to This study demonstrates that structural and functional determine the early impact of both the HIV infection and its changes of the vasculature are already present during child- treatment on the arterial wall in these children. hood in HIV-infected children. These changes were most In our study, we measured carotid IMT to assess structural pronounced in children receiving PIs but were also observed disease of the arterial wall. IMT represents the cumulative in non–PI-treated and untreated children. Our findings sup- burden of adverse influences operating from the earliest port a role for both HIV infection itself and ART, particularly stages of disease. Increased IMT, shown to reflect conven- PIs, in the pathogenesis of early vascular disease, likely to be tional risk factor burden in children as young as 10 years of relevant to future clinical atherosclerosis. age, is an independent predictor of adverse cardiovascular outcome.17 We also assessed endothelial function, which is Several studies have raised concerns about the effect of known to be a key event in the initiation and progression of HIV and ART on both progression of atherosclerosis and preclinical atherosclerosis.18 In particular, reduction in local cardiovascular events in adults.4,16 PIs in particular are nitric oxide bioavailability, reflected by the magnitude of associated with metabolic abnormalities and lipodystrophic FMD response, is associated with a proinflammatory, prolif- phenotype in adults, which may add to the background risk erative, and procoagulant phenotype that establishes a locally 5 factor profile. However, the relative impact of disease, atherogenic environment. FMD is a dynamic measure of treatment, and underlying risk factor profile remains unclear. arterial function that can be used to assess the impact of both These factors are particularly hard to disentangle in adults risk factors and interventions.19,20 Using these 2 well- because the timing of HIV infection is often unknown and validated measurements of arterial structure and function, we there may be an interaction between the presence of classic have described the impact of disease and treatment, minimiz- risk factors for atherosclerotic disease and ART. ing the effects of exposure to confounding risk factors usually The long-term consequences of HIV infection are particu- present in adults. larly important for young individuals. With newer antiretro- Only 1 other study has examined the vascular changes of viral regimens, HIV has now become a chronic illness in long-term exposure to HIV in children. FMD was impaired, 108 Circulation July 5, 2005 as in our study, but IMT was not increased.21 The more certainly warrants further prospective exploration. Other extensive vascular changes in our population may be due to mechanisms may also contribute to the adverse effect of HIV older age and worse disease. IMT was related to both age (a and ART on vascular disease, including enhanced expression surrogate for duration of HIV) and CDC stage. The observed of macrophage scavenger receptors and mitochondrial differences between HIV-infected children and control sub- toxicity.23,27 jects remained after differences in the baseline characteristics HIV infection and ART are associated with an atherogenic between the groups were accounted for, suggesting that these structural and functional arterial phenotype from early child- changes may be related to HIV disease. Our study also raised hood. Because death rates among HIV-infected children have the possibility that vascular function may be related to ART. decreased 5-fold since the introduction of the highly active Both IMT and FMD were impaired in children receiving antiretroviral treatment, careful long-term monitoring appears ART; this was most pronounced in the PI-treated children. warranted to detect emerging cardiovascular disease. Longi- However, these findings must be viewed with some caution. tudinal studies are needed to understand the cause of vascular Although PIs would not have been included or excluded on disease and to ascertain the contribution of different ART the basis of disease severity, there have been changes in the regimens. Pharmacological or physiological interventions ART regimens since triple therapy became available for may be required to prevent future vascular events in HIV- children in 1997. Perhaps most important is that triple therapy infected children. at that time would likely have included a PI. In recent years, however, there has been a trend toward the use of a non- Acknowledgments nucleoside reverse-transcriptase inhibitor as first-line therapy. We thank the Greek State Scholarship Foundation, which supported We have attempted to control for variables such as age that Marietta Charakida, and CORDA, which supported Ann Donald may have influenced our analyses, but more studies are through the Silcock legacy. We thank the MRC Childhood Nutrition required to clarify the role of ART on vascular dysfunction. Research Center for funding Clare Storry. We are extremely grateful to all the children who took part in this study and the HIV Study It is also unclear whether the observed vascular changes Team, particularly Dr Vas Novelli. We thank T. Duong, Gill Wait, that appear to be associated with ART are driven by changes and K. Dierholt at MRC Clinical Trials Unit for assistance in in the metabolic profile of these children. Insulin resistance extracting data from the CHIPS cohort. and metabolic syndrome have been independently associated with vascular disease and accelerated cardiovascular disease References in HIV-infected adults. Insulin resistance and the full meta- 1. Gortmaker SL, Hughes M, Cervia J, Brady M, Johnson GM, Seage GR, bolic syndrome are prevalent in overweight youth, and it has III, Song LY, Dankner WM, Oleske JM. Effect of combination therapy including protease inhibitors on mortality among children and adolescents been suggested that they are increasingly prevalent in high- infected with HIV-1. N Engl J Med. 2001;345:1522–1528. risk HIV-infected young people, especially blacks.22 In the 2. Palella FJ Jr, Delaney KM, Moorman AC, Loveless MO, Fuhrer J, Satten present study, we did not find major differences in BMI z GA, Aschman DJ, Holmberg SD. Declining morbidity and mortality scores between cases and controls and between HIV-infected among patients with advanced human immunodeficiency virus infection: HIV Outpatient Study Investigators. N Engl J Med. 1998;338:853–860. children receiving different antiretroviral regimens. Further 3. Carr A, Cooper DA. Adverse effects of antiretroviral therapy. Lancet. studies are required to ascertain the vascular effects of ART 2000;356:1423–1430. in childhood and to provide insight into the mechanisms 4. Friis-Moller N, Sabin CA, Weber R, d’Arminio MA, El Sadr WM, Reiss contributing to vascular disease. P, Thiebaut R, Morfeldt L, De Wit S, Pradier C, Calvo G, Law MG, Kirk O, Phillips AN, Lundgren JD. Combination antiretroviral therapy and the HIV may promote atherosclerosis by activating the vascu- risk of myocardial infarction. N Engl J Med. 2003;349:1993–2003. lar endothelium directly or indirectly by systemic cytokine 5. Walli R, Herfort O, Michl GM, Demant T, Jager H, Dieterle C, Bogner stimulation by the virus.23 In addition, opportunistic agents JR, Landgraf R, Goebel FD. Treatment with protease inhibitors associated such as cytomegalovirus or herpesvirus, often present in HIV with peripheral insulin resistance and impaired oral glucose tolerance in HIV-1–infected patients. AIDS. 1998;12:F167–F173. 24,25 may contribute to endothelial damage. In this study, 6. Carr A, Samaras K, Chisholm DJ, Cooper DA. Pathogenesis of HIV-1- HIV-infected children had elevated total cholesterol and protease inhibitor–associated peripheral lipodystrophy, hyperlipidaemia, cholesterol subfractions. In particular, it is interesting to note and insulin resistance. Lancet. 1998;351:1881–1883. that Lp(a) levels, which are highly genetically determined, 7. Stary HC. Evolution and progression of atherosclerotic lesions in coronary arteries of children and young adults. Arteriosclerosis. 1989;9: also were elevated in HIV-infected children compared with I19-I32. control subjects. This finding, however, must be interpreted 8. Bennett-Richards K, Kattenhorn M, Donald A, Oakley G, Varghese Z, with caution. The average Lp(a) level is higher in black than Rees L, Deanfield JE. Does oral folic acid lower total homocysteine in white children, so ethnic differences may contribute to the levels and improve endothelial function in children with chronic renal failure? Circulation. 2002;105:1810–1815. Lp(a) variability. Furthermore, within the HIV cohort, the 9. Jarvisalo MJ, Raitakari M, Toikka JO, Putto-Laurila A, Rontu R, Laine S, most pronounced changes in Lp(a) were noted in children Lehtimaki T, Ronnemaa T, Viikari J, Raitakari OT. Endothelial dys- receiving PI therapy. It is possible that increased synthesis or function and increased arterial intima-media thickness in children with clearance of the particle related to HIV infection and PI type 1 diabetes. Circulation. 2004;109:1750–1755. 10. Gibb DM, Duong T, Tookey PA, Sharland M, Tudor-Williams G, Novelli therapy can account for our observed differences. The distur- V, Butler K, Riordan A, Farrelly L, Masters J, Peckham CS, Dunn DT. bances of lipids [increased total cholesterol and Lp(a) and a Decline in mortality, AIDS, and hospital admissions in perinatally HIV-1 smaller LDL particle size] in our children receiving PIs are infected children in the United Kingdom and Ireland. BMJ. 2003; similar to those reported in adults but did not account for the 327:1019. 11. Freeman JV, Cole TJ, Chinn S, Jones PR, White EM, Preece MA. Cross 26 observed vascular abnormalities. Nevertheless, the impact sectional stature and weight reference curves for the UK, 1990. Arch Dis of lipid abnormalities on vascular disease progression in HIV Child. 1995;73:17–24. Charakida et al Vascular Consequences of HIV and Its Treatment 109

12. Packard CJ, O’Reilly DS, Caslake MJ, McMahon AD, Ford I, Cooney J, Miller E, Witztum JL, Mietus-Snyder M. Antioxidant vitamins C and E Macphee CH, Suckling KE, Krishna M, Wilkinson FE, Rumley A, Lowe improve endothelial function in children with hyperlipidemia: Endothelial GD. Lipoprotein-associated phospholipase A2 as an independent pre- Assessment of Risk From Lipids in Youth (EARLY) Trial. Circulation. dictor of coronary heart disease: West of Scotland Coronary Prevention 2003;108:1059–1063. Study Group. N Engl J Med. 2000;343:1148–1155. 21. Bonnet D, Aggoun Y, Szezepanski I, Bellal N, Blanche S. Arterial 13. Belo L, Caslake M, Gaffney D, Santos-Silva A, Pereira-Leite L, stiffness and endothelial dysfunction in HIV-infected children. AIDS. Quintanilha A, Rebelo I. Changes in LDL size and HDL concentration in 2004;18:1037–1041. normal and preeclamptic pregnancies. Atherosclerosis. 2002;162: 22. Bitnun A, Sochett E, Dick PT, To T, Jefferies C, Babyn P, Forbes J, Read 425–432. S, King SM. Insulin sensitivity and beta-cell function in protease inhib- 14. Leeson CP, Whincup PH, Cook DG, Donald AE, Papacosta O, Lucas A, itor–treated and –naive human immunodeficiency virus–infected Deanfield JE. Flow-mediated dilation in 9- to 11-year-old children: the children. J Clin Endocrinol Metab. 2005;90:168–174. influence of intrauterine and childhood factors. Circulation. 1997;96: 23. Lewis W. Atherosclerosis in AIDS: potential pathogenetic roles of anti- J Mol Cell Cardiol 2233–2238. retroviral therapy and HIV. . 2000;32:2115–2129. 24. Alber DG, Vallance P, Powell KL. Enhanced atherogenesis is not an 15. Centers for Disease Control. 1994 Revised classification system for obligatory response to systemic herpesvirus infection in the apoE- human immunodeficiency virus infection in children less than 13 years of deficient mouse: comparison of murine gamma-herpesvirus-68 and age. MMWR. 1994;43:RR-12. herpes simplex virus-1. Arterioscler Thromb Vasc Biol. 2002;22: 16. Rickerts V, Brodt H, Staszewski S, Stille W. Incidence of myocardial 793–798. infarctions in HIV-infected patients between 1983 and 1998: the 25. Nieto FJ, Adam E, Sorlie P, Farzadegan H, Melnick JL, Comstock GW, Frankfurt HIV-Cohort Study. Eur J Med Res. 2000;5:329–333. Szklo M. Cohort study of cytomegalovirus infection as a risk factor for 17. Jarvisalo MJ, Jartti L, Nanto-Salonen K, Irjala K, Ronnemaa T, Hartiala carotid intimal-medial thickening, a measure of subclinical atherosclero- JJ, Celermajer DS, Raitakari OT. Increased aortic intima-media thickness: sis. Circulation. 1996;94:922–927. a marker of preclinical atherosclerosis in high-risk children. Circulation. 26. Stein JH, Klein MA, Bellehumeur JL, McBride PE, Wiebe DA, Otvos JD, 2001;104:2943–2947. Sosman JM. Use of human immunodeficiency virus-1 protease inhibitors 18. Ross R. The pathogenesis of atherosclerosis: a perspective for the 1990s. is associated with atherogenic lipoprotein changes and endothelial dys- Nature. 1993;362:801–809. function. Circulation. 2001;104:257–262. 19. de Jongh S, Lilien MR, op’t Roodt J, Stroes ES, Bakker HD, Kastelein JJ. 27. Dressman J, Kincer J, Matveev SV, Guo L, Greenberg RN, Guerin T, Early statin therapy restores endothelial function in children with familial Meade D, Li XA, Zhu W, Uittenbogaard A, Wilson ME, Smart EJ. HIV hypercholesterolemia. J Am Coll Cardiol. 2002;40:2117–2121. protease inhibitors promote atherosclerotic lesion formation independent 20. Engler MM, Engler MB, Malloy MJ, Chiu EY, Schloetter MC, Paul SM, of dyslipidemia by increasing CD36-dependent cholesteryl ester accumu- Stuehlinger M, Lin KY, Cooke JP, Morrow JD, Ridker PM, Rifai N, lation in macrophages. J Clin Invest. 2003;111:389–397. Stroke

Prediction of Myocardial Infarction by N-Terminal-Pro-B–Type Natriuretic Peptide, C-Reactive Protein, and Renin in Subjects With Cerebrovascular Disease

Duncan J. Campbell, MD, PhD; Mark Woodward, PhD; John P. Chalmers, MD, PhD; Samuel A. Colman, MBios; Alicia J. Jenkins, MD; Bruce E. Kemp, PhD; Bruce C. Neal, MD, PhD; Anushka Patel, MD; Stephen W. MacMahon, PhD

Background—B-type natriuretic peptide (BNP), C-reactive protein (CRP), and renin are elevated in persons at risk for cardiovascular disease. However, data that directly compare these markers in the prediction of myocardial infarction (MI) are limited. Methods and Results—N-terminal-proBNP (NT-proBNP), CRP, and renin were measured in baseline blood samples from a nested case-control study of the 6105 participants of the Perindopril Protection Against Recurrent Stroke Study (PROGRESS), a placebo-controlled study of a perindopril-based blood pressure–lowering regimen among individuals with previous stroke or transient ischemic attack. Each of 206 subjects who experienced MI, either fatal or nonfatal, during a mean follow-up of 3.9 years was matched to 1 to 3 control subjects. Most MI cases (67%) occurred in subjects without a history of coronary heart disease. NT-proBNP, CRP, and renin each predicted MI; the odds ratio for subjects in the highest compared with the lowest quarter was 2.2 (95% CI, 1.3 to 3.6) for NT-proBNP, 2.2 (95% CI, 1.3 to 3.6) for CRP, and 1.7 (95% CI, 1.1 to 2.8) for renin. NT-proBNP and renin, but not CRP, remained predictors of MI after adjustment for all other predictors, including LDL and HDL cholesterol levels. Individuals with both NT-proBNP and renin in their highest quarters had 4.5 times the risk of MI compared with subjects with both biological markers in their lowest quarters. Conclusions—NT-proBNP and renin, but not CRP, are independent predictors of MI risk after stroke or transient ischemic attack, providing information additional to that provided by classic risk factors, and may enable more effective targeting of MI prevention strategies. (Circulation. 2005;112:110-116.) Key Words: C-reactive protein Ⅲ myocardial infarction Ⅲ natriuretic peptides Ⅲ renin Ⅲ risk factors

lthough known risk factors account for Ն90% of the persons at risk of MI, and it is unknown whether screening for Apopulation-attributable risk for myocardial infarction Ͼ1 of these biological markers provides better prognostic (MI),1 the costs of prevention strategies have led to a targeted information than screening for 1 marker. approach to prevention of coronary heart disease, whereby interventions are tailored to the estimated risk of the individ- Seep9 ual.2,3 To better identify individuals at highest risk, many We evaluated the prognostic performance of N-terminal different potential plasma markers have been assessed for (NT)-proBNP, CRP, and renin in a population with cerebro- prediction of MI risk, including C-reactive protein (CRP), vascular disease and increased risk of MI by conducting a B-type natriuretic peptides (BNP), and renin,4–14 and there nested case-control study of 206 subjects who developed MI has been recent debate about the magnitude of risk associated and 412 control subjects who did not develop MI who were with elevated CRP levels.15,16 When measured after an acute participants in the Perindopril Protection Against Recurrent coronary event, both BNP-related peptides and CRP, sepa- Stroke Study (PROGRESS). PROGRESS was a multicenter, rately and together, predict mortality, MI, and heart fail- randomized, double-blind, placebo-controlled study designed ure.17–21 However, limited data are available that directly to determine the effects of active therapy with a perindopril- compare BNP-related peptides, CRP, and renin in stable based blood pressure–lowering regimen on the risks of stroke

Received November 30, 2004; revision received December 23, 2004; accepted March 4, 2005. From St Vincent’s Institute of Medical Research (D.J.C., B.E.K.) and Department of Medicine (D.J.C., A.J.J., B.E.K.), University of Melbourne, St Vincent’s Hospital, Fitzroy, Victoria; CSIRO Health Sciences and Nutrition (B.E.K.), Parkville, Victoria; and The George Institute for International Health (M.W., J.P.C., S.A.C., B.C.N., A.P., S.W.M.), University of Sydney, Camperdown, New South Wales, Australia. Correspondence to Dr D.J. Campbell, St. Vincent’s Institute of Medical Research, 41 Victoria Parade, Fitzroy, Victoria 3065, Australia. E-mail [email protected] © 2005 American Heart Association, Inc. Circulation is available at http://www.circulationaha.org DOI: 10.1161/CIRCULATIONAHA.104.525527 110 Campbell et al Prediction of MI in Cerebrovascular Disease 111 and other major vascular events among individuals with a (Roche Diagnostics),26 and CRP was measured by high-sensitivity stroke or transient ischemic attack (TIA) within the previous nephelometry (Dade Behring) on heparin plasma that had been 5 years.22–24 This regimen substantially reduced the risk of thawed once or twice before assay. We previously showed that NT-proBNP and CRP are stable in plasma that was thawed twice stroke by 28%, major coronary events by 26%, congestive (unpublished). Plasma samples from cases and controls were assayed heart failure (CHF) by 26%, and nonfatal MI by 38%.22,24 for lipids, renin, NT-proBNP, and CRP in identical and masked fashion. The local laboratory at each study center measured plasma Methods creatinine. Patients and Study Protocol Statistical Methods We conducted a prospective nested case-control analysis among The base population for study was taken as all those who had plasma PROGRESS participants, who were a predominantly elderly popu- frozen and stored at baseline (nϭ5918). A nested case-control lation with cerebrovascular disease and were at high risk of both sample was selected from this base population. Cases (nϭ206) were coronary events and CHF.24 The design and major outcomes of anyone from the base population who developed MI during the mean PROGRESS have been described in detail elsewhere.22–24 Briefly, follow-up of the PROGRESS trial of 3.9 years. By design, 412 6105 participants were recruited from 172 collaborating centers in 10 ϫ countries from Australasia, Europe, and Asia between 1995 and (206 2) matched controls without MI during follow-up in the 1997. Participants were randomized to either placebo (nϭ3054) or PROGRESS trial were sampled, giving a total of 618 cases and active therapy (nϭ3051), comprising a flexible regimen based on the controls for biochemical analysis. For calculation of odds ratios, each ACE inhibitor perindopril (4 mg daily) with the addition of the case was matched with 1 to 3 controls randomly sampled from this diuretic indapamide at the discretion of treating physicians. The group of 618 subjects who were alive and did not have an MI institutional ethics committee of each collaborating center approved between randomization and the time of case ascertainment; anyone the trial, and all participants provided written, informed consent. who became a case after the onset of MI in the index case was Individuals were potentially eligible if they had a history of cere- eligible for selection as a match, and controls could be matched for brovascular disease (ischemic stroke, hemorrhagic stroke, or TIA but Ͼ1 case.27 Matching variables were age (within 5 years), gender, not subarachnoid hemorrhage) within the previous 5 years and no treatment allocated (perindopril based or placebo, monotherapy or clear indication for or contraindication to ACE inhibitor treatment. dual therapy), region (Australia and New Zealand, China, Japan, There were no blood pressure criteria for entry, but it was recom- France and Belgium, Italy, Sweden, United Kingdom, and Ireland), mended that the blood pressure of all hypertensive subjects be and most recent qualifying event (ischemic stroke or TIA, hemor- controlled with agents other than ACE inhibitors and angiotensin rhagic stroke, stroke of unknown type) at randomization. If cases receptor blockers before entry into the study. CHF was a study who acted as controls for other cases and controls who acted for exclusion criterion. multiple cases are included, 45 cases had 3 matched controls, 157 Before randomization, information was collected on history of had 2 matched controls, and 4 had 1 matched control. Because no vascular disease, other vascular risk factors, and current medica- matched controls were available, 3 cases were matched on incom- tions.22–24 The presence of coronary heart disease at baseline was plete criteria: Two were not matched for region, and 1 was not based on a history of MI or coronary revascularization or a history of matched for region and most recent qualifying event. angina supported by documented ECG or angiographic evidence of Baseline variables were compared between cases and controls by coronary disease. Blood samples were collected (nonfasting, 10 mL the use of ␹2, parametric, or nonparametric tests as appropriate.27 The into heparin and 10 mL into EDTA Vacutainer tubes) from 5918 of case-control data were analyzed using 2 different conditional logistic the 6105 subjects on enrollment before the run-in phase of the study, regression models to obtain odds ratios for NT-proBNP, CRP, and at which time none of the subjects was receiving ACE inhibitor or renin, according to equal quarters of the distribution of each plasma angiotensin receptor blocker therapy. After randomization, partici- marker in the total sample (cases and controls).27 Model 1 was pants were scheduled to be seen on 5 occasions in the first year and unadjusted except for the matching variables. Model 2 was adjusted every 6 months thereafter until the end of the scheduled follow-up through matching and for all baseline predictors of MI: baseline period or death. Information on all serious adverse events, including systolic blood pressure, total cholesterol, HDL cholesterol, LDL those resulting from stroke, coronary heart disease, and CHF, were cholesterol, and a history of coronary heart disease, valvular heart routinely recorded. All outcomes were coded according to the ninth disease, and peripheral arterial disease. In addition, model 2 for revision of the International Classification of Diseases (ICD-9). The NT-proBNP was adjusted for CRP and renin; model 2 for CRP was diagnosis of MI, either fatal (death within 28 days of MI) or nonfatal, adjusted for NT-proBNP and renin; and model 2 for renin was was based on the combination of an appropriate clinical history adjusted for NT-proBNP and CRP. Continuous variables are shown supported by ECG changes and/or an elevation of cardiac enzymes as meanϮSD, except for plasma levels of triglycerides, NT-proBNP, or other biochemical markers of myocardial injury (ICD-9 code CRP, and renin, which were not normally distributed and are 410.0 to 410.9). Sudden deaths were excluded from this outcome presented as median (interquartile range). Odds ratios are shown as because of uncertainty about the cause of death in this population mean (95% CI). All probability values were 2 tailed, and values of with preexisting cerebrovascular disease, as were deaths resulting PϽ0.05 were considered to indicate statistical significance. All from other acute, subacute, or chronic forms of ischemic heart analyses were performed with SAS version 8.2 (SAS Institute Inc). disease. An adjudication committee, blinded to study treatment allocation, reviewed source documentation for every potential MI, stroke, and all deaths recorded during the study follow-up period. Results The baseline clinical and biochemical characteristics of the Biochemical Analyses 206 MI cases and 412 controls without MI are given in Table Coded blood samples were shipped at 4°C to the local laboratory, 1. Most MI cases (67%) did not have a previous history of where they were centrifuged at 2000g for 10 minutes at 4°C, and the plasma was frozen in aliquots and stored at Ϫ80°C until assay. coronary heart disease. Compared with controls without MI Lipids were measured on first-thaw EDTA plasma by automated during the period of observation, MI cases had higher direct-measurement assay (Olympus AU2700, Olympus America baseline systolic blood pressure and were more likely to have Inc). LDL cholesterol was calculated for those with triglycerides a history of coronary heart disease, valvular heart disease, and Ͻ4.5 mmol/L by use of Friedewald’s equation.25 Plasma levels of active renin were measured on first-thaw heparin plasma by immu- peripheral arterial disease (Table 1). In addition, MI cases had noradiometric assay (Nichols Institute Diagnostics). NT-proBNP higher baseline total and LDL cholesterol, lower HDL cho- was measured by radioimmunoassay with an Elecsys instrument lesterol, and higher levels of NT-proBNP and CRP (Table 1). 112 Circulation July 5, 2005

TABLE 1. Baseline Clinical and Biochemical Characteristics of Patients Cases Controls Without MI Characteristic (nϭ206) (nϭ412) P Mean age, y 68Ϯ868Ϯ8 0.88* Male, n (%) 170 (83) 336 (82) 0.77* Asian,† n (%) 45 (22) 96 (23) 0.68* Randomization to perindopril-based therapy, 81 (39) 164 (40) 0.91* n (%) Qualifying event‡ Ischemic stroke, n (%) 143 (69) 286 (69) 1 Cerebral hemorrhage, n (%) 16 (8) 32 (8) 1 Unknown stroke, n (%) 20 (10) 30 (7) 0.30 TIA or amaurosis fugax, n (%) 51 (25) 105 (25) 0.84 Time since qualifying event, mo 8 (3–19) 8 (3–19) 0.97 Other vascular disease history Systolic blood pressure, mm Hg 153Ϯ19 149Ϯ19 0.03 Diastolic blood pressure, mm Hg 86Ϯ12 84Ϯ11 0.12 Hypertension,§ n (%) 109 (53) 188 (46) 0.09 Antihypertensive therapy, n (%) 112 (54) 210 (51) 0.43 History of coronary heart disease,ʈ n (%) 68 (33) 94 (23) 0.007 History of atrial fibrillation, n (%) 17 (8) 37 (9) 0.76 Valvular heart disease, n (%) 9 (4) 5 (1) 0.01 Left ventricular hypertrophy on ECG, 16 (8) 41 (10) 0.38 n (%) Peripheral arterial disease, n (%) 22 (11) 24 (6) 0.03 Other medical history Current smoker, n (%) 47 (23) 79 (19) 0.29 History of diabetes, n (%) 37 (18) 58 (14) 0.21 Body mass index, kg/m2 26.4Ϯ3.6 26.0Ϯ3.5 0.19 Current medications ␤-Blockers, n (%) 54 (26) 103 (25) 0.74 Calcium channel blockers, n (%) 90 (44) 150 (36) 0.08 Diuretics, n (%) 37 (18) 61 (15) 0.31 Other antihypertensives, n (%) 16 (8) 32 (8) 1 Antiplatelet agents, n (%) 160 (78) 313 (76) 0.64 Statins, n (%) 17 (8) 39 (9) 0.62 Oral anticoagulants, n (%) 19 (9) 49 (12) 0.32 Plasma lipids,¶ mmol/L Total cholesterol 5.8Ϯ1.1 5.5Ϯ1.1 0.002 HDL cholesterol 1.18Ϯ0.31 1.26Ϯ0.38 0.01 LDL cholesterol 3.59Ϯ1.0 3.27Ϯ0.94 0.0002 Triglycerides 2.1 (1.5–3.1) 2.0 (1.4–2.8) 0.11 Creatinine, ␮mol/L 96Ϯ24 94Ϯ23 0.34 NT-proBNP, pmol/L 23.2 (9.4–59.1) 14.8 (7.4–32.8) 0.0002 CRP, mg/L 2.7 (1.4–6.8) 2.1 (1.0–5.0) 0.0004 Renin, mU/L 27 (15–46) 23 (11–39) 0.10 Continuous variables are shown as meanϮSD or median (interquartile range). *Controls were matched for these criteria. †Participants were recruited from People’s Republic of China or Japan. ‡Subjects could have Ͼ1 qualifying event. §Baseline systolic blood pressure Ն160 mm Hg or diastolic blood pressure Ն90 mm Hg. ʈHistory of MI, coronary revascularization, or angina (supported by documented ECG or angio- graphic evidence of coronary disease). ¶To convert values for cholesterol to mg/dL, divide by 0.02586; to convert values for triglycerides to mg/dL, divide by 0.01129; to convert values for creatinine to mg/L, multiply by 0.113; to convert values for NT-proBNP to ng/L, multiply by 8.4. Campbell et al Prediction of MI in Cerebrovascular Disease 113

TABLE 2. Odds Ratios for MI According to Baseline Plasma Levels of NT-proBNP, CRP, and Renin Quarter

P for Variable 1 2 3 4 Trend NT-proBNP (pmol/L) Mean 4.2 12.0 27.2 115.4 Range 0.5–8.2 8.3–17.3 17.4–41.3 42.3–686.7 Model 1 Odds ratio 1 0.89 1.87 2.17 0.0004 95% CI 0.53–1.48 1.10–3.18 1.30–3.62 Model 2 Odds ratio 1 0.91 1.99 2.67 0.0002 95% CI 0.51–1.60 1.08–3.68 1.43–4.97 CRP, mg/L Mean 0.6 1.7 3.7 12.0 Range 0.1–1.1 1.1–2.4 2.4–5.5 5.5–62.6 Model 1 Odds ratio 1 1.72 1.75 2.18 0.004 95% CI 1.03–2.86 1.04–2.94 1.32–3.61 Model 2 Odds ratio 1 1.50 0.92 1.33 0.68 95% CI 0.85–2.65 0.50–1.66 0.75–2.35 Renin, mU/L Mean 7.6 17.5 31.6 116.0 Range 0.5–12.4 12.5–23.3 23.5–40.6 40.7–1050.8 Model 1 Odds ratio 1 1.11 1.08 1.73 0.03 95% CI 0.68–1.81 0.65–1.80 1.06–2.81 Model 2 Odds ratio 1 1.10 1.41 1.91 0.001 95% CI 0.64–1.90 0.80–2.49 1.10–3.32 Model 1 is not adjusted, except through matching. Model 2 is adjusted through matching and for baseline systolic blood pressure, total cholesterol, HDL cholesterol, LDL cholesterol, and history of coronary heart disease, valvular heart disease, and peripheral arterial disease. In addition, model 2 for NT-proBNP is adjusted for CRP and renin; model 2 for CRP is adjusted for NT-proBNP and renin; and model 2 for renin is adjusted for NT-proBNP and CRP.

The odds ratio for MI increased with increasing levels of when HDL cholesterol (Pϭ0.20), systolic blood pressure baseline NT-proBNP, CRP, and renin. For subjects in the (Pϭ0.13), or NT-proBNP (Pϭ0.12) was added separately to highest compared with the lowest quarter, the odds ratio was the model, CRP was no longer a statistically significant 2.2 (95% CI, 1.3 to 3.6) for NT-proBNP, 2.2 (95% CI, 1.3 to predictor of MI. The combination of NT-proBNP and renin 3.6) for CRP, and 1.7 (95% CI, 1.1 to 2.8) for renin (Table 2). had a multiplicative effect on the odds ratio for MI, so that NT-proBNP and renin, but not CRP, remained independent individuals with both NT-proBNP and renin in their highest predictors of MI in multivariable analysis (model 2, Table 2). quarters had 4.5 times the risk of MI compared with subjects Other variables that were predictors after multivariable anal- with both biological markers in their lowest quarters (the ysis were HDL cholesterol and LDL cholesterol. Although Figure). there was a trend for MI cases to be more likely to be taking calcium channel blocker therapy, addition of this parameter to Discussion model 2 did not significantly change the calculated odds NT-proBNP and renin were independent predictors of MI ratios. To examine why CRP failed to predict MI risk in risk, and each marker provided information additional to that multivariable analysis, we examined the effects of other obtained from established risk factors. Although CRP pre- variables in model 2. CRP predicted MI (PϽ0.05) when a dicted MI risk in an analysis adjusted for the matching history of coronary heart disease, valvular heart disease, and variables alone (model 1, Table 2), it did not predict MI in peripheral arterial disease, as well as total cholesterol, LDL multivariable analysis including all other predictors (model 2, cholesterol, and renin, were included in the model. However, Table 2). 114 Circulation July 5, 2005

these studies.7–9 Plasma BNP level predicted dopamine- induced myocardial ischemia in patients with known or suspected coronary heart disease,32 suggesting that increased BNP levels may be associated with subclinical cardiac ischemia and therefore increased risk of MI. Our study provides additional new information by demonstrating that NT-proBNP predicts MI in stable patients, most of whom did not have a history of coronary heart disease. Elevated BNP and NT-proBNP levels may indicate ventricular strain,33 and release of BNP from ventricular myocytes may be an impor- tant compensatory response. BNP has beneficial effects in heart failure34 and is reported to reduce infarct size in experimental MI.35 CRP is an established risk factor for cardiovascular events, including MI,4,5 although there has been recent debate about the magnitude of risk associated with elevated CRP lev- els.15,16 Our finding that NT-proBNP was superior to CRP in the prediction of MI risk may be a result of CRP being a nonspecific marker of acute-phase response,36 whereas NT- proBNP is a cardiac-specific marker. Increased BNP and Multivariable-adjusted odds ratio for MI according to quarters of NT-proBNP levels are associated with a broad range of NT-proBNP and renin concentration. Reference group is those 37 subjects in quarter 1 for both NT-proBNP and renin. Ranges of cardiac conditions, most of which relate to increased ven- NT-proBNP and renin levels for each quarter are shown in Table tricular strain.33 NT-proBNP and CRP were independent 2. Variables adjusted for are listed for model 2 in the footnote predictors of risk of CHF in PROGRESS participants.31 It is for Table 2. of note that 78% of MI cases and 76% of controls without MI were receiving antiplatelet agents (Table 1), predominantly We found that NT-proBNP levels above the median of 17 aspirin. Ridker et al4 showed that aspirin therapy attenuated pmol/L were associated with increased risk of MI. This prediction of risk of MI and stroke by CRP in the Physicians’ median NT-proBNP level was within the normal range for Health Study, and aspirin therapy may have similarly atten- NT-proBNP levels.28–30 Median NT-proBNP levels from uated the prediction of MI risk by CRP in our study. population studies approximate 33 pmol/L,28,29 although a Our finding that renin predicted MI risk in subjects with lower median of 16 to 19 pmol/L is reported for healthy cerebrovascular disease is in agreement with previous reports subjects without a history or symptoms of heart disease or that renin predicted coronary heart disease in hypertensive other chronic disease.29,30 We recently showed that similar subjects,10,12,14 although other studies failed to observe such levels of NT-proBNP were associated with increased risk of an association in either normotensive or hypertensive sub- CHF in PROGRESS participants.31 Wang et al6 similarly 11,13 found that BNP levels within a range currently regarded as jects. Further evidence for an association between renin normal were associated with an increased risk of death and and cardiovascular disease is its association with increased 38 first cardiovascular event, including heart failure, atrial fibril- urinary albumin excretion in essential hypertension. The lation, and stroke or TIA in the Framingham Offspring Study. relatively few studies of renin as a risk factor for cardiovas- However, BNP did not predict coronary heart disease events cular disease may be due in part to the difficulties in in the Framingham Offspring Study.6 The difference between measuring plasma renin activity. Our demonstration of renin prediction of MI risk by NT-proBNP in PROGRESS partic- as an independent predictor of MI risk using a simple ipants and failure to predict coronary heart disease events by immunoradiometric assay for active renin may lead to further BNP in the Framingham Offspring Study may have been due assessment of the potential value of renin as a risk factor for to differences between the 2 study populations. PROGRESS cardiovascular disease. participants had a mean age of 68 years, and all had A large body of data implicates angiotensin II in the established vascular disease, as evidenced by preceding pathogenesis of vascular disease,39 and angiotensin II may stroke or TIA, whereas subjects in the Framingham Offspring provide the link between increased renin levels and MI risk. Study were younger (mean age, 58 years), and most were free However, it is also possible that increased renin levels are a of cardiovascular disease at baseline. In addition, we used a marker of MI risk without playing a pathogenic role.13 For narrower definition of coronary heart disease events than example, increased MI risk and increased renin levels may Wang et al.6 both be consequences of generalized vascular disease, includ- When measured after an acute coronary event, both BNP- ing renovascular disease, which may cause increased renin related peptides and CRP, separately and together, predict secretion.40 Perindopril-based therapy reduced the incidence mortality, MI, and heart failure.17–21 Other studies of stable of major coronary events by 26% in PROGRESS partici- subjects without acute coronary syndrome show that BNP pants; however, our study did not have sufficient statistical and NT-proBNP predict all-cause mortality and cardiovascu- power to examine whether baseline plasma levels of NT- lar events, although MI risk was not specifically examined in proBNP, CRP, or renin predicted benefit from this therapy. Campbell et al Prediction of MI in Cerebrovascular Disease 115

