Volume 115, Issue 2; January 16, 2007,pp.157-285

Issue Highlights

Issue Highlights

Circulation 2007 115: 157 Editorials

The Cardiovascular World Is Definitely Not Flat P. Kaul and E.D. Peterson Circulation 2007 115: 158 - 160, doi:10.1161/CIRCULATIONAHA.106.672022

Cardiac Resynchronization Therapy in New York Heart Association Class IV Heart Failure: It Is All About Selection W.H. Wilson Tang and Gary S. Francis Circulation 2007 115: 161 - 162, doi:10.1161/CIRCULATIONAHA.106.669879

Original Articles

Congenital Heart Disease

Congenital Heart Disease in the General Population: Changing Prevalence and Age Distribution Ariane J. Marelli, Andrew S. Mackie, Raluca Ionescu-Ittu, Elham Rahme, and Louise Pilote Circulation 2007 115: 163 - 172; published online before print January 8 2007, doi:10.1161/CIRCULATIONAHA.106.627224 Coronary Heart Disease

Association of Cystatin C With Mortality, Cardiovascular Events, and Incident Heart Failure Among Persons With Coronary Heart Disease: Data From the Heart and Soul Study Joachim H. Ix, Michael G. Shlipak, Glenn M. Chertow, and Mary A. Whooley Circulation 2007 115: 173 - 179; published online before print December 26 2006, doi:10.1161/CIRCULATIONAHA.106.644286

Epidemiology

Elevated Homocysteine Is Associated With Reduced Regional Left Ventricular Function: The Multi-Ethnic Study of Atherosclerosis Khurram Nasir, Michael Tsai, Boaz D. Rosen, Veronica Fernandes, David A. Bluemke, Aaron R. Folsom, and João A.C. Lima Circulation 2007 115: 180 - 187; published online before print January 2 2007, doi:10.1161/CIRCULATIONAHA.106.633750 Greater Fish, Fruit, and Vegetable Intakes Are Related to Lower Incidence of Venous Thromboembolism: The Longitudinal Investigation of Thromboembolism Etiology Lyn M. Steffen, Aaron R. Folsom, Mary Cushman, David R. Jacobs, Jr, and Wayne D. Rosamond Circulation 2007 115: 188 - 195; published online before print December 18 2006, doi:10.1161/CIRCULATIONAHA.106.641688

Health Services and Outcomes Research

Regional Differences in Process of Care and Outcomes for Older Acute Myocardial Infarction Patients in the United States and Ontario, Canada Dennis T. Ko, Harlan M. Krumholz, Yongfei Wang, JoAnne M. Foody, Fredrick A. Masoudi, Edward P. Havranek, John J. You, David A. Alter, Therese A. Stukel, Alice M. Newman, and Jack V. Tu Circulation 2007 115: 196 - 203; published online before print December 26 2006, doi:10.1161/CIRCULATIONAHA.106.657601

Heart Failure

Effects of Cardiac Resynchronization Therapy With or Without a Defibrillator on Survival and Hospitalizations in Patients With New York Heart Association Class IV Heart Failure JoAnn Lindenfeld, Arthur M. Feldman, Leslie Saxon, John Boehmer, Peter Carson, Jalal K. Ghali, Inder Anand, Steve Singh, Jonathan S. Steinberg, Brian Jaski, Teresa DeMarco, David Mann, Patrick Yong, Elizabeth Galle, Fred Ecklund, and Michael Bristow Circulation 2007 115: 204 - 212; published online before print December 26 2006, doi:10.1161/CIRCULATIONAHA.106.629261

Hypertension

Promotion of Faster Weight Gain in Infants Born Small for Gestational Age: Is There an Adverse Effect on Later Blood Pressure? Atul Singhal, Tim J. Cole, Mary Fewtrell, Kathy Kennedy, Terence Stephenson, Alun Elias-Jones, and Alan Lucas Circulation 2007 115: 213 - 220; published online before print December 18 2006, doi:10.1161/CIRCULATIONAHA.106.617811 Cardiac and Systemic Hemodynamic Characteristics of Hypertension and Prehypertension in Adolescents and Young Adults: The Strong Heart Study Jennifer S. Drukteinis, Mary J. Roman, Richard R. Fabsitz, Elisa T. Lee, Lyle G. Best, Marie Russell, and Richard B. Devereux Circulation 2007 115: 221 - 227; published online before print January 8 2007, doi:10.1161/CIRCULATIONAHA.106.668921

Imaging

Subclinical Coronary and Aortic Atherosclerosis Detected by Magnetic Resonance Imaging in Type 1 Diabetes With and Without Diabetic Nephropathy Won Yong Kim, Anne Sofie Astrup, Matthias Stuber, Lise Tarnow, Erling Falk, René M. Botnar, Cheryl Simonsen, Lotte Pietraszek, Peter R. Hansen, Warren J. Manning, Niels T. Andersen, and Hans-Henrik Parving Circulation 2007 115: 228 - 235; published online before print December 26 2006, doi:10.1161/CIRCULATIONAHA.106.633339 Rapid Detection of Myocardial Infarction by Subsecond, Free-Breathing Delayed Contrast-Enhancement Cardiovascular Magnetic Resonance Burkhard Sievers, Michael D. Elliott, Lynne M. Hurwitz, Timothy S.E. Albert, Igor Klem, Wolfgang G. Rehwald, Michele A. Parker, Robert M. Judd, and Raymond J. Kim Circulation 2007 115: 236 - 244; published online before print January 2 2007, doi:10.1161/CIRCULATIONAHA.106.635409 Molecular Cardiology

Tumor Necrosis Factor- Induces Endothelial Dysfunction in Leprdb Mice Xue Gao, Souad Belmadani, Andrea Picchi, Xiangbin Xu, Barry J. Potter, Neera Tewari-Singh, Stefano Capobianco, William M. Chilian, and Cuihua Zhang Circulation 2007 115: 245 - 254; published online before print January 2 2007, doi:10.1161/CIRCULATIONAHA.106.650671 Hyperhomocysteinemia Alters Cardiac Substrate Metabolism by Impairing Nitric Oxide Bioavailability Through Oxidative Stress Nobuhiro Suematsu, Caroline Ojaimi, Shintaro Kinugawa, Zipping Wang, Xiaobin Xu, Akos Koller MD, Fabio A. Recchia, and Thomas H. Hintze Circulation 2007 115: 255 - 262; published online before print January 2 2007, doi:10.1161/CIRCULATIONAHA.106.652693 Controversies in Cardiovascular Medicine

The Case for Renal Artery Stenting for Treatment of Renal Artery Stenosis Christopher J. Cooper and Timothy P. Murphy Circulation 2007 115: 263 - 270, doi:10.1161/CIRCULATIONAHA.106.619015 Case Against Angioplasty and Stenting of Atherosclerotic Renal Artery Stenosis Lance D. Dworkin and Kenneth A. Jamerson Circulation 2007 115: 271 - 276, doi:10.1161/CIRCULATIONAHA.106.619031 Congenital Heart Disease for the Adult Cardiologist

Ebstein’s Anomaly Christine H. Attenhofer Jost, Heidi M. Connolly, Joseph A. Dearani, William D. Edwards, and Gordon K. Danielson Circulation 2007 115: 277 - 285, doi:10.1161/CIRCULATIONAHA.106.619338 Images in Cardiovascular Medicine

Electrocardiogram, Echocardiography, and Magnetic Resonance Imaging Characteristics in Uhl’s Disease Akli Otmani, Laurent Leborgne, Cédric Renard, Houssam Bakkour, Serge Quenum, Christophe Tribouilloy, and Jean-Luc Rey Circulation 2007 115: e11 - e12, doi:10.1161/CIRCULATIONAHA.106.630061 Magnetic Resonance Imaging of a Posttraumatic Myocardial Infarction and Ventricular Septal Defect With a Closure Device in Place Stephanie L. Jun, Nikhil K. Chanani, Phillip Moore, and Charles B. Higgins Circulation 2007 115: e13 - e15, doi:10.1161/CIRCULATIONAHA.106.631275

Sinus of Valsalva Aneurysm With Right Ventricular Outflow Tract Obstruction: Evaluation With Doppler, Real-Time 3-Dimensional and Contrast Echocardiography Eli V. Gelfand, Dorota Bzymek, and Michael T. Johnstone Circulation 2007 115: e16 - e17, doi:10.1161/CIRCULATIONAHA.106.641043 Correspondence