The potential limitations of these data merit consideration. References Our analyses were based on single baseline determinations 1. Yusuf S, Hawken S, Ounpuu S, Dans T, Avezum A, Lanas F, McQueen that may not accurately reflect NT-proBNP, CRP, and renin M, Budaj A, Pais P, Varigos J, Lisheng L. Effect of potentially modifiable risk factors associated with myocardial infarction in 52 countries (the levels over the mean follow-up of 3.9 years. This source of INTERHEART study): case-control study. Lancet. 2004;364:937–952. variability may have contributed to the failure of CRP to 2. Greenland P, Smith SC Jr, Grundy SM. Improving coronary heart disease predict MI risk in multivariable analysis, but it cannot risk assessment in asymptomatic people: role of traditional risk factors account for the observed associations between NT-proBNP, and noninvasive cardiovascular tests. Circulation. 2001;104:1863–1867. 3. National Cholesterol Education Program, National Heart, Lung, and renin, and MI risk because any random misclassification Blood Institute, National Institutes of Health. Third report of the National would bias results toward the null hypothesis. Our definition Cholesterol Education Program (NCEP) Expert Panel on Detection, Eval- of MI may have excluded cases of silent MI or those that uation, and Treatment of High Blood Cholesterol in Adults (Adult resulted in sudden death; however, our MI cases did not include Treatment Panel III): final report. Circulation. 2002;106:3143–3421. 4. Ridker PM, Cushman M, Stampfer MJ, Tracy RP, Hennekens CH. instances of death not caused by MI. In our nested case-control Inflammation, aspirin, and the risk of cardiovascular disease in apparently design, we matched for 5 baseline variables but were unable to healthy men. N Engl J Med. 1997;336:973–979. match for all predictors of MI such as known cardiac disease 5. Pearson TA, Mensah GA, Alexander RW, Anderson JL, Cannon RO 3rd, Ͼ Criqui M, Fadl YY, Fortmann SP, Hong Y, Myers GL, Rifai N, Smith SC because of difficulty in matching for 5 variables; however, we Jr, Taubert K, Tracy RP, Vinicor F. Markers of inflammation and car- were able to adjust for these baseline predictors of MI in our diovascular disease: application to clinical and public health practice: a multivariable analysis (model 2, Table 2). statement for healthcare professionals from the Centers for Disease What is the clinical value of measurement of NT-proBNP Control and Prevention and the American Heart Association. Circulation. 2003;107:499–511. and renin? We have shown that NT-proBNP and renin 6. Wang TJ, Larson MG, Levy D, Benjamin EJ, Leip EP, Omland T, Wolf provided prognostic information additional to that provided PA, Vasan RS. Plasma natriuretic peptide levels and the risk of cardio- by classic risk factors, including systolic blood pressure, total vascular events and death. N Engl J Med. 2004;350:655–663. 7. McDonagh TA, Cunningham AD, Morrison CE, McMurray JJ, Ford I, cholesterol, HDL cholesterol, LDL cholesterol, and a history Morton JJ, Dargie HJ. Left ventricular dysfunction, natriuretic peptides, of coronary heart disease, valvular heart disease, and periph- and mortality in an urban population. Heart. 2001;86:21–26. eral arterial disease. Moreover, MI cases and controls were 8. Ueda R, Yokouchi M, Suzuki T, Otomo E, Katagiri T. Prognostic value matched for age, sex, region, qualifying event, and allocated of high plasma brain natriuretic peptide concentrations in very elderly persons. Am J Med. 2003;114:266–270. treatment. Whereas measurement of NT-proBNP and renin is 9. Olsen MH, Wachtell K, Tuxen C, Fossum E, Bang LE, Hall C, Ibsen H, unlikely to influence management of individuals who already Rokkedal J, Devereux RB, Hildebrandt P. N-terminal pro-brain natriuret- qualify for intensive strategies for prevention of cardiovas- ic peptide predicts cardiovascular events in patients with hypertension cular disease, measurement of these markers may assist in the and left ventricular hypertrophy: a LIFE study. J Hypertens. 2004;22: 1597–1604. identification of individuals at increased risk of MI who may 10. Brunner HR, Laragh JH, Baer L, Newton MA, Goodwin FT, Krakoff LR, not otherwise be identified by classic risk factors. Bard RH, Buhler FR. Essential hypertension: renin and aldosterone, heart In conclusion, we performed a head-to-head evaluation of attack and stroke. N Engl J Med. 1972;286:441–449. 11. Doyle AE, Jerums G, Johnston CI, Louis WJ. Plasma renin levels and a range of potential risk factors for MI, including classic risk vascular complications in hypertension. BMJ. 1973;2:206–207. factors, in subjects with previous stroke or TIA. We found 12. Alderman MH, Madhavan S, Ooi WL, Cohen H, Sealey JE, Laragh JH. that NT-proBNP and renin, but not CRP, independently Association of the renin-sodium profile with the risk of myocardial predicted MI risk. The prognostic information provided by infarction in patients with hypertension. N Engl J Med. 1991;324: 1098–1104. NT-proBNP and renin was additional to that provided by 13. Meade TW, Cooper JA, Peart WS. Plasma renin activity and ischemic traditional cardiovascular risk factors and may enable more heart disease. N Engl J Med. 1993;329:616–619. effective targeting of MI prevention strategies. Further stud- 14. Alderman MH, Ooi WL, Cohen H, Madhavan S, Sealey JE, Laragh JH. Plasma renin activity: a risk factor for myocardial infarction in hyper- ies are required to assess whether the use of NT-proBNP and tensive patients. Am J Hypertens. 1997;10:1–8. renin levels to guide management improves patient outcomes. 15. Danesh J, Wheeler JG, Hirschfield GM, Eda S, Eiriksdottir G, Rumley A, Lowe GD, Pepys MB, Gudnason V. C-reactive protein and other circu- Acknowledgments lating markers of inflammation in the prediction of coronary heart disease. N Engl J Med. 2004;350:1387–1397. This study was funded by grants from the NIH (5-R01-HL-071685), 16. Tall AR. C-reactive protein reassessed. N Engl J Med. 2004;350: National Health and Medical Research Council of Australia, and 1450–1452. National Heart Foundation of Australia. Drs Campbell and Neal are 17. de Lemos JA, Morrow DA, Bentley JH, Omland T, Sabatine MS, recipients of Career Development Awards, and Dr Jenkins is a McCabe CH, Hall C, Cannon CP, Braunwald E. The prognostic value of recipient of a Clinical Research Fellowship from the National Heart B-type natriuretic peptide in patients with acute coronary syndromes. Foundation of Australia. Dr Kemp is an Australian Research Council N Engl J Med. 2001;345:1014–1021. Federation Fellow. PROGRESS was funded by grants from Servier, 18. Jernberg T, Stridsberg M, Venge P, Lindahl B. N-terminal pro brain the Health Research Council of New Zealand, and the National natriuretic peptide on admission for early risk stratification of patients Health and Medical Research Council of Australia. with chest pain and no ST-segment elevation. J Am Coll Cardiol. 2002; 40:437–445. Disclosure 19. Sabatine MS, Morrow DA, de Lemos JA, Gibson CM, Murphy SA, Rifai N, McCabe C, Antman EM, Cannon CP, Braunwald E. Multimarker Dr Campbell has had research contracts with Solvay Pharmaceuticals approach to risk stratification in non–ST elevation acute coronary syn- and Novartis and has received consulting fees from Novartis during dromes: simultaneous assessment of troponin I, C-reactive protein, and the past 5 years. Drs Chalmers and MacMahon hold research grants B-type natriuretic peptide. Circulation. 2002;105:1760–1763. from Servier, as Chief Investigators for PROGRESS and AD- 20. James SK, Lindahl B, Siegbahn A, Stridsberg M, Venge P, Armstrong P, VANCE, administered by the University of Sydney. Drs Chalmers, Barnathan ES, Califf R, Topol EJ, Simoons ML, Wallentin L. N-terminal Patel, Neal, Woodward, and MacMahon have also received hono- pro-brain natriuretic peptide and other risk markers for the separate raria from Servier for speaking in relation to PROGRESS and/or prediction of mortality and subsequent myocardial infarction in patients ADVANCE at scientific meetings. with unstable coronary artery disease: a Global Utilization of Strategies to 116 Circulation July 5, 2005

Open Occluded Arteries (GUSTO)-IV substudy. Circulation. 2003;108: enlargement and impaired systolic function with plasma N-terminal pro 275–281. brain natriuretic peptide concentrations. Am Heart J. 2002;143:923–929. 21. Richards AM, Nicholls MG, Espiner EA, Lainchbury JG, Troughton RW, 31. Campbell DJ, Woodward M, Chalmers JP, Colman SA, Jenkins AJ, Elliott J, Frampton C, Turner J, Crozier IG, Yandle TG. B-type natriuretic Kemp BE, Neal BC, Patel A, MacMahon SW. Prediction of heart failure peptides and ejection fraction for prognosis after myocardial infarction. by amino-terminal-pro-B-type natriuretic peptide and C-reactive protein Circulation. 2003;107:2786–2792. in subjects with cerebrovascular disease. Hypertension. 2005;45:69–74. 22. PROGRESS Collaborative Group. Randomised trial of a 32. Asada J, Tsuji H, Iwasaka T, Thomas JD, Lauer MS. Usefulness of perindopril-based blood-pressure–lowering regimen among 6,105 indi- plasma brain natriuretic peptide levels in predicting dobutamine-induced viduals with previous stroke or transient ischaemic attack. Lancet. 2001; myocardial ischemia. Am J Cardiol. 2004;93:702–704. 358:1033–1041. 33. Liang F, Wu J, Garami M, Gardner DG. Mechanical strain increases 23. PROGRESS Management Committee. Blood pressure lowering for the expression of the brain natriuretic peptide gene in rat cardiac myocytes. secondary prevention of stroke: rationale and design for PROGRESS. J Biol Chem. 1997;272:28050–28056. J Hypertens. 1996;14(suppl 2):S41–S46. 34. Marcus LS, Hart D, Packer M, Yushak M, Medina N, Danziger RS, 24. PROGRESS Collaborative Group. Effects of a perindopril-based blood Heitjan DF, Katz SD. Hemodynamic and renal excretory effects of human pressure lowering regimen on cardiac outcomes among patients with brain natriuretic peptide infusion in patients with congestive heart failure: cerebrovascular disease. Eur Heart J. 2003;24:475–484. a double-blind, placebo-controlled, randomized crossover trial. Circu- 25. Friedewald WT, Levy RI, Fredrickson DS. Estimation of the concen- lation. 1996;94:3184–3189. tration of low-density lipoprotein cholesterol in plasma, without use of the 35. D’Souza SP, Yellon DM, Martin C, Schulz R, Heusch G, Onody A, preparative ultracentrifuge. Clin Chem. 1972;18:499–502. Ferdinandy P, Baxter GF. B-type natriuretic peptide limits infarct size in 26. Karl J, Borgya A, Gallusser A, Huber E, Krueger K, Rollinger W, Schenk rat isolated hearts via KATP channel opening. Am J Physiol. 2003;284: J. Development of a novel, N-terminal-proBNP (NT-proBNP) assay with a low detection limit. Scand J Clin Lab Invest Suppl. 1999;230:177–181. H1592–H1600. 27. Woodward M. Epidemiology: Study Design and Data Analysis. 2nd ed. 36. Pepys MB, Hirschfield GM. C-reactive protein: a critical update. J Clin Boca Raton, Fla: Chapman and Hall/CRC; 2005. Invest. 2003;111:1805–1812. 28. Groenning BA, Raymond I, Hildebrandt PR, Nilsson JC, Baumann M, 37. de Lemos JA, McGuire DK, Drazner MH. B-type natriuretic peptide in Pedersen F. Diagnostic and prognostic evaluation of left ventricular cardiovascular disease. Lancet. 2003;362:316–322. systolic heart failure by plasma N-terminal pro-brain natriuretic peptide 38. Baldoncini R, Desideri G, Bellini C, Valenti M, De Mattia G, Santucci A, concentrations in a large sample of the general population. Heart. 2004; Ferri C. High plasma renin activity is combined with elevated urinary 90:297–303. albumin excretion in essential hypertensive patients. Kidney Int. 1999; 29. Raymond I, Groenning BA, Hildebrandt PR, Nilsson JC, Baumann M, 56:1499–1504. Trawinski J, Pedersen F. The influence of age, sex and other variables on 39. Dzau VJ. Tissue angiotensin and pathobiology of vascular disease: a the plasma level of N-terminal pro brain natriuretic peptide in a large unifying hypothesis. Hypertension. 2001;37:1047–1052. sample of the general population. Heart. 2003;89:745–751. 40. Brown JJ, Davies DL, Lever AF, Robertson JIS. Plasma renin concen- 30. Groenning BA, Nilsson JC, Sondergaard L, Pedersen F, Trawinski J, tration in human hypertension, III: renin in relation to complications of Baumann M, Larsson HB, Hildebrandt PR. Detection of left ventricular hypertension. BMJ. 1966;1:505–508. Vascular Medicine

Statin Treatment After Onset of Sepsis in a Murine Model Improves Survival

Marc W. Merx, MD*; Elisa A. Liehn, MD*; Jürgen Graf, MD; Annette van de Sandt; Maren Schaltenbrand; Jürgen Schrader, MD; Peter Hanrath, MD; Christian Weber, MD

Background—HMG-CoA-reductase inhibitors have been shown to exhibit pronounced immunomodulatory effects independent of lipid lowering. We have recently demonstrated that pretreatment with simvastatin profoundly improves survival in a cecal ligation and perforation (CLP) model of sepsis. Here, we studied whether treatment with simvastatin after onset of sepsis-induced hemodynamic alterations is beneficial and whether prolonged survival can also be achieved with other statins. Methods and Results—Mice were rendered septic by CLP. At 6 hours after sepsis induction, when profound hemodynamic alterations were manifest, treatment with atorvastatin, fluvastatin, pravastatin, simvastatin, or placebo was initiated. Except for fluvastatin (27Ϯ2.3 hours), survival time was extended from 23Ϯ1.2 hours for placebo-treated mice to 37Ϯ3.6 hours for simvastatin-treated, to 40Ϯ4.2 hours for atorvastatin-treated, and to 39Ϯ3.9 hours for pravastatin- treated mice. This profound improvement is based on the preservation of cardiac function and hemodynamic status in statin-treated animals, both of which are severely impaired in untreated CLP mice. As underlying mechanisms, improved susceptibility to endothelial nitric oxide synthase stimulation and reduced endothelial adhesion of leukocytes could be demonstrated after statin treatment. Conclusions—Well established in the treatment of lipid disorders and coronary artery disease, statins harbor the additional and novel potential of effective sepsis treatment. This benefit extends to several but not all statins tested. (Circulation. 2005;112:117-124.) Key Words: hemodynamics Ⅲ inflammation Ⅲ leukocytes Ⅲ sepsis

epsis, defined by consensus conference as “the systemic and microvascular function, and modulation of endothelial Sinflammatory response syndrome that occurs during in- nitric oxide synthase.15 In particular, statins have been found fection,”1 is generally viewed as a disease aggravated by the to reduce the increased endothelial adhesiveness of mono- inappropriate immune response encountered in the affected cytes from hypercholesterolemic individuals or after stimu- individual (see elsewhere for review2,3). Thus, basic research lation with cytokines under flow and static conditions.16,17 and clinical trials have focused on agents capable of blocking This appears to be partly attributable to reduced expression of steps within the inflammatory cascade.4–10 However, despite both monocytic and endothelial adhesion molecules because the multitude of therapeutic approaches evaluated, the only of selective inhibition of the integrin leukocyte function inflammation-modulating substances demonstrated to date to antigen-1 (LFA-1) by affecting Rho GTPases.17 In retrospec- benefit patients with severe sepsis are activated protein C6 tive analysis, patients with bacteremia who were concomi- and low-dose hydrocortisone.11 tantly treated with statins had a reduced overall and attribut- HMG-CoA-reductase inhibitors (statins) such as simvasta- able mortality compared with bacteremic patients not on tin have been shown to exhibit important immunomodulatory statin therapy.18 effects independent of lipid lowering.12,13 In fact, these Recently, we were able to demonstrate that pretreatment so-called pleiotropic effects are now considered to contribute with simvastatin results in profoundly improved survival in significantly to the morbidity and mortality benefit observed the clinically relevant, polimicrobial cecal ligation and per- in patients with coronary heart disease who are treated with foration (CLP) model of sepsis.19 The improvement in sur- statins. Pleiotropic effects have been demonstrated to com- vival observed after simvastatin pretreatment was based on prise antiinflammatory actions,14 improvement of endothelial the complete preservation of cardiac function and hemody-

Received August 23, 2004; revision received January 17, 2005; accepted January 20, 2005. From Medizinische Klinik I (M.W.M., J.G., A. vd S., M.S., P.H., C.W.) and Kardiovaskuläre Molekularbiologie (E.A.L., C.W.), RWTH Aachen, and Institut für Herz- und Kreislaufphysiologie (J.S.), Heinrich-Heine-Universität, Düsseldorf, Germany. *Drs Merx and Liehn contributed equally to this study. Correspondence to Dr med M.W. Merx, Medizinische Klinik I, Universitätsklinikum RWTH Aachen, Pauwelsstraße 30, 52057 Aachen, Germany. E-mail [email protected] © 2005 American Heart Association, Inc. Circulation is available at http://www.circulationaha.org DOI: 10.1161/CIRCULATIONAHA.104.502195 117 118 Circulation July 5, 2005

namic stability observed only in treated mice. As an under- after sepsis induction. The carotid artery was cannulated, and lying mechanism, we demonstrated in septic mice that in- pressure was recorded at baseline and after intraperitoneal injection ␮ creased mononuclear cell adhesiveness, an important of dobutamine (1.5 g/g body weight) and ethylthiourea (ETU; 1 ␮mol/gϭ185.1 ␮g/g body weight). contributor to sepsis pathophysiology, was reversed by statin treatment. In a clinical, prospective, observational cohort Langendorff Setup and Protocol study, Almog et al20 added strong supportive evidence with Preparation of murine hearts explanted from CLP simvastatin- reduced rate of severe sepsis in patients pretreated with a treated, CLP placebo-treated, sham simvastatin-treated, and sham statin for Ն1 month before hospital admission for acute placebo-treated mice matched for age and body weight (10 per Ϯ bacterial infection. group) was initiated 20 2 hours after sepsis induction, and retro- grade perfusion at 100 mm Hg constant pressure with modified In the present study, we approached the question of Krebs-Henseleit buffer was performed essentially as described,22 whether treatment with simvastatin after the onset of hemo- with a commercially available isolated heart apparatus (Hugo Sachs dynamic alterations might improve cardiovascular function in Elektronik). Perfusion pressure, perfusate oxygen concentration, sepsis. Because antiinflammatory properties are known to aortic flow, left ventricular developed pressure (the difference vary between individual statins, we compared the effect of between minimal diastolic pressure and maximal systolic pressure), and heart rate were measured continuously. Hearts were stimulated simvastatin with respect to cardiac output (CO) and sepsis at 600 bpm. Bradykinin and adenosine were applied at a concentra- survival with several other statins. We investigated atorva- tion of 5 and 1 ␮mol, respectively. statin as a powerful repressor of major histocompatibility class II,12 pravastatin as a hydrophilic substance without Echocardiography LFA-1 inhibitory properties,13 and fluvastatin, which can Conscious CLP statin-treated (atorvastatin, simvastatin, pravastatin, induce cyclooxygenase-2.21 To elucidate underlying mecha- or fluvastatin), CLP placebo-treated (10 per group), sham statin- treated (atorvastatin, simvastatin, pravastatin, or fluvastatin), and nisms, stimulation of endothelial nitric oxide synthase sham placebo-treated mice matched for age and body weight (6 per (eNOS) and endothelial adhesion of leukocytes were exam- group) were examined by echocardiography 20Ϯ2 hours after sepsis ined ex vivo and in vitro. induction as described previously.23 To avoid bradycardia, seen occasionally when echocardiography is performed in awake mice, Methods mice were trained to tolerate the handling associated with echocar- diographic examination on 3 to 4 separate occasions per day over a Animals period of 3 days. After this training period, all mice remained calm Male C57 mice were kept according to federal regulations. All during the examination. If image recording of sufficient quality studies were approved by the local ethics committee and the state could not be initiated within 2 minutes, the animal was allowed to animal welfare commission. rest, and images were obtained later to ensure minimal time span of All animals had ad libitum access to standard chow and water. animal handling. Mice were kept under pathogen-free conditions (specific pathogen Using a 15-MHz linear transducer connected to a Sonos 5500 free) and under a 12-hour day/night cycle. All studies were per- (Phillips Medical Systems), we obtained 2D guided M-mode images formed at the same time of day to obviate circadian influences. at the aortic root for offline aortic diameter measurements. Aortic Mice ranged in body weight from 24 to 36 g and in age from 12 flow velocity was measured with pulsed-wave Doppler. and 18 weeks. Mice were divided into subgroups matched for age Aortic flow velocity-time integral (VTI) and aortic root dimension and body weight as outlined later. (AoD) were obtained from the respective frozen images as above, and mean values from 3 to 6 heartbeats were used for further Sepsis Induction analysis. CO was calculated from the following equations: 2ϫ␲ Sepsis was induced as described previously.19 In brief, anesthesia CSAϭ(AoD/2) ,SVϭCSAϫVTI, and COϭSVϫHR, where CSA was induced by intraperitoneal administration of ketamine (60 ␮g/g is aortic cross-sectional area, SV is stroke volume, and HR is heart body weight) and xylazine (10 ␮g/g body weight). The cecum was rate. then subjected to a single “through and through” perforation with a The same procedure was followed for echocardiography involving 20-gauge needle. Sham-operated mice underwent the same proce- dobutamine (1.5 ␮g/g body weight) application. Images were ob- dure except for ligation and perforation of the cecum. Preoperatively tained within 7 and 11 minutes after intraperitoneal injection, a time and postoperatively, all mice had unlimited access to chow and frame shown to represent the plateau pharmacological effect in water. In addition, analgesia (tramadol 20 ␮g/g body weight) was preliminary experiments. applied subcutaneously immediately after the induction of sepsis and every 6 hours thereafter. At the same time points, volume support Cell Isolation, Culture, and Shear Flow (NaCl 0.9%, 0.05 mL/g body weight, prewarmed) was administered Blood samples were taken from weight- and age-matched mice through subcutaneous injection. treated by intraperitoneal injection of atorvastatin, fluvastatin, pra- Six and 18 hours after sepsis induction, age- and body weight– vastatin, or simvastatin as described above (nϭ8 per group) by matched groups of 10 mice each were treated by intraperitoneal cardiac puncture 24Ϯ2 hours after sepsis induction. Mononuclear injection of atorvastatin, fluvastatin, pravastatin, simvastatin (0.2 cells were isolated from mouse blood by gradient centrifugation with ␮g/g body weight; injected volume, 0.02 mL/g body weight), or Lympholyte-Mammalian (Cedarlain). Cell suspensions were washed placebo (carrier only, 0.02 mL/g body weight). Statins were dis- and resuspended (5ϫ105) in HHMC (HBSS, 10 mmol/L HEPES, solved in EtOH at a concentration of 10 mg/mL and diluted with 1 mmol/L Ca2ϩ, 1 mmol/L Mg2ϩ, 0.5% BSA). Viability was Ͼ97% NaCl 0.9% in a ratio of 1:1000 to yield a final concentration of 10 ␮g as determined by trypan blue dye exclusion. statin per 1 mL carrier. The NaCl 0.9%–based carrier solution for the WEHI-274.1 (mouse monocytes, ATCC) were cultured in DMEM placebo group was prepared accordingly to include EtOH only (ie, (ϩ2 mmol/L glutamineϩ0.05 mmol/L2 mercaptoethanolϩ10% FBS) without statin) at a concentration of 1:1000. as described by ATCC. For the experiments, the cells were incubated with 1 ␮mol atorvastatin, fluvastatin, or pravastatin with or without Blood Pressure Measurements the addition of 100 ␮mol mevalonic acid for 30 minutes and 24 hours Anesthesia was induced as above in CLP simvastatin-treated, CLP or were left untreated, washed, and resuspended (1ϫ106) in HHMC. placebo-treated, sham simvastatin-treated, and sham placebo-treated SV40-immortalized murine endothelial cells (kindly provided by mice matched for age and body weight (10 per group) 20Ϯ2 hours Dr H. Hengel, Berlin) were cultured in DMEM (ϩ5% FKS), grown Merx et al Statin Treatment After Onset of Sepsis 119 to confluence in 35-mm Petri dishes, stimulated with tumor necrosis factor (TNF)-␣ (200 U/mL; Sigma), and incubated with 1 ␮mol atorvastatin, fluvastatin, or pravastatin with or without 100 ␮mol mevalonic acid for 30 minutes or 24 hours or left untreated. The cells were washed, and the dishes were assembled as the lower wall in a parallel-wall flow chamber and mounted on the stage of an Olympus IX50 microscope.24,25 Mononuclear cells or WEHI-274.1 monocytes were perfused into the flow chamber at a rate of 1 dyne/cm2 for 7 minutes. After 3 minutes, the adherent cells were counted in multiple high-power fields and recorded with a JVC 3CCD video camera. The investigator responsible for cell counts in the above- mentioned assays was blinded to treatment regimen. Statistical Analysis Mean values with appropriate SEM are reported. All study groups analyzed were initially tested for normality with the Kolmogorov- Smirnov procedure with Lilliefors correction. By this assessment, all groups studied fulfilled the criteria of normal distribution. Groups were then analyzed by ANOVA, followed by Dunnett-T3 post hoc test for comparisons between groups. For cell studies, Figure 1. Survival curves and mean survival times after sepsis Newman-Keuls multiple-comparison test was used as a post hoc test induction by CLP. Mean survival time (large gray symbols) of for comparisons between groups after the initially preformed Ͻ mice treated 6 hours after sepsis induction with statins was pro- ANOVA. A value of P 0.05 was taken to indicate statistical foundly increased vs that for untreated CLP mice (PϽ0.05), significance. All statistical analyses were calculated with SPSS 12.0 except for fluvastatin treatment (Pϭ0.759). No deaths occurred (SPSS Inc). in sham-operated mice. Results after sepsis induction by CLP and evaluated responsiveness Sepsis Survival Is Extended by Statin Treatment to ␤-adrenoceptor stimulation using dobutamine. Preopera- Sepsis was induced through CLP, with sham-operated ani- tively, no differences were observed between any of the mals serving as controls. In preliminary experiments, we groups. At 20 hours after CLP (Figure 2A), CO declined from observed a hyperdynamic followed by a hypodynamic phys- 1.22Ϯ0.03 to 0.82Ϯ0.03 mL · minϪ1 ·gϪ1 in CLP placebo- iological state, similar to hemodynamic alterations in septic treated mice (PϽ0.005; nϭ10) while remaining unaltered in patients. Echocardiography revealed a significant increase in CLP mice treated with atorvastatin (1.26Ϯ0.03 mL · minϪ1 · CO of at least 20% in all septic mice 6 hours after CLP. We Ϫ Ϫ Ϫ g 1 preoperatively versus 1.28Ϯ0.03 mL · min 1 ·g1 20 thus chose to initiate statin therapy 6 hours after sepsis hours after CLP; PϭNS; nϭ10), pravastatin (1.16Ϯ0.03 mL induction by CLP, because at this time point a “clinical” Ϫ Ϫ Ϫ · min 1 ·g 1 preoperatively versus 1.12Ϯ0.02 mL · min 1 · diagnosis based on the changes in hemodynamic status could Ϫ g 1 20 hours after CLP; PϭNS; nϭ10), or simvastatin have been made. Mice were treated with atorvastatin, fluva- Ϫ Ϫ (1.20Ϯ0.03 mL · min 1 ·g 1 preoperatively versus 1.17Ϯ0.02 statin, pravastatin, simvastatin, or placebo 6 and 18 hours Ϫ Ϫ mL · min 1 ·g 1 20 hours after CLP; PϭNS; nϭ10). CLP after CLP. No deaths occurred in sham-operated animals. The fluvastatin-treated animals, however, displayed a decline in degree of sepsis induced in CLP operated mice was equal CO, albeit not to the extent of CLP placebo-treated animals across all groups studied at the time point of treatment (ie, 6 (1.18Ϯ0.03 mL · minϪ1 ·gϪ1 preoperatively versus 0.98Ϯ0.03 hours after sepsis induction) as assessed by the presence of mL · minϪ1 ·gϪ1 20 hours after CLP; PϽ0.01; nϭ10). No conjunctivitis, absence of grooming activities with resulting significant changes between preoperative and postoperative ruffled fur, no oral uptake of food or water, and lethargy. In addition, hyperdynamic cardiovascular states were docu- CO were detected in sham-operated animals. After dobuta- mented at 6 hours after sepsis induction in subsets of mice mine stimulation, CO increased in all groups preoperatively. from all CLP operated groups. Survival curves and mean A similar increase was documented after postoperative dobu- survival times for mice in which CLP was performed in tamine stimulation in sham-operated mice. Placebo-treated ␤ Figure 1 clearly delineate the survival benefit sustained from CLP mice remained refractory to -stimulation (Figure 2B). treatment. Compared with 23Ϯ1.2 hours of mean survival Notably, the responsiveness to dobutamine was restored in time in CLP placebo-treated mice, survival was extended by CLP mice by treatment with atorvastatin, pravastatin, or 70% to 39Ϯ3.9 hours with atorvastatin, by 74% to 40Ϯ4.2 simvastatin (Figure 2B), whereas CLP fluvastatin-treated hours with pravastatin, and by 61% to 37Ϯ3.6 hours with mice showed a dampened response. For all groups in which simvastatin treatment (nϭ10; PϽ0.05 versus placebo). Sur- CO was augmented by dobutamine stimulation, the effect was prisingly, fluvastatin treatment did not result in a significant secondary to a small increase in heart rate and a larger survival time increase (increase of 17% to 27Ϯ2.3 hours; increase in stroke volume (eg, in CLP atorvastatin-treated nϭ10; Pϭ0.759 versus placebo). mice, heart rate increased by 5% and stroke volume by 14% after dobutamine stimulation 20 hours after sepsis induction). Cardiac Function and Hemodynamics Are Subsequently, we performed invasive blood pressure mea- Preserved by Statin Treatment surements in simvastatin- and placebo-treated anesthetized To explore changes in CO due to sepsis, we studied conscious mice (20 hours after CLP) by cannulating the carotid artery. mice by echocardiography preoperatively and 20Ϯ2 hours Again, we applied dobutamine as inotropic stimulus and 120 Circulation July 5, 2005

Figure 2. CO as assessed by echocardiography 20Ϯ2 hours after sepsis induction. A, At 20Ϯ2 hours after sepsis induction, CO decreased significantly in CLP placebo-treated mice (clp; PϽ0.005 vs preoperation; nϭ10) whereas CO remained stable in CLP mice treated 6 hours after sepsis induction with atorva- Figure 3. Coronary flow and blood pressure measurements. A, statin, pravastatstatin, or simvastatin. However, fluvastatin treat- Baseline blood pressure measured 20Ϯ2 hours after sepsis ment was less efficient in maintaining CO of septic mice induction was decreased in CLP placebo-treated animals (CLP; (PϽ0.01 vs preoperation). B, Dobutamine stimulation increased PϽ0.001; nϭ10). Dobutamine stimulation led to an increase in CO in all groups (PϽ0.001 vs unstimulated 20 hours postopera- blood pressure in all study groups (CLPϩsimvastatin, tion for simvastatin, pravastatin, and atorvastatin; nϭ10) except shamϩsimvastatin, and shamϩplacebo dobutamine vs unstimu- in CLP placebo-treated mice (PϭNS vs CLP; nϭ10), with fluva- lated 20 hours postoperation: PϽ0.01; nϭ10) except in CLP statin-treated mice showing a dampened response (PϽ0.005 vs placebo-treated mice (CLPϩplacebo dobutamine vs unstimu- unstimulated 20 hours postoperation; nϭ10). lated 20 hours postoperation: PϭNS; nϭ10). Administration of ETU resulted in increase in blood pressure in all groups (CLPϩplacebo ETU vs unstimulated 20 hours postoperation: ETU, an unspecific inhibitor of NOS, to analyze changes in PϽ0.01; CLPϩsimvastatin, shamϩsimvastatin, and vascular resistance. Mean arterial blood pressure was de- shamϩplacebo ETU vs unstimulated 20 hours postoperation; creased by 19Ϯ7mmHg(PϽ0.01; nϭ10) in CLP placebo- PϽ0.005; nϭ10). B, After eNOS stimulation by bradykinin, coro- nary flow increased most substantially in CLP placebo-treated treated mice compared with sham-operated mice but re- mice (CLPϩplacebo vs shamϩplacebo and vs unstimulated 20 mained unaltered in CLP simvastatin-treated mice (Figure hours postoperation: PϽ0.005; nϭ10), with other groups dis- 3A). Simvastatin treatment had no effect on blood pressure in playing more moderate increase (CLPϩsimvastatin, shamϩsim- ϩ sham-operated mice (Figure 3A). Dobutamine stimulation led vastatin, and sham placebo vs unstimulated 20 hours postop- eration; PϽ0.01; nϭ10). Adenosine stimulation led to substantial to an increase in arterial blood pressure in sham-operated and increase with similar resulting flow in all groups (CLPϩplacebo, in CLP simvastatin-treated mice, whereas CLP placebo- CLPϩsimvastatin, shamϩsimvastatin, and shamϩplacebo vs treated mice displayed no significant change in arterial blood unstimulated 20 hours postoperation: PϽ0.001; nϭ10). Hearts from CLP placebo-treated mice, however, displayed significantly pressure (Figure 3A). NOS blockade with ETU led to an reduced response to adenosine in relation to their baseline flow. Ϸ50% increase in mean arterial pressure. However, the absolute value reached by CLP placebo-treated mice re- Effects on eNOS Stimulation and Coronary Flow mained significantly below the other groups (Figure 3A). Reserve This is in concordance with the lower CO in CLP placebo- Simvastatin treatment preserved contractility of isolated treated mice within the same time frame but may also hearts, which was impaired by 48% in CLP placebo-treated correspond to a reduction in peripheral vascular resistance mice compared with sham-operated animals (PϽ0.005; mediated only partly by NO. nϭ10; the Table). Similarly, left ventricular developed pres- Merx et al Statin Treatment After Onset of Sepsis 121

Contractile Function of Langendorff-Perfused Isolated Hearts

CLPϩPlacebo CLPϩSimvastatin ShamϩSimvastatin ShamϩPlacebo dP/dtmax, mm Hg/s 2654Ϯ150* 4562Ϯ320 4892Ϯ267 5139Ϯ226 LVDP, mm Hg 73.7Ϯ1.9† 97.9Ϯ5.5 104Ϯ5.4 112.5Ϯ4.9 LVDP indicates left ventricular developed pressure. Values are meanϮSEM. CLP placebo-treated mice showed significant contractile impairment (*PϽ0.005, †PϽ0.01; sample size, nϭ10 for all groups), whereas CLP simvastatin-treated mice remained at par with sham-operated mice. Simvastatin treatment alone had no influence on contractile function in sham-operated mice. sure was also decreased solely in CLP placebo-treated mice incubated with atorvastatin or fluvastatin (Figure 5A and 5B). (the Table), as were other parameters of contraction and Because the effect was evident after 30 minutes and was not relaxation (eg, time to peak pressure, relaxation half-time) reversible after addition of mevalonic acid, a short-term and oxygen consumption. exposure appears to be sufficient for inhibition of adhesion, Basal coronary flow was increased by 32% in CLP suggesting a direct interference with LFA-1 activity in placebo-treated mice compared with CLP simvastatin-treated leukocytes or stimulation of eNOS and subsequent antiadhe- mice (Figure 3B). No significant difference was detected in sive effects. This is substantiated by the fact that adhesion of basal coronary flow between the latter and sham-operated monocytes pretreated with pravastatin (known not to interact animals regardless of whether they were treated with simva- with LFA-1) for 30 minutes remained unaltered. Indeed, for statin. With a coronary flow increase of 33%, susceptibility to eNOS stimulation by bradykinin was 3 times as pronounced in CLP placebo-treated mice as in untreated sham-operated mice (Figure 3B). Treatment with simvastatin attenuated the enhanced susceptibility to eNOS stimulation in CLP mice. Maximal coronary flow, measured after adenosine applica- tion, was similar in all groups. However, coronary flow reserve was severely reduced in CLP placebo-treated mice in which adenosine stimulation resulted in a similar degree of flow increase as observed under bradykinin application.

Effects on Leukocyte-Endothelial Adhesion Absolute leukocyte count was reduced in all mice subjected to CLP compared with sham-operated respective controls (Figure 4A). No differences in leukocyte count were observed between placebo- and statin-treated septic mice with the exception that atorvastatin treatment significantly attenuated the reduction in leukocyte counts compared with placebo in CLP mice (Figure 4A). To investigate the effects of sepsis and statin treatment on leukocyte-endothelial interaction, mononuclear cells were isolated from septic and sham-operated mice that had re- ceived placebo, atorvastatin, fluvastatin, pravastatin, or sim- vastatin treatment and subjected to adhesion assays on cyto- kine-stimulated murine endothelial cells under physiological flow conditions. The adhesion of monocytes isolated from placebo-treated CLP mice was significantly increased com- pared with that of placebo-treated sham operated mice (Fig- ure 4B). Treatment with statins significantly reduced the Figure 4. Leukocyte counts and mononuclear cell adhesion to adhesion of leukocytes from sham-operated mice and septic endothelium ex vivo. Leukocyte counts were significantly lower CLP mice (Figure 4B). Thus, sepsis did not lead to increased in all septic animals (A, black bars), with atorvastatin treatment resulting in milder leukopenia (*PϽ0.05 vs placebo and other leukocyte adhesion to endothelium in statin-treated animals, statins). Mononuclear cells were isolated from blood of CLP or indicating a therapeutic inhibition of adhesion (Figure 4B). sham-operated mice treated with placebo (control), atorvastatin, In parallel to our previously published in vitro analysis of pravastatin, simvastatin, or fluvastatin and subjected to adhe- 19 sion assays on activated microvascular endothelium under flow simvastatin, the effect of atorvastatin, fluvastatin, or prava- conditions (1.5 dyne/cm2). Number of firmly attached cells was statin was also evaluated in vitro by treatment of WEHI-274.1 determined after 5 minutes. Arrest of monocytes from CLP mice cells and stimulated SV40-immortalized murine endothelial (black bars) was significantly increased vs those from sham- Ͻ ϭ cells with the statins for 30 minutes or 24 hours. After 30 operated animals (B; white bars; *P 0.05 vs sham; n 8). Treat- ment with above-mentioned statins significantly reduced arrest minutes of treatment, the number of adherent cells was of monocytes in sham-operated (*PϽ0.05 vs shamϩplacebo; reduced when monocytes, endothelial cells, or both were nϭ8) and CLP animals (*PϽ0.05 vs CLPϩplacebo; nϭ8). 122 Circulation July 5, 2005

Figure 6. Monocyte adhesion to endothelium under physiologi- cal flow conditions in vitro after 24 hours of treatment. Similar degree of adhesion inhibition is observed after 24 hours of treat- ment with atorvastatin (A) or fluvastatin (B) (*PϽ0.01; nϭ6) vs 30 minutes of treatment. However, after 24-hour treatment period, inhibition is completely reversible with mevalonate for both statins.

Discussion Following our promising recent study involving pretreatment with simvastatin to improve survival in sepsis,19 the present results give rise to even greater optimism. Significantly improved survival is demonstrated for treatment after the Figure 5. Monocyte adhesion to endothelium under physiologi- onset of sepsis, a situation more akin to clinical reality, cal flow conditions in vitro after 30 minutes of treatment. Mono- especially because a time point with clinically detectable cytes and/or endothelial cells were incubated with atorvastatin sepsis effects was chosen for treatment initiation. Further- (A), fluvastatin (B), or pravastatin (C) 1 ␮mol/L with or without mevalonate coincubation (100 ␮mol/L) for 30 minutes as indi- more, this benefit is not limited to simvastatin but extends to cated and subjected to adhesion assays. Number of monocytes 2 of 3 additional statins currently evaluated, namely atorva- firmly adherent to activated endothelium at 1.5 dyne/cm2 was statin and pravastatin. As with pretreatment, the improvement determined after 5 minutes. For pravastatin, treatment of both monocytes and endothelium is required to achieve adhesion in survival observed here stems from the complete preserva- inhibition, whereas treatment of monocytes or endothelium tion of cardiac function and hemodynamic stability observed alone is sufficient to inhibit adhesion with atorvastatin or fluva- in mice treated 6 hours after sepsis induction with atorvasta- statin. Inhibition of adhesion by above-mentioned statins is not tin, pravastatin, or simvastatin. Increased mononuclear cell reversible after addition of mevalonate (*PϽ0.01; **PϽ0.001). adhesiveness in septic mice, an important contributor to sepsis pathophysiology, is reversed by treatment with all studied statins and represents 1 of the underlying pravastatin, treatment of both monocytes and endothelium mechanisms. was necessary to achieve reduced endothelial adhesion (Fig- Because the CLP model of sepsis closely resembles the ure 5C). pathophysiology of human sepsis,26 it provided us with the After 24 hours, the inhibitory effect of atorvastatin and opportunity to define a time point at which hemodynamic pravastatin was completely reversible (ie, no longer signifi- alterations secondary to sepsis induction became apparent. cantly reduced compared with control) by coincubation with These hemodynamic alterations, namely a hyperdynamic mevalonic acid of monocytes, endothelial cells, or both cell followed by a hypodynamic physiological state, that also types (Figure 6A and 6B). This indicates that interference typically are found in patients affected by sepsis were readily with the mevalonic acid–dependent pathway by atorvastatin detectable in our model by serial echocardiographic studies. and pravastatin (eg, inhibition of Rho-GTPase membrane A hyperdynamic increase in excess of 20% above baseline localization and activity) is predominantly responsible for was observed in all septic mice 6 hours after sepsis induction adhesion effects after 24 hours of treatment with the above by CLP; thus, this time point was chosen for the beginning of statins. statin therapy. Merx et al Statin Treatment After Onset of Sepsis 123

In concordance with our previous studies of simvastatin the absence of pravastatin effects on monocyte adhesion after pretreatment in sepsis,19 impaired CO is secondary to reduced 30 minutes of incubation might be interpreted as a pharma- contractility of septic hearts, and this acute septic cardiomy- cokinetic effect, especially because pravastatin was the only opathy is refractory to catecholamine stimulation, a fact hydrophilic statin studied. Under this hypothesis, we would paralleled in human pathophysiology.27 The reduced arterial expect uptake of pravastatin to be delayed compared with blood pressure of septic mice and the limited rise in blood more lipophilic statins, so treatment of both monocytes and pressure observed after nonselective NOS inhibition suggest endothelial cells would be required to achieve the observed that reduced peripheral vascular resistance is mediated by NO significant inhibition by additive effects. This is also sup- and further vasodilatory agents. The pronounced cardiac ported by the tendency toward reduced adhesion after treat- dysfunction and hypodynamic circulation present severe ment of endothelial cells alone with pravastatin for 30 manifestations of sepsis contributing to the poor outcome and minutes. In addition, fluvastatin did not improve survival short survival time documented in our mouse model of sepsis. despite blocking LFA-1–mediated adhesion, so interference Although not reaching the Ϸ4-fold increase in median sur- with LFA-1 does not seem to be the sole mechanism required vival time observed in simvastatin-pretreated septic mice,19 to provide effective sepsis therapy with statins. our median survival time was increased by Ͼ60% for Incubation of endothelial cells with atorvastatin and fluva- atorvastatin, pravastatin, and simvastatin treatment. Indeed, statin (and as previously reported with simvastatin19) for 30 therapeutic benefit from these statins was so robust that minutes also led to a decline in monocyte adhesion. Because cardiac function and hemodynamic status remained com- this incubation period is too short to modify the cellular sterol pletely unaffected 20 hours after sepsis induction. pool or the function of Rho, mechanisms operating indepen- Because TNF-␣ mimics sepsis by enhancing the endothe- dently of the cholesterol synthesis pathway must exist. One of lial expression of adhesion molecules, chemokines,28 and these mechanisms could be NO release from endothelium cytokines,29 we isolated peripheral blood mononuclear cells observed as early as 8 minutes after exposure to statins.34 from mouse blood and studied their adhesion to TNF-␣– Indeed, the increase in coronary flow observed after brady- stimulated endothelial cells. As expected, adhesion of mono- kinin stimulation of hearts from sham-operated animals after cytes isolated from septic animals was increased. However, treatment with simvastatin is likely mediated by increased monocytes from statin-treated septic animals showed mark- endothelial NO release. Statins have been described to edly attenuated adhesion to endothelium. decrease the phosphorylation of p44 (encoded by ERK1 Because mevalonate is the precursor not only of cholester- gene), which is partially enhanced by the addition of meva- ol but also of many other nonsteroidal isoprenoid products, lonate, while suppressing the phosphorylation p42 (encoded HMG-CoA reductase inhibition might affect several other by ERK 2 gene), which is not restored by the addition of cellular functions. Pleiotropic actions of statins include sup- mevalonate.35 Although this mechanism was described in pression of T-cell responses,30 reduced expression of class II smooth muscle cells, these proteins have also been implicated major histocompatibility complexes on antigen-presenting in endothelial cell responses to shear stress.36 In light of our cells,12 and reduced chemokine synthesis in peripheral blood experiments, these signal elements might be more promi- mononuclear cells.31 The above-mentioned observa- nently involved in endothelial responses to statins than tions12,30,31 could all be reversed by the addition of meval- previously appreciated. onate, indicating that they are causally related to the inhibi- In summary, atorvastatin, pravastatin, and simvastatin, tion of HMG-CoA reductase. However, Weitz-Schmidt et al13 well established in the treatment of lipid disorders and demonstrated that several statins are capable of blocking the coronary artery disease, might have the additional potential of LFA-1–ICAM-1 interaction, providing a mevalonate- and being effective agents in the treatment of sepsis. We believe thus HMG-CoA reductase–independent pathway for antiin- that the promising results presented here, established in a flammatory and immunomodulatory statin actions. In accor- clinical relevant disease model and applying a clinically dance with these findings, in our in vitro experiments, even a feasible therapeutic regimen, in conjunction with abundant very short incubation of monocytes with atorvastatin and safety data resulting from the widespread application of fluvastatin (and simvastatin, as published previously19) but statins and encouraging clinical retrospective18 and observa- not pravastatin (with pravastatin known to differ from ator- tional data,20 warrant prospective phase III trials. vastatin and fluvastatin in that it does not interact with LFA-1) resulted in reduced adhesion of monocytes to endo- Acknowledgment thelium not reversible by mevalonic acid. However, other This work was supported in part by Deutsche Forschungs Gemein- mechanisms such as decreased CD11b expression and re- schaft grant ME 1821 to Dr Merx. We thank S. Becher and I. duced CD-11b–dependent adhesion of monocytes on endo- Moshkova for excellent technical assistance. thelium also demonstrated for simvastatin32 cannot be ex- cluded. Moreover, effects of lovastatin on ␤ -integrin– References 1 1. Bone RC, Balk RA, Cerra FB, Dellinger RP, Fein AM, Knaus WA, 33 mediated adhesion have been demonstrated. Because both Schein RM, Sibbald WJ. Definitions for sepsis and organ failure and of these effects are reversible with mevalonic acid, an guidelines for the use of innovative therapies in sepsis. Chest. 1992;101: involvement of geranylgeranylated proteins has been postu- 1644–1655. lated.33 Such mechanisms are indeed more likely because 2. Hotchkiss RS, Karl IE. The pathophysiology and treatment of sepsis. N Engl J Med. 2003;348:138–150. pravastatin was effective in reducing monocyte adhesiveness 3. Riedemann NC, Guo R-F, Ward PA. Novel strategies for the treatment of after 24 hours in a mevalonate-reversible fashion. In addition, sepsis. Nat Med. 2003;9:517–524. 124 Circulation July 5, 2005

4. Bone RC, Fisher CJ Jr, Clemmer TP, Slotman GJ, Metz CA, Balk RA. A improves survival in a murine model of sepsis. Circulation. 2004;109: controlled clinical trial of high-dose methylprednisolone in the treatment 2560–2565. of severe sepsis and septic shock. N Engl J Med. 1987;317:653–658. 20. Almog Y, Shefer A, Novack V, Maimon N, Barski L, Eizinger M, Friger 5. Annane D. Corticosteroids for septic shock. Crit Care Med. 2001;29: M, Zeller L, Danon A. Prior statin therapy is associated with a decreased S117–S120. rate of severe sepsis. Circulation. 2004;110:880–885. 6. Bernard GR, Vincent JL, Laterre PF, LaRosa SP, Dhainaut JF, Lopez- 21. Chen JC, Huang KC, Wingerd B, Wu WT, Lin WW. HMG-CoA Rodriguez A, Steingrub JS, Garber GE, Helterbrand JD, Ely EW, Fisher reductase inhibitors induce COX-2 gene expression in murine macro- CJ. The recombinant human activated protein: efficacy and safety of phages: role of MAPK cascades and promoter elements for CREB and recombinant human activated protein C for severe sepsis. N Engl J Med. C/EBPbeta. Exp Cell Res. 2004;301:305–319. 2001;344:699–709. 22. Merx MW, Flogel U, Stumpe T, Godecke A, Decking UK, Schrader J. 7. Fisher CJ, Agosti JM, Opal SM, Lowry SF, Balk RA, Sadoff JC, Myoglobin facilitates oxygen diffusion. FASEB J. 2001;15:1077–1079. Abraham E, Schein RMH, Benjamin E, for the Soluble TNF Receptor 23. Heger J, Godecke A, Flogel U, Merx MW, Molojavyi A, Kuhn-Velten WN, Schrader J. Cardiac-specific overexpression of inducible nitric oxide Sepsis Study Group. Treatment of septic shock with the tumor necrosis synthase does not result in severe cardiac dysfunction. Circ Res. 2002; factor receptor: Fc fusion protein. N Engl J Med. 1996;334:1697–1702. 90:93–99. 8. Fisher CJ Jr, Slotman GJ, Opal SM, Pribble JP, Bone RC, Emmanuel G, 24. Ostermann G, Weber KS, Zernecke A, Schroder A, Weber C. JAM- 1 is Ng D, Bloedow DC, Catalano MA. Initial evaluation of human recom- a ligand of the beta (2) integrin LFA-1 involved in transendothelial binant interleukin-1 receptor antagonist in the treatment of sepsis syn- migration of leukocytes. Nat Immunol. 2003;3:151–158. drome: a randomized, open-label, placebo-controlled multicenter trial: the 25. Baltus T, Weber KSC, Johnson Z, Proudfoot AEI, Weber C. Oligomer- IL-1RA Sepsis Syndrome Study Group. Crit Care Med. 1994;22:12–21. ization of RANTES is required for CCR1-mediated arrest but not CCR5- 9. Ziegler EJ, Fisher CJ Jr, Sprung CL, Straube RC, Sadoff JC, Foulke GE, mediated transmigration of leukocytes on inflamed endothelium. Blood. Wortel CH, Fink MP, Dellinger RP, Teng NN. Treatment of gram- 2003;102:1985–1988. negative bacteremia and septic shock with HA-1A human monoclonal 26. Deitch EA. Animal models of sepsis and shock: a review and lessons antibody against endotoxin: a randomized, double-blind, placebo- learned. Shock. 1998;1–11. controlled trial: the HA-1A Sepsis Study Group. N Engl J Med. 1991; 27. Parrillo JE, Parker MM, Natanson C, Suffredini AF, Danner RL, Cunnion 324:429–436. RE, Ognibene FP. Septic shock in humans: advances in the understanding 10. Bernard GR, Wheeler AP, Russel JA, Schein R, Summer WR, Steinberg of pathogenesis, cardiovascular dysfunction, and therapy. Ann Intern KP, Fulkerson WJ, Wright PE, Christmann BW, Dupont WD, Higgins Med. 1990;113:227–242. SB, Swindell BB. The effects of ibuprofen on the physiology and survival 28. Carlos TM, Harlan JM. Leukocyte-endothelial adhesion molecules. of patients with sepsis. N Engl J Med. 1997;912–918. Blood. 1994;84:2068–2101. 11. Annane D, Bellissant E, Bollaert PE, Briegel J, Keh D, Kupfer Y. 29. RajavashisthTB, Andalibi AS, Territo MC, Berliner JA, Navab M, Corticosteroids for severe sepsis and septic shock: a systematic review Fogelman AM, Lusis AJ. Induction of endothelial cell expression of and meta-analysis. BMJ. 2004;329:480. granulocyte and macrophage colony stimulation factors by modified 12. Kwak B, Mulhaupt F, Myit S, Mach F. Statins as a newly recognized type low-density lipoprotein. Nature. 1990;344:254–257. of immunomodulator. Nat Med. 2000;6:1399–1402. 30. Kurakata S, Kada M, Shimada Y, Komai T, Nomoto K. Effects of 13. Weitz-Schmidt G, Welzenbach K, Brinkmann V, Kamata T, Kallen J, different inhibitors of 3-hydroxy-3-methylglutaryl coenzyme A Bruns C, Cottens S, Takada Y, Hommel U. Statins selectively inhibit (HMG-CoA) reductase, pravastatin sodium and simvastatin, on sterol leukocyte function antigen-1 by binding to a novel regulatory integrin synthesis and immunological functions in human lymphocytes in vitro. site. Nat Med. 2001;7:687–692. Immunopharmacology. 1996;34:51–61. 14. Pruefer D, Makowski J, Schnell M, Buerke U, Dahm M, Oelert H, 31. Romano M, Diomede L, Sironi M, Massimiliano L, Sottocorno M, Polen- Sibelius U, Grandel U, Grimminger F, Seeger W, Meyer J, Darius H, tarutti N, Guglielmotti A, Albani D, Bruno A, Fruscella P, Salmona M, Buerke M. Simvastatin inhibits inflammatory properties of Staphylo- Vecchi A, Pinza M, Mantovani A. Inhibition of monocyte chemotactic protein-1 synthesis by statins. Lab Invest. 2000;80:1095–1100. coccus aureus alpha toxin. Circulation. 2003;15:2104–2110. 32. Weber C, Erl W, Weber KSC, Weber PC. HMG-CoA reductase inhibitors 15. Laufs U, Liao JK. Post-transcriptional regulation of endothelial nitric decrease CD11b expression and CD11b-dependent adhesion of oxide synthase mRNA stability by Rho GTPase. J Biol Chem. 1998;273: monocytes to endothelium and reduce increased adhesiveness of 24266–24271. monocytes isolated from patients with hypercholesterolemia. J Am Coll 16. Serrano J, Yoshida VM, Venturinelli ML, D’Amico E, Monteiro HP, Cardiol. 1997;30:1212–1217. Ramires JA, da Luz PL. Effect of simvastatin on monocyte adhesion 33. Li X, Liu L, Tupper JC, Bannerman DD, Winn RK, Sebti SM, Hamilton molecule expression in patients with hypercholesterolemia. Atheroscle- AD, Harlan JM. Inhibition of protein geranylgeranylation and rosis. 2001;157:505–512. RhoA/RhoA kinase pathway induces apoptosis in human endothelial 17. Yoshida M, Sawada T, Ishii H, Gerszten RE, Rosenzweig A, Gimbrone cells. J Biol Chem. 2002;277:15309–15316. MA Jr, Yasukochi Y, Numano F. HMG-CoA reductase inhibitor mod- 34. Kaesemeyer WH, Caldwell RB, Huang J, Caldwell RW. Pravastatin ulates monocyte-endothelial cell interaction under physiological flow sodium activates endothelial nitric oxide synthase independent of its conditions in vitro: involvement of Rho GTPase-dependent mechanism. cholesterol-lowering actions. J Am Coll Cardiol. 1999;33:234–241. Arterioscler Thromb Vasc Biol. 2001;21:1165–1171. 35. Sindermann JR, Schmidt A, Breithardt G, Buddecke E. Lovastatin 18. Liappis AP, Kan VL, Rochester CG, Simon GL. The effect of statins on controls signal transduction in vascular smooth muscle cells by modu- mortality in patients with bacteremia. Clin Infect Dis. 2001;33: lating phosphorylation levels of mevalonate-independent pathway. Basic 1352–1357. Res Cardiol. 2001;96:283–289. 19. Merx MW, Liehn EA, Janssens U, Lütticken R, Schrader J, Hanrath P, 36. Tseng H, Peterson TE, Berk BC. Fluid shear stress stimulates mitogen- Weber C. HMG-CoA reductase inhibitor simvastatin profoundly activated protein kinase in endothelial cells. Circ Res. 1995;77:869–878. Contemporary Reviews in Cardiovascular Medicine

Valvular Heart Disease Aortic Regurgitation

Raffi Bekeredjian, MD; Paul A. Grayburn, MD

Abstract—Aortic regurgitation (AR) is characterized by diastolic reflux of blood from the aorta into the left ventricle (LV). Acute AR typically causes severe pulmonary edema and hypotension and is a surgical emergency. Chronic severe AR causes combined LV volume and pressure overload. It is accompanied by systolic hypertension and wide pulse pressure, which account for peripheral physical findings, such as bounding pulses. The afterload excess caused by systolic hypertension leads to progressive LV dilation and systolic dysfunction. The most important diagnostic test for AR is echocardiography. It provides the ability to determine the cause of AR and to assess the severity of AR and its effect on LV size, function, and hemodynamics. Many patients with chronic severe AR may remain clinically compensated for years with normal LV function and no symptoms. These patients do not require surgery but can be followed carefully for the onset of symptoms or LV dilation/dysfunction. Surgery should be considered before the LV ejection fraction falls below 55% or the LV end-diastolic dimension reaches 55 mm. Symptomatic patients should undergo surgery unless there are excessive comorbidities or other contraindications. The primary role of medical therapy with vasodilators is to delay the need for surgery in asymptomatic patients with normal LV function or to treat patients in whom surgery is not an option. The goal of vasodilator therapy is to achieve a significant decrease in systolic arterial pressure. Future therapies may focus on molecular mechanisms to prevent adverse LV remodeling and fibrosis. (Circulation. 2005;112: 125-134.) Key Words: aorta Ⅲ echocardiography Ⅲ valves Ⅲ ventricles

ortic regurgitation (AR) is characterized by diastolic associated with modest increases in aortic root size but not Areflux of blood from the aorta into the left ventricle (LV) AR when age is included in the model.3,4 The Strong Heart due to malcoaptation of the aortic cusps. Its clinical presen- Study5 showed an overall prevalence of AR of 10% in a tation is variable and depends on a complex interplay of a Native American population. Most cases were of mild sever- number of factors, including acuity of onset, aortic and LV ity; age and aortic root diameter, but not gender, were compliance, hemodynamic conditions, and severity of the independent predictors of AR in this study. lesion. Although chronic AR is generally well tolerated for many years, acute AR may lead to rapid cardiac decompen- Etiology sation and, if untreated, to early death.1 This review focuses AR results from malcoaptation of the aortic leaflets due to on the clinical manifestations of AR, evaluation of its severity abnormalities of the aortic leaflets, their supporting structures and hemodynamic consequences, and its treatment. (aortic root and annulus), or both. Diseases that primarily affect the leaflets include bicuspid aortic valve and other Prevalence congenital abnormalities, atherosclerotic degeneration, infec- The prevalence of chronic AR and incidence of acute AR are tive endocarditis, rheumatic heart disease, connective tissue not precisely known. Singh et al2 reported the prevalence of or inflammatory diseases, antiphospholipid syndrome, and chronic AR detected by color Doppler echocardiography in a use of anorectic drugs.6–12 The leaflets can also be affected by large unselected adult population (the Framingham Offspring trauma, due either to chest wall or deceleration injury, or a jet Study). The overall prevalence AR in men was 13% and in lesion, due to dynamic or fixed subaortic stenosis. Diseases women 8.5%. However, most of the AR in this population that primarily affect the annulus or aortic root include was trace or mild in severity; moderate or severe AR was rare idiopathic aortic root dilation, aortoannular ectasia, Marfan (Table 1). Multiple logistic regression analysis revealed age syndrome, Ehlers-Danlos syndrome, osteogenesis imperfecta, and male gender to be predictors of AR. Interestingly, aortic dissection, syphilitic aortitis, or various connective hypertension did not predict AR on multivariate analysis, tissue diseases.13 A bicuspid aortic valve is commonly asso- confirming results of earlier studies that hypertension is ciated with dilation of the aortic root in addition to the

From the Department of Cardiology, University of Heidelberg, Heidelberg, Germany (R.B.), and Department of Internal Medicine, Cardiology Section, Baylor University Medical Center, Dallas, Tex (P.A.G.). Correspondence to Paul A. Grayburn, MD, Baylor Heart and Vascular Institute, 621 N Hall St, Suite H030, Dallas, TX 75226. E-mail [email protected] © 2005 American Heart Association, Inc. Circulation is available at http://www.circulationaha.org DOI: 10.1161/CIRCULATIONAHA.104.488825 125 126 Circulation July 5, 2005

TABLE 1. Prevalence of AR in the Framingham In early, compensated severe AR, the LV adapts to the Offspring Study volume overload by eccentric hypertrophy, in which sarco- Age, y meres are laid down in series and myofibers are elongat- ed.25,26 Eccentric hypertrophy preserves LV diastolic compli- 26–39 40–49 50–59 60–69 70–83 ance, such that LV filling pressures remain normal or mildly Men (nϭ91) (n-352) (nϭ433) (nϭ359) (nϭ91) increased despite a large regurgitant volume. In addition, None 96.7% 95.4% 91.1% 74.3% 75.6% eccentric hypertrophy increases LV mass, such that the LV Trace 3.3% 2.9% 4.7% 13.0% 10.0% volume/mass ratio is normal, and LV ejection fraction Mild 0% 1.4% 3.7% 12.1% 12.2% (LVEF) is maintained by increased preload. The slope of the LV pressure volume relationship (elastance or E ), a load- ՆModerate 0% 0.3% 0.5% 0.6% 2.2% max independent measure of myocardial function, is normal.27 ϭ ϭ ϭ ϭ ϭ Women (n 93) (n 451) (n 515) (n 390) (n 90) Over time, progressive LV dilation and systolic hypertension None 98.9% 96.6% 92.4% 86.9% 73.0% increase wall stress and the volume/mass ratio. As this Trace 1.1% 2.7% 5.5% 6.3% 10.1% occurs, there is a phase during which LVEF is still normal,

Mild 0% 0.7% 1.9% 6.0% 14.6% but Emax decreases, indicating early myocardial dysfunction ՆModerate 0% 0% 0.2% 0.8% 2.3% that is largely masked by increased preload. At this stage, 27 By multivariate analysis, only age and gender predicted AR prevalence. LVEF still increases after successful valve replacement. Adapted from Singh et al.2 Eventually, the increase in wall stress leads to overt LV systolic dysfunction, manifested by a decline in LVEF and severely reduced E . In chronic severe AR, end-systolic congenital leaflet abnormality.14,15 Ankylosing spondylitis max wall stress can be as high as in aortic stenosis.28 Marked LV can cause disease of both the leaflets and the aortic root. hypertrophy (cor bovinum) develops with increased LV Finally, chronic severe AR of any cause can lead to progres- volume and mass and spherical geometry.29 sive enlargement of the aortic root and further worsening of In decompensated severe AR, LV systolic dysfunction is AR over time. accompanied by decreased LV diastolic compliance as a Acute AR is most commonly caused by bacterial endocar- result of hypertrophy and fibrosis, leading to high filling ditis, aortic dissection, or blunt chest trauma.16–18 Other less pressures and heart failure symptoms. Exertional dyspnea is common causes of acute AR include nonbacterial endocardi- the most common manifestation, but angina can also occur 19 20 tis, laceration of the aorta, and complications of invasive because of a reduction in coronary flow reserve with predom- procedures such as aortic valvuloplasty and percutaneous inantly systolic coronary flow.30,31 In experimental animals, 21 balloon dilatation of aortic coarctation. Fortunately, acute the transition from a compliant (chronic compensated AR) to AR, which has a poor prognosis, is rare. a stiff (decompensated AR) LV chamber appears to involve The prevalence of chronic AR is much higher, and its upregulation of several cardiac fibroblast genes.32,33 Acute causes are different. In a prospective study of 104 patients AR leads to rapid decompensation due to low forward cardiac with chronic AR, 35% had unknown causes, 26% idiopathic output and pulmonary congestion. There is not time for root dilation, 13% congenital abnormalities, 12% rheumatic compensatory LV dilation to occur, and severe hypotension heart disease, 10% infective endocarditis, and 7% degenera- occurs rather than the systolic hypertension that is character- tive valve disease.22 A different study of 246 patients dem- istic of chronic severe AR. The different stages of AR are onstrated 40% degenerative causes, 28% congenital causes, shown in Figure 1. 19% aortic root enlargement, 6% rheumatic causes, 3% aortitis, and 3% endocarditis.23 These numbers only represent Physical Findings a rough estimate because demographic changes in population A variety of physical signs have been described for AR. On age, geographic location, and socioeconomic status may auscultation, a high-frequency, decrescendo diastolic murmur affect prevalence of different diseases, such as rheumatic is typically heard over the third or fourth intercostal space at heart disease. the left sternal border. In some patients, a mid and late diastolic apical rumble (Austin-Flint murmur) is heard, pos- Pathophysiology sibly because of vibration of the anterior mitral leaflet as it is Chronic severe AR imposes a combined volume and pressure struck by a posteriorly directed AR jet.34 A systolic ejection overload on the LV. The volume overload is a consequence of murmur due to high ejection volumes should be present in the regurgitant volume itself and is therefore directly related significant AR. Further findings on auscultation are soft or to the severity of the leak. Thus, whereas mild AR produces absent second heart sound and presence of a third heart only minimal volume overload, severe AR can produce sound. In acute AR, the diastolic murmur may be absent massive LV volume overload and progressive chamber dila- because of rapid equilibration of aortic and LV diastolic tion. The pressure overload results from systolic hyperten- pressures. The only clue may be an absent second heart sound sion, which occurs as a result of increased total aortic stroke in the setting of severe hypotension and pulmonary edema. volume, because both the regurgitant volume and the forward In chronic severe AR, the elevated stroke volume and stroke volume are ejected into the aorta during systole.24 systolic hypertension produce a variety of interesting physical Systolic hypertension can contribute to a cycle of progressive findings. Among these are the bounding carotid pulse (Cor- dilation of the aortic root and subsequent worsening of AR. rigan’s pulse), head bobbing (de Musset’s sign), pulsation of Bekeredjian and Grayburn Aortic Regurgitation 127

Figure 1. Different stages of AR. Top left, In mild AR, LV size, function, and hemodynamics are nor- mal. Top right, In acute severe AR, there is equili- bration of aortic and LV pressures (80/40 mm Hg in this example). Left atrial pressure is elevated, leading to pulmonary edema. Bottom left, In chronic severe, compensated AR, the LV may begin to dilate, but LVEF is often maintained in the normal range by increased preload. There is sys- tolic arterial hypertension and a wide pulse pres- sure. However, LV filling pressures are normal or only slightly elevated, such that dyspnea is absent. Bottom right, In decompensated chronic severe AR, the LV is dilated and hypertrophied, and LV function is often depressed as a result of afterload excess. Forward output is decreased, leading to fatigue and other low-output symptoms. Fibrosis and hypertrophy decrease LV compliance, leading to increased filling pressures and dyspnea.

the uvula (Muller’s sign), and pistol shot sounds over the functional class, comorbidity index, atrial fibrillation, LV femoral artery with compression (Traube’s sign). During end-systolic diameter, and ejection fraction (EF). As shown compression with a glass slide, capillary pulsations can be in Figure 2, patients with greater NYHA functional class or seen on the fingernail (Quincke’s sign). LV end-diastolic diameters Ͼ25 mm/m2 had an adverse prognosis. Taken together, these studies indicate that asymp- Progression and Natural History tomatic patients with normal LV function generally have a Progression of AR involves a complicated interaction of favorable prognosis and indicate that decline in LVEF with several variables, including AR severity, aortic root pathol- exercise or serial follow-up may identify patients who will ogy, and the adaptive response of the LV. AR severity may worsen as a result of progressive leaflet pathology and/or further dilation of the aortic root. In addition, LV dilation occurs gradually and progressively, depending on the severity of AR, hemodynamic factors, and the degree of eccentric hypertrophy and remodeling, which may vary from patient to patient and may be related to genetic factors. Reimold et al35 have shown that quantitative measures of AR severity by echocardiography worsen over time. Padial et al36 showed that patients with more rapidly progressive increases in aortic root size also tend to have significant worsening of AR severity and LV dilation. A few studies have investigated the mortality and morbid- ity of chronic AR if left without surgical treatment. Bonow et al37 studied 104 asymptomatic patients with severe AR and normal LVEF. The rate of attrition (defined as death, symp- toms, or asymptomatic LV dysfunction) was Ͻ5%/y over 11-year follow-up. The rate of sudden death was only 0.4%/y. At 11 years, 58% of patients remained asymptomatic with normal LV systolic function. Borer et al22 found similar results in 104 different patients monitored for a mean of 7.3 years. The rate of attrition was 6.2%/y and was predicted by the change in LVEF or LVEF adjusted for wall stress from rest to exercise. At 5 years, 75% of patients remained free of Figure 2. Top, Survival of patients with chronic severe AR by 23 symptoms (NYHA class). Survival in asymptomatic patients (class I) death, symptoms, or LV dysfunction. Dujardin et al inves- is no different than expected (Pϭ0.38). However, patients with tigated the fate of 246 patients with moderately severe or class II symptoms have a significantly worse survival (Pϭ0.02), and severe AR with a mean follow-up time of 7 years. Unlike the patients with class II to IV symptoms have a markedly worse sur- Ͻ 2 prior studies, these patients were not all asymptomatic with vival (P 0.001). Bottom, Survival for patients stratified by LV end- systolic dimension (LVESD). Patients with LV end-systolic dimen- normal LV systolic function. The 10-year mortality rate was sion Ͻ25 mm/m2 have a markedly worse survival (PϽ0.001). 34%, with independent predictors of survival being age, Adapted from Dujardin et al.23 128 Circulation July 5, 2005

TABLE 2. Natural History of AR Anatomy of the Aortic Root and Leaflets Echocardiographic evaluation of the anatomy of the aortic Asymptomatic patients with normal LV systolic function root, annulus, and leaflets is important in defining the Progression to symptoms and/or LV dysfunction Ͻ6%/y etiology and severity of AR. As noted earlier, disorders such Ͻ Progression to asymptomatic LV dysfunction 3.5%/y as aortic root dilation, bicuspid aortic valve, endocarditis, Sudden death Ͻ0.2%/y degenerative aortic valve disease, and dissection of the Asymptomatic patients with LV systolic dysfunction ascending aorta have different implications with regard to Progression to symptoms Ͼ25%/y treatment. Although it is common to see mild AR with a Symptomatic patients structurally normal aortic valve and supporting apparatus, it is Mortality rate Ͼ10%/y rare for severe AR to occur without major lesions of the Adapted with permission from ACC/AHA guidelines.38 leaflets or the aortic root. Figure 3 shows echocardiographic examples of different causes of AR. require surgical intervention. Patients with even moderate Color Flow Mapping symptoms or evidence of LV dilation are at higher risk and Doppler color flow mapping is widely used to identify the should be considered for early intervention. The American presence of AR and estimate its severity. In general, color College of Cardiology/American Heart Association Guide- flow jets are composed of 3 distinct segments. The proximal lines for Management of Patients with Valvular Heart Dis- flow convergence zone is the area of flow acceleration into ease have nicely summarized the natural history of chronic the orifice, the vena contracta is the narrowest and highest- AR (Table 2).38 velocity region of the jet at or just downstream from the Echocardiography orifice, and the jet itself occurs distal to the orifice in the LV The most important diagnostic test for evaluation of AR is cavity in the case of AR. Measurement of jet area or echocardiography. It allows (1) assessment of the anatomy of penetration into the LV cavity is not accurate in assessing AR the aortic leaflets and the aortic root, (2) detection of the severity. Perry et al40 compared the ratio of AR jet width to presence and severity of AR, and (3) characterization of LV LV outflow tract (LVOT) width in a parasternal long-axis size and function. The American Society of Echocardiogra- view to angiography. A jet width/LVOT width Ͻ25% is phy guidelines for quantification of valvular regurgitation specific for mild AR, whereas a jet width/LVOT width ratio emphasize the need to integrate all of this information to Ͼ65% is specific for severe AR (Figure 4). This works best properly evaluate patients with AR.39 when the regurgitant orifice is relatively round in shape.

Figure 3. Echocardiographic images from different patients with AR due to different pathologies. Top left, Parasternal long-axis view showing a dilated aortic root (arrows) due to aortoannular ectasia. Top right, Parasternal long-axis view showing large, mobile vegeta- tion (arrow) on the aortic valve in a patient with infective endocarditis. Bottom left, Parasternal short-axis view showing a bicuspid aor- tic valve with characteristic elliptical opening (arrow). Bottom right, Parasternal long-axis view of a patient with acute AR due to aortic dissection. Intimal flap is shown by arrows. Bekeredjian and Grayburn Aortic Regurgitation 129

Figure 4. Color flow images from parasternal long-axis views in patients with mild (left) and severe (right) AR. Jet width is Ͻ25% of LVOT width in mild AR. This jet is eccentric; width is mea- sured at the origin of the jet adjacent to the leaflets. In severe AR, jet width is usually Ͼ50% of LVOT width. A jet width/LVOT width Ͼ65% (as in this patient) is specific for severe AR.39

When it is elliptical, as in bicuspid aortic valves, this ratio can proximal isovelocity surface area (PISA) region that is as lead to underestimation of AR severity.41 The short-axis view large as possible. The surface area of the PISA region is 2␲r2, is helpful in identifying such cases. where r is the radius from the alias line to the orifice. Peak regurgitant flow is obtained by multiplying this value by the Vena Contracta Imaging aliasing velocity, and effective regurgitant orifice area is the Vena contracta is defined as the narrowest central flow region peak regurgitant flow divided by the peak velocity obtained of a jet. In AR, it can be measured in a parasternal long-axis by continuous wave Doppler. The PISA method has been or short-axis view in a color Doppler mode. Animal studies shown to work in AR but is less accurate in eccentric jets or have shown good correlation of vena contracta width and aortic root dilation.46 severity of AR.42 Clinical studies have confirmed the useful- 43,44 ness of this measurement for judging AR severity. Quantitative Doppler Flow Measurements Tribouilloy et al43 demonstrated in a study with 79 patients AR volume and fraction can be calculated by comparing flow that a vena contracta width of Ն6 mm correlates well with at the aortic level (total stroke volume) with that at the mitral severe AR, having a sensitivity of 95% and a specificity of valve level (forward stroke volume).39 The total stroke 90%. Conversely, a vena contracta width Ͻ0.3 cm is specific volume is generally measured in the LVOT by multiplying for mild AR. Willett et al44 compared vena contracta width by the LVOT area times the velocity time integral of pulsed transesophageal echocardiography to simultaneous aortic Doppler LVOT flow. The mitral stroke volume is measured flow probe measurements of regurgitant volume and fraction in similar fashion but is more prone to error because of in an intraoperative setting. Figure 5 shows an example of the difficulty in accurately measuring the mitral annulus and vena contracta in a patient with moderate AR. placing the pulsed Doppler sample volume at the level of the Jet Eccentricity annulus. Effective regurgitant orifice area can be calculated Eccentricity of the regurgitant jet may contribute to the by dividing the regurgitant volume by the velocity time understanding of mechanisms of aortic valve dysfunction.45 A integral of the AR jet obtained from continuous wave centrally directed jet entrains fluid on all sides and generally Doppler. This method, although tedious, provides quantita- appears larger and wider than eccentric jets directed anteri- tive measures of AR severity. The cut points for AR severity orly toward the ventricular septum or posteriorly toward the measured by regurgitant volume, regurgitant fraction, and anterior mitral leaflet. This should be taken into account when effective regurgitant orifice area are shown in Table 3.39 AR severity is graded. Supportive Findings Proximal Isovelocity Surface Area Method A number of echocardiographic findings provide supporting It is less common to identify a clear proximal flow conver- evidence for AR severity. By M-mode echocardiography, gence in AR compared with MR. However, when it is early mitral valve closure indicates increased LV filling present, the Nyquist velocity should be shifted toward the pressures and is often present in severe AR, unless masked by direction of the jet to produce a clearly visible, round tachycardia.47 The continuous wave Doppler spectral signal

Figure 5. Vena contracta images of AR jet by transesophageal echocardiography in long-axis (left) and short-axis (right) views. The vena contracta is seen as the narrow- est part of the jet as it emerges from the regurgitant orifice. The short-axis view is difficult to orient precisely in the plane of the vena contracta but is useful in deter- mining whether the jet is central and round (in which case the long-axis vena con- tracta accurately describes AR severity) or markedly elliptical, as in bicuspid aortic valves (in which the long-axis vena con- tracta may underestimate AR severity). Reprinted from Willett et al,44 copyright 2001, with permission from the American College of Cardiology Foundation. 130 Circulation July 5, 2005

TABLE 3. Application of Specific and Supportive Signs, and Quantitative Parameters in the Grading of Aortic Regurgitation Severity

Mild Moderate Severe Specific signs for AR severity Central jet, width Ͻ25% of LVOT† Signs of AR Ͼmild present Central jet, width Ն65% of LVOT† Vena contracta Ͻ0.3 cm† but no criteria for severe AR Vena contracta Ͼ0.6 cm† No or brief early diastolic flow reversal in descending aorta Supportive signs Pressure half-time Ͼ500 ms Intermediate values Pressure half-time Ͻ200 ms Normal LV size* Holodiastolic aortic flow reversal in descending aorta Moderate or greater LV enlargement‡ Quantitative parameters§ RVol, mL/beat Ͻ30 30–44 45–59 Ն60 RF, % Ͻ30 30–39 40–49 Ն50 EROA, cm2 Ͻ0.10 0.10–0.19 0.20–0.29 Ն0.30 *LV size applied only to chronic lesions. †At a Nyquist of 50–60 cm/s. ‡In the absence of other etiologies of LV dilatation. §Quantitative parameters can help sub-classify the moderate regurgitation group into mild-to-moderate and moderate-to-severe regurgitation as shown. AR indicates aortic regurgitation; EROA, effective regurgitant orifice area; LV, left ventricle; LVOT, left ventricular outflow tract; RVol, regurgitant volume; and RF, regurgitant fraction. Table reprinted with permission of the American Society of Echocardiography from Zoghbi et al,39 Table 6. of the AR jet provides clues to the severity of the leak. With the pressure half-time. Conversely, moderate AR into a stiff severe AR, diastolic pressure will decrease rapidly in the LV, especially in the acute or subacute setting, may signifi- aorta, thus leading to a shorter pressure half-time or more cantly shorten pressure half-time. Thus, pressure half-time rapid deceleration slope (Figure 6).48,49 As a general rule, an and early mitral closure should be considered markers of the AR pressure half-time Ͻ200 ms indicates severe AR, hemodynamic consequences of AR rather than the regurgitant whereas a pressure half-time Ͼ500 ms suggests mild AR.39 volume itself. A complete echocardiographic study provides LV end-diastolic pressure can be calculated as the diastolic measurements of the severity of the leak (regurgitant volume, blood pressure minus the end-diastolic pressure gradient fraction, and orifice area) and the hemodynamic effects of AR calculated from the modified Bernoulli equation (Figure 6).48 (LV volumes, pressure half-time, LV end-diastolic pressure). Importantly, the rate of deceleration of AR velocities simply Another important supportive sign of severe AR is diastol- reflects the rate of equilibration of the diastolic pressure ic flow reversal in the descending aorta. Although brief early gradient between the aorta and LV. In chronic compensated diastolic flow reversal is often seen in normal subjects, AR, a large regurgitant volume may not significantly shorten holodiastolic flow reversal usually indicates at least moderate

Figure 6. Continuous wave Doppler of AR jet in a patient with moderate AR and a long-standing history of hypertension. The slope of velocity deceleration is fairly steep, with a pressure half-time (PHT) of 315 ms. LV end-diastolic pressure (LVEDP) can be calculated by converting end-diastolic velocity (measured at the R wave peak) to pressure gradient by 4V2 and subtracting this value from the dia- stolic blood pressure (BP). Patients with chronic compensated AR may have a relatively flat slope, reflecting a compliant LV with a normal or only slightly elevated LVEDP. Bekeredjian and Grayburn Aortic Regurgitation 131

AR.50 Diastolic flow reversal in the descending aorta is best sedentary lifestyle. In such patients, exercise testing may be measured with pulsed-wave Doppler from a suprasternal very useful in eliciting symptoms or determining functional probe position. capacity. Some studies have suggested that an exercise- induced decrease in LVEF is a predictor of poor outcome that LV Size and Geometry warrants surgery.22,55–57 However, most of these studies Echocardiography is useful in measuring LV dimensions, included patients who already had symptoms, LV dilation, or volumes, and LVEF, all of which are important determinants decreased resting LVEF. Thus, it is not clear that exercise of the need for surgery in chronic severe AR. Serial progres- LVEF is helpful in determining the need for surgery in sion of LV dilation predicts the need for surgery.37 Because asymptomatic patients with normal LV size and function.38 LV chamber dilation and systolic dysfunction can occur from other causes (ie, cardiomyopathy), it is important to establish Surgical Treatment a link between severity of AR and LV dysfunction. This can In acute AR, immediate surgical intervention is necessary be difficult at times and underscores the need for accurate, because the acute volume overload results in life-threatening careful quantification of AR severity. Repeated echocardiog- hypotension and pulmonary edema.1 Vasodilator therapy with raphy to assess progression of LV dilation and severity of AR sodium nitroprusside may stabilize the patient during trans- is recommended every 2 to 3 years in stable asymptomatic port to the operating department. Aortic balloon counterpul- patients with normal LV size and function.38 In asymptomatic sation is contraindicated because it worsens AR. ␤-Blockers patients with LV dilation, more frequent echocardiography should be avoided in acute AR because they prolong diastole (every 6 to 12 months) is indicated.38 and may worsen AR. Atrial pacing to increase heart rate might be of theoretical benefit58,59; however, this does not Cardiac Catheterization have an established role in clinical practice. Several studies Even if echocardiography accurately identifies severity of AR have demonstrated that emergency aortic valve replacement and degree of LV function, catheterization may be needed to can be performed with low operative mortality and good evaluate coronary anatomy in patients requiring surgical long-term results in acute AR.60–62 intervention. As a general rule, men aged Ͼ35 years, pre- In contrast to acute AR, patients with chronic AR may be menopausal women aged Ͼ35 years with risk factors for asymptomatic for many years or even their entire life. coronary artery disease, or postmenopausal women should Therefore, the critical issue is to determine if and when undergo preoperative coronary arteriography.38 Supravalvu- surgical intervention is required. There are no randomized lar aortography provides a semiquantitative approach to grade controlled trials to guide surgical decision making. However, AR during heart catheterization. Visual grading of AR reasonable guidelines have been proposed on the basis of the severity is based on the amount of contrast that appears in the aforementioned natural history of AR, retrospective studies, LV after aortography. Mild or 1ϩ AR is contrast appearing in and expert opinion.38 The operative mortality for isolated the LV but clearing with each beat. Moderate or 2ϩ AR is aortic valve replacement is Ϸ4%.63–65 It is higher with faint opacification of the entire LV over several cardiac concomitant aortic root replacement or coronary bypass cycles. Moderately severe or 3ϩ AR is opacification of the surgery or if there are substantial comorbidities, including entire LV with the same intensity as in the aorta. Severe or advanced age. As shown in Table 2, the death rate for 4ϩ AR is opacification of the entire LV on the first heart beat asymptomatic patients with normal LV size and function is with an intensity higher than in the aorta. Unfortunately, this Ͻ0.2%/y. Therefore, asymptomatic patients with normal LV method is subjective, depends on the amount of contrast size and systolic function do not require surgery but should be injected and the size of the LV, and correlates poorly with monitored carefully for development of symptoms, LV dys- regurgitant volume, particularly in patients with dilated function, or progressive LV dilation. In contrast, symptomatic LVs.51 patients with chronic severe AR have a mortality Ͼ10%/y Cine MRI can also be used to detect and quantify AR.52–54 and therefore should undergo surgery unless there are exces- Phase velocity encoding is used to calculate forward stroke sive comorbidities or a condition with a known short life volume through the aortic valve. Total LV stroke volume is expectancy. The more difficult issue is when to operate on determined from LV end-diastolic and end-systolic volumes, asymptomatic patients to prevent irreversible LV dysfunction which are measured by summing the volumes of a stack of from occurring. Outcomes are better in patients with an slices of known thickness (typically 8 to 10 mm) through the LVEF Ͼ55% or an end-systolic LV diameter Ͻ55 mm (or LV from base to apex. The difference between aortic and LV Ͻ25 mm/m2).23,38,66,67 This has been termed the “55 rule.”67 stroke volumes is the regurgitant volume. Although cine Careful, serial echocardiographic follow-up is necessary to magnetic resonance is not as well validated as echocardiog- identify patients for surgery before their LV values reach raphy for quantification of AR severity, it provides highly these thresholds. accurate measurements of LV volumes, mass, and EF and Surgery for symptomatic patients with severe AR has been therefore could be useful for detecting progressive LV dila- shown to reduce LV volumes, LV mass, and wall stress and tion and timing of operation for asymptomatic severe AR. to increase LVEF.68–71 Even patients with dilated LV or low LVEF can benefit from surgery. Chaliki et al72 reported the Role of Exercise Testing results of surgery in 450 patients with severe AR. Operative Many asymptomatic patients with valvular heart disease have mortality was 14%, 6.7%, and 3.7% for those with LVEF gradually and imperceptibly reduced their activities or lead a Ͻ35%, 36% to 49%, and Ն50%, respectively (Figure 7). 132 Circulation July 5, 2005

Figure 7. Data from Chaliki et al72 show survival in patients after aortic valve replacement as a function of preopera- tive LVEF. LoEF indicates markedly reduced LV function; MedEF, moderately reduced LV function; and Nl EF, normal LV function. Reprinted with permission.