Letter by Ben-Dov and Bursztyn Regarding Article, "Role of Diuretics in the Prevention of Heart Failure: The Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial" Iddo Z. Ben-Dov and Michael Bursztyn Circulation 2007 115: e18, doi:10.1161/CIRCULATIONAHA.106.639617 Response to Letter Regarding Article, "Role of Diuretics in the Prevention of Heart Failure: The Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT)" Barry R. Davis, Linda B. Piller, Kay Dunn, Jeffrey A. Cutler, Michael Proschan, Curt Furberg, David Goff, Stanley Franklin, Frans Leenen, Syed Mohiuddin, Vasilios Papademetriou, Allan Ellsworth, John Golden, Pedro Colon, and Richard Crow Circulation 2007 115: e19, doi:10.1161/CIRCULATIONAHA.106.651539 Acknowledgment of Reviewers

Acknowledgment of Reviewers

Circulation 2007 115: e20 - e30 Correction

Correction

Circulation 2007 115: e31, doi:10.1161/CIRCULATIONAHA.106.180989 European Perspectives

European Perspectives

Circulation 2007 115: 5f - 8f

Circulation JOURNAL OF THE AMERICAN HEART ASSOCIATION IIssssuuee HHiigghhlliigghhttss Vol 115, No 2, January 16, 2007

REGIONAL DIFFERENCES IN PROCESS OF CARE SUBCLINICAL CORONARY AND AORTIC AND OUTCOMES FOR OLDER ACUTE MYOCARDIAL ATHEROSCLEROSIS DETECTED BY MAGNETIC INFARCTION PATIENTS IN THE UNITED STATES RESONANCE IMAGING IN TYPE 1 DIABETES AND ONTARIO, CANADA, by Ko et al. WITH AND WITHOUT DIABETIC NEPHROPATHY, by Kim et al. Contrasting the care of patients in different countries can provide insights into the effect of clinical strategies and Patients with type 1 diabetes and nephropathy maintain higher systems on patient outcomes. Several studies have investigated cardiovascular mortality compared with diabetic patients with nor- whether the more expensive care in the United States for moalbuminuria. Risk assessment in asymptomatic patients with type patients with acute myocardial infarction provides better out- 1 diabetes is less stringent than for those with type 2 diabetes, and the comes than those achieved in Canada, where care is less effect of diabetic nephropathy on atherosclerosis in type 1 diabetes is expensive. These studies, however, have generally been lim- not well described. In this cross-sectional study of 136 asymptomatic ited by the use of administrative claims data, which lack type 1 diabetic patients with longstanding diabetes, Kim et al use clinical detail. Investigators from both countries have now cardiovascular magnetic resonance imaging to reveal greater coro- made use of data sets derived from extensive medical record nary plaque burden and higher prevalence of coronary artery stenoses abstractions to evaluate how the practice and outcomes for in patients with diabetic nephropathy compared with those with patients in Ontario compare with various regions of the United normoalbuminuria. These data suggest that coronary atherothrombo- States. The study by Ko and colleagues reveals striking sis may play an important role in the overall high cardiovascular variations in practice, along with some similarities, and raises mortality in patients with type 1 diabetes and nephropathy. See p 228. questions about whether the contrasting styles translate into differences in mortality rates. See p 196 (editorial p 158).

Visit http://circ.ahajournals.org: EFFECTS OF CARDIAC RESYNCHRONIZATION Images in Cardiovascular Medicine THERAPY WITH OR WITHOUT A Electrocardiogram, Echocardiography, and Magnetic Reso- DEFIBRILLATOR ON SURVIVAL AND nance Imaging Characteristics in Uhl’s Disease. See p e11. HOSPITALIZATIONS IN PATIENTS WITH NEW YORK HEART ASSOCIATION CLASS IV HEART Magnetic Resonance Imaging of a Posttraumatic Myocar- FAILURE, by Lindenfeld et al. dial Infarction and Ventricular Septal Defect With a Closure Device in Place. See p e13. Cardiac resynchronization therapy and the implantable defi- Sinus of Valsalva Aneurysm With Right Ventricular Outflow brillator, alone or in combination, for the appropriate patients, Tract Obstruction: Evaluation With Doppler, Real-Time have been proven to be a lifesaving adjunct to medical therapy 3-Dimensional and Contrast Echocardiography. See p e16. in patients with heart failure. Although the clinical benefits of these sophisticated electrophysiological devices are unques- tioned, patient selection to optimize risk-benefit and cost effectiveness remains a difficult challenge. In this issue of Circulation, Lindenfeld et al probe the Comparison of Medical Therapy, Pacing, and Defibrillation in Heart Failure (COM- PANION) study data and show that these devices can reduce rates of death and hospitalizations in ambulatory New York Heart Association class IV patients. See p 204 (editorial p 161). Correspondence See p e18.

Downloaded from circ.ahajournals.org 157at Mohammed Mahboob on June 10, 2007 Editorial

The Cardiovascular World Is Definitely Not Flat

P. Kaul, PhD; E.D. Peterson, MD, MPH

homas Friedman’s revolutionary book The World Is overall use of these medicines, these contemporary data show Flat describes how recent changes in technology are regional variability that mirror results found by Ko et al. AMI Ttransforming the world.1 The ubiquitous nature of the patients hospitalized in the northeastern and midwestern US Internet now facilitates rapid information exchange, thus regions remain more likely to receive ␤-blockers than their empowering the consumer. As a result, the world is becoming western or southern peers, with mean hospital rates of 90% more “flat”; ie, traditional regional boundaries are breaking and 86% versus 83% and 81%, respectively. Third, and down, and opportunities are more evenly spread. In stark perhaps most important, the data demonstrate that interhos- contrast to this new world of commerce, Ko and colleagues2 pital variability in care dwarfs regional variability. For demonstrate just how “nonflat” the cardiovascular world has example, within a western US region, the standard deviation been. Rather than homogenized delivery of evidence-based in hospital-to-hospital ␤-blocker use was 22%, whereas the care, the authors observe remarkable intercountry and intra- 25th to 75th hospital distribution ranged from 79% to 98%, country variability in both treatments and outcomes. twice as large as seen among the regional averages. Thus, although Ko et al emphasize that national summary statistics Article p 196 may be misleading because of regional differences, these data The study by Ko et al, published in this week’s Circula- point out that even regional classifications mask remarkable tion, was well performed. The authors had access to detailed local site-to-site variability. clinical information on a large random sample of US Medi- Beyond using evidence-based medicines, the Ko et al study care acute myocardial infarction (AMI) patients and a similar also examined variability in access to invasive cardiac pro- cohort of elderly AMI patients from Ontario, Canada, all cedures, including cardiac catheterization, percutaneous cor- hospitalized between 1998 and 2001. The authors paid special onary intervention, and coronary artery bypass surgery attention to creating similar entry criteria and examining (CABG). All would agree that eligible AMI patients have comparable process indicators. Short and intermediate out- access to and should get an aspirin, a ␤-blocker, an angio- comes also were compared only after risk adjustment. Over- tensin-converting enzyme inhibitor, and a statin. Access to all, the study found that AMI patients hospitalized in the invasive facilities is limited, however, and there exists diver- United States were slightly less likely to receive aspirin, gent clinical opinion on who should receive such procedures ␤-blockers, or angiotensin-converting enzyme inhibitors but after an AMI. As the findings of Ko et al clearly demonstrate, much more likely to receive invasive cardiac procedures than regional access to invasive procedures tended to be a remark- their Canadian peers. Yet, the study also demonstrates that ably strong predictor of subsequent use. the term, US AMI care, is a misnomer in that regional Such analyses always leave a reader wondering which rate differences within the United States were as large as or larger is right and whether any of these differences matter. Are the than those seen across international boundaries. tighter controls on technology seen in Ontario resulting in An important first question is whether these findings are of more efficient cardiac care? Or are they depriving AMI historical interest or applicable to current day practice. Since the patients of needed, lifesaving procedures? Although observa- 1990s, many publications have highlighted the problems of tional studies cannot fully determine these answers, regional underuse of evidence-based AMI therapies. Thankfully, recent analyses can provide provocative yet conflicting suggestions. data demonstrate that the gap between evidence and practice First, it is interesting to compare regional patterns of patient is narrowing.3–6 Using data available from the US Health and selection for catheterization. While randomized trials have Human Services Web site (http://www.hospitalcompare. consistently demonstrated that an invasive treatment strategy hhs.gov) we display 2004 to 2005 post–AMI discharge is most beneficial among older and higher risk AMI patients,7 ␤ -blocker use from 3462 US hospitals divided into the 4 the current study finds paradoxically that younger and health- major census regions. The data provide several messages. ier patients receive more interventions in all US regions as ␤ First, overall community use of -blockers appears to have well as in Ontario (Table 4). Thus, physicians in all regions risen since the study by Ko et al. Second, despite higher may have opportunities to improve their procedure selection criteria to maximize use among those most in need.8 Regional data also allow one to examine how differences The opinions expressed in this article are not necessarily those of the editors or of the American Heart Association. in use of a diagnostic procedure affect subsequent treatment. From Duke Medical Center, Durham, NC. A goal of cardiac catheterization is to identify those with Correspondence to Eric Peterson, MD, Duke Medical Center, Durham, severe coronary disease, a setting in which revascularization NC 27715. E-mail [email protected] 9 (Circulation. 2007;115:158-160.) can improve survival relative to conservative care. If regions © 2007 American Heart Association, Inc. with higher catheterization rates have lower efficiency for Circulation is available at http://www.circulationaha.org detecting severe disease, then the word “flat” may more DOI: 10.1161/CIRCULATIONAHA.106.672022 accurately describe “flat of the curve” medical care.10 Al-