Moreover, surgical survivors with low preoperative LVEF hypertrophy and preserve LV systolic function. Endocarditis had improved symptoms and LV function. Thus, it is almost prophylaxis is important for all patients with AR. never “too late” to operate in chronic severe AR, although Future developments in interventional cardiology may patients with severe LV dysfunction and a systolic blood offer new alternatives for patients with severe AR who are not pressure Ͻ120 mm Hg may be at particularly high risk.73 considered surgical candidates. Percutaneous transcatheter implantation of a heart valve prosthesis may be possible in Medical Therapy such patients, although this is still investigational at this The regurgitant volume in AR is determined by the regurgi- time.84 tant orifice area, the square root of the diastolic pressure gradient across the valve, and the duration of diastolic flow Conclusions (which may not be holodiastolic if the LV is stiff and pressure On the basis of available evidence and consensus opinion, equilibrates early).74 Medical therapy is not able to signifi- surgery is indicated for patients with severe AR who either cantly reduce regurgitant volume in chronic severe AR (1) are symptomatic or (2) have evidence of increasing LV because the regurgitant orifice area is relatively fixed and the size or decreasing LVEF. It appears that it is best to operate diastolic blood pressure is already low.74 Further reducing before LV end-diastolic diameter increases to Ͼ55 mm or diastolic blood pressure might adversely affect coronary 25 mm/m2 or before LVEF falls to Ͻ55%. This underscores perfusion and should be avoided. Moreover, the square root the importance of careful quantification of AR severity and function dictates that a 25% reduction in diastolic pressure LV function. The role of medical therapy, particularly vaso- gradient would only achieve a 13% reduction in regurgitant dilators, is primarily to decrease systolic hypertension and volume.74 Therefore, the main goal of medical therapy is to delay the onset of LV dysfunction in asymptomatic patients. reduce the systolic hypertension associated with chronic severe AR and thereby reduce wall stress and improve LV References function.74,75 A number of small studies have investigated the 1. Cohn LH, Birjiniuk V. Therapy of acute aortic regurgitation. Cardiol effects of various vasodilators on hemodynamics and LV Clin. 1991;9:339–352. 2. Singh JP, Evans JC, Levy D, Larson MG, Freed LA, Fuller DL, Lehman 76–82 function in chronic AR. Only 2 randomized, placebo- B, Benjamin EJ. Prevalence and clinical determinants of mitral, tricuspid, controlled studies have demonstrated significant reductions in and aortic regurgitation (the Framingham Heart Study) [published cor- LV end-diastolic diameter and an increase in LVEF with rection appears in Am J Cardiol. 1999;84:1143]. Am J Cardiol. 1999;83: 77 897–902. vasodilator therapy using hydralazine in 45 patients and 3. Kim M, Roman MJ, Cavallini MC, Schwartz JE, Pickering TG, Devereux nifedipine in 72 patients.82 Medical therapy with nifedipine RB. Effect of hypertension on aortic root size and prevalence of aortic has been shown to delay the need for surgery compared with regurgitation. Hypertension. 1996;28:47–52. digoxin in a randomized trial.83 Thus, medical therapy may be 4. Vasan RS, Larson MG, Levy D, Larson MG, Freed LA, Fuller DL, Lehman B, Benjamin EJ. Determinants of echocardiographic aortic root beneficial in delaying the need for surgery in asymptomatic size. Circulation. 1995;91:734–740. patients with normal LV function. It may also be useful in 5. Lebowitz NE, Bella JN, Roman MJ, Liu JE, Fishman DP, Paranicas M, patients with severe AR who are not considered candidates Lee ET, Fabsitz RR, Welty TK, Howard BV, Devereux RB. Prevalence and correlates of aortic regurgitation in American Indians: the Strong for surgery. Importantly, the goal of medical therapy is to Heart Study. J Am Coll Cardiol. 200;36:461–467. significantly reduce systolic blood pressure to relieve the 6. Olson LJ, Subramanian R, Edwards WD. Surgical pathology of pure afterload mismatch that burdens the LV in chronic severe AR. aortic insufficiency: a study of 225 cases. Mayo Clin Proc. 1984;59: It is conceivable that further insights into molecular mecha- 835–841. 7. Waller BF, Howard J, Fess S. Pathology of aortic valve stenosis and pure nisms of myocardial adaptation to volume overload may yield aortic regurgitation: a clinical-morphologic assessment: part II. Clin new therapeutic targets to reduce myocardial fibrosis and Cardiol. 1994;17:15–16. Bekeredjian and Grayburn Aortic Regurgitation 133

8. Waller BF, Taliercio CP, Dickos DK, Howard J, Adlam JH, Jolly W. Rare pectoris and normal coronary arteries. J Am Coll Cardiol. 1988;11: or unusual causes of chronic, isolated, pure aortic regurgitation. Clin 478–486. Cardiol. 1990;13:577–581. 31. Ardehali A, Segal J, Cheitlin MD. Coronary blood flow reserve in acute 9. Guiney TE, Davies MJ, Parker DJ, Leech GJ, Leatham A. The aetiology aortic regurgitation. J Am Coll Cardiol. 1995;25:1387–1392. and course of isolated severe aortic regurgitation: a clinical, pathological, 32. Borer JS, Truter S, Herrold EM, Falcone DJ, Pena M, Carter JN, Dumlao and echocardiographic study. Br Heart J. 1987;58:358–368. TF, Lee JA, Supino PG. Myocardial fibrosis in chronic aortic regurgita- 10. Roberts WC, Morrow AG, McINtosh CL, Jones M, Epstein SE. Congen- tion: molecular and cellular responses to volume overload. Circulation. itally bicuspid aortic valve causing severe, pure aortic regurgitation 2002;105:1837–1842. without superimposed infective endocarditis. Am J Cardiol. 1981;47: 33. Truter SL, Goldin D, Kolesar J, Dumlao TF, Borer JS. Abnormal gene 206–209. expression of cardiac fibroblasts in experimental aortic regurgitation. 11. Tarasoutchi F, Grinberg M, Spina GS, Sampaio RO, Cardoso LF, Rossi Am J Ther. 2000;7:237–243. EG, Pomerantzeff P, Laurindo F, da Luz PL, Ramires JA. Ten-year 34. Rahko PS. Doppler and echocardiographic characteristics of patients clinical laboratory follow-up after application of a symptom-based ther- having an Austin Flint murmur. Circulation. 1991;83:1940–1950. apeutic strategy to patients with severe chronic aortic regurgitation of 35. Reimold S, Orav EJ, Come PC, Caguioa ES, Lee RT. Progressive predominant rheumatic etiology. J Am Coll Cardiol. 2003;41:1316–1324. enlargement of the regurgitant orifice in patients with chronic aortic 12. Michel PL, Acar J, Chomette G, Iung B. Degenerative aortic regurgita- regurgitation. J Am Soc Echocardiogr. 1998;11:259–265. tion. Eur Heart J. 1991;12:875–882. 36. Padial LR, Oliver A, Sagie A, Weyman AE, King ME, Levine RA. 13. Roman MJ, Devereux RB, Niles NW, Hochreiter C, Kligfield P, Sato N, Two-dimensional echocardiographic assessment of the progression of Spitzer MC, Borer JS. Aortic root dilatation as a cause of isolated, severe aortic root size in 127 patients with chronic aortic regurgitation: role of aortic regurgitation. Ann Intern Med. 1987;106:800–807. the supraaortic ridge and relation to the progression of the lesion. Am 14. Hahn RT, Roman MJ, Mogtader AH, Devereux RB. Association of aortic Heart J. 1997;134:814–821. dilation with regurgitant, stenotic and functionally normal bicuspid aortic 37. Bonow RO, Lakatos E, Maron BJ, Epstein SE. Serial long-term valves. J Am Coll Cardiol. 1992;19:283–288. assessment of the natural history of asymptomatic patients with chronic 15. Ferencik M, Pape LA. Changes in size of ascending aorta and aortic valve aortic regurgitation and normal left ventricular systolic function. Circu- function with time in patients with congenitally bicuspid aortic valves. lation. 1991;84:1625–1635. Am J Cardiol. 2003;92:43–46. 38. Bonow RO, Carabello B, de Leon AC Jr, Edmunds LH Jr, Fedderly BJ, 16. Mann T, McLaurin L, Grossman W, Craige E. Assessing the hemody- Freed MD, Gaasch WH, McKay CR, Nishimura RA, O’Gara PT, namic severity of acute aortic regurgitation due to infective endocarditis. O’Rourke RA, Rahimtoola SH, Ritchie JL, Cheitlin MD, Eagle KA, N Engl J Med. 1975;293:108–113. Gardner TJ, Garson A Jr, Gibbons RJ, Russell RO, Ryan TJ, Smith SC Jr. 17. Gustavsson CG, Gustafson A, Albrechtsson U, Larusdottir H, Stahl E, ACC/AHA guidelines for the management of patients with valvular heart Olin C. Diagnosis and management of acute aortic dissection, clinical and disease: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Committee on Man- radiological follow-up. Acta Med Scand. 1988;223:247–253. agement of Patients With Valvular Heart Disease). J Am Coll Cardiol. 18. Obadia JF, Tatou E, David M. Aortic valve regurgitation caused by blunt 1998;32:1486–1588. chest injury. Br Heart J. 1995;74:545–547. 39. Zoghbi WA, Enriquez-Sarano M, Foster E, Grayburn PA, Kraft CD, 19. Kardaras FG, Kardara DF, Rontogiani DP, Sioras EP, Christopoulou- Levine RA, Nihoyannopoulos P, Otto CM, Quinones MA, Rakowski H, Cokkinou V, Lolas CT, Anthopoulos LP. Acute aortic regurgitation Stewart WJ, Waggoner A, Weissman NJ. Recommendations for eval- caused by non-bacterial thrombotic endocarditis. Eur Heart J. 1995;6: uation of the severity of native valvular regurgitation with two- 1152–1154. dimensional and Doppler echocardiography. J Am Soc Echocardiogr. 20. Yeo TC, Ling LH, Ng WL, Chia BL. Spontaneous aortic laceration 2003;16:777–802. causing flail aortic valve and acute aortic regurgitation. J Am Soc Echo- 40. Perry GJ, Helmcke F, Nanda NC, Byard C, Soto B. Evaluation of aortic cardiogr. 1999;12:76–78. insufficiency by Doppler color flow mapping. J Am Coll Cardiol. 1987; 21. McCrindle BW, for the Valvuloplasty and Angioplasty of Congenital 9:952–959. Anomalies (VACA) Registry Investigators. Independent predictors of 41. Taylor AL, Eichhorn EJ, Brickner ME, Eberhart RC, Grayburn PA. immediate results of percutaneous balloon aortic valvotomy in children. Aortic valve morphology: an important in vitro determinant of proximal Am J Cardiol. 1996;77:286–293. regurgitant jet width by Doppler color flow mapping. J Am Coll Cardiol. 22. Borer JS, Hochreiter C, Herrold EM, Supino P, Aschermann M, Wencker 1990;16:405–412. D, Devereux RB, Roman MJ, Szulc M, Kligfield P, Isom OW. Prediction 42. Ishii M, Jones M, Shiota T, Yamada I, Heinrich RS, Holcomb SR, of indications for valve replacement among asymptomatic or minimally Yoganathan AP, Sahn DJ. Quantifying aortic regurgitation by using the symptomatic patients with chronic aortic regurgitation and normal left color Doppler-imaged vena contracta: a chronic animal model study. ventricular performance. Circulation. 1998;97:525–534. Circulation. 1997;96:2009–2015. 23. Dujardin KS, Enriquez-Sarano M, Schaff HV, Bailey KR, Seward JB, 43. Tribouilloy CM, Enriquez-Sarano M, Bailey KR, Seward JB, Tajik AJ. Tajik AJ. Mortality and morbidity of aortic regurgitation in clinical Assessment of severity of aortic regurgitation using the width of the vena practice: a long-term follow-up study. Circulation. 1999;99:1851–1857. contracta: a clinical color Doppler imaging study. Circulation. 2000;102: 24. Carabello BA. Aortic regurgitation: a lesion with similarities to both 558–564. aortic stenosis and mitral regurgitation. Circulation. 1990;82:1051–1053. 44. Willett DL, Hall SA, Jessen ME, Wait MA, Grayburn PA. Assessment of 25. Ricci DR. Afterload mismatch and preload reserve in chronic aortic aortic regurgitation by transesophageal color Doppler imaging of the vena regurgitation. Circulation. 1982;66:826–834. contracta: validation against an intraoperative aortic flow probe. JAm 26. Ross J Jr, McCullagh WH. Nature of enhanced performance of the dilated Coll Cardiol. 2001;37:1450–1455. left ventricle during chronic volume overloading. Circ Res. 1972;30: 45. Cohen GI, Duffy CI, Klein AL, Miller DP, Cosgrove DM, Stewart WJ. 549–556. Color Doppler and two-dimensional echocardiographic determination of 27. Starling MR, Kirsh MM, Montgomery DG, Gross MD. Mechanism for the mechanism of aortic regurgitation with surgical correlation. JAmSoc left ventricular systolic dysfunction in aortic regurgitation: importance for Echocardiogr. 1996;9:508–515. predicting the functional response to aortic valve replacement. J Am Coll 46. Tribouilloy CM, Enriquez-Sarano M, Fett SL, Bailey KR, Seward JB, Cardiol. 1991;17:887–897. Tajik AJ. Application of the proximal flow convergence method to 28. Wisenbaugh T, Spann JF, Carabello BA. Differences in myocardial per- calculate the effective regurgitant orifice area in aortic regurgitation. JAm formance and load between patients with similar amounts of chronic Coll Cardiol. 1998;32:1032–1039. aortic versus chronic mitral regurgitation. J Am Coll Cardiol. 1984;3: 47. Meyer T, Sareli P, Pocock WA, Dean H, Epstein M, Barlow J. Echocar- 916–923. diographic and hemodynamic correlates of diastolic closure of mitral 29. Magid NM, Young MS, Wallerson DC, Goldweit RS, Carter JN, valve and diastolic opening of aortic valve in severe aortic regurgitation. Devereux RB, Borer JS. Hypertrophic and functional response to exper- Am J Cardiol. 1987;59:1144–1148. imental chronic aortic regurgitation. J Mol Cell Cardiol. 1988;20: 48. Grayburn PA, Handshoe R, Smith MD, Harrison MR, DeMaria AN. 239–246. Quantitative assessment of the hemodynamic consequences of aortic 30. Nitenberg A, Foult JM, Antony I, Blanchet F, Rahali M. Coronary flow regurgitation by means of continuous wave Doppler recordings. JAm and resistance reserve in patients with chronic aortic regurgitation, angina Coll Cardiol. 1987;10:135–141. 134 Circulation July 5, 2005

49. Teague SM, Heinsimer JA, Anderson JL, Sublett K, Olson EG, Voyles 65. Henry Wl, Bonow RO, Borer JS, Ware JH, Kent KM, Redwood DR, WF, Thadani U. Quantification of aortic regurgitation utilizing con- McIntosh CL, Morrow AG, Epstein SE. Observations on the optimum tinuous wave Doppler ultrasound. J Am Coll Cardiol. 1986;8:592–599. time for operative intervention for aortic regurgitation, I: evaluation of the 50. Touche T, Prasquier R, Nitenberg A, de Zuttere D, Gourgon R. results of aortic valve replacement in symptomatic patients. Circulation. Assessment and follow-up of patients with aortic regurgitation by an 1980;61:471–483. updated Doppler echocardiographic measurement of the regurgitant 66. Borer JS, Bonow RO. Contemporary approach to aortic and mitral regur- fraction in the aortic arch. Circulation. 1985;72:819–824. gitation. Circulation. 2003;108:2432–2438. 51. Croft CH, Lipscomb K, Mathis K, Firth BG, Nicod P, Tilton G, 67. Carabello BA, Crawford FA. Valvular heart disease. N Engl J Med. Winniford MD, Hillis LD. Limitations of qualitative angiographic 1997;337:32–41. grading in aortic or mitral regurgitation. Am J Cardiol. 1984;53: 68. Ishii K, Hirota Y, Suwa M, Kita Y, Onaka H, Kawamura K. Natural 1593–1598. history and left ventricular response in chronic aortic regurgitation. 52. Kozerke S, Schwitter J, Pedersen EM, Boesiger P. Aortic and mitral Am J Cardiol. 1996;78:357–361. regurgitation: quantification using moving slice velocity mapping. J 69. Bonow RO, Dodd JT, Maron BJ, O’Gara PT, White GG, McIntosh CL, Magn Reson Imaging. 2001;14:106–112. Clark RE, Epstein SE. Long-term serial changes in left ventricular 53. Chatzimavroudis GP, Oshinski JN, Franch RH, Pettigrew RI, Walker PG, function and reversal of ventricular dilatation after valve replacement for Yoganathan AP. Quantification of the aortic regurgitant volume with chronic aortic regurgitation. Circulation. 1988;78:1108–1120. magnetic resonance phase velocity mapping: a clinical investigation of 70. Taniguchi K, Nakano S, Kawashima Y, Sakai K, Kawamoto T, Sakaki S, the importance of imaging slice location. J Heart Valve Dis. 1998;7: Kobayashi J, Morimoto S, Matsuda H. Left ventricular ejection per- 94–101. formance, wall stress, and contractile state in aortic regurgitation before 54. Krombach GA, Kuhl H, Bucker A, Mahnken AH, Spuntrup E, Lipke C, and after aortic valve replacement. Circulation. 1990;82:798–807. Schroder J, Gunther RW. Cine MR imaging of heart valve dysfunction 71. Roman MJ, Klein L, Devereux RB, Kligfield P, Niles NW, Hochreiter C, with segmented true fast imaging with steady state free precession. J Isom OW, Borer JS. Reversal of left ventricular dilatation, hypertrophy, Magn Reson Imaging. 2004;19:59–67. and dysfunction by valve replacement in aortic regurgitation. Am Heart J. 55. Lewis SM, Riba AL, Berger HJ, Davies RA, Wackers FJ, Alexander J, 1989;118:553–563. Sands MJ, Cohen LS, Zaret BL. Radionuclide angiographic exercise left 72. Chaliki HP, Mohty D, Avierinos J-F, Scott CG, Schaff HV, Tajik AJ, ventricular performance in chronic aortic regurgitation: relationship to Enriquez-Sarano M. Outcomes after aortic valve replacement in patients resting echographic ventricular dimensions and systolic wall stress index. with severe aortic regurgitation and markedly reduced left ventricular Am Heart J. 1982;103:498–504. function. Circulation. 2002;106:2687–2693. 56. Goldman ME, Packer M, Horowitz SF, Meller J, Patterson RE, Kukin M, 73. Carabello BA. Is it ever too late to operate on the patient with valvular Teichholz LE, Gorlin R. Relation between exercise-induced changes in heart disease? J Am Coll Cardiol. 2004;44:376–383. ejection fraction and systolic loading conditions at rest in aortic regurgi- 74. Levine HJ, Gaasch WH. Vasoactive drugs in chronic regurgitant lesions tation. J Am Coll Cardiol. 1984;3:924–929. of the mitral and aortic valves. J Am Coll Cardiol. 1996;28:1083–1091. 57. Greenberg B, Massie B, Thomas D, Bristow JD, Cheitlin M, Broudy D, 75. Grayburn PA. Vasodilator therapy for chronic aortic and mitral regurgi- Szlachcic J, Krishnamurthy G. Association between the exercise ejection tation. Am J Med Sci. 2000;320:202–208. fraction response and systolic wall stress in patients with chronic aortic 76. Kleaveland JP, Reichek N, McCarthy DM, Chandler T, Priest C, insufficiency. Circulation. 1985;71:458–465. Muhammed A, Makler PT Jr, Hirshfeld J. Effects of six-month afterload 58. Meyer TE, Sareli P, Marcus RH, Patel J, Berk MR. Beneficial effect of reduction therapy with hydralazine in chronic aortic regurgitation. atrial pacing in severe acute aortic regurgitation and role of M-mode Am J Cardiol. 1986;57:1109–1116. echocardiography in determining the optimal pacing interval. 77. Greenberg B, Massie B, Bristow JD, Cheitlin M, Siemienczuk D, Topic Am J Cardiol. 1991;67:398–403. N, Wilson RA, Szlachcic J, Thomas D. Long-term vasodilator therapy of 59. Firth BG, Dehmer GJ, Nicod P, Willerson JT, Hillis LD. Effect of chronic aortic insufficiency: a randomized double-blinded, placebo- increasing heart rate in patients with aortic regurgitation: effect of incre- controlled clinical trial. Circulation. 1988;78:92–103. mental atrial pacing on scintigraphic, hemodynamic and thermodilution 78. Dumesnil JG, Tran K, Dagenais GR. Beneficial long-term effects of measurements. Am J Cardiol. 1982;49:1860–1867. hydralazine in aortic regurgitation. Arch Intern Med. 1990;150:757–760. 60. Pompilio G, Brockmann C, Bruneau M, Buche M, Amrani M, Louagie Y, 79. Lin M, Chiang HT, Lin SL, Chang MS, Chiang BN, Kuo HW, Cheitlin Eucher P, Rubay J, Jamart J, Dion R, Schoevaerdts JC. Long-term MD. Vasodilator therapy in chronic asymptomatic aortic regurgitation: survival after aortic valve replacement for native active infective endo- enalapril versus hydralazine therapy. J Am Coll Cardiol. 1994;24: carditis. Cardiovasc Surg. 1998;6:126–132. 1046–1053. 61. Vogt PR, von Segesser LK, Jenni R, Niederhauser U, Genoni M, Kunzli 80. Wisenbaugh T, Sinovich V, Dullabh A, Sareli P. Six month pilot study of A, Schneider J, Turina MI. Emergency surgery for acute infective aortic captopril for mildly symptomatic, severe isolated mitral and isolated valve endocarditis: performance of cryopreserved homografts and mode aortic regurgitation. J Heart Valve Dis. 1994;3:197–204. of failure. Eur J Cardiothorac Surg. 1997;11:53–61. 81. Schon HR, Dorn R, Barthel P, Schomig A. Effects of 12 months quinapril 62. Ergin MA, McCullough J, Galla JD, Lansman SL, Griepp RB. Radical therapy in asymptomatic patients with chronic aortic regurgitation. replacement of the aortic root in acute type A dissection: indications and J Heart Valve Dis. 1994;3:500–509. outcome. Eur J Cardiothorac Surg. 1996;10:840–844. 82. Scognamiglio R, Fasoli G, Ponchia A, Dalla-Volta S. Long-term nifedi- 63. Edwards FH, Peterson ED, Coombs LP, DeLong ER, Jamieson WR, pine unloading therapy in asymptomatic patients with chronic severe Shroyer ALW, Grover FL. Prediction of operative mortality after valve aortic regurgitation. J Am Coll Cardiol. 1990;16:424–429. replacement surgery. J Am Coll Cardiol. 2001;37:885–892. 83. Scognamiglio R, Rahimtoola SH, Fasoli G, Nistri S, Dalla Volta S. 64. Florath I, Rosendahl UP, Mortasawi A, Bauer SF, Dalladaku F, Ennker Nifedipine in asymptomatic patients with severe aortic regurgitation and IC, Ennker JC. Current determinants of operative mortality in 1400 normal left ventricular function. N Engl J Med. 1994;331:1417–1423. patients requiring aortic valve replacement. Ann Thorac Surg. 2003;76: 84. Boudjemline Y, Bonhoeffer P. Steps toward percutaneous aortic valve 75–83. replacement. Circulation. 2002;105:775–778. New Drugs and Technologies

Frontiers in Cardiovascular Magnetic Resonance

Valentin Fuster, MD, PhD; Raymond J. Kim, MD

ardiovascular magnetic resonance (MR) is emerging as a elements such as crystalline cholesterol and cholesteryl esters Cmultipurpose imaging modality for the assessment of intermixed with monocyte-derived macrophages and other cells; cardiovascular disease in general and ischemic heart disease (3) thrombotic, or deposition of platelets and/or fibrin; and (4) in particular. Currently, the pace of innovation is rapid, and calcification, usually related to fibrous rather than to lipid-rich the modality is changing from one that is used primarily as a plaques.1,5,6 Varying proportions of these components occur in research tool to one that is increasingly used in routine different plaques, thus giving rise to a heterogeneity or spectrum clinical practice. The process of innovation includes not only of lesions. These components mainly affect the intima, but improvements in scanner hardware, such as coil and gradient secondary changes also occur in the media and adventitia; these technology, and the development of new contrast agents but presumably include growth of vasa vasorum and extravasated also the development of novel pulse sequences. The concept erythrocytes.7,8 As examples of the heterogeneity of lesions, of the pulse sequence, in which programming changes at the disruption-prone plaques in the coronary arteries, the so-called scanner can lead to fundamental changes in activating tissue, vulnerable plaques, tend to have a thin fibrous cap (cap thickness is unique to MR and gives this modality the potential to Ϸ65 to 150 ␮m ) and a large lipid core (Ͼ40% of the total lesion assess a vast number of biological parameters. area).9 About two thirds of the acute coronary syndromes (ACS) Cardiovascular MR promises to play an important clinical result from disruption of a modestly stenotic vulnerable plaque, and investigational role in both vascular and cardiac systems. not visible by x-ray angiography, which triggers an acute Current and potential future applications of cardiovascular thrombus formation that may result in a thrombotic occlusion.9 MR will be discussed with a particular focus on ischemic Similar observations have been made of small lipid-rich lesions heart disease. Multidetector-row computed tomography, an- of the thoracic aorta, which, after disruption and thrombosis, other promising and complementary noninvasive imaging may result in stroke.10 Unlike coronary and aortic vulnerable technology, will be discussed briefly in relation to cardiovas- plaques, carotid plaques prone to disruption and thrombosis are cular MR for the assessment of atherothrombotic disease. predominantly fibrotic and severely stenotic.10 Similar observa- tions have been made of severely stenotic and fibrotic plaques The Vasculature and leading to thrombotic complications (presumably favored by a Atherothrombotic Disease hypercoagulant state) that affect the peripheral arteries and Nomenclature and Evolving Imaging Assessment occasionally the coronary arteries, which explains approximately Atherothrombosis is a systemic or multiterritory arterial one third of the ACS.10 Therefore, in atherothrombotic disease, it disease that primarily affects the large- and medium systemic has been proposed that the term “high-risk plaque” may be used arteries, including the aorta and the carotid, coronary, and interchangeably with the classic term “vulnerable plaque,” which peripheral arteries. Although the epicardial coronary arteries traditionally implies the presence of a lipid-rich core.2 appear to be the most susceptible to atherothrombosis,1,2 Accordingly, reliable noninvasive imaging modalities able to intramyocardial arteries are relatively resistant. The concept detect atherothrombotic disease in the various stages and regions of multiterritory atherothrombosis has been addressed in 2 and to characterize plaque composition are clinically desirable. large studies of symptomatic patients that showed that at Additionally, the availability of such imaging modalities will entry into the studies, 3% to 8% had symptomatic athero- improve our understanding of the pathophysiological mecha- thrombotic disease in all 3 main arterial districts and 23% to nisms underlying the atherothrombotic processes and allow us to 32% had disease in 2 districts.3,4 better risk-stratify the “burden” of disease. Moreover, such tools From a structural point of view, the 4 main components of the may permit optimal tailoring of treatment and allow direct atherothrombotic plaques are as follows: (1) fibrocellular, or monitoring of the vascular response.10 Most invasive techniques, extracellular matrix of various fibril types intermixed with such as coronary angiography and intravascular ultrasound, smooth muscle cells and other cells; (2) lipid-cellular, or lipid identify luminal diameter or stenosis, wall thickness, and plaque

From the Zena and Michael A. Wiener Cardiovascular Institute and The Marie-Josee and Henry R. Kravis Cardiovascular Health Center (V.F.), The Mount Sinai School of Medicine, New York, NY, and Duke Cardiovascular Magnetic Resonance Center (R.J.K.), Duke University Medical Center, Durham, NC. The online-only Data Supplement, which contains Movies I through V, can be found with this article at http://www.circulationaha.org. Correspondence to Raymond J. Kim, MD, Duke Cardiovascular MRI Center, DUMC-3934, Durham, NC 27710 (e-mail [email protected]), or Valentin Fuster, MD, PhD, Mount Sinai School of Medicine, Box 1030, New York, NY 10029 (e-mail [email protected]). (Circulation. 2005;112:135-144.) © 2005 American Heart Association, Inc. Circulation is available at http://www.circulationaha.org DOI: 10.1161/01.CIR.0000155618.37779.A0 135 136 Circulation July 5, 2005

MR and CT Angiography of the Coronary Arteries

Patients, Sensitivity, Specificity, Reference Technique n % % Regenfus et al67 MRA 61 85 90 Plein et al68 MRA 40 74 88 Watanabe et al69 MRA 22 80 85 Kim et al70 (non-CE) MRA 109 93 42 Nikolaou et al71 MRA 40 72 60 Ropers et al72 MDCTA (16) 77 92 93 Nieman et al73 MDCTA (16) 59 95 86 Knez et al74 MDCTA (4) 44 78 98 Nieman et al75 MDCTA (4) 35 81 97 Achenbach et al76 MDCTA (4) 64 85 76 Kuettner et al77 MDCTA (4) 66 37 99 MDCTA indicates MDCT angiography; CE, contrast enhanced. volume.11 However, 2 emerging and most promising tech- niques—computed tomography (CT) and MR—are likely to gain wider acceptance by the medical community and to be Figure 1. In vivo black-blood MR cross-sectional T2-weighted image of patient with significant plaque in right carotid artery applied to larger populations groups because they are noninva- (arrow). Magnified image (bottom left) shows complex lipid-rich sive and can both evaluate luminal stenosis (CT) and character- plaque. Courtesy Z.A. Fayad, Mount Sinai School of Medicine. ize plaque composition (MR).12 water content, physical state, molecular motion, or diffusion. MR Imaging Specifically, recent improvements in MR techniques (eg, black- Because atherothrombotic disease affects the entire arterial blood MRI, faster imaging and detection coils), conducive to system, simultaneous assessment from supra-aortic arteries to high-resolution and contrast imaging, have permitted the study the distal runoff vessels has been proposed with contrast- of the various plaque components with multicontrast MR, enhanced, whole-body MR angiography (MRA). Most im- generated by T1- and T2-weighted, proton-density–weighted, and portantly, high-resolution MRI has emerged as the potential time-of-flight imaging.11,12,15,16 Moreover, MR provides imag- leading noninvasive in vivo imaging modality for atheroscle- ing without ionizing radiation and can be repeated over time. rotic plaque characterization. MRI Studies of Carotid Artery Plaques Whole-Body, Contrast-Enhanced MRA The superficial location and relative absence of motion of the MRA is highly specific and sensitive compared with x-ray carotid arteries allows excellent delineation of plaque by MR angiography for the detection of luminal narrowing Ͼ50%.13 techniques (Figure 1). Thus far, MR studies have shown the Whole-body MRA excludes the intracranial and coronary arter- characterization of normal and pathological arterial walls,12,16 ies, for which a dedicated examination is still required. Several the quantification of plaque size and therapeutic regression,17,18 coronary MRA techniques have been proposed for the assess- and the detection of fibrous cap integrity, as well as disruption- ment of coronary stenosis, anomalies, and patency of bypass related transient ischemic attack or stroke.19 Thus, it can be grafts. Thus far, the sensitivity and specificity of coronary MRA predicted that MRA, which demonstrates the severity and with later-generation 3D techniques are quite reasonable in distribution of stenotic plaques, and high-resolution MRI, which select patient cohorts (Table).12–14 Coronary MRA, however, is 12 technically challenging, and currently there are limitations in characterizes such plaques, will eventually be combined. spatial coverage and resolution, temporal resolution, and image MRI Studies of Aortic Plaques quality that preclude the routine use of coronary MRA for The principal challenges associated with high-resolution MRI of everyday clinical application. the thoracic aorta are that attainment of sufficient sensitivity for Overall, it is possible that in the near future, contrast-enhanced submillimeter imaging and the exclusion of artifacts caused by MRA with the use of gadolinium-based contrast agents will provide respiratory motion and blood flow. Matched MRI and trans- complete assessment of the systemic arterial tree, whereas nonin- esophageal echocardiography cross-sectional aortic segments vasive CT with intravenous injection of contrast medium may showed a strong correlation for plaque thickness, whereas MRI 12 replace conventional diagnostic coronary angiography in part. was the best contributor to plaque characterization and therapeu- However, contrast-enhanced molecular MRA, when added to re- tic regression (Figure 2).17,20 In a recent study of asymptomatic gional high-resolution MRI (ie, coronary arteries), may provide subjects, the Framingham Heart Study showed by MRI that additional information in plaque characterization. aortic plaque prevalence and burden (ie, plaque volume/aortic Regional High-Resolution MRI for volume) significantly increased with age and were higher in the Plaque Characterization abdominal aorta than in the thoracic aorta.21 Importantly, the MR differentiates plaque components on the basis of biophysical Framingham Heart Study coronary risk score was strongly and biochemical parameters such as chemical composition, associated with asymptomatic aortic plaques as detected by Fuster and Kim Frontiers in Cardiovascular Magnetic Resonance 137

Figure 3. In vivo MR black-blood cross-sectional images of human coronary arteries demonstrating plaque, presumably with deposition of fat (arrow, A), concentric fibrotic lesion (B) in left anterior descending artery (LAD), and ectatic but atherosclerotic right coronary artery (C). RV indicates right ventricle; LV, left ventricle. Modified with permission from Fayad et al.25