158

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Kaul and Peterson The Cardiovascular World Is Definitely Not Flat 159

Variability in AMI discharge ␤-blocker use by US regions. nϭ3462 hospitals. Box demonstrates 25th to 75th regional percentile. Line indicates median; ϩ, mean; whiskers, 1.5ϫinterquartile range; and *, outlier hospitals.

though the Ko et al study does not give a breakdown by sons. First, they appear to be contradictory to earlier works region of coronary anatomy findings, others have studied this done by this group. In a similar analysis of patients hospital- question and have found the diagnostic efficiency of US and ized for AMI in 1991 to 1992, Tu et al11 found slightly lower Canadian catheterization strategies to be identical.8 30-day mortality rates in US patients compared with Cana- As a surrogate, the Ko et al study provides data on the dian patients and no difference in 1-year mortality rates. percentage of patients receiving subsequent CABG or percu- Interestingly, during the interval between the 2 study popu- taneous coronary intervention per diagnostic procedure per- lations, differences in cardiac catheterization rates in the 2 formed. On the basis of data from the Ko et al Table 3, the countries actually attenuated, more as a result of the increase rate of CABG per catheterization performed in Ontario is in rates in Canada (from 6.7% to 16.8%) than in the United higher than that seen in the overall US population (23.8% States (34.9% in 1991–1992 to 38.7% in 1998–2001). versus 16.5%). The US regional analysis gives a much less Second, these findings are in contrast to those observed consistent message, however. The Northeast (with the most from the Global Utilization of Streptokinase and t-PA for frugal catheterization rate) paradoxically also has the lowest Occluded Coronary Arteries (GUSTO-I) trial. In GUSTO, CABG yield per catheterization performed (13.7%). And if 1-year mortality rates were equivalent between the United the metric is changed to any revascularization procedure, then States and Canada, and 5-year mortality actually favored the the Midwest (with one of the most liberal catheterization United States (19.6% versus 21.4%, PϽ0.0001; risk-adjusted strategies) also has the highest rate of percutaneous coronary Canadian hazard ratio, 1.17; 95% CI, 1.07 to 1.28).12 There intervention or CABG per diagnostic study performed (71%), are many possible explanations for the differences in out- compared with 62.5% in Ontario. Thus, it appears that the set comes observed among these cohort studies. The GUSTO point for revascularization and catheterization appears to vary population represents those selected for a clinical trial, yet it markedly among regions.8 also is representative of findings in an all-ages patient Beyond such surrogates, the authors appropriately empha- population and those treated in all regions of Canada. In size that patient outcomes should be the ultimate metric for contrast, the Ko et al cohort represents an unselective, comparing differing care strategies. Despite the varied use of elderly-only population from Ontario. Interestingly, in the medicines and invasive procedures, Ko and colleagues found GUSTO population, the revascularization differential be- that 30-day risk–adjusted mortality rates in the United States tween United States and Canada was also more marked than and Canada were not different. In contrast, within the United seen in the present study (the GUSTO US-Canadian CABG States itself, 30-day mortality rates actually did differ among rates of 13.1% versus 4% compared with US-Ontario rates of regions, from a low of 15.3% (14.9% to 16.5%) in the 6.4% versus 4%). Additionally, there were differences in the Northeast to a high of 19.0% (18.4% to 19.6%) in the South. modeling approaches used. Ko et al were unable to combine By 1 year, the international comparisons had changed. Stan- the data across countries and therefore calculated an esti- dardized mortality rates became significantly higher in the mated standardized mortality rate for both countries using a United States compared with Canada (31.9% versus 27.7%) published short-term mortality model.13 In contrast, the and remained so at 3 years (45.9% in the United States, GUSTO analysis identified factors associated with 5-year 40.3% in Canada). mortality in a combined database and directly assessed These differences between the United States and Canada in intercountry risks. Unfortunately, the Ko et al study does not intermediate mortality rates are interesting for several rea- provide observed crude mortality rates, so it remains unclear

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 160 Circulation January 16, 2007