Figure 2. T2-weighted MR images at 2 different time points (base- line and 24 months after initiation of lipid-lowering therapy by enhanced CT, high-resolution MRI coupled with contrast- statins) from same patient. Details of descending aorta are shown. enhanced molecular MRA promises to fulfill this goal. At 24 months after lipid lowering, MRI shows thinner plaque and smaller lipid area (hypointense signal from 1 to 4 o’clock) com- Contrast-Enhanced Molecular MRI for pared with baseline (arrows). Bar scale indicates 10 mm. Plaque Characterization An alternative approach to high-resolution MRI for plaque MRI. Such an approach may turn out to be very valuable for characterization is to image plaque through the introduction of identification, quantification, and the therapeutic management of contrast agents that are targeted to specific cells, molecules, or plaque burden, particularly in asymptomatic individuals with a processes that can be precisely localized and quantified.15,28–31 12 high risk factor profile. Examples might include the following (Figure 4): adhesion MRI Studies of Peripheral Arteries molecules (vascular cell adhesion molecule-1, intercellular ad- High-resolution MR of the femoral and popliteal arteries and of hesion molecule, and selectins), macrophages within the context the response to balloon angioplasty has been reported.22 The of apoptosis (phosphatidylserine and synaptotagmin), fibrous cap extent of the plaques could be defined such that even in within the context of proteolysis (matrix metalloproteinases), angiographically normal segments of vessel, lesions with cross- lipid core (nonspecific lipophilic; Figure 5), angiogenesis (inte- ␣ ␤ ␣ ␤ 15,28–31 sectional areas ranging from 49% to 76% of potential lumen area grin V 3), and thrombosis (fibrin and integrin IIb 3). were identified. After angioplasty, plaque fissuring and local Targeted imaging agents are generally created by chemically dissection were identified easily, and serial changes in lumen attaching an affinity ligand, such as an antibody, peptide, or small molecule, to a magnetic compound, such as superpara- diameter, blood flow, and lesion size were documented. In the magnetic particles of iron oxide or gadolinium chelates.15,28 Our future, this technology, when combined with contrast-enhanced expanding understanding of cellular and molecular events within molecular MRA as discussed later, may be of great value for the atherosclerotic plaque has been accompanied by imaginative postinterventional assessment of different therapeutic strategies application of imaging tools, which has led to the new field of such as new antithrombotic or antifibrotic drugs. contrast-enhanced molecular MRI. Such molecular technology, MRI Studies of Coronary Artery Plaques when combined with high-resolution MRI, promises comple- With a combination of multicontrast MR imaging sequences, mentary structural and biological information and, therefore, differentiation of fibrocellular, lipid-rich, and calcified re- more detailed plaque characterization. In addition, thinner slices, gions of the atherosclerotic coronary plaque is feasible, as such as those obtained with 3D MR acquisition techniques, and shown in an ex vivo study on human coronary arteries other evolving MR technologies, such as water diffusion weight- correlated to histopathology.23 Black-blood MR methods ing, magnetization transfer weighting, and steady-state free- used in the human carotid artery and aorta have been applied precession (SSFP) sequences, all promise to further improve to imaging of the coronary arterial lumen and wall. The artery wall structural and biological imaging. method was validated in swine coronary lesions induced by Functional Vascular MRI balloon angioplasty.24 High-resolution black-blood MR of Noninvasive imaging techniques such as CT and MRA allow both normal and atherosclerotic human coronary arteries was assessment of vascular anatomy but do not provide information performed for direct assessment of coronary wall thickness about blood flow. For clinical purposes, flow information is and the visualization of focal atherosclerotic plaque in the important, because anatomy and function may not be directly wall (Figure 3).25 To alleviate the need for breath holding, related. Global coherent free precession (GCFP) is a new real-time navigator for respiratory gating and real-time slice- concept in MRI that can be used to produce images that depict position correction have been reported.26 Near-isotropic vascular anatomy simultaneously with vascular function (blood spatial-resolution black-blood imaging may provide a quick flow).32 Protons in moving blood are “tagged” every few way to image a long segment of the coronary artery wall and milliseconds as they travel through an arbitrary region in space. may be useful for rapid coronary plaque burden measure- Simultaneous with tagging of new blood, previously tagged ments.27 A crucial ultimate goal of cardiovascular noninva- blood is maintained in the GCFP state, which allows acquisition sive imaging is to have reliable technology for plaque of consecutive movie frames as the heart pushes blood through characterization of the coronary arteries. Guided by contrast- the vasculature. Body tissue surrounding the moving blood is 138 Circulation July 5, 2005

Figure 4. Examples of molecular targets and technological modalities relevant to cardiovascular imaging. AHA indicates American Heart Association. Modified with permission from Choudhury et al.31

never excited and is invisible. With this approach, pulsating Future Integration of Noninvasive Coronary CT blood can be seen flowing through 3D space for distances of up and MRI to 16 cm (Figure 6). Although additional technical development Today, 2 different modes of CT are available.12 One uses will be required before the full potential of GCFP MRI can be nonmechanical movement of the x-ray source (ie, electron- recognized, the current data demonstrate that GCFP MRI can be beam CT), and another involves the motion of the x-ray used to produce cine angiograms that are remarkably similar to source and table, combined with multiple detectors to acquire those produced by invasive X-ray angiography, but noninva- the data in spiral or helical fashion (ie, multidetector-row CT sively and without the need for contrast agents or ionizing [MDCT]). Although electron-beam CT has been considered radiation. the “gold standard” for the assessment of calcified plaques, MDCT usually includes an initial nonenhanced scan for the screening and quantification of coronary artery calcium followed by CT angiography for direct visualization of coronary artery disease.12 Results of a number of studies concerning the use of contrast-enhanced MDCT for noninva- sive coronary angiography have been published. It appears that the diagnostic accuracy is reasonable (Table), but com- plete assessment can be hindered by calcium deposits in the vessel wall and by motion artifacts, particularly in patients with high heart rates.12 Two studies with 16-slice scanners have been reported recently, each with improved accuracy

Figure 5. A, In vivo T1-weighted MR transverse image of abdominal aorta, 24 hours after gadofluorine injection. B, Mag- nification shows plaque enhancement after injection. C, Corre- sponding histological section. Combined Masson elastin stain- Figure 6. Single movie frames from GCFP MRI (left) and inva- ing allows characterization of different components of plaque sive catheterization (right) in same patient. GCFP images did not (original magnification ϫ10). Appearance of MR image corre- require invasive procedure, contrast agent, or ionizing radiation. lates closely with matched histological section shown in C. Ad Total GCFP acquisition time was 15 seconds. Full-motion mov- indicates adventitia; FC, fibrous cap; L, lumen; and LC, lipid ies can be viewed in the online-only Data Supplement (select core. Modified with permission from Sirol et al.29 Movie I). Modified with permission from Rehwald et al.32 Fuster and Kim Frontiers in Cardiovascular Magnetic Resonance 139

Figure 7. MR images of morphology (a), motion (b), perfusion (c), and delayed enhancement (d) in patient with left atrial mass (arrows). Biopsy demonstrated recurrent invasive thymoma (note several extracardiac masses are also present). Perfusion is reduced compared with left ventricular myocardium. Hyperenhance- ment is present in heterogeneous fash- ion. Full-motion movies can be viewed in the online-only Data Supplement (select Movie II).

compared with reports with 4-slice scanners (Table). The next future. Additionally, we will discuss how some recent findings generation of MDCT scanners will almost certainly allow for by cardiac MR provide insights into cardiac pathophysiology. even faster gantry rotation and simultaneous acquisition of Ͼ16 slices. The breath-hold time may decrease to Ͻ10 Morphology and Function seconds, thus reducing the volume of contrast media needed MR provides arguably the best and most comprehensive ap- for sufficient enhancement of the coronary arteries. Temporal proach to evaluating the structure and function of the heart. A and spatial resolution may further improve, ideally to 100 ms number of techniques have been developed, including those that and 0.6-mm slice thickness. These enhancements may help in can render fat or flowing blood invisible, allow rapid imaging the detection, differentiation, and reliable quantification of that is free of motion artifacts even during free breathing, and, calcified and noncalcified coronary artery plaques. Improve- with the addition of gadolinium contrast, provide information ment of spatial resolution and new image-reconstruction regarding tissue perfusion, necrosis, and fibrosis. The rapid pace algorithms should further reduce beam-hardening artifacts of innovation, however, raises some issues that are perhaps more and partial volume effects caused by calcifications and unique to MR than to the other imaging modalities. For instance, improve the assessment of complex mixed plaques. Further there is often a discrepancy between the newer techniques that optimization of multisegmental reconstruction algorithms are quickly adopted in clinical practice and those that are may allow the investigation of patients with higher heart rates described in the published literature. This problem is com- without any loss in image quality. pounded by the fact that new techniques in MR often include CT and MRI together may provide unique information, such new relationships between image intensity and the underlying as assessment of subclinical disease, the study of atherothrom- physiology rather than just the provision of improved signal-to- botic progression, and response to therapy. CT may first be used noise ratio or improved resolution. It is important to realize that to localize suspicious atherothrombotic lesions in the coronary concepts or algorithms associated with older techniques may not arteries within a short scan time. MRA does the same in the apply to the newer techniques. systemic arteries but within a much longer scan time. High- Consider MR for the assessment of cardiac masses. This resolution MRI and contrast-enhanced molecular MRI can then literature is extensive; however, the vast majority of the data, be used for structural and biological plaque characterization of even those from recent publications,33 were acquired by early the problem sites. Furthermore, the role of MRI in the in vivo spin-echo techniques that have several limitations. They are slow monitoring of therapies can be pivotal for the better understand- (several minutes per image) and prone to motion artifacts due to ing of new pharmacological agents before clinical trials are free breathing during image acquisition, and they provide limited undergone. It can also serve as a guide to assess the vascular wall T1 weighting. Much of this literature involved attempts at tissue response by individual patients to proven beneficial therapies. characterization by comparison of image intensities on T1-, T2-, and proton-density–weighted images. Differentiation between The Myocardium benign and malignant masses from image intensity features, The purpose of this section is to highlight some of the recent however, was usually poor.33 technical advances in cardiac MR, and in particular, to focus on These older spin-echo techniques are no longer used in how these advances may affect the clinical assessment of clinical practice. Instead, at present, a typical protocol for the patients. Rather than providing a comprehensive review of the evaluation of a cardiac mass would consist of the following: (1) literature, we will speculate on how these new techniques could 1 or more stacks of single-shot imaging that combines rapid be optimally used in clinical practice, currently and in the near (0.25 seconds/slice) image acquisition with comprehensive an- 140 Circulation July 5, 2005

Future effort should be directed toward testing whether changes in cardiac parameters as measured by MR indeed translate into differences in patient outcome.

Infarction and Viability Recently, numerous studies have demonstrated the effective- ness of a segmented inversion-recovery gradient-echo se- quence after the intravenous administration of gadolinium contrast for detecting myocardial infarction (MI) and deter- Figure 8. Comparison of long-axis cine views acquired with mining viability. This technique, termed delayed contrast- conventional gradient-recalled echo (GRE) and SSFP tech- niques. Slow flow of blood inhibits delineation of endocardial enhanced MRI (DE-MRI), was first described fewer than 5 border of left ventricular apex on GRE cine. With SSFP, there is years ago,37 but there is already consideration that “DE-MRI overall higher signal-to-noise ratio and excellent contrast may well represent the new gold standard in the detection of between myocardium and blood. Full-motion cines can be irreversibly damaged myocardium.”38 viewed in the online-only Data Supplement (select Movie III). The results in the literature with DE-MRI should not be grouped with the results with older MR techniques. A major atomic coverage to quickly delineate morphology (Figure 7a); limitation of the initial techniques was insufficient image con- (2) cine imaging to view motion during the cardiac cycle (Figure trast between normal and infarcted myocardium. DE-MRI, on 7b); (3) first-pass perfusion imaging during the transit of an the other hand, provides image intensity differences that are intravenous bolus of gadolinium contrast (Figure 7c); and (4) 39 postcontrast delayed-enhancement imaging, which accentuates 10-fold greater. Suboptimal image quality was a major factor differences in contrast uptake between the mass and normal in leading to the erroneous conclusion that chronic infarcts do myocardium and between different regions of the mass (Figure not hyperenhance and, conversely, the speculation that viable 7d). Each of these steps consists of pulse sequences that have myocytes could exhibit hyperenhancement. improved significantly, even in the last 2 to 3 years. For instance, There is a wealth of data in animal models of ischemic injury 37,40,41 delineation of morphology with black-blood HASTE (half Fou- that directly compares DE-MRI to histopathology. These rier single-shot turbo spin-echo34) is currently of sufficient data demonstrate that DE-MRI can delineate between reversible quality that segmented black-blood sequences, which require 8- and irreversible myocardial injury independent of wall motion, to 10-second breath holds per slice, are used sparingly. infarct age, or reperfusion status. Human studies demonstrate We expect that tissue characterization of cardiac masses will be that DE-MRI is effective in identifying the presence, location, 39,42 improved substantially by the performance of first-pass perfusion and extent of MI in both the acute and chronic settings. and delayed-enhancement imaging (also see following sections for Additionally, DE-MRI provides scar-size measurements that are further description of these techniques). In particular, there will be closely correlated with positron emission tomography in patients excellent discrimination between cardiac thrombus and tumors. We with ischemic cardiomyopathy,43,44 provides results superior to speculate that these techniques will become the standard approach single-photon emission computed tomography (SPECT) in pa- for the noninvasive assessment of cardiac masses. Several centers tients with subendocardial infarctions,45 and can be used to are currently investigating these strategies. predict reversible myocardial dysfunction in those undergoing One of the most significant advances in MR has been the revascularization procedures.44,46,47 robust implementation of SSFP sequences. SSFP (also desig- A major advantage of DE-MRI is the high spatial resolution. nated as TrueFISP, FIESTA, or balanced FFE) provides sub- With a standard implementation, a group of 10 hyperenhanced stantially higher signal-to-noise ratio than can be obtained by pixels (voxel, 1.9ϫ1.4ϫ6 mm) in a typical image would represent conventional gradient-echo techniques, along with excellent an infarction of 0.16 g, or a region one thousandth of the left contrast between myocardium and blood35 (Figure 8). Imple- ventricular myocardial mass.42 This level of resolution, more than mentations include single-shot 2D versions with or without 40-fold greater than SPECT, allows visualization of even microin- prepulses to provide rapid snapshot images to delineate mor- farcts that cannot be detected by other imaging techniques.48 phology, first-pass perfusion, or delayed enhancement; 3D Besides spatial resolution, DE-MRI is different from radio- sequences to provide angiogram-like views of the vasculature nuclide imaging in that it provides direct visualization of both without the need for contrast media; and multiphase segmented nonviable and viable myocardium. For instance, rather than 2D sequences to provide high-resolution cine images of the simply identifying a region of acute infarction as nonviable heart.35 The latter is currently the “gold standard” approach to because of reduced tracer activity, DE-MRI can distinguish assess ventricular volumes, mass, and function. between acute infarcts with necrotic myocytes and acute infarcts The accuracy and reproducibility of MR in assessing cardiac with necrotic myocytes and damaged microvasculature. The morphology and function leads to low interstudy variability in latter, termed the “no-reflow phenomenon,” indicates compro- quantifying these parameters. This is turn translates into significant mised tissue perfusion despite restoration of epicardial artery reductions in sample sizes that are required to test the efficacy of patency. The incidence and extent of early no-reflow appears to therapeutic interventions on these parameters.36 It is expected that be associated with worse left ventricular remodeling and out- the number of drug and device trials that use cardiac MR parameters come. Although the initial MR studies of no-reflow used as study end points will increase substantially in the future. Ulti- single-shot perfusion sequences 1 to 2 minutes after contrast mately, however, patient outcome is the relevant clinical issue. injection,49 DE-MRI performed 5 to 10 minutes after contrast Fuster and Kim Frontiers in Cardiovascular Magnetic Resonance 141

Figure 9. No-reflow phenomenon revealed by DE-MRI. Labels refer to time after administration of gadolinium contrast. Subendocardial black zone surrounded by hyperenhancement corresponds to region of no-reflow (arrow) within acute infarction. This region can be dis- tinguished from normal myocardium because it is encompassed in 3D space by hyperenhanced myocardium or left ventricular cavity and by the fact that it slowly becomes hyperenhanced over time. Reprinted with permission from Kim et al.50 provides higher image quality and delineates regions with more infarction that akinesia or dyskinesia is expected if infarction profound microvascular damage (Figure 9).50 involves Ն20% of the wall thickness. Evaluation by DE-MRI in The ability to simultaneously visualize nonviable and viable humans, however, suggests that a threshold phenomenon does myocardium provides additional advantages. For example, DE- not exist.56,57 These data suggest that it is unwise to extrapolate MRI can accurately assess ventricular remodeling after acute MI the results of Lieberman et al,55 who did not consider the effects at an early time point before measurements of ventricular of stunning or ongoing ischemia, to humans with MI who may volumes, internal dimensions, and ventricular mass have not have residual stunning, ischemia, or hibernation. Additional changed. This is possible because DE-MRI can assess serially, studies will be needed in these controversial topics. concurrent directionally opposite changes such as resorption of infarcted tissue and hypertrophy of viable myocardium.51 Ischemia When only viable myocardium can be visualized, the percent- There are a variety of MR techniques that can be used to age of viability in a given segment is assessed indirectly and detect myocardial ischemia. Whereas coronary MRA can generally refers to the amount of viability in the segment provide detail concerning anatomy, stress testing with imag- normalized to the segment with the maximum amount of ing of myocardial contraction or perfusion can provide viability or to data from a gender-specific database of controls. information concerning the presence and functional signifi- Conversely, when both viable and infarcted myocardium can be cance of coronary lesions. Dobutamine MR to detect ische- visualized, the percentage of viability can be assessed directly and expressed as the amount of viability in the segment normal- ized to the amount of viability plus infarction in the same segment (Figure 10a). These differences in the way in which viability is measured can alter clinical interpretation. Figure 10b demonstrates MR images in a patient with chronic coronary disease and an akinetic anterior wall. Although the anterior wall is thinned, only a small subendocardial portion of the anterior wall is infarcted. In this case, the indirect method would show that the anterior wall is only 39% viable (compared with the remote region), whereas the direct method would show that the anterior wall is 70% viable. The indirect method would predict no recovery of wall motion after revascularization, whereas the direct method would predict recovery. The postrevascularization images (bottom right of Figure 10b) demonstrate in this patient that the direct method is correct. The ability of DE-MRI to directly visualize the transmural extent of infarction (and viability) has led to some recent observations that appear to refute certain traditional concepts regarding cardiac pathophysiology. For example, prior studies Figure 10. a, Cartoon highlighting differences between direct indicate that in patients with coronary disease and ventricular and indirect method of quantifying regional viability. Viable myo- dysfunction, regions with thinned myocardium represent scar cardium is black, and infarct is white. “Remote” zone represents segment with maximum amount of viability. b, Long-axis MR tissue and cannot improve in contractile function after coronary images of patient before and 2 months after revascularization. revascularization.52 The patient example in Figure 10, along with Although akinetic anterior wall is “thinned” (diastolic wall thick- data from an ongoing pilot study,53 however, suggest that ness 5 mm; remote zone 9 mm), DE-MRI demonstrates that there is only subendocardial infarction (1.5 mm thick). Direct thinning should not be equated with the lack of viability and that assessment of viability would show that anterior wall is predom- in some patients, these regions can improve after revasculariza- inately viable (3.5/5 mmϭ70% viable), whereas indirect method tion.54 Likewise, it is commonly assumed that a threshold would show that anterior wall is predominately nonviable ϭ phenomenon exists between the transmural extent of infarction (3.5/9 mm 39% viable). Cine views after revascularization demon- strate recovery of wall motion and diastolic wall thickness. Full- and systolic wall thickening. This assumption is based on results motion movies can be viewed in the online-only Data Supplement by Lieberman et al,55 who demonstrated in a dog model of acute (select Movie IV). Modified with permission from Kim et al.54 142 Circulation July 5, 2005 mia-induced wall-motion abnormalities is an established technique for the diagnosis of coronary disease. It yields higher diagnostic accuracy than dobutamine echocardiogra- phy58 and can be effective in patients not suited for echocar- diography because of poor acoustic windows.59 Since the publication of these studies, MR image quality has improved with the widespread availability of SSFP imaging. Parallel imaging techniques that use spatial information from arrays of radiofrequency detector coils to accelerate imaging are Figure 11. Short-axis view of stress perfusion (a), rest perfusion expected to improve image quality further. Nonetheless, (b), function (c), and delayed enhancement (d) in patient with left anterior descending coronary artery disease. Perfusion images logistic issues regarding patient safety and adequate monitor- were acquired with saturation recovery gradient-echo sequence ing are nontrivial matters that require thorough planning and with parallel imaging acceleration. Note large anteroseptal perfu- experienced personnel. sion defect (arrow) is present only at stress. Wall motion and Currently, stress perfusion MR is less established for clinical delayed enhancement are normal. Full-motion movies can be viewed in the online-only Data Supplement (select Movie V). application. There are convincing data that correlate MR indexes of perfusion with tissue perfusion in animal models60,61 and excellent correlations with radionuclide imaging and invasive disease. On this point, it should be noted that perfusion imaging x-ray angiography in humans.62–64 However, the published data is quite demanding in terms of scanner hardware. Images are Ϸ so far do not demonstrate the feasibility of stress perfusion MR acquired in 100 ms rather than built up over several cardiac for everyday clinical use. The current studies are limited in that cycles, which is the case for conventional cine and DE-MRI they were either retrospective (for patient enrollment and data imaging. Thus, the signal-to-noise ratio is substantially lower for analysis), required central venous catheters, imaged only 1 to 2 perfusion imaging, and even with the latest improvements, slices per heartbeat, or excluded patients with diabetes. Addi- artifacts can obscure diagnosis. One of our current strategies to tionally, most studies required extensive interactive postprocess- improve the accuracy of MR for the detection of coronary ing after data collection, which reduces the applicability of this disease is to incorporate rest perfusion (performed 15 minutes technique for routine clinical diagnosis. after stress perfusion) and DE-MRI findings with the stress Despite these limitations, we speculate that stress perfusion perfusion results in a proscribed manner. This algorithmic MR will not only become a routine clinical procedure but also approach (Figure 12) is based on the assumption that DE-MRI is the dominant stress MR modality in the future. Perfusion MR is the most sensitive and specific MR technique for the detection of promising for a number of reasons. Decreased perfusion is the MI and that hyperenhancement patterns on DE-MRI can be first step in the ischemic cascade. Techniques that assess accurately classified as ischemic or nonischemic.65,66 Conceptu- perfusion have the potential to be more sensitive than techniques ally, it then follows that perfusion defects that have similar that assess later steps. Regarding logistics, stress perfusion intensity and extent during both stress and rest (“fixed defect”) imaging is quick and simple. For example, we perform adeno- but do not have hyperenhancement (no infarct) are artifactual sine stress imaging as follows. After cine imaging, the patient and should not be considered as caused by coronary disease. table is pulled out partially to allow full access to the patient; This approach needs to be tested in large prospective clinical adenosine is then infused at 140 ␮g·kgϪ1 · minϪ1 for 2 minutes. trials. We anticipate that clinical MR examinations will become At this time, the perfusion sequence is applied, which automat- increasingly comprehensive (eg, coronary MRA, cine imaging, ically centers the patient back into the scanner and commences image acquisition. Gadolinium contrast followed by a saline flush is infused rapidly by a peripheral vein at this time as well. On the console, real-time updates of myocardial perfusion images are shown as the images are acquired. Once the gado- linium bolus has transited the left ventricular myocardium, the adenosine is stopped, and imaging is completed. The patient table then can be pulled back out of the scanner bore if necessary. The total time of imaging for stress perfusion is 30 to 45 seconds, and the total time of adenosine infusion is Ͻ3 minutes. The pulse sequences used for stress perfusion imaging are undergoing rapid evolution. SSFP and parallel imaging tech- niques continue to improve image quality. These improvements are expected to allow quick visual interpretation of the perfusion images for routine clinical diagnosis (Figure 11). Moreover, there is no reason to interpret the stress perfusion images in isolation. We expect that a multiprotocol approach with incor- Figure 12. Flow chart for potential algorithm incorporating poration of cine and DE-MRI results with the perfusion findings stress and rest perfusion MR along with delayed enhancement imaging (DE-MRI) for detection of coronary disease (CAD). *Pos- will not only provide a comprehensive cardiac evaluation but itive DE-MRI study would be presence of hyperenhancement in also improve the accuracy of MR for the detection of coronary ischemic pattern.65,66 Fuster and Kim Frontiers in Cardiovascular Magnetic Resonance 143 stress and rest perfusion, and DE-MRI) in the near future. We 13. Ruehm SG, Goyen M, Barkhausen J, et al. Rapid magnetic resonance foresee that effort will be needed not only to improve the angiography for detection of atherosclerosis. Lancet. 2001;357: 1086–1091. imaging technology but to categorize and understand the discor- 14. Kim WY, Danias PG, Stuber M, et al. Coronary magnetic resonance dant results that may occur among the different MR protocols for angiography for the detection of coronary stenoses. N Engl J Med. a given patient. 2001;345:1863–1869. 15. Choudhury RP, Fuster V, Badimon JJ, et al. MRI and characterization of atherosclerotic plaque: emerging applications and molecular imaging. Summary Arterioscler Thromb Vasc Biol. 2002;22:1065–1074. Cardiovascular MR encompasses a variety of different tech- 16. Yuan C, Mitsumori LM, Ferguson MS, et al. In vivo accuracy of multi- niques that provide a comprehensive evaluation of the range spectral magnetic resonance imaging for identifying lipid-rich necrotic of cardiovascular disorders. Atherothrombosis throughout the cores and intraplaque hemorrhage in advanced human carotid plaques. Circulation. 2001;104:2051–2056. vascular system can be directly imaged, quantified, and 17. Corti R, Fuster V, Fayad ZA, et al. Lipid lowering by simvastatin induces characterized according to plaque components. By providing regression of human atherosclerotic lesions: two years’ follow-up by information about vascular blood flow concurrent with vas- high-resolution noninvasive magnetic resonance imaging. Circulation. 2002;106:2884–2887. cular anatomy, the functional significance of stenotic lesions 18. Zhao XQ, Yuan C, Hatsukami TS, et al. Effects of prolonged intensive can be determined. The morphology and function of the lipid-lowering therapy on the characteristics of carotid atherosclerotic cardiac system can be viewed in exquisite detail that rivals plaques in vivo by MRI: a case-control study. Arterioscler Thromb Vasc any other imaging modality. Pathophysiological processes Biol. 2001;21:1623–1629. 19. Yuan C, Zhang SH, Polissar NL, et al. Identification of fibrous cap such as MI and ischemia, stunning and hibernation, and rupture with magnetic resonance imaging is highly associated with recent scarring and fibrosis can be identified easily using quick and transient ischemic attack or stroke. Circulation. 2002;105:181–185. simple protocols. Published studies, however, describe a variety 20. Fayad ZA, Nahar T, Fallon JT, et al. In vivo MR evaluation of athero- of different pulse sequences and protocols for similar imaging sclerotic plaques in the human thoracic aorta: a comparison with TEE. Circulation. 2000;101:2503–2509. purposes and consist of relatively small numbers of patients. The 21. Jaffer FA, O’Donnell CJ, Larson MG, et al. Age and sex distribution of goals of future investigation will be to refine the technology, subclinical aortic atherosclerosis: a magnetic resonance imaging exami- establish standard protocols for image acquisition and interpre- nation of the Framingham Heart Study. Arterioscler Thromb Vasc Biol. 2002;22:849–854. tation, address the issue of cost-effectiveness, and validate a 22. Coulden RA, Moss H, Graves MJ, et al. High resolution magnetic res- range of clinical applications in large-scale clinical trials. onance imaging of atherosclerosis and the response to balloon angio- plasty. Heart. 2000;83:188–191. Acknowledgments 23. Itskovich VV, Samber DD, Mani V, et al. Quantification of human atherosclerotic plaques using spatially enhanced cluster analysis of This work was supported in part by National Institutes of Health, multicontrast-weighted magnetic resonance images. Magn Reson Med. Specialized Center of Research HL54469 (Dr Fuster), NHLBI 2004;52:515–523. RO1-HL61801 (Dr Fuster), and NHLBI R01-HL64726 (Dr Kim). 24. Worthley SG, Helft G, Fuster V, et al. Noninvasive in vivo magnetic resonance imaging of experimental coronary artery lesions in a porcine References model. Circulation. 2000;101:2956–2961. 1. Fuster V, Badimon L, Badimon JJ, et al. The pathogenesis of coronary 25. Fayad ZA, Fuster V, Fallon JT, et al. Noninvasive in vivo human artery disease and the acute coronary syndromes: parts 1 and 2. N Engl coronary artery lumen and wall imaging using black-blood magnetic J Med. 1992;326:242–250;310–318. resonance imaging. Circulation. 2000;102:506–510. 2. Maseri A, Fuster V. Is there a vulnerable plaque? Circulation. 2003;107: 26. Botnar RM, Stuber M, Kissinger KV, et al. Free-breathing 3D coronary 2068–2071. MRA: the impact of “isotropic” image resolution. J Magn Reson Imaging. 3. Aronow WS, Ahn C. Prevalence of coexistence of coronary artery 2000;11:389–393. disease, peripheral arterial disease, and atherothrombotic brain infarction 27. Kim WY, Stuber M, Bornert P, et al. Three-dimensional black-blood in men and women Ͼ or ϭ 62 years of age. Am J Cardiol. 1994;74: cardiac magnetic resonance coronary vessel wall imaging detects positive 64–65. arterial remodeling in patients with nonsignificant coronary artery 4. CAPRIE Steering Committee. A randomised, blinded, trial of clopidogrel disease. Circulation. 2002;106:296–299. versus aspirin in patients at risk of ischaemic events (CAPRIE). Lancet. 28. Jaffer FA, Weissleder R. Seeing within: molecular imaging of the car- 1996;348:1329–1339. diovascular system. Circ Res. 2004;94:433–445. 5. Stary HC. Natural history and histological classification of atherosclerotic 29. Sirol M, Itskovich VV, Mani V, et al. Lipid-rich atherosclerotic plaques lesions: an update. Arterioscler Thromb Vasc Biol. 2000;20:1177–1178. detected by gadofluorine-enhanced in vivo magnetic resonance imaging. 6. Virmani R, Kolodgie FD, Burke AP, et al. Lessons from sudden coronary Circulation. 2004;109:2890–2896. death: a comprehensive morphological classification scheme for athero- 30. Wickline SA, Lanza GM. Nanotechnology for molecular imaging and sclerotic lesions. Arterioscler Thromb Vasc Biol. 2000;20:1262–1275. targeted therapy. Circulation. 2003;107:1092–1095. 7. Moreno PR, Purushothaman KR, Fuster V, et al. Intimomedial interface 31. Choudhury RP, Fuster V, Fayad ZA. Imaging atherosclerosis and damage and adventitial inflammation is increased beneath disrupted ath- thrombosis in drug discovery. Nat Drug Develop. In press. erosclerosis in the aorta: implications for plaque vulnerability. Circu- 32. Rehwald WG, Chen EL, Kim RJ, et al. Noninvasive cineangiography by lation. 2002;105:2504–2511. magnetic resonance global coherent free precession. Nat Med. 2004;10: 8. Kolodgie FD, Gold HK, Burke AP, et al. Intraplaque and progression of 545–549. coronary atheroma. N Engl J Med. 2003;349:2316–2325. 33. Hoffmann U, Globits S, Schima W, et al. Usefulness of magnetic res- 9. Falk E, Shah PK, Fuster V. Coronary plaque disruption. Circulation. onance imaging of cardiac and paracardiac masses. Am J Cardiol. 2003; 1995;92:657–671. 92:890–895. 10. Fayad ZA, Fuster V. Clinical imaging of the high-risk or vulnerable 34. Semelka RC, Kelekis NL, Thomasson D, et al. HASTE MR imaging: atherosclerotic plaque. Circ Res. 2001;89:305–316. description of technique and preliminary results in the abdomen. J Magn 11. MacNeill ED, Lowe HC, Takano M, et al. Intravascular modalities for Reson Imaging. 1996;6:698–699. detection of vulnerable plaque: current status. Arterioscler Thromb Vasc 35. Barkhausen J, Ruehm SG, Goyen M, et al. MR evaluation of ventricular Biol. 2003;23:1333–1342. function: true fast imaging with steady-state precession versus fast 12. Fayad ZA, Fuster V, Nikolaou K, et al. Computed tomography and low-angle shot cine MR imaging: feasibility study. Radiology. 2001;219: magnetic resonance imaging for noninvasive coronary angiography and 264–269. plaque imaging: current and potential future concepts. Circulation. 2002; 36. Grothues F, Smith GC, Moon JC, et al. Comparison of interstudy repro- 106:2026–2034. ducibility of cardiovascular magnetic resonance with two-dimensional 144 Circulation July 5, 2005

echocardiography in normal subjects and in patients with heart failure or 58. Nagel E, Lehmkuhl HB, Bocksch W, et al. Noninvasive diagnosis of left ventricular hypertrophy. Am J Cardiol. 2002;90:29–34. ischemia-induced wall motion abnormalities with the use of high-dose 37. Kim RJ, Fieno DS, Parrish TB, et al. Relationship of MRI delayed dobutamine stress MRI: comparison with dobutamine stress echocardi- contrast enhancement to irreversible injury, infarct age, and contractile ography. Circulation. 1999;99:763–770. function. Circulation. 1999;100:1992–2002. 59. Hundley WG, Hamilton CA, Thomas MS, et al. Utility of fast cine 38. Shan K, Constantine G, Sivananthan M, et al. Role of cardiac magnetic magnetic resonance imaging and display for the detection of myocardial resonance imaging in the assessment of myocardial viability. Circulation. ischemia in patients not well suited for second harmonic stress echocar- 2004;109:1328–1334. diography. Circulation. 1999;100:1697–1702. 39. Simonetti OP, Kim RJ, Fieno DS, et al. An improved MR imaging 60. Wilke N, Jerosch-Herold M, Wang Y, et al. Myocardial perfusion technique for the visualization of myocardial infarction. Radiology. 2001; reserve: assessment with multisection, quantitative, first-pass MR 218:215–223. imaging. Radiology. 1997;204:373–384. 40. Fieno DS, Kim RJ, Chen EL, et al. Contrast-enhanced magnetic res- 61. Klocke FJ, Simonetti OP, Judd RM, et al. Limits of detection of regional onance imaging of myocardium at risk: distinction between reversible and differences in vasodilated flow in viable myocardium by first-pass irreversible injury throughout infarct healing. J Am Coll Cardiol. 2000; magnetic resonance perfusion imaging. Circulation. 2001;104: 36:1985–1991. 2412–2416. 41. Rehwald WG, Fieno DS, Chen EL, et al. Myocardial magnetic resonance 62. Al-Saadi N, Nagel E, Gross M, et al. Noninvasive detection of myocardial imaging contrast agent concentrations after reversible and irreversible ischemia from perfusion reserve based on cardiovascular magnetic res- ischemic injury. Circulation. 2002;105:224–229. onance. Circulation. 2000;101:1379–1383. 42. Wu E, Judd RM, Vargas JD, et al. Visualisation of presence, location, and 63. Schwitter J, Nanz D, Kneifel S, et al. Assessment of myocardial perfusion transmural extent of healed Q- wave and non-Q-wave myocardial in coronary artery disease by magnetic resonance: a comparison with infarction. Lancet. 2001;357:21–28. positron emission tomography and coronary angiography. Circulation. 43. Klein C, Nekolla SG, Bengel FM, et al. Assessment of myocardial 2001;103:2230–2235. viability with contrast-enhanced magnetic resonance imaging: com- 64. Nagel E, Klein C, Paetsch I, et al. Magnetic resonance perfusion mea- parison with positron emission tomography. Circulation. 2002;105: surements for the noninvasive detection of coronary artery disease. Cir- 162–167. 44. Knuesel PR, Nanz D, Wyss C, et al. Characterization of dysfunctional culation. 2003;108:432–437. myocardium by positron emission tomography and magnetic resonance: 65. McCrohon JA, Moon JC, Prasad SK, et al. Differentiation of heart failure relation to functional outcome after revascularization. Circulation. 2003; related to dilated cardiomyopathy and coronary artery disease using 108:1095–1100. gadolinium-enhanced cardiovascular magnetic resonance. Circulation. 45. Wagner A, Mahrholdt H, Holly TA, et al. Contrast-enhanced MRI and 2003;108:54–59. routine single photon emission computed tomography (SPECT) perfusion 66. Kim RJ, Judd RM. Gadolinium-enhanced magnetic resonance imaging in imaging for detection of subendocardial myocardial infarcts: an imaging hypertrophic cardiomyopathy: in vivo imaging of the pathologic substrate study. Lancet. 2003;361:374–379. for premature cardiac death? J Am Coll Cardiol. 2003;41:1568–1572. 46. Kim RJ, Wu E, Rafael A, et al. The use of contrast-enhanced magnetic 67. Regenfus M, Ropers D, Achenbach S, et al. Diagnostic value of resonance imaging to identify reversible myocardial dysfunction. N Engl maximum intensity projections versus source images for assessment of J Med. 2000;343:1445–1453. contrast-enhanced three-dimensional breath-hold magnetic resonance 47. Choi KM, Kim RJ, Gubernikoff G, et al. Transmural extent of acute coronary angiography. Invest Radiol. 2003;38:200–206. myocardial infarction predicts long-term improvement in contractile 68. Plein S, Jones TR, Ridgway JP, Sivananthan MU. Three-dimensional function. Circulation. 2001;104:1101–1107. coronary MR angiography performed with subject-specific cardiac acqui- 48. Ricciardi MJ, Wu E, Davidson CJ, et al. Visualization of discrete micro- sition windows and motion-adapted respiratory gating. AJR Am J infarction after percutaneous coronary intervention associated with mild Roentgenol. 2003;180:505–512. creatine kinase-MB elevation. Circulation. 2001;103:2780–2783. 69. Watanabe Y, Nagayama M, Amoh Y, et al. High-resolution selective 49. Lima JA, Judd RM, Bazille A, et al. Regional heterogeneity of human three-dimensional magnetic resonance coronary angiography with myocardial infarcts demonstrated by contrast-enhanced MRI; potential navigator-echo technique: segment-by-segment evaluation of coronary mechanisms. Circulation. 1995;92:1117–1125. artery stenosis. J Magn Reson Imaging. 2002;16:238–245. 50. Kim RJ, Choi KM, Judd RM. Assessment of myocardial viability by 70. Kim WY, Danias PG, Stuber M, et al. Coronary magnetic resonance contrast enhancement. In: Higgins CB, de Roos A, eds. Cardiovascular angiography for the detection of coronary stenoses. N Engl J Med. MRI & MRA. Philadelphia, Pa: Lippincott Williams & Wilkins; 2003: 2001;345:1863–1869. 209–237. 71. Nikolaou K, Huber A, Knez A, et al. Navigator echo-based respiratory 51. Fieno DS, Hillenbrand HB, Rehwald WG, et al. Infarct resorption, com- gating for three-dimensional MR coronary angiography: reduction of scan pensatory hypertrophy, and differing patterns of ventricular remodeling time using a slice interpolation technique. J Comput Assist Tomogr. following myocardial infarctions of varying size. J Am Coll Cardiol. 2001;25:378–387. 2004;43:2124–2131. 72. Ropers D, Baum U, Pohle K, et al. Detection of coronary artery stenoses 52. Cwajg JM, Cwajg E, Nagueh SF, et al. End-diastolic wall thickness as a with thin-slice multi-detector row spiral computed tomography and mul- predictor of recovery of function in myocardial hibernation: relation to tiplanar reconstruction. Circulation. 2003;107:664–666. rest-redistribution T1-201 tomography and dobutamine stress echocardi- 73. Nieman K, Cademartiri F, Lemos PA, et al. Reliable noninvasive ography. J Am Coll Cardiol. 2000;35:1152–1161. coronary angiography with fast submillimeter multislice spiral computed 53. Shah DJ, Kim HW, Elliott M, et al. Contrast MRI predicts reverse tomography. Circulation. 2002;106:2051–2054. remodeling and contractile improvement in akinetic thinned myocardium. 74. Knez A, Becker CR, Leber A, et al. Usefulness of multislice spiral Circulation. 2003;108(suppl IV):IV-697. Abstract. computed tomography angiography for determination of coronary artery 54. Kim RJ, Shah DJ. Fundamental concepts in myocardial viability stenoses. Am J Cardiol. 2001;88:1191–1194. assessment revisited: when knowing how much is “alive” is not enough. 75. Nieman K, Oudkerk M, Rensing BJ, et al. Coronary angiography with Heart. 2004;90:137–140. multi-slice computed tomography. Lancet. 2001;357:599–603. 55. Lieberman AN, Weiss JL, Jugdutt BI, et al. Two-dimensional echocar- 76. Achenbach S, Giesler T, Ropers D, et al. Detection of coronary artery diography and infarct size: relationship of regional wall motion and stenoses by contrast-enhanced, retrospectively electrocardiographically- thickening to the extent of myocardial infarction in the dog. Circulation. Circulation 1981;63:739–746. gated, multislice spiral computed tomography. . 2001;103: 56. Mahrholdt H, Wagner A, Parker M, et al. Relationship of contractile 2535–2538. function to transmural extent of infarction in patients with chronic 77. Kuettner A, Kopp AF, Schroeder S, et al. Diagnostic accuracy of multi- coronary artery disease. J Am Coll Cardiol. 2003;42:505–512. detector computed tomography coronary angiography in patients with 57. Nelson C, McCrohon J, Khafagi F, et al. Impact of scar thickness on the angiographically proven coronary artery disease. J Am Coll Cardiol. assessment of viability using dobutamine echocardiography and thallium 2004;43:831–839. single-photon emission computed tomography: a comparison with contrast-enhanced magnetic resonance imaging. J Am Coll Cardiol. 2004; KEY WORDS: magnetic resonance imaging Ⅲ atherosclerosis Ⅲ myocardial 43:1248–1256. infarction Ⅲ coronary disease Ⅲ vasculature Special Report