whether differential adjustment methodologies contributed to 3. Krumholz HM, Chen J, Rathorse SS, Wang Y, Radford MJ. Regional these divergent conclusions. variation in the treatment and outcomes of myocardial infarction: inves- tigating New England’s advantage. Am Heart J. 2003;146:242–249. Although survival is certainly an important end point, it 4. O’Conner GT, Quinton HB, Traven ND, Ramunno LD, Dodds TA, may not be a sensitive or specific metric for assessing the Marciniak TA, Wennberg JE. Geographic variation in the treatment of downstream impact of differential treatment strategies. In acute myocardial infarction: the Cooperative Cardiovascular Project. JAMA. 1999;281:627–633. particular, beyond cardiovascular care, subtle differences in 5. Krumholz HM, Radford MJ, Wang Y, Chen J, Heiat A, Marciniak TA. baseline comorbidity or social and environmental factors can National use and effectiveness of beta-blockers for the treatment of significantly influence one’s lifespan after an AMI. As an elderly patients after acute myocardial infarction: National Cooperative Cardiovascular Project. JAMA. 1998;280:623–629. alternative outcome assessment, patient symptom status and 6. Jencks SF, Huff ED, Cuerdon T. Change in the quality of care delivered functional outcomes may be considered when the impact of to Medicare beneficiaries, 1998–1999 to 2000–2001. JAMA. 2003;289: differential treatment patterns is measured. Although the Ko 305–312. et al study did not provide such data, the GUSTO trial has 7. Cannon CP, Weintraub WS, Demopoulos LA, Vicari R, Frey MJ, Lakkis N, Neumann FJ, Robertson DH, DeLucca PT, DiBattiste PM, Gibson previously reported that lower revascularization rates in CM, Braunwald E, for the TACTICS (Treat Angina With Aggrastat and Canada were associated with greater functional impairment at Determine Cost of Therapy With an Invasive or Conservative Strat- 1 year.14 Similarly, another study found that a more conser- egy)–Thrombolysis in Myocardial Infarction 18 Investigators. Com- parison of early invasive and conservative strategies in patients with vative use of revascularization in blacks relative to whites unstable coronary syndromes treated with the glycoprotein IIb/IIIa inhib- also was associated with a higher burden of angina symptoms itor tirofiban. N Engl J Med. 2001;344:1879–1887. and cardiac impairment among the black patients.15 Although 8. Batchelor WB, Mark DB, Knight JD, Granger CB, Armstrong PW, Califf RM, Peterson ED. Development and validation of a simple model to none of these works are definitive, they stress the need for a predict severe coronary artery disease after myocardial infarction: broad assessment of the downstream consequences of differ- potential impact on cardiac catheterization use in the United States and ential care. Canada. Am Heart J. 2003;145:349–355. In summary, the work by Ko et al provides important 9. CASS Investigators. Myocardial infarction and mortality in the Coronary Artery Surgery Study (CASS) randomized trial. N Engl J Med. 1984;310: insights. Their study clearly demonstrates that the 3 key 750–758. determinants of one’s AMI care may not be the evidence but 10. Enthoven AC. Shattuck Lecture: cutting cost without cutting the quality rather, as in real estate, “location, location, location.” But will of care. N Engl J Med. 1978;298:1229–1238. 11. Tu JV, Pashos CL, Naylor CD, Chen E, Normand SL, Newhouse JP, the next generation’s cardiovascular care be different? Med- McNeil BJ. Use of cardiac procedures and outcomes in elderly patients icine is just beginning to enter Friedman’s brave new world with myocardial infarction in the United States and Canada. N Engl of technology-assisted open access to information. As evi- J Med. 1997;336:1500–1505. 12. Kaul P, Armstrong PW, Chang WC, Naylor CD, Granger CB, Lee KL, denced by the Figure, the public can, with a click of a mouse, Peterson ED, Califf RM, Topol EJ, Mark DB. Long-term mortality of freely view their local hospital’s AMI care quality. In an even patients with acute myocardial infarction in the United States and Canada: more revolutionary development, payers are rapidly attaching comparison of patients enrolled in the Global Utilization of Streptokinase and t-PA for Occluded Coronary Arteries (GUSTO-I). Circulation. 2004; strong financial incentives to such quality indicator informa- 110:1754–1760. tion under various “pay for performance” programs.16 Will 13. Granger CB, Goldberg RJ, Dabbous O, Pieper KS, Eagle KA, Cannon this new era of open accountability drive down the long- CP, Van de Werf F, Avezum A, Goodman SG, Flather MD, Fox KA. witnessed variability in medical care and thereby “flatten the Global registry of acute coronary events, I: predictors of hospital mor- tality in the global registry of acute coronary events. Arch Intern Med. cardiovascular world?” Stay tuned.... 2003;163:2345–2353. 14. Mark DB, Naylor CD, Clapp-Channing NE, Sutherland W, Pilote L, Disclosures Armstrong PW. Use of medical resources and quality of life after acute myocardial infarction in Canada and the United States. N Engl J Med. None. 1994;331:1130–1135. 15. Kaul P, Lytle BL, Spertus JA, DeLong ER, Peterson ED. Influence of References racial disparities in procedure use on functional status outcomes among 1. Freidman TL. The World is Flat: A Brief History of the Twenty-First patients with coronary artery disease. Circulation. 2005;111:1284–1290. Century. New York, NY: Farrar, Straus and Giroux; 2005. 16. Rosenthal MB, Landon BE, Normand SLT, Frank RG, Epstein AM. Pay 2. Ko DT, Krumholz HM, Wang Y, Foody JM, Masoudi FA, Havranek EP, for performance in commercial HMOs. N Engl J Med. 2006;355: You JY, Alter DA, Stukel TA, Newman AM, Tu JV. Regional differences 1985–1902. in process of care and outcomes of older acute myocardial infarction patients in the United States and Ontario, Canada. Circulation. 2007;115: KEY WORDS: Editorials Ⅲ catheterization Ⅲ myocardial infarction 196–203. Ⅲ quality of health care Ⅲ survival

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Editorial

Cardiac Resynchronization Therapy in New York Heart Association Class IV Heart Failure It Is All About Selection

W.H. Wilson Tang, MD; Gary S. Francis, MD

n recent years, cardiac resynchronization therapy (CRT) ventricular tachyarrhythmias). In fact, the 217 NYHA class has emerged as an important therapeutic strategy in IV subjects in the COMPANION trial encompassed almost Ipatients with advanced heart failure. It has become half (48%) of all NYHA class IV subjects enrolled in common practice to use CRT in combination with implant- multicenter CRT trials published to date. The COMPANION able cardioverter-defibrillator (so-called CRT-D) in patients investigators observed that CRT and CRT-D significantly with impaired left ventricular systolic function and New York improved time to the combined end point of all-cause death Heart Association (NYHA) class III symptoms.1 Evidence is or heart failure hospitalization compared with optimal med- also mounting regarding the potential benefit of CRT in ical therapy.9 Some interesting insights become obvious, delaying disease progression in patients with symptomatic however, after careful scrutiny of the data presented by heart failure.2,3 However, the role of CRT and CRT-D is less Lindenfeld and colleagues.9 clear in patients with very advanced heart failure, especially First, the line between the NYHA classes is not distinct; it in those with NYHA class IV symptoms (those with symp- is always more or less in the eyes of the beholder and can be Ͻ toms at rest and worsening with exertion). In fact, 5% of all highly variable.10 In general, NYHA classification is judged heart failure subjects enrolled in large multicenter mortality clinically by the patient’s symptom burden either by self- device trials fulfilled this category.4 Although there has been report or by physician assessment. It does not necessarily increasing debate over this issue,5–7 patients in the intensive reflect the underlying chronicity or stability of the clinical care unit requiring inotropic and mechanical support are still presentation. In fact, determining disease severity in heart not considered suitable candidates for CRT or CRT-D “sal- failure still requires a wide range of clinical, biochemical, and 8 vage” therapy. Some patients with very advanced, refractory functional parameters. A “staging” process has been used in heart failure are ambulatory, however; in these patients, transplantation evaluation, but universally accepted and de- neither cardiac transplantation nor permanent mechanical finable measures are still lacking. Therefore, drawing com- support devices are appropriate or imminent. Many of these parisons between patients, let alone between trials, can be a patients are still considered as approaching “stage D heart great challenge. For now, the true testimony of heart failure failure,” in which they can be best characterized by experi- severity in any trial lies in the direct comparison of annual- encing end-stage disease refractory to optimal medical ther- ized mortality with other trials, which can be vastly different apy. Even now, we have few data in such patients because even when patients were selected within the same NYHA they have been systematically excluded or avoided in most III-IV class range. As described in the Lindenfeld et al article, CRT trials, perhaps because of a presumed shortened the 1-year mortality rate for the NYHA class IV optimal lifespan. medical therapy cohort in COMPANION (44%), even in the Article p 204 best-case scenario with CRT-D (30%), was higher than most contemporary neurohormonal antagonist and oral inotropic With this in mind, the Comparison of Medical, Pacing, and Defibrillation Therapies in Heart Failure (COMPANION) drug trials in severe heart failure (12% to 26%) and only trial investigators retrospectively reviewed their experience slightly lower than that observed in the intravenous inotropic with this severely symptomatic but ambulatory cohort to drug trials and destination therapy trials (51% to 76%). This assess the benefits of both CRT and CRT-D.9 As described, is perhaps because of the requirement of a wide QRS duration COMPANION is the single largest clinical experience of and a heart failure hospitalization in the preceding 12 months CRT/CRT-D devices implanted in patients with NYHA class at the time of enrollment, both markers of advanced diseases IV heart failure (without spontaneous or inducible sustained and powerful predictors of mortality. Hence, the NYHA class IV experience in COMPANION is perhaps the closest to what we would consider a stage D heart failure cohort. The opinions expressed in this article are not necessarily those of the Previous studies also have suggested that patients in editors or of the American Heart Association. From the Department of Cardiovascular Medicine, Cleveland Clinic, NYHA class IV may derive significant benefit from a Cleveland, Ohio. reduction of sudden cardiac death, presumably by receiving Correspondence to Gary S. Francis, MD, 9500 Euclid Ave, Desk F15, more frequent but appropriate delivery of defibrillator thera- Cleveland, OH 44195. E-mail [email protected] 11 (Circulation. 2007;115:161-162.) py. This is confirmed in this COMPANION analysis, in © 2007 American Heart Association, Inc. which adjusted rates for sudden cardiac death were found to Circulation is available at http://www.circulationaha.org be significantly lower in the CRT-D group compared with the DOI: 10.1161/CIRCULATIONAHA.106.669879 CRT or optimal medical therapy groups in the NYHA class

161

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 162 Circulation January 16, 2007