A Vision for the Future Opportunities and Challenges Notes From the Director of the National Heart, Lung, and Blood Institute

Elizabeth G. Nabel, MD

t is a pleasure and a privilege to address the readership of in heart, lung, blood, and sleep diseases and that give these ICirculation in my new capacity as Director of the National creative scientists the intellectual freedom to pursue their Heart, Lung, and Blood Institute (NHLBI). Our Institute has ideas and follow them in unexpected or serendipitous direc- a long and distinguished record of scientific progress in tions. By bringing unconventional perspectives and original- cardiovascular, lung, and blood diseases and sleep disorders, ity to bear on key research questions, awardees may develop and the present transition affords an opportunity for reflection seminal theories or technologies that will propel fields for- and critical assessment of our future directions. In this first ward and facilitate the translation of discovery into treatments edition of what I hope will be many “Notes From the to improve human health. Director,” I want to share my vision for the Institute. This The Institute also will pursue funding approaches that vision is based on a fundamental set of values—excellence, make it easier for scientists to conduct interdisciplinary innovation, integrity, respect, and compassion—that will research. For instance, the National Institutes of Health (NIH) permeate all activities of the NHLBI. I believe that scientific is considering granting principal investigator status not just to discovery provides the basis for progress and that the NHLBI a single investigator, as is the norm, but to all key members is uniquely positioned to catalyze changes that must be made of a research team. Integrated reviews of NHLBI-solicited to transform our new scientific knowledge into tangible programs would take into account the melding of various improvements in health. Within this framework, let me disciplines to address the problem at hand and provide outline some priorities for the coming years—priorities that encouragement for interdisciplinary teams to evolve in both will, of course, undergo reevaluation and reformulation as we directed and unexpected ways. seek the input of our grantees, constituents, and advisors. The NHLBI Division of Intramural Research is a special program that has the resources to conduct bold, innovative, Basic Research distinctive basic and clinical research. The Division is well Basic research provides the foundation of the NHLBI port- positioned to take on high-risk, cutting-edge projects that folio and has been one of its great strengths. The typical complement work performed in the extramural community, model of investigation—research conducted by single inves- and we are committed to maintaining and nurturing this tigators or small groups of investigators on projects of their extraordinary scientific resource. own inspiration—accounts for most of the unanticipated and major scientific discoveries in this country. I believe strongly Clinical Investigations, Trials, and Networks that we must protect and nurture investigator-initiated re- Clinical research is critical if we are to translate basic search. The NHLBI will continue to invest in the most discoveries into the reality of better health. Such work is often talented scientists conducting the highest-caliber research. time consuming and inefficient, however, and is increasingly In addition to renewed support of investigator-initiated burdened by regulatory requirements. Our challenge is to research, the NHLBI must exert national leadership in cap- expand clinical research to complement the exciting basic turing research opportunities, taking risks, and developing an science discoveries, while making it more efficient and innovative and distinctive research portfolio that is science cost-effective. driven. We intend to make the most of exciting and unprec- We intend to develop a translational research agenda edented opportunities to support emerging scientific fields. supported by clinical trials, clinical networks, and clinical One approach is to develop funding mechanisms (eg, for workforce training. Clinical trials must be driven by science support of high-risk research) that encourage innovative and designed to foster evidence-based decision-making in thinkers to turn their attention to the major current challenges clinical practice. Key components should focus on increasing

From the National Heart, Lung, and Blood Institute, Bethesda, Md. Correspondence to Elizabeth G. Nabel, MD, Director, National Heart, Lung, and Blood Institute, National Institutes of Health, Bldg 10/8C103, 10 Center Dr, Bethesda, MD 20892. E-mail [email protected] This article has been copublished in the July 1, 2005, issue of SLEEP and will be copublished in the July 15, 2005, issue of Blood and the August 1, 2005, issue of the Journal of Respiratory and Critical Care Medicine. (Circulation 2005;112:145-146.) © 2005 American Heart Association, Inc. Circulation is available at http://www.circulationaha.org DOI: 10.1161/CIRCULATIONAHA.105.564252 145 146 Circulation July 5, 2005 interactions between basic and clinical investigators and committed and sustained efforts by many in our government, easing the movement of new tools from the laboratory to the health centers, and society. clinic. An infrastructure that comprises core facilities to provide clinical researchers access to sophisticated manufac- Outreach and Communication turing capacity, along with expert advice to ensure that drug Our mission extends beyond research alone; we have an obli- development regulations are observed, could expedite the gation to translate our research findings into education and translational process. dissemination programs, particularly to address the health needs The NHLBI has a rich history of developing and support- of at-risk populations in underserved communities. We will ing clinical research networks, and we plan to build on this continue to work collaboratively with our federal colleagues, strength to develop new partnerships among organized pa- including the Centers for Disease Control and Prevention (CDC) tient communities, community-based physicians, and aca- and Health Resources and Services Administration (HRSA), to demic researchers. Integrated clinical networks of academic support prevention and treatment programs. In addition, we have health centers that include community practices will enhance an unprecedented opportunity to work with the relevant profes- our ability to conduct clinical trials. Such efforts also will sional organizations that have a large stake in developing and require improved bioinformatics and clinical databases, better implementing practice guidelines and monitoring their effective- standards for clinical research protocols, and cooperation ness, and with patient advocacy groups. Education of our between patient advocacy groups and the NHLBI. Thought patients and the public at large regarding prevention and treat- also must be given to the development of metrics to improve ment of heart, lung, blood, and sleep disorders must be one of the ascertainment of clinical outcomes as well as quality our highest priorities. assessment. This research will require the tools and expertise Rates of cardiovascular disease, asthma, chronic obstructive of many fields, including those focused on health education, pulmonary disease, and blood-borne diseases are rising world- outcomes, healthcare delivery, and healthcare economics. wide, and I am committed to our involvement in global health The NHLBI also must cultivate clinical researchers who have issues. We will take this opportunity to review the NHLBI skills commensurate with the complexity and needs of our portfolio in international programs in light of changing global research enterprise. Clinicians must be trained to work in demographics and to establish priorities and goals for these interdisciplinary, team-oriented environments and must possess programs so that Institute resources are used most effectively. skills in an array of relevant disciplines, including genetics, As co-chair (with Dr Allen Spiegel, Director of the epidemiology, biostatistics, and behavioral medicine. National Institute of Diabetes and Digestive and Kidney Diseases) of the NIH Obesity Research Task Force, I am Training, Mentoring, and Education working to enhance obesity research and education across the We intend to conduct a careful review of NHLBI training NIH. My vision is to bring to the Task Force an emphasis on programs with an eye toward improving their ability to equip basic research into the mechanisms of obesity-induced car- emerging scientists with the knowledge and skills needed for diovascular and pulmonary disease development and progres- success in an ever-changing and complex research environ- sion; on clinical investigations of cardiovascular, pulmonary, ment. During these times of tight budgets, we will focus on and sleep complications of obesity; and on education pro- helping our new investigators make the transition from grams to prevent onset and progression of obesity, especially fellowships to independent faculty positions (for instance, by among our youth. Our NIH efforts will be coordinated with designing portable mentored awards that provide more flex- the Department of Health and Human Services, other federal ibility and control in pursuing their research interests). I agencies, professional societies, and consumer groups to believe strongly that skills-development programs should be achieve synergy in our efforts. I am fortunate, indeed, to be included in all program projects, specialized centers of able to draw on the many productive experiences of the research, and other large multicomponent grants. Opportuni- NHLBI in the field of obesity, as well as the Institute’s ties to develop research interests and skills should be made proven models for outreach and education, to share successful available to students at all levels, beginning with high school, approaches that might be applied at the NIH level. and should focus special attention on underrepresented groups, such as racial and ethnic minorities and individuals The Challenge from disadvantaged backgrounds. In summary, I foresee an array of opportunities to build and diversify the strengths of the NHLBI. Our challenge is to take Health Disparities the Institute to the next level of excellence. The realization of Disparities in health status constitute a significant global issue this vision will require the advice, wisdom, and efforts of and a long-standing concern of the NHLBI. Research is essential many. I look forward to working with you to achieve these to understand the contributions of genetics, health behavior, diet, goals. We are engaged in a special form of public service— socioeconomic status, culture, and environmental exposures to that is, the promotion of patient and public health. Be assured health disparities of relevance to the NHLBI and to formulate, that I will work diligently to preserve public trust in our evaluate, and disseminate intervention programs. This work will Institute, the NIH, and the biomedical research enterprise, and necessarily entail a vigorous effort to increase the representation to ensure that the NHLBI serves the public with the highest of minorities in the ranks of NHLBI researchers. A full resolu- level of integrity. I hope you will join me in this exciting tion of the health disparities problem will only occur through venture. Images in Cardiovascular Medicine

Detection of Luminal-Intimal Border and Coronary Wall Enhancement in Intravascular Ultrasound Imaging After Injection of Microbubbles and Simultaneous Sonication With Transthoracic Echocardiography

Manolis Vavuranakis, MD, FESC; Ioannis A. Kakadiaris, PhD; Sean M. O’Malley, BS; Christodoulos Stefanadis, MD, FESC; Sophia Vaina, MD; Maria Drakopoulou, MD; Ioannis Mitropoulos, MD; Stephane Carlier, MD, PhD; Morteza Naghavi, MD

61-year-old man presented with unstable angina (Braun- intracoronary ultrasound probe, an enhancement of the entire Awald class 2B). Coronary angiography revealed a mild plaque and adventitia was seen. The luminal-intimal lesion on the very proximal segment of the left anterior boundary appeared to show a ring-like enhancement, descending coronary artery (LAD) and a significant stenosis which clearly defined the inner borders of the coronary (80%) in the mid-segment. Intracoronary ultrasound was used arterial wall (Figure 2). The precise mechanism of this to further evaluate proximal coronary artery stenosis. It was observation is not clearly defined, although the adhesion of found to be a soft plaque without significant luminal stenosis microbubbles to inflamed endothelial cells was reported but without clear definition of the luminal-intimal boundary. previously. Sonication of the microbubbles by the external Intravenous injection of gas-filled microbubble ultrasound acoustic energy may facilitate adherence to endothelium contrast agents have been used for endocardial border detec- for a short time, and then may be washed out by the tion, especially when they are sonicated by acoustic power forthcoming blood. Acute coronary syndromes are the result of and produce harmonics. We performed continuous intracoro- plaque rupture or endothelial erosion in the majority of cases. nary ultrasound recordings (EndoSonics; 20 MHz) in the Therefore, techniques that can help define the integrity of the proximal left anterior descending coronary artery before luminal-intimal border and intraplaque leakage of blood through (Figure 1, A), during (Figure 1, B) and after injection (Figure the vasa vasorum or plaque cap could be of major importance for 1, C) of 4 mL of SonoVue (ultrasound contrast agent with detecting vulnerable plaque and understanding the pathophysi- lyophilized capsule filled with sulfurhexafluoride). Simulta- ology of acute coronary syndromes. neously with contrast injection, ultrasound acoustic power of 0.6 mechanical index was delivered via a transthoracic transducer (2.5 MHz) toward the left main to sonicate the Disclosure delivered microbubbles. Immediately after the passage of the Dr Naghavi is a share-holder in and consultant to Volcano Corpora- microbubble contrast agent, which was clearly detected by the tion and Endothelix Inc and is a scientific advisor to Pfizer Inc.

From Department of Cardiology, University of Athens, Hippokration Hospital, Athens, Greece (M.V., C.S., S.V., M.D., I.M.); Department of Computer Science, University of Houston, Houston, Tex (I.A.K., S.M.O.); Intravascular Imaging & Physiology, Cardiovascular Research Foundation, New York, NY (S.C.); and Association for Eradication of Heart Attack, Houston, Tex (M.N.). Correspondence to Manolis Vavuranakis, MD, for the Ultimate IVUS at University of Houston Collaborative Project, Haimanda 24-26, Marousi 15122, Greece. E-mail [email protected] (Circulation. 2005;112;e1-e2.) © 2005 American Heart Association, Inc. Circulation is available at http://www.circulationaha.org DOI: 10.1161/CIRCULATIONAHA.104.479915 e1 e2 Circulation July 5, 2005

Figure 1. A, Baseline IVUS image of a non–flow-limiting plaque (35%) Figure 2. Differential IVUS images showing the subtracted in the proximal LAD. Arrows indicate areas where the luminal-intimal postinjection signals from baseline signals. A, Black and white border is not clearly defined. B, Detection of gas-filled microbubble (signal intensity of Figure 1AϪFigure 1C); B, color-coded panel ultrasound contrast agent passage around the intracoronary ultra- A; C, thresholded to show most significant areas of sound catheter at the imaging site. C, Significant changes in the signal enhancement. intensity of the entire plaque area, including the adventitia, are observed. Note the luminal-intimal interface ring, indicating an echo- reflectant halo by microbubbles. Images in Cardiovascular Medicine

Detection of Carotid Atherosclerotic Plaque Ulceration, Calcification, and Thrombosis by Multicontrast Weighted Magnetic Resonance Imaging

Baocheng Chu, MD, PhD; Marina S. Ferguson, MT; Hunter Underhill, MD; Norihide Takaya, MD, PhD; Jianming Cai, MD, PhD; Michel Kliot, MD; Chun Yuan, PhD; Thomas S. Hatsukami, MD

62-year-old man presented to the emergency department sequences confirmed the stenosis and demonstrated ulceration Awith a chief complaint of severe headache and decreased and calcification in both carotids and mural thrombus formation vision in his left eye. Initial physical examination demonstrated in the left carotid (Figures 1 and 2). After completely recovering a new-onset left homonymous hemianopsia, which warranted a from the stroke, the patient underwent staged bilateral carotid stroke workup. The patient’s head CT was significant for a endarterectomy. Histological examination of the specimens con- 2.3ϫ3.7-cm acute hemorrhage in the right posterior parietal and firmed the MRI findings of bilateral ulceration and left mural occipital lobes. Conventional angiography was performed and thrombus formation (Figure 3). interpreted as Ͼ90% stenosis of both internal carotid arteries without ulceration. The remainder of the work-up, including Disclosure echocardiogram, was negative. An ensuing carotid magnetic Dr Kliot is cofounder of a company, UltraImage Corp, which is now resonance examination with a phased-array carotid coil and part of Pathway Medical Technologies and develops and makes MRI high-resolution (0.3ϫ0.3 mm pixel size) multicontrast weighted phase-array coils similar to those used in this article.

Figure 1. 3-D time-of-flight (TOF) image shows a surface ulcer (long arrow) in the distal right common carotid artery. Black-blood, fast-spin echo, T1-weighted (T1W), postcontrast-enhanced T1W (T1W-CE), proton density weighted (PDW), and T2-weighted (T2W) images confirm surface discontinuity. The hypointense areas on all 4 weightings correspond to calcifications (short arrows).

From the Departments of Radiology (B.C., M.S.F., H.U., N.T., J.C., C.Y.) and Surgery (M.K., T.S.H.), University of Washington, Seattle, and VA Puget Sound Health Care System (T.S.H.), Seattle, Wash. Correspondence to Baocheng Chu, MD, PhD, Department of Radiology, Box 357115, University of Washington, 1959 NE Pacific St, Seattle, WA 98195. E-mail [email protected] (Circulation. 2005;112;e3-e4.) © 2005 American Heart Association, Inc. Circulation is available at http://www.circulationaha.org DOI: 10.1161/CIRCULATIONAHA.104.494419 e3 e4 Circulation July 5, 2005

Figure 2. All weightings of the left carotid artery show a distinct ulcer (long arrows) and a small calcification (short arrows). The striking hyperintensity on the ulcer surface in the T2-weighted (T2W) image indicates the presence of a mural thrombus. Focal contrast enhance- ment on postcontrast-enhanced T1W (T1W-CE) indicates vasculature at the base of the thrombus.

Figure 3. H&E stain of right and left carotid endarterectomy specimens. The right plaque contained extensive calcifi- cations and fibrosis. A well-defined pen- etrating ulcer extends 4 mm from the lumen surface through a fibrotic matrix. A penetrating ulcer with mural thrombus formation is seen in the left carotid end- arterectomy specimen. Asterisks are placed in the lumen of the common and the internal arteries of both carotids. Location indicators are millimeter dis- tance to the bifurcation. ϩ indicates locations in the internal carotid artery; –, locations in the common carotid artery. Images in Cardiovascular Medicine

Primary Lymphoma of the Heart

Jeffrey Kuvin, MD; Nisha Parikh, MD; Robert Salomon, MD; Arthur Tischler, MD; Philip Daoust, MD; Yevgeniy Arshanskiy, MD; Karl Coyner, MD; Philip Carpino, PA; Natesa G. Pandian; Carey Kimmelstiel, MD; Caroline Foote, MD; John Erban, MD; Hassan Rastegar, MD

previously healthy 65-year-old woman presented with wall mass on the CT images. The distribution of F-18 FDG in Apalpitations and positional chest discomfort 3 weeks whole-body imaging was otherwise unremarkable, and no after she sustained chest wall trauma in a motor vehicle additional abnormal masses were identified. accident. Physical examination revealed occasional prema- The patient underwent surgery. The pericardium was thin, ture ventricular beats and low-grade fever. Her erythrocyte with a small amount of cloudy pericardial fluid. A round, firm sedimentation rate was elevated (66 mm/h). Transthoracic mass was detected, originating from the atrioventricular groove, and transesophageal echocardiography revealed a 3ϫ3-cm, attaching to the right atrium and free wall of the right ventricle, well-demarcated, homogeneous, round mass moving with the and encasing the right coronary artery (Figure 3A). The mass heart adjacent to the right atrium (Figures 1A, B). There was was excised without removing any portion of right atrium or invagination of nearby cardiac chambers but no obstruction to ventricle. The right coronary artery was preserved by peeling the tumor from the vessel after the mass was divided into 2 halves right heart filling. MRI showed a circumscribed mass with (Figure 3B). Immunophenotyping by flow cytometry and tissue dense tissue characterization (isointense to myocardium) not staining demonstrated that the malignant cells were positive for consistent with blood or fat (Figure 2A). There was minimal CD45, CD19, and CD20, and were negative for surface immu- enhancement of the mass after gadolinium injection. Coro- noglobulin. These findings were diagnostic of a diffuse large nary angiography was normal. Two weeks later, the patient’s B-cell lymphoma (Figure 4). The pericardial fluid also showed symptoms improved. Repeat transthoracic echocardiogram evidence of lymphoma. A bone marrow biopsy obtained from showed no change in size or consistency of the mass. A the sternum at time of resection revealed benign lymphoid PET/CT (combined positron emission tomography and com- nodules and a mediastinal node adjacent to the pericardium was puted tomography) scan after administration of F-18 fluoro- negative for lymphoma. The patient made an uneventful recov- deoxyglucose (FDG) revealed focal intense activity adjacent ery from surgery, received systemic chemotherapy, and 1 year to the right heart (Figure 2B), which correlated with a cardiac later remains in complete remission.

Figure 1. A, Transthoracic echocardio- gram, apical 4-chamber view. Round mass adjacent to right heart (arrow) with invagination of right atrial wall and right ventricular inflow. B, Transesophageal echocardiogram: Well-demarcated mass (arrow) adjacent to the right atrium (RA). A linear structure with color Doppler flow (dashed arrow) is noted within the mass.

From the Departments of Medicine (J.K., N.P., C.K., N.G.P., C.F., J.E.), Pathology (R.S., A.T., P.D.), Radiology (Y.A.), and Surgery (P.C., H.R.), Tufts-New England Medical Center, Tufts University School of Medicine, Boston, Mass; and the Department of Radiology (K.C.), Brigham and Women’s Hospital, Harvard Medical School, Boston, Mass. Correspondence to Jeffrey Kuvin, MD, Division of Cardiology, Department of Medicine, Tufts-New England Medical Center, 750 Washington St, Box 315, Boston, MA 02111. E-mail [email protected] (Circulation. 2005;112:e5-e6.) © 2005 American Heart Association, Inc. Circulation is available at http://www.circulationaha.org DOI: 10.1161/CIRCULATIONAHA.104.495135 e5 e6 Circulation July 5, 2005

Figure 2. A, MRI, coronal multislice sin- gle shot. Round mass in the region of the right atrioventricular groove (arrow), isointense to the myocardium. B, F-18 FDG PET imaging, coronal view. Intense focal FDG activity (arrow) adjacent to the right atrium correlates with the round mass shown on the accompanying MRI image.

Figure 4. Microscopic pathology (hematoxylin-eosin stain, origi- nal magnification ϫ300) reveals packed sheets of large cells with round, lobulated nuclear contours, fine punctate chromatin, and distinct nucleoli. Frequent apoptotic cells, moderate number of mitoses, and areas of coagulation necrosis were present (not all represented in this image). Immunostaining for CD20 (insert) demonstrates a large B-cell lymphoma.

Figure 3. A, Intraoperative photograph of the well-circumscribed epicardial mass adherent to the right heart (arrow), enveloping the right coronary artery (dashed arrow). This vessel corre- sponds to the color Doppler signal noted on echocardiography in Figure 1B. B, Excised lesion (cut into halves) shows a smooth outer surface; the inner surface is indented (arrows) from pas- sage of the right coronary artery. Book Review

Computed Tomography of the Coronary Arteries examination including coronary imaging, is more demanding at Pim J. de Feyter, MD, PhD, and Gabriel P. Krestin, MD, the image-acquisition stage. The book then covers coronary imaging, beginning with PhD, eds. normal coronary anatomy (Chapter 3) and a review of the 208 pp. London, UK: Taylor & Francis; 2005. $85.00. histological basis of coronary disease with illustrations by CT ISBN 1-84184-439-X (Chapter 4). Chapters 5 through 12 progress through topics Nearly 5 decades have passed since F. Mason Sones, MD, systematically including stenosis/occlusion, plaque imaging, performed the first coronary angiogram. Despite transformative stent patency and collateral flow assessment, imaging of coro- advances in the treatment of coronary atherosclerosis, this nary anomalies, and bypass-graft imaging. There is also a chapter traditional procedure remains the foundation on which state-of- that reviews the data relevant to coronary calcification (Chapter the-art diagnosis and management rests. Noninvasive imaging 7) and the Agatston calcium scoring algorithm. Although a modalities such as scintigraphy and echocardiography have calcium score can be obtained with multidetector CT, typically, contributed immeasurably to the care of the patient with coronary an electron-beam CT system (EBCT) is used for this test. disease, but these techniques only indirectly determine coronary Noncoronary applications of CT, including evaluation of cardiac patency. During the past few years, imaging methods that masses and imaging of the great vessels, are discussed in approach the accuracy of x-ray angiography have become in- Chapters 13 to 16. There is an invaluable review of artifacts and creasingly available with the advent of the latest generation of their manifestations (Chapter 18), beautifully illustrated in sharp, MRI and multidetector/multislice CT scanners. clear examples. Chapter 19, which reviews the timing techniques Although these 2 modalities use a common vocabulary, during the iodinated contrast bolus, seems better placed at the considerably more attention has been focused recently on CT, beginning of the book because it reviews the technical aspects of primarily because of its strikingly clear image quality. Not to be image acquisition. The text concludes with a brief discussion of left ventricular function assessment and presents the editors’ overlooked are the economic implications of CT in terms of perspective on the future of coronary CT, including a few reimbursement and control of technology. The arrival of coro- 64-slice studies. nary CT has opened a new front in the seemingly perpetual to The straightforward explanations of potentially complex con- and fro between the cardiology and radiology camps, with battles cepts and striking illustrations of volume-rendered and multipla- now being fought at the state and national regulatory levels. nar reconstructed images, nearly always presented with the Coverage of coronary CT in the lay press and aggressive correlative gold standard, x-ray angiography, make this an marketing of CT imaging technology to cardiologists and pa- excellent foundation book. Furthermore, the crisp cartoon-like tients only serves to increase tension. A recently posted animated illustrations reiterate points made in the accompanying text and advertisement on The New York Times web site beckons patients reinforce the concepts. to “ask their doctor” about the latest-generation CT scanners that Multidetector CT technology has rapidly evolved during the can scan the heart in 8 to 10 seconds and provide an instant past 4 years from 4- to 8- to 16- and now 64-slice devices, each diagnosis. Widespread awareness necessitates that the general capable of imaging the coronary arteries in less time. The cardiologist familiarize him- or herself with this new imaging majority of examples in this book, and in the literature, are taken language. In fact, it is now necessary to understand the funda- with 16-slice systems. Although individual institutions with mentals of MRI and CT imaging to critically review the access to the newest systems (including the editors of this text) contemporary literature. have reported high sensitivities/specificities as compared with It is in this context that Drs de Feyter and Krestin have crafted cardiac catheterization for the detection of significant coronary an exquisitely illustrated and clearly written text entitled Com- stenoses, the rapid pace of change has not permitted large-scale, puted Tomography of the Coronary Arteries. This is neither a multicenter demonstrations of efficacy. In spite of the potential technical manual nor an exhaustive reference text, but rather a for CT to be used as a screening tool for low- to intermediate-risk concise overview of CT technique for the general cardiologist or patients, nearly all of the published studies have examined non–CT-educated specialist in cardiac imaging. The text reflects patients with a high clinical suspicion of coronary artery disease. the collaboration between a radiologist (Dr Krestin) and cardi- Data from an intermediate-risk population for which CT may be ologist (Dr de Feyter) from Erasmus Medical Center, Rotterdam, most beneficial have yet to be reported. The editors do include a The Netherlands, and is representative of the importance of chapter entitled “The Emergency Department” (Chapter 14), collaboration across specialties. Well aware of the turf wars this which seems to suggest that coronary CT has a role in the technology has incited, the editors note in the preface that this evaluation of patients with chest pain in the emergency depart- book would not have been written “without the faithful and ment; however, the chapter reviews the data supporting stratifi- friendly cooperation between cardiologists and radiologists in cation of risk based on CT-determined calcium score and not our institution.” angiography. CT angiography may prove useful in this setting The book is organized as a comprehensive introductory text, and the high negative predictive value (Ͼ97%) reported with with 20 concise chapters. The first 2 chapters cover the basic 16-slice devices does suggest that CT may be good at screening. principles of CT imaging and specific relevant aspects of image The text presents CT angiography as a safe procedure with processing. There is less emphasis on CT technique and more on minimal risk as compared with x-ray angiography, and in the image interpretation. This is appropriate, for in practice, CT acute setting, this is undoubtedly true; however, a risk–benefit technology has advanced to a degree that the acquisition of data analysis must also include consideration of radiation exposure is largely automated. In contrast, cardiovascular MRI (CMR), because with 16- and 64-slice CT, the dose may be at least 4-fold which has the potential to provide a comprehensive functional greater than diagnostic x-ray angiography. The long-term risk

(Circulation. 2005;112:e7-e8.) © 2005 American Heart Association, Inc. Circulation is available at http://www.circulationaha.org DOI: 10.1161/CIRCULATIONAHA.105.553347 e7 e8 Book Review associated with this level of exposure, perhaps multiplied by viously unrecognized and clinically silent stenoses. Although we repeats scans, is undefined. are potentially nearing a time when a patient’s coronary anatomy That being said, noninvasive coronary imaging by CT is likely can be defined in Ͻ10 seconds, we as cardiologists will need to here to stay. Its importance to the future of cardiology is understand the applications and limitations of CT techniques. evidenced by the preparation of Core Cardiology Training Computed Tomography of the Coronary Arteries is a concise and Symposium (COCATS) guidelines to specify fundamental and easy-to-follow overview of the relevant concepts of CT coronary advanced training requirements in cardiology fellowship pro- angiography. The authors and editors should be commended for grams. Whoever ultimately reads CT scans, and this reviewer producing this exceptional introduction to a complex yet prom- strongly advocates a cooperative model as put forth in Computed Tomography of the Coronary Arteries, interpretation of the scan ising technique. results in the context of the patient’s overall clinical presentation Frederick L. Ruberg, MD will be essential. The general cardiologist will need to integrate Section of Cardiology imaging information with functional stress testing. Interestingly, Boston University School of Medicine interventional cardiologists may receive more referrals for pre- Boston, Mass Correspondence

Letter Regarding Article by McNair et al, “SCN5A 2. Groenewegen WA, Firouzi M, Bezzina CR, Vliex S, van Langen IM, Mutation Associated With Dilated Cardiomyopathy, Sandkuijl L, Smits JP, Hulsbeek M, Rook MB, Jongsma HJ, Wilde AA. A cardiac sodium channel mutation cosegregates with a rare Conduction Disorder, and Arrhythmia” connexin40 genotype in familial atrial standstill. Circ Res. 2003;92: To the Editor: 14–22. In their article on the SCN5A mutation D1275N, McNair et al1 3. Bezzina CR, Rook MB, Groenewegen WA, Herfst LJ, van der Wal claim that the mutation is the primary cause of a familial form of AC, Lam J, Jongsma HJ, Wilde AA, Mannens MM. Compound het- dilated cardiomyopathy (DCM) with conduction disorder and su- erozygosity for mutations (W156X and R225W) in SCN5A associated praventricular arrhythmias. Although DCM has been reported with with severe cardiac conduction disturbances and degenerative changes both SCN5A and other ion channel mutations, McNair et al claim in the conduction system. Circ Res. 2003;92:159–168. their family is remarkable because “there is a strong correlation between the penetrance of a conduction disorder and the manifes- tation of dilation.” Response A close look at their data reveals that ventricular dysfunc- Groenewegen and Wilde question the strength of the “link” tion (shortening fraction Ͻ28%) is primarily present in the between the dilated cardiomyopathy (DCM) phenotype and the older members of the family, 4 of 5 also with atrial fibrilla- SCN5A mutation (D1275N) reported in our article.1 In their tion. In only one of these individuals is the ventricular study, they proposed that the cardiac phenotype (atrial standstill) dimension given, and this dimension is within normal limits. was caused by the concurrence of the same mutation (D1275N) In our article on the same SCN5A mutation in a family with with a connexin40 polymorphism.2 atrial standstill2 (co-inherited with polymorphisms in the In addition to the 6 subjects with ventricular dilation and/or atrial-specific gap junction channel protein Connexin40), only dysfunction in the original article by Olson et al,3 we provided one atrial standstill patient showed left atrial enlargement and evidence for ventricular dilation and/or dysfunction in 3 other none had DCM. subjects (III:4, III:10, IV:2), as shown in our table.1 The On the basis of the data provided, there is no doubt that the autopsy report of subject II:1 confirmed “4-chamber dilation.” conduction disorders, the various atrial arrhythmias, and the Furthermore, shortly after the publication of our article, Olson associated (bi-) atrial dilatation are linked to the SCN5A et al4 reported the results of their survey of the SCN5A gene mutation. These features (and stroke at young age) largely in a large DCM population (including the same family), citing determine the proposed score.1 This phenotype is part of the additional cases in which SCN5A mutations cosegregate with plethora of described sodium channelopathies, which in more the DCM phenotype. These data, together with the previous extensively affected patients may also include the develop- identification of incidental DCM cases in LQT3 and HERG ment of fibrosis.3 The link to DCM by this single mutation, mutation carriers,1 and the recent identification of ABCC9 which in expression studies performed thus far does not mutations (SUR2A subunit of the cardiac KATP channel) in a significantly affect sodium channel function, has not been DCM population,5 strongly suggest a link between ion chan- proven, however. First, a LOD score for linkage between nel mutations and myocardial dysfunction. patients with ventricular dysfunction and D1275N and/or Groenewegen points out that if the phenotype of DCM (a D3S1211 allele 1 (also presumably D1275N carriers, although late-onset finding) is used in isolation, then only a modest this has not been proven) is estimated to be at best Ϸ1.2 (ie, LOD score could be obtained. We argue that it is inappropri- an Ϸ10% chance of a false-positive result), and is therefore ate to score only patients with DCM as affected when the far less than the LOD score of 3 usually required for statistical SCN5A mutation in this family produces a complex phenotype evidence of genetic linkage. Second, DCM is not present in that includes arrhythmias and conduction disease. His sugges- the above-described family with the same mutation.2 The tion to use a LOD score of 3.0 in a candidate gene study is an predominant presence in older adult patients favors DCM to incorrect approach (that criterion applies to genome-wide be secondary to long-lasting atrial arrhythmias. Hence, these approaches); also, in a candidate approach, a LOD score of 1.2 data do not permit the listing of “mild” SCN5A mutations as is closer to a 1%, not a 10%, false-positive rate. This is of DCM causing genetic aberrations. Alternatively, another gene little consequence, in fact, because Olson et al recently with reduced penetrance, in linkage disequilibrium with the reported a LOD score of 5.8 in this family.4 The lack of DCM D1275N mutation, plays a role in the DCM phenotype. in the family reported by Groenwegen and associates carrying the identical mutation,2 who were younger at the time of Acknowledgment reporting (late teens to 33 years in their study versus 24 to 88 Supported in part by the Netherlands Heart Foundation years in our study), does not disprove the DCM hypothesis (M96.001). because additional genetic or environmental factors could account for the differences between the 2 families. Whether W.A. Groenewegen, PhD the observed late-onset DCM is a direct consequence of the Department of Medical Physiology underlying SCN5A mutation or a reflection of chronic arrhyth- University Medical Center Utrecht mia remains to be determined. Utrecht, The Netherlands William P. McNair, BA A.A.M. Wilde, MD, PhD Lisa Ku, MS Department of Cardiology Matthew R.G. Taylor, MD Academic Medical Center Pam R. Fain, PhD Amsterdam, The Netherlands Eugene Wolfel, MD 1. McNair WP, Ku L, Taylor MR, Fain PR, Dao D, Wolfel E, Mestroni L; Luisa Mestroni, MD Familial Cardiomyopathy Registry Research Group. SCN5A mutation University of Colorado Cardiovascular Institute associated with dilated cardiomyopathy, conduction disorder, and ar- University of Colorado Health Sciences Center rhythmia. Circulation. 2004;110:2163–2167. Denver, Colo e9 e10 Correspondence

Note A cardiac sodium channel mutation cosegregates with a rare connexin40 Please note that one of the original authors of our article,1 Dmi Dao, genotype in familial atrial standstill. Circ Res. 2003;92:14–22. 3. Olson TM, Keating MT. Mapping a cardiomyopathy locus to chro- could not be reached to provide her consent to this response; mosome 3p22-p25. J Clin Invest. 1996;97:528–532. therefore, she has not been listed as an author on this response. 4. Olson TM, Michels VV, Ballew JD, Reyna SP, Karst ML, Herron KI, Horton SC, Rodeheffer RJ, Anderson JL. Sodium channel mutations and suscepti- 1. McNair WP, Ku L, Taylor MR, Fain PR, Dao D, Wolfel E, Mestroni L; bility ot heart failure and atrial fibrillation. JAMA. 2005;293:447–454. Familial Cardiomyopathy Registry Research Group. SCN5A mutation 5. Bienengraeber M, Olson TM, Selivanon VA, Kathmann EC, O’Cochlain associated with dilated cardiomyopathy, conduction disorder, and ar- F, Gao F, Karger AB, Ballew JD, Hodgson DM, Zingman LV, Pang Y-P, rhythmia. Circulation. 2004;110:2163–2167. Alekseev AE, Terzic A. ABCC9 mutations identified in human dilated

2. Groenewegen WA, Firouzi M, Bezzina CR, Vliex S, van Langen IM, cardiomyopathy disrupt catalytic KATP channel gating. Nat Genet. 2004; Sandkuijl L, Smits JP, Hulsbeek M, Rook MB, Jongsma HJ, Wilde AA. 36:382–387. Correspondence

Letter Regarding Article by Galbreath et al, We also agree with Wilson and Linden that disease management will “Long-Term Healthcare and Cost Outcomes of have the greatest impact when there is substantial room for improve- Disease Management in a Large, Randomized, ment in therapies being applied as well as in compliance with those Community-Based Population With Heart Failure” therapies. We were impressed with the level of treatment at baseline in our population and concluded that the lack of reduction in healthcare To the Editor: utilization and costs resulted in part from that fact. All of the patients in We commend Dr Galbreath and colleagues on a well-designed study this study had an established diagnosis of congestive heart failure and evaluating the impact of a disease management program on health almost all had primary care physicians at the time of study entry. This status, utilization, and cost.1 There is, however, one extremely important differs from many trials of disease management in which study entry point that requires comment. The authors indicate that the results of this occurred during the initial hospitalization for heart failure in patients study are generalizable because the “catchment” area spanned 70 000 who may have had little previous access to health care. It is important mi2 in south Texas. This study has strong “internal validity,” but we to point out, however, that because of the nature of the intervention we would argue that the “external validity”2 of the findings will apply only tested, a disease management program that provided feedback and to similar people, settings, treatments, and outcomes.3 Such a “gener- alizable” population would be one with a high level of baseline recommendations to patients and primary care physicians but did not compliance with recommended treatment guidelines. In this study, the have independent prescriptive authority, medication doses were not authors state that 77% of the patients were taking either an angioten- routinely maximized and blood pressure control at the end of the trial sin-converting enzyme inhibitor or angiotensin receptor blockers at the was often suboptimal. Thus, there remains further room for improve- time of enrollment, as compared with the 30% to 50% rate noted in ment in the management of congestive heart failure in our patients, but other trials. This is not a trivial issue because drug therapy is a central the disease management program we tested was not successful in part of recommended guidelines.3 The authors correctly noted that bringing it about. under circumstances of high adherence to guidelines “the value added To realize this ultimate goal, requiring the participation of large by a disease management program will be more difficult to groups of patients in adjunct programs such as the one we tested will not demonstrate.” suffice. The significant challenge remains in developing systems that This factor alone may account for why no reductions in utilization or allow more complete control of patient management in such large costs were realized in the study group when compared with the groups, so that treatment end points are achieved without interfering reference. Disease management by its very name implies that partici- with a patient’s relationship with a primary care provider who must pants will be guided toward improving control of their condition. The manage the patient’s overall health care (often encompassing numerous mainstay of these programs is in bringing the individual and his or her medical problems) and who must concur with specific therapy recom- physician in line with evidence-based practice guidelines. If individuals mendations of the disease management program. Only when this goal already adhere to self-management behaviors, then there is little gain to has been reached will truly optimal disease management be possible. be expected from a disease management program. Disclosure Ariel Linden, DrPH, MS The Department of Defense funded this study; its interests were Linden Consulting Group Hillsboro, Ore in determining whether disease management when applied to [email protected] federal beneficiaries with congestive heart failure would be effective in reducing mortality, morbidity, and healthcare costs. Thomas Wilson, DrPH, PhD If significant improvements over usual care were observed, then Wilson Research, LLC this or a similar model would have been instituted on a local and Loveland, Ohio perhaps national level in military treatment facilities. [email protected] Autumn Dawn Galbreath, MD 1. Galbreath AD, Krasuski RA, Smith B, Stajduhar KC, Kwan MD, Ellis R, Gregory L. Freeman, MD Freeman GL. Long-term healthcare and cost outcomes of disease man- Division of Cardiology agement in a large, randomized, community-based population with heart University of Texas Health Science Center failure. Circulation. 2004;110:3518–3526. University of Texas Disease Management Center 2. Wilson TW, MacDowell M. Framework for assessing causality in disease man- San Antonio, Tex agement programs: principles. Dis Manag. 2003;6:143–158. 3. Linden A, Adams JL, Roberts N. Generalizing disease management program Brad Smith, PhD results: how to get from here to there. Manag Care Interface. 2004;17:38–45. University of Texas Disease Management Center Altarum Institute Response San Antonio, Tex We agree with Drs Wilson and Linden that before applying a study’s results to one’s own patient population, the similarity between the Richard A. Krasuski, MD patient groups must be carefully considered. Unlike previous studies Division of Cardiology that have evaluated the impact of disease management strategies in Wilford Hall Medical Center small populations of patients from a single hospital or health mainte- San Antonio, Tex nance organization, we studied a large group of patients recruited from Karl C. Stajduhar, MD a wide spectrum of healthcare systems and demographics. We enrolled Michael D. Kwan, MD patients from large cities (San Antonio, Austin), from small towns Division of Cardiology (Beeville, Three Rivers), and from isolated rural settings in Texas. Brooke Army Medical Center Although it could be argued that a ranch in south Texas is different from San Antonio, Tex a dairy farm in New England, we again emphasize that this was a broadly heterogeneous population from a range of backgrounds. Thus, Robert Ellis, MD we feel justified in stating that our results are more generally applicable Tricare Southwest than those of less diverse previous studies. San Antonio, Tex

e11 Acknowledgment of Reviewers

The Editors express appreciation to the following referees who served from April 1, 2004, to December 31, 2004.