IV cohort. Furthermore, there was improvement in self- toms? The precise algorithm for this selection process to reported quality-of-life scores and changes in NYHA class in achieve long-term clinical benefits still needs to be defined. both NYHA class III and IV cohorts, although this was practically an unblinded comparison between optimal medi- Disclosures cal therapy (no device implantation) and CRT/CRT-D device Dr Tang is a consultant for Medtronic, Inc, and Boston Scientific, implantation. However, it is more difficult to explain the fact Inc. Dr Francis reports no conflicts. that the incidence of heart failure death (heart failure caused References by “pump failure”) or time to heart failure death was not 1. Daubert JC, Leclercq C, Donal E, Mabo P. Cardiac resynchronisation significantly affected by CRT/CRT-D therapies in this therapy in heart failure: current status. Heart Fail Rev. 2006;11:147–154. COMPANION analysis. It is important to emphasize that 2. Rivero-Ayerza M, Theuns DA, Garcia-Garcia HM, Boersma E, Simoons while a mortality benefit was observed with the use of M, Jordaens LJ. Effects of cardiac resynchronization therapy on overall mortality and mode of death: a meta-analysis of randomized controlled CRT-only devices in the Cardiac Resynchronization Heart trials. Eur Heart J. 2006;27:2682–2688. Failure (CARE-HF) trial, most patients enrolled had predom- 3. Vanderheyden M, Wellens F, Bartunek J, Verstreken S, Walraevens M, inantly NYHA class III symptoms, and the annualized mor- Geelen P, De Proft M, Goethals M. Cardiac resynchronization therapy delays heart transplantation in patients with end-stage heart failure and tality rate was only 12%.12 Although the sample size was too mechanical dyssynchrony. J Heart Lung Transplant. 2006;25:447–453. small to draw any conclusions, it is conceivable that the apparent 4. Lehmann MH, Aaronson KD. CRT-D therapy in heart failure: how much diminishing returns of CRT in reducing heart failure deaths can do NYHA class IV patients benefit? J Cardiovasc Electrophysiol. 2006; be explained by the difference in the COMPANION population 17:491–494. 5. Cowburn PJ, Patel H, Jolliffe RE, Wald RW, Parker JD. Cardiac resyn- rather than the degree of symptom burden (NYHA class IV), chronization therapy: an option for inotrope-supported patients with whereby reverse remodeling in this stage D cohort may not be end-stage heart failure? Eur J Heart Fail. 2005;7:215–217. realistically possible. That being said, there is currently no 6. James KB, Militello M, Barbara G, Wilkoff BL. Biventricular pacing for heart failure patients on inotropic support: a review of 38 consecutive evidence to demonstrate that defibrillators alone can be as cases. Tex Heart Inst J. 2006;33:19–22. effective as CRT-D in this population. 7. Konstantino Y, Iakobishvili Z, Arad O, Ben-Gal T, Kusniec J, Mazur A, What is reassuring from this COMPANION post hoc Porter A, Strasberg B, Battler A, Hasdai D. Urgent cardiac resynchroni- analysis is that appropriately selected ambulatory patients zation therapy in patients with decompensated chronic heart failure receiving inotropic therapy: a case series. Cardiology. 2006;106:59–62. with severe but stable heart failure and no need for inotropic 8. Strickberger SA, Conti J, Daoud EG, Havranek E, Mehra MR, Pina IL, support can have acceptable risks for procedure-related com- Young J. Patient selection for cardiac resynchronization therapy: from the plications. These observations have clearly raised some Council on Clinical Cardiology Subcommittee on Electrocardiography and Arrhythmias and the Quality of Care and Outcomes Research Inter- important challenges, however, particularly in light of limited disciplinary Working Group, in collaboration with the Heart Rhythm published NYHA class IV and/or stage D heart failure Society. Circulation. 2005;111:2146–2150. experience. Clearly, the role of CRT in the management of 9. Lindenfeld J, Feldman AM, Saxon L, Boehmer J, Carson P, Ghali JK, advanced heart failure with NYHA class IV symptoms Anand I, Singh S, Steinberg JS, Jaski B, DeMarco T, Mann D, Yong P, Galle E, Ecklund F, Bristow M. Effects of cardiac resynchronization depends largely on appropriate patient selection, much the therapy with or without a defibrillator on survival and hospitalizations in same way as determining the risk-to-benefit ratio of various patients with New York Heart Association class IV heart failure. Circu- other stage D treatment options. The goals and potential lation. 2007;115:204–212. 10. Hauptman PJ, Masoudi FA, Weintraub WS, Pina I, Jones PG, Spertus JA. benefits of CRT/CRT-D therapy can be very different as Variability in the clinical status of patients with advanced heart failure. symptom burden increases, and Lindenfeld and colleagues J Card Fail. 2004;10:397–402. have presented insightful findings that call for more research 11. Desai AD, Burke MC, Hong TE, Kim S, Salem Y, Yong PG, Knight BP. in this area. But how can we better distinguish a stage C from Predictors of appropriate defibrillator therapy among patients with an implantable defibrillator that delivers cardiac resynchronization therapy. a stage D heart failure patient in the presence of overt NYHA J Cardiovasc Electrophysiol. 2006;17:486–490. class IV symptoms? Is there enough evidence to justify the 12. Cleland JG, Daubert JC, Erdmann E, Freemantle N, Gras D, morbidity and mortality benefits of CRT for all NYHA class Kappenberger L, Tavazzi L. The effect of cardiac resynchronization on morbidity and mortality in heart failure. N Engl J Med. 2005;352: IV patients beyond an implantable defibrillator? What is the 1539–1549. best way to determine a clinical benefit in this population with advanced disease: deterring death or improving symp- KEY WORDS: Editorials Ⅲ heart failure Ⅲ pacemakers Ⅲ prognosis

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Congenital Heart Disease

Congenital Heart Disease in the General Population Changing Prevalence and Age Distribution

Ariane J. Marelli, MD; Andrew S. Mackie, MD, SM; Raluca Ionescu-Ittu, MSc; Elham Rahme, PhD; Louise Pilote, MD, MPH, PhD

Background—Empirical data on the changing epidemiology of congenital heart disease (CHD) are scant. We determined the prevalence, age distribution, and proportion of adults and children with severe and other forms of CHD in the general population from 1985 to 2000. Methods and Results—Where healthcare access is universal, we used administrative databases that systematically recorded all diagnoses and claims. Diagnostic codes conformed to the International Classification of Disease, ninth revision. Severe CHD was defined as tetralogy of Fallot, truncus arteriosus, transposition complexes, endocardial cushion defects, and univentricular heart. Prevalence of severe and other CHD lesions was determined in l985, 1990, 1995, and 2000 using population numbers in Quebec. Children were subjects Ͻ18 years of age. The prevalence was 4.09 per 1000 adults in the year 2000 for all CHD and 0.38 per 1000 (9%) for those with severe lesions. Female subjects accounted for 57% of the adult CHD population. The median age of all patients with severe CHD was 11 years (interquartile range, 4 to 22 years) in 1985 and 17 years (interquartile range, 10 to 28 years) in 2000 (PϽ0.0001). The prevalence of severe CHD increased from 1985 to 2000, but the increase in adults was significantly higher than that observed in children. In the year 2000, 49% of those alive with severe CHD were adults. Conclusions—The prevalence in adults and median age of patients with severe CHD increased in the general population from 1985 to 2000. In 2000, there were nearly equal numbers of adults and children with severe CHD. (Circulation. 2007;115:163-172.) Key Words: adults Ⅲ children Ⅲ epidemiology Ⅲ heart defects, congenital Ⅲ prevalence

he reported prevalence of congenital heart disease (CHD) Methods 1–5 Tvaries between 4 and 10 per 1 000 live births. Ad- Data Acquisition vances in pediatric cardiac care have resulted in an increasing In Quebec (Canada), every individual is assigned a unique Medicare number of adults with CHD being followed up in tertiary care number in the first year of life. All diagnoses and health services centers.1,6,7 These data have been important in generating rendered are systematically recorded until death. Demographic interest in adult CHD as a new subspecialty of cardiology.6,7 records on all patients include the date of birth and death. Two administrative databases were used: the physician’s services and Clinical Perspective p 172 claims database (Re´gie de l’Assurance Maladie du Que´bec) from l983 to 2000 and the hospital discharge summary database (Med- Although estimates of the prevalence of CHD in adults Echo) from l987 to 2000. Through the use of encrypted numbers, have been generated,2,8 no empirical data are available on the both databases were linked for each patient. Diagnoses conform to the International Classification of Disease, ninth revision (ICD-9), epidemiology of CHD in the general adult population. To which has 24 codes to designate CHD anomalies. We created a date, no studies have documented the prevalence of CHD in population-based CHD database containing data files of all patients adults and children in the same population. Such data are diagnosed with CHD who came into contact with the healthcare important determinants of how healthcare resources should system in Quebec from January 1, 1983, to December 31, 2000 (inclusive). At every point in time, patients were divided on the basis be allocated to patients with CHD. of age; children were Ͻ18 years of age, and adults were Ն18 years Where access to health care is universal, diagnoses are of age. recorded for the population at large.9 For congenital cardiac Permission for the study was obtained from the McGill University lesions, we used administrative databases in a geographically Health Center ethics board and government agencies responsible for privacy of access to information in Quebec. focused region to determine the prevalence of CHD. Our goals were to measure the observed prevalence, age, and Diagnostic Codes and Classification of Disease proportions of adults relative to children with CHD in the Imitating the Report of the New England Regional Infant Cardiac general population from 1985 to 2000. Program, we established a hierarchy of CHD diagnoses and surgical