Einari Aavik Peter Alexandersen David Antoniucci Robert J. Bache Nicola Abate Khaled Alfakih Jovan P. Antovic Jean E. Bachet Amr E. Abbas Ottavio R. Alfieri Charles Antzelevitch Walter Backes Antonio Abbate Francois Alhenc-Gelas Piero Anversa Peter Backx Kevin C. Abbott Ziad A. Ali L.J. Appel Larry Baddour Robert D. Abbott Etienne M. Aliot Ian Appleton David Badesch Koji Abe Hussein R. Al-Khalidi Andrew E. Arai Juan J. Badimon E. Dale Abel Lindsey D. Allan Maria Rosario G. Araneta Lina Badimon George S. Abela Yves Allemann Stephen L. Archer Cornel Badorff Benjamin S. Abella Maurits A. Allessie Moshe Arditi Stephen F. Badylak Aiden Abidov Kristina Allikmets Ross Arena Alexei Y. Bagrov Alexandre Abizaid Matthew A. Allison Thomas Arentz Ajay Bahl M. Roselle Abraham Thomas G. Allison Laurent Argaud Ferdinand H. Bahlmann Pierre Abraham Kevin C. Allman Gary C. Armitage Colin Baigent Theodore P. Abraham Laura Almasy Paul W. Armstrong Steven Bailey Dan Abramov Jesus Almendral Donna K. Arnett Donald S. Baim Charles S. Abrams Carlos Alonso-Villaverde Leonard F. Arnolda Alison E. Baird Jerome L. Abramson Joseph S. Alpert Wilbert S. Aronow S. Paul Bajaj Hugues Abriel Martin A. Alpert Marie Arsenault Arvind Bakhru Stephan Achenbach David A. Alter Margaret A. Arstall Patricia F. Bakker Michael Acker Peter Alter Michael Artman Stephan J.L. Bakker Michael J. Ackerman Guy Alvarez Yujiro Asada George L. Bakris Stamatis Adamopoulos John A. Ambrose Takayuki Asahara Prabhakaran Balagopal Robert J. Adams Pierre Ambrosi Masanori Asakrua Stephan Baldus Volker Adams Peter Ammann Raimondo Ascione Christie M. Ballantyne Ian Adatia Ezra Amsterdam Alexzander Asea Jean-Luc Balligand Philip A. Ades Ping An Arlene S. Ash Scott W. Ballinger Jennifer Adgey Inder S. Anand Euan A. Ashley Enzo Ballotta Marina Afanasyeva G.M. Anantharamaiah Muhammad Ashraf James A. Balschi Vahid Afshar-Kharghan Burt Anderson Nick Ashton Ko Bando Stefan Agewall Gitte Andersen Gregory K. Asimakis José R. Banegas Tetsuro Ago H. Vernon Anderson Samuel J. Asirvatham Mary A. Banerji Piergiuseppe Agostoni Jeffrey L. Anderson Gerd Assmann Adrian P. Banning Pietro Agricola Kelley P. Anderson Birgit Assmus David A. Baran David Aguilar Mark E. Anderson Brad Astor Eddy Barasch Seyedhossein Aharinejad Neil Anderson Bela F. Asztalos Giuseppe Barbaro Amrita Ahluwalia Page A.W. Anderson Dan Atar John C. Barbato Ismayil Ahmet Peter G. Anderson Vasilios G. Athyros Paule C. Barbeau Enrico Aidala Robert H. Anderson Dianne L. Atkins Jean T. Barbey Ken-ichi Aihara Todd J. Anderson Larry D. Atwood Robert Bard Masanori Aikawa Stefan Andreas Andrew M. Atz Edit Bardi Barbara Ainsworth Arne K. Andreassen Johann W. Auer Philip M. Barger William C. Aird Felicita Andreotti Pal Aukrust Stephen G. E. Barker Masazumi Akahoshi Douglas Andres Angelo Auricchio S. Serge Barold Olakunle O. Akinboboye Vicente Andres Gerard P. Aurigemma Alain D. Baron Masahiro Akishita Ramaroson Andriantsitohaina Melissa A. Austin Lili A. Barouch Junya Ako Rajesh K. Aneja Richard C. Austin Jose A. Barrabes Christine M. Albert Giovanni Anfossi Michael V. Autieri Elizabeth Barrett-Connor Jeffrey Albert Annalisa Angelini Pablo Avanzas William H. Barry Michelle Albert Gianni D. Angelini Abraham Aviv Robyn J. Barst Gabriel Aldea Dominick J. Angiolillo Philip E. Aylward Thomas Bartel Michael H. Alderman E. Angles-Cano Michel Azizi Philip Barter Alexey N. Aleshin Stefan D. Anker Angelo Azzi Jürgen Barth Marie-Christine Alessi Brian H. Annex Jan Baan John R. Bartholomew John H. Alexander Jack Ansell Vladimir R. Babaev Matthias Barton M. Yvonne Alexander Riitta L. Antikainen Gerard Babatasi Benico Barzilai Mark E. Alexander Tarek F. Antonios Fritz H. Bach Theodore A. Bass e12 Acknowledgment of Reviewers e13

Jean-Pierre Bassand Lars Berglund Peter F. Bodary Sorin J. Brener Craig T. Basson Martin W. Bergmann Christoph Bode Kate M. Brett Shari S. Bassuk Steven R. Bergmann William E. Boden Christoph Brezinka Eric R. Bates Bradford C. Berk Gerd Bodlaj J. Timothy Bricker Anjan Batra Lisa Berkman Manfred Boehm Charles R. Bridges Robert Bauernschmitt Javier Bermejo Michael Boehm Regina Brigelius-Flohe Johann Bauersachs Jose M. Bernal Jolanda M. Boer Michele Brignole Kenneth L. Baughman Michael C. Berndt Guido Boerrigter Ralph G. Brindis Iris Baumgartner Gerald J. Berry Eric Boersma Charles Brink Ralf W. Baumgartner Mark F. Berry Eric Boerwinkle Eliot A. Brinton William Baumgartner Donald M. Bers Rainer H. Boger Michael R. Bristow Christophe Bauters Giuseppe S. Berton Richard G. Bogle Steven L. Britton Jeroen J. Bax Alain G. Bertoni Frank Bogun Bruce R. Brodie B. Timothy Baxter Michel E. Bertrand David F. Bohr Sergey V. Brodsky Gary F. Baxter Charles I. Berul William Boisvert Ulrich Broeckel Kirk W. Beach Patricia J.M. Best Joerg Bojunga Alessandra Brofferio M. Flint Beal Reinaldo B. Bestetti Paula M. Bokesch Robert D. Brook Alvarez Beatriz Christer Betsholtz Thomas Boland Maria M. Brooks Jean-Louis Beaudeux Saroja Bharati Roberto Bolli James M. Brophy Christoph R. Becker Aruni Bhatnagar Victoria Bolotina M. Julia Brosnan Diane M. Becker Deepak Bhatnagar Marvin O. Boluyt K. Bridget Brosnihan Lance B. Becker Deepak L. Bhatt Massimo Bonacchi Brigitta C. Brott Lewis C. Becker Italo Biaggioni Raoul Bonan Margaret E. Brousseau Richard C. Becker Cesario Bianchi Nikolaos Bonaros Jan-Leendert P. Brouwer Frank Beckers Marco E. Bianchi Lawrence I. Bonchek Gregory L. Brower Joshua A. Beckman Giorgio M. Biasi Meredith Bond B. Greg Brown Bettina Beech Hans K. Biesalski Diana Bonderman David L. Brown Juerg H. Beer D.H. Biesma Hendrik Bonnemeier Kathleen K. Brown Philipp Beerbaum Erik A.L. Biessen Enzo Bonora Nancy J. Brown Richard H. Behrman J. Thomas Bigger Robert O. Bonow Warren S. Browner Berthold Bein Nick R. Bijsterveld Maria R. Bonsignore Michael Brownlee Sean C. Beinart Diane E. Bild Piet W. Boonstra Ian N. Bruce Alexa Beiser Jacques Billette Jeffrey Boord Christian Bruch Romualdo Belardinelli George Billman George W. Booz Josep Brugada Robert M. Bell Feng B. Bin Nicolas Borenstein Ramon Brugada Jonathan N. Bella Philip F. Binkley Jeffrey S. Borer Lyndia Brumback George A. Beller John D. Birkmeyer Martin Borggrefe Corinna B. Brunckhorst Michelle P. Bendeck Yochai Birnbaum John Boscardin Frank Martin Brunkhorst David G. Benditt Eva Biro René M. Botnar Eric J. Brunner Martin Bendszus Nanette H. Bishopric Michiel L. Bots Hans R. Brunner Athanase Benetos John A. Bittl Chantal M. Boulanger Helle Bruunsgaard R. Benezra Vera Bittner Anne Bouloumie Robert M. Bryan Frank M. Bengel Edward O. Bixler Henri M. Bounameaux Michael Bryer-Ash Jean-Pierre Bénitah Daniel J. Blackman Pierre Boutouyrie Rosemary S. Bubien Ivor J. Benjamin Eugene H. Blackstone Daniel F. Bowen-Pope Paolo Bucciarelli Ralf Benndorf Steven N. Blair Neil E. Bowles Timothy G. Buchman Joel S. Bennett Jean-Jacques Blanc Penelope A. Boyden John Buckwalter Martin R. Bennett William S. Blaner Mark R. Boyett Matthew J. Budoff William M. Bennett Stefan Blankenberg Biykem Bozkurt Werner Budts Rondelet Benoit W. Matthijs Blankesteijn David J. Bradley Arno Buecker Neal Benowitz Andrew D. Blann T. Douglas Bradley Raffaele Bugiardini D. Woodrow Benson Alex Blatt Randy W. Braith L. Maximilian Buja Lee N. Benson Andrew D. Blaufox Frederick Brancati Jens Bulow Merrill D. Benson Erwin Blessing Ralf P. Brandes Burkhard D. Bültmann Aloys Berg Peter C. Block Ruediger C. Braun-Dullaeus Cecil M. Burchfiel Robert A. Berg David A. Bluemke Eugene Braunwald Lora Burke Knut E. Berge Friedrich C. Blumberg Alan C. Braverman John C. Burnett Alan K. Berger Roger S. Blumenthal Molly S. Bray Jane C. Burns Peter B. Berger Alex Bobik Claudia Bregonzio Paul B. Burton Ronald Berger Edimar A. Bocchi Roger E. Breitbart Ivo Buschmann Rudolf Berger Jorge B. Boczkowski Ole A. Breithardt David W. Busija e14 Acknowledgment of Reviewers

Rudi Busse Paola Casanello Jina Choo Domenico Corrado Javed Butler Juan P. Casas Tz-Chong Chou Javier Corral Alfred E. Buxton Wayne E. Cascio Robin P. Choudhury Dalila Corry Brian F. Buxton Ivan P. Casserly Benjamin J.W. Chow Alberto Corsini Peter H. Byers Lisa A. Cassis Judith C. Chow Marshal A. Corson Benjamin F. Byrd, III Agustin Castellanos Timothy F. Christian Roberto Corti Melissa J. Byrne Edoardo Castelli Mina Chung Joseph S. Coselli Christopher H. Cabell Alessandro Cataliotti Taylor Chung Francesco Cosentino Candido Cabo Marco Cattaneo Mariantonietta Cicoira Francisco G. Cosio Howard Cabral David S. Celermajer Marilyn J. Cipolla Marco A. Costa Kevin S. Cahill Bojan Cercek Francesco Cipollone Lisa C. Costello-Boerrigter Hua L. Cai Antonio Ceriello Kieran Clarke William G. Cotts Michael E. Cain Manuel Cerqueira Robert Clarke Paul J.M. Coucke Paolo Calabro Matteo Cesari William R. Clarke Thierry Couffinhal Antonio M. Calafiore Claudia U. Chae Catherine M. Clase David Couper S.C. Calaghan Mohamed Chahine John G.F. Cleland Adrian Covic Angelino Calderone Alan Chait Paula R. Clemen Dermot Cox Mary Caldwell Bernard R. Chaitman Ton J. Cleophas Jill P. Crandall David A. Calhoun Aravinda Chakravarti Aldo Clerico Michael H. Crawford Robert M. Califf Lorraine Chalifour Angela Clerk Sybil Crawford Hugh Calkins Angel Chamorro Alexander Clowes Filippo Crea David J. Callans Hunter C. Champion William T. Clusin Mark A. Creager Francois A. Cambien Bysani Chandrasekar Ronald I. Clyman Harry J.G.M. Crijns Richard P. Cambria Anthony C. Chang Andrew J. Coats Janet B. Croft Paolo G. Camici Chih-Jen Chang Leonard A. Cobb Kevin D. Croft Vito M. Campese Ruey-Kang R. Chang William A. Coetzee Rachelle H. Crosbie Umberto Campia Kevin S. Channer Thomas M. Coffman Carroll E. Cross Antonio C. Campos de Keith M. Channon David J. Cohen John R. Crouse Carvalho M. John Chapman Eric A. Cohen Richard S. Crow Paul Canner Nora M. Chapman Hillel W. Cohen J. Kennedy Cruickshank Christopher P. Cannon Israel F. Charo Howard A. Cohen Bruce F. Culleton Richard O. Cannon John C. Chatham Marc Cohen Juraj Culman Charles E. Canter Lee-Young Chau Michael V. Cohen R. William Currie John G. Canto Sarwat Chaudhry Richard A. Cohen Jesse W. Currier Warren J. Cantor Gautam Chaudhuri Jay N. Cohn Anne B. Curtis John M. Canty Aurea J. Chaves Lawrence H. Cohn Mary Cushman Noel M. Caplice Melvin D. Cheitlin William E. Cohn Daniele M. Cusi Maurizo C. Capogrossi Alex F. Chen Steven D. Colan Jeffrey A. Cutler Riccardo Cappato Horng H. Chen Jamie Coleman Maria-Cristina Cuturi Thomas P. Cappola Ian Y. Chen Désiré Collen Myron I. Cybulsky Alessandro M. Capponi Jiu-an Chen Barry S. Coller Tillmann Cyrus Joseph A. Caprini Peng-Sheng Chen Robert W. Colman Sébastien Czernichow Sonia Caprio Shih-Ann Chen Antonio Colombo Mat J.A.P. Daemen Alessandro Capucci Edouard Cheneau Maria Giovanna Colombo Michael Daffertshofer Massimo Caputo Debbie Cheng David Colquhoun Ralph B. D’Agostino Blase A. Carabello James H. Chesebro Catherine Communal Mohamed R. Daha Brian J. Carey Mordechai Chevion Gianluigi Condorelli Hiroyuki Daida Stephane G. Carlier Derek P. Chew William E. Connor James E. Dalen Mark D. Carlson Elena Chiarpotto Robert V. Considine Ronald L. Dalman Edward Carmeliet Masaaki Chiku C. Richard Conti Edward R. Damiano Pizzi Carmine John S. Child Elena Conti Patricia D’Amore Mercedes R. Carnethon William M. Chilian David J. Cook Nicolas Danchin Robert M. Carney Kazuo Chin George A. Cook Paresh Dandona Clara Carpeggiani Michael T. Chin Nancy R. Cook Haim D. Danenberg Oscar A. Carretero Marcello Chinali John P. Cooke George Dangas John D. Carroll Randolph W. Chitwood, Jr Joshua M. Cooper Peter G. Danias Joseph Carrozza Ray C-J Chiu Leslie T. Cooper Werner G. Daniel Jeffrey L. Carson Leslie Cho Mark E. Cooper Jean-Marie Daniel Lamaziere Andrew J. Carter Aram V. Chobanian Richard S. Cooper Stephen R. Daniels Lori Carter-Edwards Anand Chockalingam Josef Coresh A.H. Jan Danser Wayne Carver Eric T. Choi James Coromilas Victor Darley-Usmar Acknowledgment of Reviewers e15

Anthony M. Dart Ranjan Deka Torsten Doenst Mark J. Eisenberg Dipak K. Das Jacqueline M. Dekker Pieter A. Doevendans Mickey S. Eisenberg Undurti N. Das Federica del Monte Hisham Dokainish Graeme Eisenhofer Jean C. Daubert Etienne Delacrétaz Anna F. Dominiczak Daniel T. Eitzman Harold L. Dauerman Joris Delanghe William T. Donahoo Mikael Elam Alan Daugherty Claude Delcayre J. Kevin Donahue John A. Elefteriades Anthony P. Davenport Christophe Delclaux Rosario Donato Michael T. Eliasziw Tirone E. David Thomas E. Delea Peter Doris Uri Elkayam Sandra T. Davidge Jose A. Delgado Alves Gerald W. Dorn, II Kenneth A. Ellenbogen Charles J. Davidson E. Patchen Dellinger David E. Dostal Myrvin Ellestad Karina W. Davidson R. Phillip Dellinger Pamela S. Douglas Oyvind Ellingsen Barry R. Davis Louis J. Dell’Italia James M. Downey Patrick Ellinor Patricia H. Davis Kevin C. Dellsperger Stephen W. Downing Justine A. Ellis Roger A. Davis Mario Delmar Ramona Doyle Stephen G. Ellis Roger B. Davis Michael D. Delp Kathleen Dracup Kristin E. Ellison Russell C. Davis Anthony N. DeMaria Luciano F. Drager Nabil El-Sherif Robin L. Davisson Yi Deng Mark H. Drazner Mary Emond Vedat Davutoglu Donald R. Dengel Wim Drenthen Noriaki Emoto Kevin P. Davy Christophe Depre Helmut Drexler Masao Endoh Dana Dawson Dimitri Deserranno Daniel L. Dries Matthias Endres Jonathan R.S. Day Alessandro Desideri Ahmed Ben Driss Richard M. Engelman Sharlene M. Day Christopher A. DeSouza Jie Du David M. Engman Jacques de Bakker George Despotis Terry W. Du Clos Gunnar Engstrom Fred De Beer Jean-Pierre Despres Dayue Duan Andrew E. Epstein Bernard De Bruyne Zeruesenay Desta Raghvendra K. Dubey Frederick H. Epstein Raffaele De Caterina Anita DeStefano Anne M. Dubin Laurence M. Epstein Ulf de Faires Robert C. Detrano Anique Ducharme Stephen E. Epstein Pim J. de Feyter Katherine M. Detre Samuel C. Dudley Raimund R. Erbel Giovanni de Gaetano Tobias Deuse Stephen J. Duffy John M. Erikson Peter de Jaegere Mohan Devbhandari Dirk J. Duncker Einar E. Eriksson Paul E. de Jong Richard B. Devereux Daniel Duprez Wolfgang Erl Marlies de Lange Mieke Dewerchin Jocelyn Dupuis Thomas P. Erlinger Peter W. de Leeuw Marc Dewey Josee Dupuis Thomas Eschenhagen James A. de Lemos Mandeep Dhadly David T. Durack Charles T. Esmon Michel de Lorgeril Naranjan S. Dhalla Carlos M.G. Duran Ricardo J. Esper Giuseppe De Luca Stefan Dhein William Durante Christine Espinola-Klein Moniek P.M. de Maat Gerald F. Di Bona J. Larry Durstine Katherine Esposito Ebo D. de Muinck Marcelo F. Di Carli Firat Duru N.A. Mark Estes Ramon de Nooijer Carlo Di Mario Susan Duval Zeev Estrov Anne De Paepe Mario Di Napoli Kim A. Eagle Masato Eto Giovanni De Pergola Pietro Di Pasquale Elaine D. Eaker Tanenao Eto Dominique de Prost Michaela Diamant Robert T. Eberhardt Paulo Roberto B. Evora Giovanni de Simone George A. Diamond Franz R. Eberli Justin A. Ezekowitz Johan H. De Sutter David A. Dichek Lynn E. Eberly Michael D. Ezekowitz Robbert J. de Winter Wolfgang Dichtl Steven N. Ebert John W. Fabre Dick de Zeeuw Kenneth Dickstein Shah Ebrahim Rosalind Fabunmi Barbara J. Deal Sean P. Didion Dwain L. Eckberg Gianpaolo Fadini John E. Deanfield Andre M. Diedrich Robert H. Eckel Erling Falk Arjun Deb Jutta Dierkes Jay M. Edelberg Rodney H. Falk Robert F. DeBusk Javier Diez Elazer R. Edelman Bonita Falkner G. William Dec Wolfgang H. Dillmann Robert R. Edelman James C. Fang William M. DeCampli Joseph S. Dillon Thomas S. Edgington James I. Fann Jeanne M. DeCara Vasken Dilsizian L. Henry Edmunds Frank Faraci Robert S. Decker John P. DiMarco Igor R. Efimov Andrew Farb Gordon R. DeFoe Stefanie Dimmeler Kensuke Egashira Harrison W. Farber Curt G. DeGroff Anh Tuan Dinh-Xuan Satoru Eguchi Jawed Fareed Gregory J. Dehmer Donald J. DiPette Marek P. Ehrlich William B. Farquhar Jan T. Deichen Michael Diringer Benjamin W. Eidem Vladimir G. Fast Elisabeth Deindl Abhinav Diwan John F. Eidt Khalil Fattouch Carolin Deiner Sanjay Dixit John W. Eikelboom Elda Favari Elisabetta Dejana Douglas W. Dockery Howard J. Eisen David P. Faxon e16 Acknowledgment of Reviewers

William P. Fay John S. Floras Mingui Fu A. Martin Gerdes Zahi A. Fayad James Floyd Shmuel Fuchs Marie D. Gerhard-Herman Jocelyne Fayn Martin Fluck Bianca Fuhrman Lucie Germain Franz Fazekas Joseph Flynn Masashi Fujita Guido Germano Sergio Fazio Robert Fogel Naomi Fukagawa Bernard J. Gersh William F. Fearon Alan M. Fogelman Tohru Fukai M. Eric Gershwin Paul W.M. Fedak Antonio F. Folino Keiichi Fukuda Myron C. Gerson Eugenia Fedoseyeva Franco Folli Kesiuke Fukuo Welton M. Gersony Jeffrey A. Feinstein Gregg C. Fonarow Pino Fundarò Edward P. Gerstenfeld Mary F. Feitosa Guo-Hua Fong John W. Funder Robert E. Gerszten Robert Felder Ignatius W. Fong Colin D. Funk S. David Gertz G. Michael Felker Vivian A. Fonseca Curt D. Furberg Leonard S. Gettes Michael P. Feneley JoAnne M. Foody Raffaello Furlan Tal Geva Qingping Feng Thomas Force Mark I. Furman Henry Gewirtz Peter Ferdinandy Earl S. Ford Masato Furuhashi Michael Gewitz James J. Ferguson, III George D. Ford Valentin Fuster Jalal K. Ghali T. Bruce Ferguson Myriam Fornage William H. Gaasch William A. Ghali Francisco Fernandez-Aviles James S. Forrester Christian Gachet Mihai Gheorghiade Jose M. Fernandez-Real Ulrich Forstermann David R. Gagnon Lorenzo Ghiadoni Olivier Feron Trudy M. Forte James V. Gainer Hossein A. Ghofrani Markus Ferrari Elyse Foster Fiorenzo Gaita Carlo Giansante Victor A. Ferrari F. Gerald Fowkes Patrick J. Gallagher Gary H. Gibbons Victor A. Ferraris Caroline S. Fox Claudio Galli Raymond J. Gibbons Paolo Ferrazzi Ervin Fox Augusto Gallino C. Michael Gibson Robert E. Ferrell Alain Fraisse James M. Galloway Samuel S. Gidding Andreas Festa Mark W. Frampton Jonas B. Galper Stephan Gielen Stephan Fichtlscherer Silvia Franceschi Apoor S. Gami Martine Gilard Anette Fiebeler Charles W. Francis Santhi Ganesh Ian C. Gilchrist Loren J. Field Gary S. Francis Peter Ganz Thomas D. Giles David S. Fieno Veronica Franco Mario J. Garcia Wayne R. Giles Michael A. Fifer Dignat-George Françoise Julius M. Gardin Linda D. Gillam Hans R. Figulla Nikolaos G. Frangogiannis Helena M. Gardiner Jonathan Gillard Janos G. Filep Markus H. Frank Sheila M. Gardiner A. Marc Gillinov Antonio P. Filipe, Jr. Stanley S. Franklin David G. Gardner Anne M. Gillis Jeffrey R. Fineman Wayne J. Franklin Martin J. Gardner Matthew W. Gillman Mitchell S. Finkel Michael R. Franz Roy S. Gardner Richard F. Gillum Toren Finkel Robert P. Frantz Alan Garfinkel Robert F. Gilmour Dianne M. Finkelstein Stefan Frantz Philippe Garot Larry C. Gilstrap Marcus Fischer Maria Grazia Franzosi Peter Garred Frank J. Giordano Thorsten Fischer Nancy Frasure-Smith Jean-Michel T. Gaspoz Domenico Girelli Richard D. Fish Robert W.M. Frater Michael A. Gatzoulis Cynthia J. Girman Michael C. Fishbein David S. Freedman Kimberlee Gauvreau Anselm K. Gitt Edward A. Fisher Bruce A. Freeman Timothy P. Gavin Dario Giugliano John D. Fisher Balz Frei Haralambos P. Gavras Gregory R. Giugliano John T. Fisher John K. French Irene Gavras Alexandre Giusti-Paiva Patrick W. Fisher Michael P. Frenneaux Meinrad Gawaz Michael M. Givertz Steven A. Fisher Ulrich H. Frey Steffen Gay David Gjertson Glenn I. Fishman Matthias Frick J. William Gaynor Mark T. Gladwin Desmond J. Fitzgerald Linda F. Fried J. Michael Gaziano Stanton A. Glantz Garret A. FitzGerald Susan K. Fried Raul J. Gazmuri Stephen P. Glasser Peter J. Fitzgerald R.P. Friedland Carmine Gazzaruso Stephen J. Glatt Frank A. Flachskampf Deborah M. Friedman Robert L. Geggel Nicola Glorioso Greg C. Flaker Paul A. Friedman Bruce D. Gelb Donald D. Glower Scott D. Flamm Jefferson Frisbee Jacques Genest Charles J. Glueck Marcus D. Flather Victor Froelicher Yong-Jian Geng Robert J. Glynn Jerome L. Fleg Jiri J. Frohlich Alfred L. George Alan S. Go Kirsten E. Fleischmann Alberto Froio Sarah J. George Ulrich Göbel Ingrid Fleming Peter C. Frommelt Demetrios Georgiou Andrew Gogbashian Richard M. Fleming Andrea Frustaci Alexander Geppert Noyan Gokce Markus Flesch Robert L. Frye Lior Gepstein Diane R. Gold Gerald F. Fletcher Ryan M. Fryer Hanspeter Gerber Jeffrey P. Gold Acknowledgment of Reviewers e17

Michael R. Gold Eugene A. Grossi Anthony J. Hanley Keith Henry Ira J. Goldberg Ehud Grossman James A. Hanley Timothy D. Henry Ronald B. Goldberg Blair P. Grubb Edward L. Hannan Heike M. Hermanns Jeffrey J. Goldberger Eberhard Grube William H. Hansen Ramon C. Hermida Sherita H. Golden Scott M. Grundy Goran K. Hansson Adrian F. Hernandez Ira D. Goldfine Gary L. Grunkemeier Akiyoshi Hara Miika Hernelahti Joshua I. Goldhaber Eliseo Guallar Shuntaro Hara Victoria L.M. Herrera Samuel Z. Goldhaber Maurizio D. Guazzi Joshua M. Hare David M. Herrington Lee Goldman Vilmundur G. Gudnason Robert A. Harrington Howard C. Herrmann Pascal J. Goldschmidt Peter G. Guerra William S. Harris Ray E. Hershberger Steven R. Goldsmith Gerard M. Guiraudon David G. Harrison Charles A. Herzog Larry B. Goldstein Martha Gulati Alison L. Harte David C. Hess Michael S. Goligorsky Giosue Gulli Paula J. Harvey Otto M. Hess Paolo Golino Hakan Gullu Rodrigo Hasbun Gerd F. Heusch Jonathan Golledge Julian Gunn David Hasdai Karsten Heusser Celso E. Gomez-Sanchez Mahesh P. Gupta Naoyuki Hasebe Steven B. Heymsfield Philimon Gona Narendra K. Gupta Gerd Hasenfuss William R. Hiatt Mario D. Gonzalez Paul A. Gurbel J. Michael Hasenkam Yukihito Higashi Elizabeth Goodman Enrique P. Gurfinkel Naotake Hashimoto Charles B. Higgins Lawrence T. Goodnough Geoffrey C. Gurtner Paul M. Hassoun Denise Hilfiker-Kleiner Theodore A. Gooley Swaminatha V. Gurudevan Thomas S. Hatsukami Joseph A. Hill John Gorcsan Arjun Gururaj Daniel C. Hatton Gerhard Hindricks Neil Gordon Matthias Gutberlet Richard N.W. Hauer Thomas H. Hintze Joel M. Gore David D. Gutterman Paul J. Hauptman Shuji Hinuma Tommaso Gori Przemyslaw Guzik Elizabeth R. Hauser Masayasu Hiraoka Mark W. Gorman Tomasz J. Guzik Richard J. Havel Loren F. Hiratzka Robert C. Gorman Stefano Guzzetti Axel Haverich Karen K. Hirschi Shinya Goto Katrina Gwin-Hardy Edward P. Havranek Valeria Hirschler Roberta A. Gottlieb Donald C. Haas Robert A. Haworth John W. Hirshfeld Stephen S. Gottlieb Felix Haas Ilan Hay Keiichi Hishikawa Antonio M. Gotto Helmut Habazettl Nissim Hay Mark A. Hlatky K. Lance Gould Robert H. Habib Junichiro Hayano Donald Hnatowich Luis Henrique W. Gowdak Rory Hachamovitch David L. Hayes Carolyn Ho Kristof Graf Walter E. Haefeli Sharonne N. Hayes Helen H. Hobbs Patricia M. Grambsch Judith Haendeler Daniel Hayoz Didier Hober Juan F. Granada Hans U. Haering Stanley L. Hazen Robert W. Hobson Christopher B. Granger Steven M. Haffner Jiang He Judith S. Hochman D. Neil Granger David E. Haines Ka He Hanoch Hod Augustus O. Grant William D. Haire Tongrong He Julien I. Hoffman Henk L. Granzier Michel Haissaguerre Anthony M. Heagerty Udo Hoffmann Guido Grassi Roger J. Hajjar Harvey S. Hecht Peter Höglund David Gray Charles A. Hales Susan R. Heckbert Thomas Hohlfeld Paul A. Grayburn Kathleen J. Haley Peter S. Heeger Stefan H. Hohnloser J. Thomas Grayston Michael E. Halkos Timothy Heeren Brian D. Hoit Daniel J. Green Jennifer L. Hall Christopher Heeschen John E. Hokanson Darren C. Greenwood Par Hallberg Linda J. Heffner Fernando Holguin Edward W. Gregg Hermann Haller Robert A. Hegele Judd E. Hollander Michelle Ann Grenier Jozsef Haller Paul A. Heidenreich Morley Hollenberg Kathy K. Griendling Kevin J. Hallock Jörg Heierhorst Thomas A. Holly Brian P. Griffin Perry V. Halushka Albert Heim William L. Holman Helen R. Griffiths Rainer Hambrecht Robert J. Heine David R. Holmes Clarence E. Grim Mohamed H. Hamdan Gerardo Heiss Paul Holvoet Cindy L. Grines Pavel Hamet Alan W. Heldman Michael Holzer Steven K. Grinspoon John A. Hamilton Gary V. Heller Shunichi Homma Jean Ann Grisso Christian W. Hamm Joan Heller Brown Myeong-Ki Hong Francine Grodstein H. Kirk Hammond Barbara L. Hempstead Yuling Hong William J. Groh Zhong C. Han Robert Henderson Rocio S. Honigmann Marie-Louise M. Gronholdt Wayne W. Hancock Marc Hendrikx Jane L. Hoover-Plow Robert Gropler Diane E. Handy Peter Henke Richard Hopkins Garrett J. Gross Claude E. Hanet Charles H. Hennekens William E. Hopkins Oliver Gross Graeme J. Hankey Michael Hennerici Uta C. Hoppe e18 Acknowledgment of Reviewers

Masatsugu Hori Masahiro Ito Shuichi Jono Philipp A. Kaufmann Lisa K. Hornberger Wulf D. Ito Jens Jordan Sanjay Kaul Benjamin D. Horne Toshiyuki Itoi Jacob Joseph Sanjiv Kaul Burkhard Hornig Susan L. Ivey Mark E. Josephson Koji Kawahito John D. Horowitz D. Dunbar Ivy Kaumudi J. Joshipura Chuichi Kawai Lawrence D. Horwitz Yuichi Iwaki Janna Journeycake David M. Kaye Steven R. Houser Tohru Izumi Pekka Jousilahti Teruhisa Kazui Barbara V. Howard Christopher L. Jackson Aleksandar Jovanovic Elsadig Kazzam George Howard EdwinK. Jackson Ian R. Jowsey Mark T. Keating T. Howard Howell Graham Jackson Michael J. Joyner Craig A. Keebler Henry H. Hsia Shaun P. Jackson Bodh I. Jugdutt Beate E. Kehrel Frank B. Hu Alice K. Jacobs J. Wouter Jukema Aaron S. Kelly Peifeng Hu David R. Jacobs David N. Juurlink Daniel P. Kelly Paul L. Huang Donald W. Jacobsen Stefan Kaab Ralph A. Kelly Sally A. Huber Paul Jacques Jens J. Kaden Malte Kelm Whady A. Hueb Tazeen H. Jafar Alan H. Kadish Anita M. Kelsey Joerg Huelsken Allan S. Jaffe Yutaka Kagaya Byron W. Kemper Chris C. Hughes Thomas Jahnke Henry S. Kahn Richard D. Kenagy David Y. Hui Mukesh K. Jain Richard Kahn Thomas A. Kent Heikki V. Huikuri Rajan Jain Fumihiko Kajiya Richard E. Kerber P.P. Hujoel Pierre Jaïs Gabor Kaley Dean J. Kereiakes Russell D. Hull Jose Jalife Christoph Kalka Karl B. Kern Per M. Humpert Jorge E. Jalil Klaus Kallenbach Morton J. Kern Karin H. Humphries Ik-Kyung Jang Jonathan M. Kalman William S. Kerwin Stephen E. Humphries Joseph S. Janicki Lalit Kalra Steven J. Keteyian Thomas Hund Warren R. Janowitz Grzegorz L. Kaluza Paul Khairy Joseph Hung Michiel J. Janse Balaraman Kalyanaraman Bijoy Khandheria Kelly J. Hunt Ian Janssen Vaijinath S. Kamanna Ashwani Khanna Steven C. Hunt Stefan P. Janssens Timothy J. Kamp Stefan Kiechl Patrick R. Hunziker Craig T. January Junji Kanda Jan T. Kielstein Winston L. Hutchinson James L. Januzzi Takeshi Kanda Shinji Kihara Adolph M. Hutter Rudolf Jarai David E. Kandzari Tatsuya Kiji Guido Iaccarino Mikko J. Jarvisalo Laura B. Kane Dae Jung Kim Mark Iafrati Patrick Y. Jay Nicholas Kang Hyo-Soo Kim Sahoko Ichihara Goy Jean-Jacques William B. Kannel InKyeom Kim Raymond E. Ideker David J. Jenkins Norman M. Kaplan Jason K. Kim Richard G. Ijzerman Rolf Jenni Tomas Kara Raymond J. Kim Uichi Ikeda Allen Jeremias Richard H. Karas Thomas R. Kimball Katsunori Ikewaki Michael Jerosch-Herold Johan Karlberg Stephen E. Kimmel John S. Ikonomidis Paula Jerrard-Dunne Joel S. Karliner Carey D. Kimmelstiel Sabino Iliceto Xavier Jeunemaitre Morris Karmazyn Akinori Kimura Armin Imhof Ashish K. Jha Aly Karsan George L. King Akihiro Inazu Ishwarlal Jialal Karl R. Karsch Spencer B. King Sandro Inchiostro Canwen Jiang Soji Kasayama John G. Kingma Ciro Indolfi Huang Jianhua Carlos S. Kase Scott Kinlay Julie R. Ingelfinger Bernd Jilma Juan-Carlos Kaski Ulrich Kintscher David A. Ingram, Jr. Hanjoong Jo Edward K. Kasper Kevin E. Kip Joanne S. Ingwall Mark A. Jobling David A. Kass Charles J.H.J. Kirchhof Nobutaka Inoue Edward J. Johns Robert S. Kass James Kirklin Teruo Inoue B. Delia Johnson Ghassan Kassab Lorrie A. Kirshenbaum Cecilia Invitti Bruce D. Johnson John J.P. Kastelein Chiharu Kishimoto D.P. Inwald Jason L. Johnson Adnan Kastrati Brett M. Kissela Dan-Dominic G. Ionescu Richard J. Johnson Naoto Katakami Toru Kita Kaikobad J. Irani Robert L. Johnson Sekar Kathiresan Akira Kitabatake Mitsuhiro Isaka Daniel W. Jones Masahiko Kato Masafumi Kitakaze Shun Ishibashi Gregory T. Jones Tomohiro Katsuya Kazuo Kitamura Junnichi Ishii Peter L. Jones Hugo A. Katus Richard N. Kitsis Kikuo Isoda Robert H. Jones Zvonimir S. Katusic Andre G. Kleber Eric M. Isselbacher Steven P. Jones Arnold M. Katz Robert Kleemann Takaaki Isshiki W. Keith Jones Stuart D. Katz Neal S. Kleiman Hiroshi Ito Habo J. Jongsma Marc P. Kaufman Allan L. Klein Acknowledgment of Reviewers e19

George J. Klein Irving L. Kron Torben B. Larsen Jerrold H. Levy Lloyd W. Klein Marvin W. Kronenberg Warren K. Laskey Wilbur Y. Lew Paul D. Kligfield Itzhak Kronzon Robert D. Lasley Martin M. LeWinter James R. Klinger Florian Krotz Larry A. Latson Alan B. Lewis Elizabeth S. Klings Paul Kubes Jo-Dee L. Lattimore Klaus F. Ley Francis J. Klocke Nils Kucher Joseph Lau Andrew C. Li Robert A. Kloner Karl-Heinz Kuck Wei C. Lau Jennifer S. Li Bradley P. Knight Nino Kuenzli Michael S. Lauer Jian Li John L. Knight Hartmut Kuhn Ulrich Laufs Jianhua Li Anne A. Knowlton Michaela Kuhn Jari A. Laukkanen Na Li Kirk U. Knowlton Helena Kuivaniemi Stephane Laurent Shengxu Li Sarah S. Knox Marrick L. Kukin Kenneth R. Laurita Yan C. Li Merril L. Knudtson Rakesh C. Kukreja Debbie A. Lawlor Bruce T. Liang Juhani Knuuti Lewis H. Kuller Daniel A. Lawrence Chang-seng Liang Dennis T. Ko Iftikhar J. Kullo Lesley Lawrenson James K. Liao Yoshio Kobayashi Premkumari Kumarathasan Louise Lawson Ronglih Liao Colleen G. Koch Richard E. Kuntz Jennifer S. Lawton Youlian Liao Walter J. Koch Calvin J. Kuo Harold L. Lazar Peter Libby Itsuo Kodama Lih Kuo Ronald M. Lazar Joseph R. Libonati Wolfgang Koenig Christian Kupatt Dominique Le Guludec David S. Liebeskind Theo Kofidis Dhandapani Kuppuswamy Alexander Leaf Philip R. Liebson Kwang K. Koh Masahiko Kurabayashi Sam D. Leary Choong-Chin Liew Frank Kolodgie Tobias Kurth Robert J. Lederman Stephen B. Liggett Issei Komuro Theodore Kurtz Amanda J. Lee Kathleen C. Light Takahisa Kondo Kengo F. Kusano Hon-Chi Lee Yean L. Lim Marvin A. Konstam Jeffrey T. Kuvin Richard T. Lee Chee Chew Lim Stavros V. Konstantinides William A. Kuziel Thomas H. Lee Joao A. Lima Igor E. Konstantinov Tatiana Kuznetsova C.P.M. Leeson Marian C. Limacher Anatol Kontush Kevin F. Kwaku David J. Lefer Ming T. Lin Marianne Eline Kooi Raymond Y. Kwong Jean-Francois Legare Shien-Fong Lin Willem J. Kop Michael Kyller Jacopo M. Legramante Michael Lincoff Stephen L. Kopecky Maria T. La Rovere Michael H. Lehmann Bertil Lindahl Bruce A. Koplan David E. Laaksonen Stephan E. Lehnart JoAnn Lindenfeld Ran Kornowski Arthur Labovitz Leslie Leinwand Marshall D. Lindheimer Mikhail Kosiborod Louis M. Labrousse Norbert Leitinger Jonathan R. Lindner Andreas Koster Roger J. Laham Thierry H. LeJemtel Volkhard Lindner Rudolph W. Koster Shenghan Lai Paul LeLorier Jerry B. Lingrel Sawa Kostin John G. Lainchbury Giuseppe Lembo MacRae F. Linton Theodore A. Kotchen Edward G. Lakatta Pedro A. Lemos Gregory Y.H. Lip Hans Kottkamp Hanna-Maaria Lakka Steven R. Lentz William C. Little Nicholas T. Kouchoukos Timo A. Lakka David A. Leon Sheldon E. Litwin Petri T. Kovanen Jules Y.T. Lam Antonio Maria Leone Jun Liu Peter R. Kowey Stephen C.T. Lam Jonathan Leor Kiang Liu Jun Koyama Benoit Lamarche Amir Lerman Peter P. Liu Andrew D. Krahn John J. Lamberti Bruce B. Lerman Simin Liu Aldi Kraja Rachel Lampert Edward J. Lesnefsky Yongge Liu Jonathan Krakoff Katja H. Lampinen Philippe F. Lesnik Eng H. Lo Christopher M. Kramer Kathryn G. Lamping Heather S. Lett Amanda Lochner Evangelia G. Kranias Hui Y. Lan Donald Y. Leung James E. Lock William E. Kraus Gary Landreth Michael C. H. Leung Warren E. Lockette Ronald M. Krauss Donald W. Landry Marcel M. Levi Ian M. Loftus Andrew J. Krentz Michael J. Landzberg Roberto Levi Anne-Marie Lompre Nancy R. Kressin David A. Lane Adeera Levin Barry London Reinhold Kreutz Roberto M. Lang Benjamin D. Levine Gérard M. London Jorg Kreuzer Jonathan J. Langberg Glenn N. Levine Carlin S. Long P.A. Krieg Bas Langeveld Robert A. Levine Eva M. Lonn Murali C. Krishna B. Lowell Langille Sidney Levitsky Gary D. Lopaschuk Rajesh Krishnamurthy David Langleben Bernard I. Levy John J. Lopez Eswar Krishnan Alexandra J. Lansky Daniel Levy Patricio López-Jaramillo Leonard Kritharides Harris Lari Robert J. Levy Francisco Lopez-Jimenez Michael H. Kroll John C. LaRosa Andrew P. Levy Christine H. Lorenz e20 Acknowledgment of Reviewers

David J. Loskutoff Michael J. Mann William M. McClellan Michael Miller Douglas W. Losordo Stewart Mann Seth McClennen Nancy H. Miller Eric B. Loucks Peter B. Manning Michael V. McConnell Todd D. Miller Charles J. Lowenstein Warren J. Manning James McCord Virginia M. Miller Gerald Luc Pier M. Mannucci Brian W. McCrindle Tohru Minamino Lee Lucas Teri A. Manolio Peter A. McCullough Erich Minar Benedict R. Lucchesi Moussa Mansour James D. McCully Gary S. Mintz Pamela A. Lucchesi Michael S. Marber Mary M. McDermott Israel Mirsky Andreas Luchner Keith L. March Theresa A. McDonagh Yoshio Misawa John Ludbrook Simona Marchesi Daniel McGee Seema Mital David Ludwig Francis E. Marchlinski John C. McGiff Brett M. Mitchell Russell V. Luepker Frank I. Marcus Henry C. McGill Gary F. Mitchell Friedrich C. Luft Maurizio Margaglione Michael McGoon Jere H. Mitchell Esther Lutgens Ali J. Marian Thomas M. McIntyre Richard N. Mitchell Aernout L. Luttun Allyn L. Mark William J. McKenna R. Scott Mitchell Robert L. Lux Daniel B. Mark Timothy A. McKinsey Arnold Mitnitski Bruce W. Lytle Andrew R. Marks Tracey L. McLaughlin Suneet Mittal Christoph Maack Barry J. Maron Vallerie V. McLaughlin Murray A. Mittleman David M. Maahs Luc Maroteaux Julie R. McMullen Kunio Miyatake Renke Maas Michel Marre John J.V. McMurray Kohei Miyazono Peter S. Macdonald Mario B. Marrero Elizabeth M. McNally Emile R. Mohler, III Christopher K. Macgowan Oscar C. Marroquin Coleen A. McNamara Nicanor I. Moldovan Guy A. MacGowan Philip A. Marsden Patrick H. McNulty Peter Molenaar Francois Mach Audrey C. Marshall Tim C. McQuinn Ernesto Molina Stella M. Macin Steven P. Marso Charles F. McTiernan David J. Moliterno Christopher Mack Fabio Martelli Gary E. McVeigh Jeffery D. Molkentin Michael J. Mack Douglas Martin Roger Mee Tom E. Mollnes Wendy J. Mack Jack L. Martin Mandeep R. Mehra Kevin M. Monahan Isla S. Mackenzie Paul T. Martin Roxana Mehran Laurent Monassier Rachel H. Mackey Wim Martinet Jawahar L. Mehta Gilles Montalescot William R. MacLellan Yukio Maruyama Rajendra H. Mehta Joan Montaner Michal Maczewski Thomas H. Marwick Shamir R. Mehta Nicola Montano Paolo Madeddu Gerald R. Marx James B. Meigs Alan R. Moody Mohammad Madjid Nikolaus Marx Cynthia J. Meininger James C. Moon Joren C. Madsen Steven O. Marx Gerhard W. Meissner David F. Moore Koji Maemura Attilio Maseri Jan R. Mellembakken Phillip Moore Aldo P. Maggioni Peter J. Mason Philippe Menasche Martin Morad Kenneth W. Mahaffey Robert J. Mason Michael E. Mendelsohn Fred Morady Michael C. Mahaney Frederick A. Masoudi Carlos F. Mendes de Leon Christine S. Moravec Lynn Mahony Joseph M. Massaro Armando J. Mendez Henning Morawietz Heimo Mairbaurl Barry M. Massie Maurizio Menichelli Kerrie L. Moreau Bernhard Maisch Bashir M. Matata George A. Mensah Pierre Moreau Alan S. Maisel Ellisiv B. Mathiesen James O. Menzoian Pedro R. Moreno William H. Maisel Hiroaki Matsubara Jean-Jacques Mercadier Raul Moreno Mark W. Majesky Hikaru Matsuda Anwar T. Merchant Thomas M. Morgan Amy S. Major Reiko Matsui Yahye Merhi Peter M. Morganelli Robert T. Mallet Kanji Matsukawa Ilse L. Mertens Anthony P. Morise Ziad Mallat Akira Matsumori Franz H. Messerli Ryuichi Morishita Alberto Malliani Hidehiro Matsuoka Ruben Mestril Toshisuke Morita Giuseppe Mancia Hiroaki Matsuoka Luisa Mestroni Nicholas W. Morrell G.B. John Mancini Rumiko Matsuoka Heiko Methe Joel Morrisett Pitchaiah Mandava Masunori Matsuzaki Philippe Meurin John A. Morrison Olivia Manfrini Christian M. Matter Martijn Meuwissen Sean J. Morrison Dennis T. Mangano Ray V. Matthews Theo E. Meyer David A. Morrow Arduino A. Mangoni Kimmo J. Mattila Evangelos D. Michelakis Jason D. Morrow Venkatesh Mani Clive N. May Holly R. Middlekauff Richard F. Mortensen Calin V. Maniu Bongani M. Mayosi Michele Mietus-Snyder Lori J. Mosca Douglas L. Mann Melanie Maytin Richard V. Milani Ralph S. Mosca Giovanni E. Mann Todor N. Mazgalev D. Craig Miller Mauro Moscucci Johannes F. Mann Nathalie M. Mazure D. Douglas Miller Jeffrey W. Moses Kenneth G. Mann Eileen McCall Leslie W. Miller Arthur J. Moss Acknowledgment of Reviewers e21

Richard L. Moss Andrea Natale Christopher J. O’Donnell Richard L. Page Evangeline D. Motley Rama Natarajan Erwin N. Oechslin Ramdas G. Pai Karen S. Moulton Viswanathan Natarajan Patrick T. O’Gara Rosario Palacios Jean-Jacques Mourad Hendrik Nathoe Hisao Ogawa Wulf Palinski Issam D. Moussa Stanley Nattel Yoshihiro Ogawa Julio C. Palmaz Gilbert H. Mudge Matthew T. Naughton Toshio Ogihara Lyle J. Palmer Christian Mueller Mohamad Navab Jae K. Oh Sebastian Palmeri Thomas Muenzel Frank Naya Ann M. O’Hare Hui-Lin Pan Alessandro Mugelli Krassen Nedeltchev Takayoshi Ohkubo Demosthenes Panagiotakos Andreas Mugge Ilka Nemere Tomoko Ohkusa Natesa G. Pandian Joseph B. Muhlestein Dario Neri Erik M. Ohman James S. Pankow Debabrata Mukherjee Shawna D. Nesbitt Veronica Ojetti Julio A. Panza Rupak Mukherjee Aleksandar N. Neskovic Akinlolu O. Ojo Nicholas F. Paoni Barbara J.M. Mulder Paul J. Nestel Peter M. Okin Carlo Pappone James E. Muller Stefan Neubauer Katashi Okoshi Gilles Paradis Jochen Muller-Ehmsen Ellis J. Neufeld Jeffrey E. Olgin Patrick Parfrey Janet M. Mullington Gishel New Jobien K. Olijhoek Michael Parides Michael J. Mulvany Andrew C. Newby Jeffrey W. Olin Stephen Paridon Neal I. Muni David E. Newby Brian Olshansky Paolo Parini Jorg Muntwyler L. Kristin Newby Timothy M. Olson Jeong-Euy Park Joanne Murabito Anne B. Newman Patrick G. O’Malley Seung-Jung Park David Murdoch John H. Newman Reed A. Omary Donna Parker Toyoaki Murohara Gary E. Newton Jeffrey H. Omens John D. Parker Elizabeth Murphy Christopher H. Newton-Cheh Torbjorn Omland Ira A. Parness Philip M. Murphy Ludwig Neyses Steve R. Ommen Juan C. Parodi Timothy P. Murphy Graham Nichol Altan Onat Alessandro Parolari Charles E. Murry Stephen J. Nicholls Marie S. O’Neill Steve W. Parry Timothy I. Musch Wilmer W. Nichols William W. O’Neill Vincenzo Pasceri Rene J. Musters Andrew C. Nicholson Takayuki Ono Ares D. Pasipoularides Steven E. Mutsaers Georg Nickenig Koji Onoda Gerard Pasterkamp Bulent Mutus Martin J. Nicklin Henry Ooi Ayan Patel Robert J. Myerburg Pascal H. Nicod Suzanne Oparil Lisa Patel Daniel D. Myers Christoph A. Nienaber Tobias Opthof Rakesh P. Patel Jonathan Myers Michael R. Nihill Hakan Oral Vickas V. Patel Elizabeth G. Nabel Seppo T. Nikkari John F. Oram David J. Paterson Christoph K. Naber Dimitar Nikolov E. John Orav Paola Patrignani Bernardo Nadal-Ginard Richard M. Niles Trevor J. Orchard Carlo Patrono Zurab G. Nadareishvili Rick A. Nishimura Jose M. Ordovas Richard D. Patten Koonlawee Nademanee Mari K. Nishizaka Donald Orlic Cam Patterson Abraham Nader Steven E. Nissen John A. Ormiston Peter M. Pattynama Vinay Nadkarni Tianhua Niu Joseph P. Ornato Walter J. Paulus Ryozo Nagai Koichi Node Brian O’Rourke Jeffrey M. Pearl Hideaki Nagase Constance T. Noguchi Michael F. O’Rourke Justin D. Pearlman Noritoshi Nagaya Eisei Noiri Robert A. O’Rourke Mary A. Peberdy Eike Nagel Georg Noll Tetsuya Oshima Ole D. Pedersen Sherif F. Nagueh Borge G. Nordestgaard Clive Osmond Susanne S. Pedersen Hiroshi Nakagawa Mikael Norman Jan Ostergren Patrick Peeters Hajime Nakamura Kari E. North David Ott Antonio Pelliccia Takeshi Nakano Gavin R. Norton Fillipo Ottani Patricia A. Pellikka Kanji Nakatsu Michel Noutsias Catherine M. Otto Theo Pelzer Gilles Nalbone Gian M. Novaro Feifan Ouyang Michael Pencina Brahmajee K. Nallamothu Ulrike Nowak-Gottl Michel Ovize Marc S. Penn Byung-Ho Nam Evgeny Nudler Mehmet C. Oz Dudley J. Pennell Navin C. Nanda William C. Nugent Susan E. Ozanne William Penny Manasi Nandi Carole Ober Pal Pacher Carl J. Pepine Raffaele Napoli Martin Oberhoff Chris J. Packard Mark B. Pepys Girish Narayan Edward R. O’Brien Douglas L. Packer Mark A. Pereira Sanjiv M. Narayan Christopher J. Occleshaw Francis D. Pagani Francisco Perez-Vizcaino Craig R. Narins Ira S. Ockene Massimo Pagani Emerson C. Perin Krzysztof Narkiewicz Christopher M. O’Connor Patrick J. Pagano Harris Perlman Jagat Narula Gerald T. O’Connor Pierre Page Joseph K. Perloff e22 Acknowledgment of Reviewers

Eduardo R. Perna Wendy S. Post K. Srinath Reddy Campbell Rogers Thomas V. Perneger Robert S. Poston Vivek Y. Reddy Ariel Roguin Francesco Perticone Tina S. Poulsen Margaret M. Redfield John Rogus Arkadii M. Pertsov Neil Poulter Andrew N. Redington Mary J. Roman Inga Peter Janet T. Powell Judith G. Regensteiner Mats Rönnback Annette Peters Andrew J. Powell Enrique Regidor Dieter Ropers Nicholas S. Peters William S. Powell Jalees Rehman Emilio Ros Eric D. Peterson Scott K. Powers Johan Reiber Wayne D. Rosamond Kirk L. Peterson Henry J. Pownall Nathaniel Reichek Jonathan Rosand J.Thomas Peterson Abhiram Prasad Muredach Reilly Noel R. Rose Eva Petkova Francesco Prati Sharon C. Reimold Michael R. Rosen Patricia A. Peyser Domenico Pratico Steven E. Reis David S. Rosenbaum Marc A. Pfeffer Josef Prchal Peter J. Reiser Frits R. Rosendaal Ivan Philip Stephen M. Prescott Michael J. Reiter Clive Rosendorff George J. Philippides Russell L. Prewitt P.H. Reitsma Michael E. Rosenfeld Bradley G. Phillips Beth F. Printz Jian-Fang Ren Bruce R. Rosengard Christopher O. Phillips Frits W. Prinzen Jun Ren Todd K. Rosengart Richard P. Phipps Silvia G. Priori Helaine E. Resnick David N. Rosenthal Francesco Piarulli Kirkwood A. Pritchard Ariel J. Reyes Anthony Rosenzweig Philippe Pibarot Linda L. Pritchard Dwight Reynolds Bernard Rosner Eugenio Picano Eric N. Prystowsky Matthew R. Reynolds Allan M. Ross Michael H. Picard Bruce M. Psaty Shereif H. Rezkalla John Ross J. Geoffrey Pickering William Pu Jonathan Rhodes Robert Ross Galen M. Pieper Vladimir Pucovsky Flavio Ribichini Andrea Rossi Luc A. Pierard John D. Puskas Paul M. Ribisl Gian Paolo D. Rossi Grant N. Pierce Pirkko J. Pussinen Romeo Ricci Ranieri Rossi Burkert Pieske Reed Pyeritz Peter A. Rice Thomas Rostock Bill A. Pietra Kalevi Pyorala Lawrence Rice Michael Roth Gabriele Piffaretti Stuart F. Quan Michael W. Rich Richard B. Rothman Frank A. Pigula Thomas Quaschning Vincent Richard Steven A. Rothman Nico H.J. Pijls Thomas Quertermous A. Mark Richards Peter M. Rothwell Louise Pilote Miguel A. Quiñones Paul M. Ridker Philippe Rouet Ileana Pina Ton J. Rabelink Barbara Riegel Jean-Lucien Rouleau Theodore Pincus Marlene Rabinovitch Walter F. Riesen Anna V. Roux David J. Pinsky Miriam T. Rademaker Nader Rifai Anne H. Rowley Duane S. Pinto Daniel J. Rader Giorgio Rigatelli Prabir Roy-Chaudhury Yigal M. Pinto Marek W. Radomski Eric B. Rimm Alan Rozanski Tobias Pischon Shahin Rafii Gilles Rioufol Hong Ruan Federico Piscione Paolo Raggi Rebecca H. Ritchie Melvyn Rubenfire Cristina Pislaru Shahbudin H. Rahimtoola James M. Ritter Frederick L. Ruberg Bertram Pitt Elaine W. Raines Eberhard Ritz Lewis J. Rubin Maria V. Pitzalis Olli T. Raitakari Alain Rivard Terrence D. Ruddy Manel Pladevall Satish R. Raj Jeffrey Robbins Lawrence L. Rudel Jonathan F. Plehn Sanjay Rajagopalan Robert C. Robbins Neil B. Ruderman Johannes Pleiner Nalini M. Rajamannan Robert Roberts Yoram Rudy Jorge Plutzky Sumathi Ramachandran David Robertson Marc Ruel A. Graham Pockley Kenneth S. Ramos Sander J. Robins Wolfram Ruf Stuart J. Pocock J. Scott Rankin Richard B. Robinson Zaverio M. Ruggeri Bruno K. Podesser Dabeeru C. Rao Simon C. Robson Jean-Bernard Ruidavets Philip J. Podrid L.Vijay Rao Albert P. Rocchini Luis M. Ruilope Paul Poirier Elliot Rapaport Luc Rochette John A. Rumberger Roberto Pola Tienush Rassaf Howard E. Rockett John S. Rumsfeld Don Poldermans Saif S. Rathore Howard A. Rockman Marschall S. Runge Victoria Polyakova Peter B. Raven Dan M. Roden Heinz Rupp Philip A. Poole-Wilson Ursula Ravens David Rodman Frank Ruschitzka Clive A. Pope Katya Ravid Beatriz L. Rodriguez James W.E. Rush Jeffrey J. Popma Chester A. Ray Leonardo Rodriguez Jeremy N. Ruskin Richard L. Popp Reza S. Razavi Fernando Rodriguez-Artalejo Kerry S. Russell J. David Port Fabio A. Recchia Alicia Rodriguez-Pla Mary E. Russell Francesco Portaluppi Rita F. Redberg Marco Roffi Raymond R. Russell Thomas R. Porter Alluru S. Reddi Veronique L. Roger Wolfgang Rutsch Acknowledgment of Reviewers e23

Martin K. Rutter Saul Schaefer Miki L. Schwartzman Allan A. Shor Peter N. Ruygrok Juergen R. Schaefer Robert A. Schweikert Angela C. Shore Thomas J. Ryan Klaus P. Schafers Robert H.G. Schwinger Linda Shore-Lesserson Jack Rychik Hartzell V. Schaff Juerg Schwitter Ashfaq Shuaib Lars Ryden Martin J. Schalij Alan Scott Robert J. Siegel Tobias Saam Jutta Schaper Russell S. Scott Hans-Hinrich Sievers Manel Sabate Wolfgang Schaper Miran Sebestjen Ulrich Sigwart Marc S. Sabatine Christoph Scharf Udo Sechtem Michael J. Silka Roger A. Sabbadini Gina Schatteman Artyom Sedrakyan Marc A. Silver Hani N. Sabbah Robert G. Schaub Ellen W. Seely Donald S. Silverberg Joseph F. Sabik Andre J. Scheen Harry Segall David I. Silverman Luigi Sacca James M. Scheiman Pravin B. Sehgal Gregg J. Silverman Ralph L. Sacco Dierk Scheinert Christine E. Seidman Norman H. Silverman Michael N. Sack Melvin M. Scheinman Jonathan G. Seidman Jean-Sebastien Silvestre Jonathan D. Sackner-Bernstein Heinrich R. Schelbert Christian Seiler Robert D. Simari Frank M. Sacks Benjamin J. Scherlag Frank W. Sellke R. John Simes H. Mehrdad Sadeghi Ralph T. Schermuly Joseph B. Selvanayagam Paolo Simioni Junichi Sadoshima Urs Scherrer Luyi Sen Daniel I. Simon Michel E. Safar Marielle Scherrer-Crosbie Laureen Sena Scott I. Simon Jeffrey E. Saffitz Deborah A. Scheuer Shoichi Senda Orlando P. Simonetti Kiran B. Sagar James Scheuer Roxy Senior Leon A. Simons David J. Sahn Ernesto L. Schiffrin Thomas D. Sequist Michael Simons Yoshifumi Saijo Nelson B. Schiller Susan M. Sereika Maarten L. Simoons Yoshihiko Saito Mark D. Schluchter Patrick W. Serruys Paul C. Simpson Tomohiro Sakamoto Klaus-Dieter Schluter William C. Sessa John E. Sims Ichiro Sakuma Alvin Schmaier Howard D. Sesso Alan R. Sinaiko Tomas A. Salerno John F. Schmedtje Peter S. Sever Jürgen R. Sindermann Veikko Salomaa Chris Schmid Robert E. Shaddy Pawan K. Singal Koen J. Salu Holger Schmid Ajay M. Shah Krishna Singh Carlo Salvarani Ann-Marie Schmidt Dipen C. Shah Sanjay Singh Flora Sam Carsten B. Schmidt-Weber Pravin M. Shah Lawrence I. Sinoway Habib Samady Gerd Schmitz Prediman K. Shah Albert J. Sinusas Afshin Samali David J. Schneider Robin Shandas Karin Sipido Nilesh J. Samani Michael D. Schneider Richard P. Shannon David S. Siscovick Gianmario Sambuceti Andreas Schober Oz M. Shapira Samuel C. Siu Jonathan M. Samet Gabriele Schoedon Arya M. Sharma Deborah A. Siwik Willis K. Samson Albert Schoemig Frank R. Sharp Carsten Skurk Prashanthan Sanders Frederick J. Schoen Norman Sharpe Cornelis J. Slager John E. Sanderson David A. Schoenfeld A. Richey Sharrett Mara Slawsky David C. Sane Paul Schoenhagen Michael J. Shattock Peter Sleight Anthony J. Sanfilippo Peter M. Scholz Philip W. Shaul Marvin J. Slepian L. Fernando Santana Uwe Schonbeck Leslee J. Shaw Karen Sliwa Massimo Santini Ronald Schondorf Amanda M. Shearman Gregory Sloop Marisa Santos Wilhelm Schoner Michael Shechter Joost P.G. Sluijter Maria-Jesus Sanz Rolf Schroeder Imad Sheiban Richard W. Smalling John L. Sapp Stephen Schroeder James Shepherd Eric J. Smart Maurice E. Sarano Valarie Schroeder Warren Sherman Otto A. Smiseth Ian J. Sarembock Karsten Schrör Mark V. Sherrid Alberto Smith Mark J. Sarnak Joerg B. Schulz Sanjay Shete Felicity B. Smith Masataka Sata Rainer Schulz Weibin Shi George D. Smith Toshiaki Sato Richard Schulz Rei Shibata Gordon C.S. Smith Naveed Sattar P. Christian Schulze Mei-Chiung Shih Jonathan Smith J. Philip Saul Paul T. Schumacker Kazuyuki Shimada Nicholas L. Smith Elijah Saunders Holger J. Schunemann Wataru Shimizu Scott A. Smith Kurt W. Saupe Markus Schwaiger Hiroaki Shimokawa Sidney C. Smith Bernhard Sauter Lee H. Schwamm Ken Shinmura Steven R. Smith Motoji Sawabe Gregory G. Schwartz Satoshi Shintani Ulf Smith Tatsuya Sawamura Ketty Schwartz Ichiro Shiojima Warren M. Smith Leslie A. Saxon Peter J. Schwartz Kalyanam Shivkumar William M. Smith James W. Sayre Stephen M. Schwartz Michael G. Shlipak Pieter C. Smits Angelo M. Scanu David S. Schwartzman Ralph V. Shohet Allan D. Sniderman e24 Acknowledgment of Reviewers

Marieke B. Snijder Christoph Stellbrink Istvan Szokodi Rong Tian Burton E. Sobel Kurt R. Stenmark Roman F. Sztajzel Uwe J.F. Tietge Kenji Sobue David W. Stepp Ira A. Tabas Laurence Tiret Stefan Soderberg Andrew Steptoe Stefano Taddei Marc D. Tischler Kyoko Soejima David M. Stern Heinrich Taegtmeyer Susan Tiukinoy Constantinos T. Sofocleous Naftali Stern Peter Taggart Jonathan M. Tobis Raija Soininen Lynne Warner Stevenson Kazuhiro Takahashi Geoffrey H. Tofler Minna Soinio Duncan J. Stewart Masato M. Takahashi Stevan P. Tofovic John Solaro Kerry J. Stewart Bonpei Takase Naoki Tokita Steven J. Sollott Jim Stewart Hiroshi Takayama Eran Toledo Scott D. Solomon Julian M. Stewart Yoshiyu Takeda Douglas M. Tollefsen Prem Soman Ralph A. H. Stewart Satoshi Takeo Robert J. Tomanek John Somberg Roland Stocker Akira Takeshita Gordon F. Tomaselli Virend K. Somers Jean-Claude Stoclet Renato Talamini Marcello Tonelli Paul D. Sorlie Katarzyna Stolarz William T. Talman Peter Tontonoz Farzaneh Aghdassi Sorond Claudia Stollberger Rasa Tamosiuniene Eric Topol J. Eduardo Sousa Gregg W. Stone Chee Eng Tan Jan H.M. Tordoir P.C. Souverein Neil J. Stone Walter A. Tan Per Tornvall James R. Sowers Peter H. Stone Toshihiro Tanaka Olga H. Toro-Salazar Madison S. Spach George A. Stouffer Weihong Tang Christian Torp-Pedersen Rainer Spanbroek Vibeke Strand W.H. Wilson Tang Guillermo Torre-Amione Carl P. Sparrow Timo E. Strandberg Yao Liang Tang Jan Torzewski Christian M. Spaulding John R. Stratton Rajendra K. Tangirala Tor D. Tosteson J. David Spence Bodo E. Strauer Ataru Taniguchi Peter P. Toth William H. Spencer Arnold Strauss Laszlo B. Tanko Florence Toti John A. Spertus William B. Strawn Felix C. Tanner Arturo G. Touchard Philip Spevak S. Adam Strickberger Jean-Claude Tardif Rhian M. Touyz Lukas E. Spieker Jack P. Strong Robert B. Tate Jeffrey A. Towbin Francis G. Spinale Allan D. Struthers Hideki Tatewaki Dwight A. Towler David H. Spodick Matthias Stuber Allen J. Taylor Jonathan N. Townend David Spragg Jorg Stypmann Andrew M. Taylor Paul A. Townsend Joachim Spranger Ding-Feng Su Anne L. Taylor Maurizio Trevisan Deepak Srivastava Isabella Sudano Joan M. Taylor Richard W. Troughton Martin G. St. John Sutton Peter H. Sugden W. Robert Taylor Nathan A. Trueblood Eugenio Stabile Galina K Sukhova James E. Tcheng Shen K. Tsai Austin Stack Yao Sun Guillermo J. Tearney Min-Fu Tsan Jan A. Staessen Zhonghua Sun Alain Tedgui Teresa S.M. Tsang Diana M. Stafforini Thoralf M. Sundt, III Usha Tedrow Philip S. Tsao Gregory L. Stahl Ruey J. Sung Paul S. Teirstein Hung-Fat Tse Anton F.H. Stalenhoef H. Robert Superko David F. Teitel Etsuko Tsuda Bruce S. Stambler Howard K. Surks George Tellides Gozoh Tsujimoto Jonathan S. Stamler Mark A. Sussman Marije ten Wolde Katsuhiko Tsujioka Meir J. Stampfer Thomas M. Suter Koon K. Teo Hiroyuki Tsukui Kenneth Stanley Fraser W.H. Sutherland Oren M. Tepper Jack V. Tu William C. Stanley George R. Sutherland Gail R. ter Haar Michael L. Tuck Alice V. Stanton John L. Sutko Hiroki Teragawa Rubin M. Tuder Randall C. Starling Richard Sutton Enrique Teran Paul A. Tunick Brian L. Stauffer Kim Sutton-Tyrrell Dellara F. Terry Zoltan G. Turi Charles Steenbergen Hiroshi Suzuki Daniel Teupser Craig D. Turnbull Philippe G. Steg Ken Suzuki David Thaler Stephen T. Turner Coen D. Stehouwer Alan F. Sved Pierre Theroux Alexander G.G. Turpie Evan A. Stein Lars G. Svensson Aravinda Thiagalingam Katherine R. Tuttle Kenneth M. Stein Madhav Swaminathan Perumal Thiagarajan E. Murat Tuzcu Paul D. Stein Lorna Swan Chris Thiemermann Marcel Twickler Francene M. Steinberg Karl Swedberg Gaetano Thiene Suresh C. Tyagi Helmut O. Steinberg G. Sweeney Victor L.J.L. Thijssen Toshimitsu Uede Susan F. Steinberg Michael O. Sweeney Anita C. Thomas Per M. Ueland Julia Steinberger Charles D. Swerdlow James D. Thomas Renan Uflacker Gustav Steinhoff Bernard Swynghedauw MaryLou Thompson Cuno S.P.M. Uiterwaal Robin H. Steinhorn Christer Sylven Paul D. Thompson Gudrun Ulrich-Merzenich Steve R. Steinhubl Zoltán Szabó Kent Thornburg Shin-Ichiro Umemura Acknowledgment of Reviewers e25

Paul M. Underwood Renaud Vincent Mark W.I. Webster Heinrike Wilkens Roger H. Unger Jakob Vinten-Johansen William Weglicki Ian B. Wilkinson Thomas Unger Francesco Violi Chiming Wei Bruce L. Wilkoff Zoltan Ungvari Maria L. Virella Li Wei Walter C. Willett Joseph L. Unthank Renu Virmani Max H. Weil David O. Williams Gilbert R. Upchurch Sami Viskin Hartmut Weiler Kevin J. Williams Zsolt Urbán Eric Vittinghoff Janice Weinberg Mark A. Williams Fumitaka Ushikubi Barbara Voetsch Tanja Weinbrenner Paul T. Williams Viola Vaccarino Michael Vogel Andrew R. Weintraub Roberta G. Williams Marco Valgimigli Robert A. Vogel William S. Weintraub R. Sanders Williams Patrick J.T. Vallance Paul G.A. Volders Michael Weis Allison E. Willing Jesus G. Vallejo Stefano Volpato Richard D. Weisel Scott R. Willoughby Eric Van Belle Klaus von Bergmann Mary C. Weiser-Evans Emily Wilson Gerald van Belle Jan H. von der Thüsen Myron L. Weisfeldt Peter W. Wilson Marc van Bilsen Arnold von Eckardstein Daiana Weiss Gayle L. Winters Luc M. Van Bortel Robert Voswinckel Guenter Weiss Andrew L. Wit Frans J. Van de Werf Atsuyuki Wada James N. Weiss Hanspeter Witschi Johanna G. van der Bom Carol Wadham Robert G. Weiss Maarten Witsenburg Willem J. van der Giessen Bernard Waeber Neil J. Weissman Jacqueline C.M. Witteman Bernd van der Loo Lynne Wagenknecht Jeffrey I. Weitz Rochus Witthaut Freek J. van der Meer Andreas H. Wagner Babette B. Weksler Fred H.M. Wittkampf Irene M. van der Meer Denisa D. Wagner Hein J. Wellens Joseph L.Witztum Yvonne T. van der Schouw Galen S. Wagner Ian J. Welsby Stephen D. Wiviott Jolanda van der Velden Louis K. Wagner Frederick G. Welt J. Frederick Woessner Ger J. van der Vusse Peter D. Wagner Francine K. Welty Wojciech Wojakowski Miranda Van Eck Shawn Wagner Stephen E. Welty Philip A. Wolf George F. Van Hare Ron Waksman Nanette K. Wenger Michael S. Wolin C. Heleen van Ommen Albert L. Waldo Bruce M. Wentworth Robert Wolk Niels van Royen Brian R. Walker Jolanda J. Wentzel Kai C. Wollert Dirk J. van Veldhuisen Lars Wallentin Rene R. Wenzel Ernst Wolner David R. Van Wagoner B. Gunnar Wallin Volker Wenzel Cheuk-kit Wong Anne M. VanBuskirk John Wallwork Gerald S. Werner LennieWong Mani A. Vannan Edward P. Walsh Rainer Wessely Nathan D. Wong Nerea Varo Peter N. Walsh Malcolm West John C. Wood Sudesh Vasdev Dirk H. Walter Rudi G. Westendorp Mark A. Wood Giuseppe Vassalli Thomas Walther Justin Westhuyzen Hermann Wrigge Theodoros Vassilakopoulos Bingcheng Wang Charles V. Wetli Jackson T. Wright, Jr. Guy Vassort Donna H.Wang Glenn T. Wetzel R. Scott Wright Stephen F. Vatner Thomas J. Wang Lewis Wexler Ed X. Wu Matteo Vatta Wei Wang Cornelia M. Weyand Gordon D. Wu Douglas E. Vaughan Xiaohong Wang Arthur E. Weyman Joseph C. Wu William K. Vaughn Yibin Wang Andrew S. Weyrich Kenneth K. Wu Mark A. Veazie Zhiguo Wang Christopher J. White D. George Wyse James L. Velianou Carole A. Warnes C. Roger White Guohua Xi Richard C. Venema Karl Wasserman Halina White Lei Xiao Paolo Verdecchia David D. Waters Richard H. White Rui-Ping Xiao Pieter D. Verdouw Hugh Watkins William B. White Chengjie Xiong Stefan Verheye Steve P. Watson Patrick L. Whitlow Magdi H. Yacoub Petra Verhoef Wendy A. Wattigney J. Lindsay Whitton Jay S. Yadav Jean Philippe Verhoye Gerald F. Watts Mark H. Wholey Yoshiji Yamada Subodh Verma Sergio Waxman Lawrence Wickerham Kazuhiro Yamamoto Richard L. Verrier W. Douglas Weaver Samuel A. Wickline Yoshiharu Yamamoto Francesco Versaci Catherine Webb Petr Widimsky Atsushi Yamashita Giorgio A. Vescovo David J. Webb Susan E. Wiegers Gan-Xin Yan George W. Vetrovec Gary Webb Wouter Wieling Xinhua Yan Aristidis Veves Steven A. Webber FrankWiesmann Clyde W. Yancy G. Wesley Vick Christian Weber William Wijns Qiong Yang Neill Videlefsky Karl T. Weber David J. Wilber Xiao-Ping Yang Flordeliza S. Villanueva Michael A. Weber Arthur A.M. Wilde Zhihong Yang Francisco Villarreal Nina C. Weber Rachel P. Wildman Hirofumi Yasue D. Geoffrey Vince Keith A. Webster Markus J. Wilhelm Frank Yatsu e26 Acknowledgment of Reviewers

Richard Ye Pierre Y. Youinou Marc Zee Felix Zijlstra Jerry Yee James B. Young Robert Y. Zee Michael R. Zile Edward T.H. Yeh Lawrence H. Young Kenton J. Zehr Peter J. Zimetbaum Mao-Hsiung Yen Martin E. Young Andreas M. Zeiher Marc Zimmermann Midori A. Yenari Pampee P. Young Darryl C. Zeldin Jean-Marc Zingg Shaw-Fang Yet Chun Yuan Andrey G. Zenovich Douglas P. Zipes Alan C. Yeung Sun Yuhua Uwe Zeymer Carmine Zoccali Seppo Yla-Herttuala Salim Yusuf Cuihua Zhang William A. Zoghbi Agneta Yngve Susanne Zadelaar Yingyi Zhang Ai-Ping Zou Paul G. Yock Kenneth G. Zahka Guixiang Zhao Ming Zou Junji Yodoi Osama O. Zaidat Zhi-Jie Zheng Irving H. Zucker Young-sup Yoon Alberto Zanchetti Guangming Zhong Bram D. Zuckerman Chaim Yosefy Faiez Zannad Robert Zhong Mahmoud Zureik Hiroshi Yoshida Wojciech Zareba Jianhui Zhu Jay L. Zweier Masayuki Yoshida Barry L. Zaret Xinsheng Zhu Noriko Yoshida Alan M. Zaslavsky Brenda K. Zierler Dimitri E. Zylberstein