Received March 15, 2006; accepted November 10, 2006. From the McGill Adult Unit for Congenital Heart Disease Excellence (MAUDE Unit), Montreal, Quebec, Canada. Reprint requests to Ariane Marelli, MD, FRCP(C), FACC, Associate Professor of Medicine, McGill University, Director, McGill Adult Unit for Congenital Heart Disease, McGill University Health Center, 1650 Cedar Ave, Montreal, Canada H3G 1A4. E-mail [email protected] © 2007 American Heart Association, Inc. Circulation is available at http://www.circulationaha.org DOI: 10.1161/CIRCULATIONAHA.106.627224

163

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 164 Circulation January 16, 2007

TABLE 1. Hierarchy of CHD Diagnostic Codes Inclusion and Exclusion Criteria We included patients who had at least 1 ICD-9 CHD code and/or a Categorical surgical procedure specific to these codes between 1983 and 2000. Hierarchy Block Categorical Diagnosis ICD-9 Codes Diagnostic codes were cross-referenced between the hospital dis- 1 AVCD 7456 charge summary database, those found in the physician’s services TOF 7452 and claims database, and those inferred from the surgical procedure. When Ͼ1 ICD-9 code was recorded, we chose the most frequent Univentricular heart 7453 diagnosis in the diagnostic hierarchical block with the lowest Transposition complex* 7451 numerical value (Table 1), giving priority to the diagnosis made by a CV specialist. When diagnoses were discrepant between databases, Truncus arteriosus 7450 the records were manually reviewed by 1 pediatric cardiologist Hypoplastic left heart syndrome 7467 (A.S.M.) and 1 adult CHD specialist (A.J.M.). 2 ASD 7455 We excluded patients who had severe (block 1) or nonspecific (block 5) CHD codes with a diagnosis that was not made by a CV VSD 7454 specialist and who had not been hospitalized for CHD (Table 1). We PDA 7470 excluded patients with diagnoses in blocks 2 through 4 that had not Aortic coarctation 7471 been made by a primary care physician, CV specialist, or echocar- diographer (Table 1). We also excluded patients in whom a CHD Ebstein’s anomaly 7462 procedure had been claimed by a physician other than a CV surgeon 3 Unspecified defect of septal closure 7459 without an ICD-9 CHD diagnosis (Table 2). 4 Anomalies of the pulmonary artery 7473 Statistical Analysis Anomalies of the pulmonary valve 7460 Prevalence is defined as the proportion of individuals in a population Congenital tricuspid valve disease 7461 who have a disease at a specific point in time.10 We measured the Congenital aortic stenosis 7463 observed prevalence of CHD corresponding to the number of patients observable in our database between 1983 and 2000 by way Congenital aortic insufficiency 7464 of an ICD-9 CHD diagnosis in a given year divided by the Quebec Congenital mitral stenosis 7465 population in that year. Because CHD is present since birth, a patient Congenital mitral insufficiency 7466 contributed to the prevalence of all years he or she was alive regardless of when the diagnosis was made during that person’s life. Anomalies of great veins 7474 When a patient died, he or she no longer contributed to the 5 Other unspecified anomalies of the heart 7468/9 prevalence. Other unspecified anomalies of circulation 7479 Absolute numbers, proportions, and prevalence of adults and children are reported on the basis of the subject’s age for the years Other unspecified anomalies of the aorta 7472 1985, 1990, 1995, and 2000. For age- and sex-related calculations, *Complete or congenitally corrected transposition. both the CHD and Quebec populations had the same age and sex distribution. In 2000, the total population in Quebec was 7 357 029, of whom 5 760 295 were adults.11 Prevalence reported among interventions for patients with several diagnostic and surgical codes.5 children and adults for both severe and other CHD represents We classified the 24 ICD-9 diagnostic codes in hierarchical blocks 1 weighted averages of prevalence in more specific age groups and through 5 (Table 1). We designated as specific CHD procedures CHD lesions. To investigate trends in subgroups, we measured CHD those from which a CHD diagnosis could be inferred (Table 2). For prevalence in age groups of 1 to 12, 13 to 17, 18 to 25, 26 to 40, and each diagnosis made and/or procedure (invasive and noninvasive) Ͼ40 years and in atrial septal defect (ASD), ventricular septal defect performed, physicians were classified by specialty as a cardiovascu- (VSD), endocardial cushion defect (AVCD), and tetralogy of Fallot lar (CV) specialist (pediatric, medical or surgical), primary care (TOF). physician (general or family physician, pediatrician or internist), or Poisson bivariate models were used to quantify the change in CHD echocardiographer. prevalence from 1985 to 2000 and differences between males and Two categories of CHD severity were defined. We classified as females. Changes in prevalence were measured as prevalence ratios severe those lesions with an ICD-9 diagnosis that reflected the and 95% CIs. Age and sex distributions of different cohorts were highest probability of being associated with cyanosis or requiring summarized using medians (interquartile range [IQR]) and propor- surgical intervention early in life. These included all patients with a tions. Medians and proportions in different CHD cohorts were block 1 diagnosis (Table 1). All remaining lesions were classified as compared through the use of Wilcoxon and ␹2 tests. other CHD. To investigate the likelihood of misclassification of patients with unspecified defects, we compared subjects with diagnostic codes in TABLE 2. Hierarchy of CHD Surgical Codes block 5 with those having severe CHD (block 1) and those with diagnoses in blocks 2 through 4 (Table 1). We performed 2 logistic Surgical Procedures Billed by a regression models using cardiac hospitalization and CHD specific Cardiovascular Surgeon Inferred CHD Diagnosis surgery from 1983 to 2000 as dependent variables and age, sex, and Atrioventricular canal repair AVCD duration of follow-up as covariates. Results are reported using odds Atrial switch Transposition complex ratios (ORs) and 95% CIs. Analyses were performed with SAS statistical software (version 8.02, Cary, NC). Arterial switch Transposition complex The authors had full access to the data and take full responsibility Blalock-Hanlon procedure Transposition complex for their integrity. All authors have read and agree to the manuscript. Coarctation repair Aortic coarctation Results Fontan procedure Univentricular heart Fallot repair TOF Derivation of Patient Cohort Healthcare records were obtained from the administrative Patent ductus ligation PDA databases on 61 386 patients. Of the 45 960 patients included Shunt closure Unspecified defect of septal closure in our CHD database, 32 481 (71%) had a unique ICD-9

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Marelli et al Population Prevalence of Congenital Heart Disease 165

Figure 1. Derivation of study cohort.

CHD diagnosis used between 1983 and 2000. After manual excluded 1141 patients (7%) who had a CHD surgical review of discrepant diagnoses, the diagnosis was modified in procedure billed by a non-CV surgeon (Table 2). The num- 2861 patients, of whom 499 had surgical codes used to bers of excluded patients in each diagnostic block are modify the ICD-9 diagnosis. summarized in Figure 1. In the year 2000, 42 542 patients were alive: 18 979 children (45%) and 23 563 adults (55%). Of these, we Prevalence of CHD included 11 207 patients with unspecified CHD codes (block In 2000, the prevalence of CHD was 11.89 per 1000 children, 5) as patients with “other” CHD. The patients with unspeci- 4.09 per 1000 adults, and 5.78 per 1000 in the general fied CHD codes had diagnoses made by a CV specialist or population. The prevalence of severe CHD was 1.45 per 1000 had been hospitalized with a CHD diagnosis. Comparing children and 0.38 per 1000 adults, accounting for 12% and these patients with the 4521 patients classified as severe CHD 9% of all CHD lesions in children and adults, respectively (block 1) shows ORs and 95% CIs as follows: cardiac (Table 3). In children and adults, conotruncal anomalies and hospitalization, 0.20 (95% CI, 0.18 to 0.23); and CHD- AVCD were the most prevalent lesions among those with specific surgery, 0.02 (95% CI, 0.02 to 0.03). Comparing severe lesions, whereas ASD, VSD, and patent ductus arte- these patients with the 26 814 patients with diagnoses in riosus (PDA) were the most common lesions among those blocks 2 through 4 shows the following ORs and 95% CIs: with other CHD (Table 3). cardiac hospitalization, 0.57 (95% CI, 0.53 to 0.60); and CHD-specific surgery, 0.04 (95% CI, 0.03 to 0.05). There- Sex Distribution of CHD fore, patients with unspecified defects had cardiac hospital- In 2000, females accounted for 52% and 57% of the CHD izations and CHD rates that were 80% and 98% lower, population in children and adults, respectively. The propor- respectively, than those with severe CHD, indicating that tion of females in the CHD population was significantly misclassification was unlikely. higher than that of the Quebec population: 57% versus 51% We excluded 15 426 (25%) patients whose median year of in adults (PϽ0.0001) and 52% versus 49% in children birth was 1945 (IQR, 1925 to 1967). Of these, 14 285 (93%) (PϽ0.0001). The prevalence of all CHD in adults was had never been hospitalized for CHD and had a diagnosis in significantly higher in females than in males (4.55 per 1000 block 1 or 5 never made by a CV specialist or a diagnosis in females versus 3.61 per 1000 males; PϽ0.0001). In those blocks 2 through 4 never made by a CV specialist, primary with severe disease, the prevalence was higher in female care physician, or echocardiographer (Table 1). We also adults (0.41 per 1000 versus 0.35 per 1000; Pϭ0.0001) but

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 166 Circulation January 16, 2007

TABLE 3. Prevalence of Severe and Other CHD in a Population of 5 760 295 Adults and 1 596 734 Children in the Year 2000

Adults Alive in 2000 Children Alive in 2000

Prevalence Prevalence n (%) per 1000 Adults n (%) per 1000 Children All congenital heart lesions* 23 563 (100) 4.09 18 979 (100) 11.89 Severe lesions TOF or truncus arteriosus 1001 0.17 778 0.49 AVCD 834 0.14 914 0.57 Transposition complex 235 0.04 424 0.27 Univentricular hearts 150 0.03 213 0.13 All severe lesions 2205 (9) 0.38 2316 (12) 1.45 Other lesions ASD 5076 0.88 6205 3.89 VSD 4486 0.78 6709 4.20 PDA 103 0.02 493 0.31 Unspecified defect of septal closure 161 0.03 9 0.01 Unspecified congenital anomalies† 9621 1.67 1586 0.99 Congenital aortic stenosis or insufficiency‡ 619 0.11 425 0.27 Anomalies of pulmonary artery or valve 698 0.12 798 0.50 Aortic coarctation 389 0.07 396 0.25 Congenital mitral or tricuspid valve disease 178 0.03 60 0.04 Ebstein’s anomaly 50 0.01 29 0.02 Unknown congenital heart lesions 91 0.02 35 0.02 Anomalies of the great veins 25 0.00 24 0.01 All other lesions 21 358 (91) 3.71 16 663 (88) 10.44 *There are 154 adults (15 with severe CHD, 139 with other CHD) and 199 children (13 with severe and 106 with other CHD) to whom we have assigned 2 final CHD diagnoses. †Includes ICD-9 codes 7468, 7469, and 7479 (see Table 1). ‡Includes ICD-9 codes 7463, 7464, and 7472 (see Table 1).

not in female children. Shunt lesions (ASD, VSD, PDA, and Changing Age Distribution of Adults and Children AVCD) were more common in female adults (2.13 per 1000 With CHD versus 1.46 per 1000, PϽ0.0001) and children (9.95 per 1000 The median age of all patients with severe CHD was 11 years versus 7.92 per 1000; PϽ0.0001). Transposition complexes (IQR, 4 to 22 years) in 1985 and 17 years (IQR, 10 to 28 and coarctation were more common in males in adults (0.05 years) in 2000 (PϽ0.0001). Among adults in the year 2000, per 1000 versus 0.03 per 1000, Pϭ0.01; 0.08 per 1000 versus the median age of the CHD population was 40 years (IQR, 27 0.05 per 1000, PϽ0.0001, respectively) and children (0.31 to 60 years) with a median age of 29 years (IQR, 22 to 39 per 1000 versus 0.22 per 1000, Pϭ0.0002; 0.30 per 1000 years) in those with severe disease versus 42 years (IQR, 28 versus 0.19 per 1000, PϽ0.0001, respectively). to 62 years) in those with other CHD. In 1985, CHD of all ages was also more prevalent in Changing Proportions of Adults and Children females than in males (4.83 per 1000 versus 3.94 per 1000; With CHD PϽ0.0001). For severe lesions, the prevalence was higher in When all CHD lesions were considered, there have been more female adults (0.25 per 1000 versus 0.16 per 1000; adults than children with CHD since 1985 (Figure 2A). For PϽ0.0001) but not in female children. Shunts, including severe CHD, the ratio of adults to children increased so that ASD, VSD, PDA, and AVCD, were more common in female in 2000, 49% (95% CI, 47 to 50) of patients with severe Ͻ adults (1.54 per 1000 versus 0.93 per 1000; P 0.0001) and lesions were adults (Figure 2B). children (3.82 per 1000 versus 3.03 per 1000; PϽ0.0001). Although a male predominance was observed for children Changing Prevalence of CHD in Adults with coarctation (0.24 per 1000 versus 0.16 per 1000; and Children Pϭ0.0001) and transposition complexes (0.19 per 1000 The prevalence of all CHD increased from 1985 to 2000 in versus 0.11 per 1000; PϽ0.0001), a male predominance was both adults and children (Figure 3A). However, the preva- not observed in adults for either coarctation or transposition lence of severe CHD in adults increased by 85% (prevalence complexes in l985. ratio for year 2000 versus year 1985, 1.85; 95% CI, 1.72 to

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Marelli et al Population Prevalence of Congenital Heart Disease 167

Figure 2. Numbers and proportion of adults and children with all CHD (A) and severe CHD (B) in 1985, 1990, 1995, and 2000.

2.00) from 0.21 to 0.38 per 1000 adults, whereas the infants using surveillance registries.3–5,12 Reported rates var- prevalence of severe disease in children increased by only ied from 4 to 10 per 1000 live births.3,4 For severe CHD, the 22% (prevalence ratio for year 2000 versus year 1985, 1.22; New England Infant Cardiac Registry reported a prevalence 95% CI, 1.15 to 1.30) from 1.19 to 1.45 per 1000 children of 1.5 per 1000.5 Referral-based participation may result in over the same time period (Figure 3B). The highest increase overascertainment of severe disease and underascertainment in prevalence occurred among adolescents 13 to 17 years of of mild lesions more likely to become manifest after the first age, followed by adults in the age group of 18 to 40 years year of life. Not surprisingly, the population-based preva- (Figure 4). lence of CHD in children in this study is higher than The prevalence of ASD increased at higher rates in previously found. For severe CHD, which is expected to children compared with adults (Figure 5A), whereas that of become manifest in the first year of life regardless of the AVCD and TOF increased at higher rates in adults compared ascertainment method, our results for children are in line with with children (Figure 5B). The prevalence of VSD increased the findings of others.5 at the same rate in both children and adults. The highest The prevalence of CHD in adults has previously been increase in prevalence was among children with ASD and estimated1,2,8 but not measured. For the year 2000, we adults with AVCD, with a Ͼ2-fold increase in prevalence determined a population prevalence of 4.09 per 1000 for all from 1985 to 2000 (prevalence for children with ASD, 3.3; CHD in adults. If we assume that the 15 425 patients 95% CI, 3.1 to 3.4; prevalence for adults with AVCD, 2.5; excluded from our sample represent a margin of error, 95% CI, 2.2 to 2.8) (Figure 5). including them increases the overall prevalence of CHD in adults to 6.02 per 1000. Extrapolating a prevalence of 4.09 Discussion per 1000 to a Canadian population of 24 million adults11 and We report a population-based prevalence of CHD in children a US population of 209 million adults13 corresponds to observed for up to 18 years of life: 11.89 per 1000 for all Ϸ96 000 patients in Canada and Ϸ856 000 patients in the CHD and 1.45 per 1000 for severe disease in the year 2000. United States. Estimates of the number of adults with CHD Previous studies documented the prevalence of CHD among have been generated using prevalence at birth and estimated

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 168 Circulation January 16, 2007

Figure 3. Change in CHD prevalence of children and adults from 1985 to 2000 among patients with all CHD (A) and severe CHD (B).

changes in survival.1,2,8 For the US population, the range of (9%). Extrapolating a prevalence of 0.38 per 1000 corre- estimates varies from 787 8001 to 1.3 million.8 Therefore, our sponds to Ϸ9000 patients in Canada and Ϸ80 000 patients in findings in terms of overall numbers of adults with CHD are the United States who in the year 2000 can be expected to in line with published estimates.1,2,8 have had severe CHD. Based on requirements for expert The prevalence of severe CHD in a general population of care,1 the range of published estimated number of adults with adults was 0.38 per 1000 or 9% of the total in this study. complex CHD in the United States in 2000 varied from Without exclusion of the 15 425 patients, the prevalence of 117 000 to 180 000, depending on the denominator used.1,8 adults with severe CHD in our population increases from 0.38 The proportion of adults with complex CHD1 varied from per 1000 to 0.53 per 1000 in 2000, and the proportion of 14% to 16% when survival estimates of referred patients in subjects with severe CHD remains constant at 0.53 of 6.02 the first year of life were used.1,2,8 Approximating the

Figure 4. Change in prevalence of CHD from 1985 to 2000 among patients of different age groups.

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Marelli et al Population Prevalence of Congenital Heart Disease 169

Figure 5. Change in prevalence of specific lesions in children vs adults from 1985 to 2000 among subjects with septal defects (A) and severe congenital heart lesions (B).

combination of lesions that we have in common with the common lesions found in infants were VSD, D-transposition categories of “complex” and “moderate” disease used at the of the great arteries, and TOF.5 In this study, at all ages, the 32nd Bethesda Conference (TOF, truncus arteriosus, AVCD, most common defects were conotruncal anomalies and transposition of the great arteries, univentricular heart, aortic AVCD for severe CHD and ASD and VSD for other CHD. coarctation, Ebstein’s anomaly, PDA, pulmonary artery and We found that ASD was the most common congenital lesion pulmonary vein anomalies, VSD)1 amounts to a prevalence of in adults. Consistent with what has been published in chil- 1.38 out of 4.09, or 34% of adult patients in our study. dren,2 conotruncal anomalies, including TOF, were the most Published estimates on the combined proportion of adults common severe congenital lesions in adults. with complex and moderate CHD vary: 31%,2 44%,8 and New data on age and sex differences in the CHD popula- 53%,1 with a wide range of denominators. Our population tion are presented. The median age of patients with severe prevalence rates for severe and moderate CHD are therefore CHD increased from late childhood to late adolescence from in the lower range of estimates derived from tertiary care 1985 to 2000. In 2000, there were significantly more females centers.2 The fact that our numbers can be observed in the in the adult and pediatric CHD populations. For subjects with general population is an indication that adult CHD will severe disease, a female predominance was observed in 1985 increasingly come to general medical attention. and 2000 in adults but not in children. Mortality among males We report the frequency of specific CHD in children and has been reported to be 5% greater than in females in a adults in the same population. In previous studies, the 3 most population of high-risk CHD infants.5 Our results in terms of

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 170 Circulation January 16, 2007

sex differences for specific lesions in children are consistent This study has limitations related to the detection of CHD, with prevalence-at-birth studies showing that transposition of depending on contact with the healthcare system. We may the great arteries complexes, aortic coarctation, and lesions of have excluded subjects who did not require health services the left ventricular outflow tract are more prevalent in males, from 1983 to 2000. However, for the population at large, the whereas septal defects are more prevalent in females.5,14 proportion of people using health services in Quebec is high: Further studies are required to determine whether sex differ- 80% to 81% per year from 1998 to 2004.11 Even so, we may ences can be demonstrated in mortality rates in CHD patients. have excluded adults with lesions that can be clinically silent The rising prevalence of severe CHD in a predominantly such as ASD, congenitally corrected transposition of the great female population, in conjunction with higher transmission arteries, or Ebstein’s anomaly. This may result in the under- rates,15,16 may change the incidence of CHD in years ahead. estimation of the prevalence of overall CHD. The potential To the best of our knowledge, this study is the first to for migration into and out of the province of Quebec may document the changing prevalence of adults with CHD in a have influenced the number of patients receiving health care North American, predominantly white population during a within this jurisdiction. During the study period, there was a Ϸ period of major progress in management of CHD.6 The rise in net immigration into the province of an average of 24 000 absolute numbers and prevalence of severe CHD in adults people per year.11 If we use the overall prevalence of 5.78 per exceeded that observed in children. By the year 2000, adults 1000 for CHD of all ages determined in this study, we may and children constituted nearly equal proportions of those thereby have overestimated the number of patients with CHD alive with severe CHD, with an 85% increase in prevalence by only 0.05%. Our data came from 1 of 13 Canadian observed in adults between 1985 and 2000. The largest provinces and territories. However, the population of Quebec 11 increase in prevalence was in adolescents and young adults, accounts for 23% to 25% of Canada’s children and adults, suggesting that the prevalence of severe CHD has yet to peak making our sample large enough to strengthen the interpre- in adults. tation of our results. The rise in observed prevalence of CHD is likely multi- Although each person born in Quebec is assigned a unique factorial. A true rise in the overall number of patients with all medical identifier within days or weeks of birth, the timing of this designation may take up to 1 year. Therefore, infants who CHD may be due to changing birth prevalence rates related to died before obtaining a Medicare number may not be in our increasing maternal age or exposure to medication during the database. In Quebec, in the year 2000, 91 infants died with first trimester of pregnancy.3,17,18 More likely, this finding any congenital malformation in the first year of life, the can be attributed to the development of cardiac ultrasound registration of which is a legal requirement (Vital Statistics since the 1980s, maximizing the possibility of detecting less Act or equivalent legislation).20 If we assume that all had severe CHD. Our observations in prevalence rates of specific CHD and that all were missed in our database, including them lesions are consistent with this supposition. The prevalence of would increase the prevalence of CHD only from 11.89 to ASD increased at higher rates in children, whereas that of 11.94 per 1000 children in the year 2000. Even so, the AVCD and TOF increased at higher rates in adults. Ascer- prevalence rates of specific lesions that we report in children tainment bias is more likely to affect the detection of ASDs; cannot be directly compared with those previously reported in survival is more likely to affect TOF and AVCD in adult- infants in the first year of life.5 hood. It is unlikely that ascertainment bias would account for We relied on administrative databases, which are prone to the increase in prevalence of severe CHD observed in adults misclassification21 and lack of specificity. We excluded 25% in excess of that observed in children because there is no of our initial sample because our data suggested that they had reason to believe that echocardiography has been more miscoded CHD and we did not want to inaccurately inflate accessible to adults than children. The observed rise in the overall prevalence of CHD. We decided to include 11 207 prevalence of VSDs in children and adults is consistent with subjects who had an unspecified CHD with a diagnosis made increasing detection rates at all ages with Doppler interroga- by a CV specialist or during a hospitalization. Cardiologists tion. Studies examining changing birth rates of CHD and the may not be familiar with specific ICD-9 CHD codes, and use of cardiac ultrasound are needed to completely under- there are no designated codes for certain lesions. In these stand the rise in observed prevalence of CHD. subjects, no other information was available to specify a CHD The rise in prevalence and numbers of adults with severe diagnosis. Although some patients with unspecified anoma- CHD in excess of that observed in children supports, although lies may have been misclassified in other rather than severe does not prove, that surgical progress has resulted in de- CHD, we demonstrated significantly lower rates of CHD creased mortality1,8,19 because patients with severe lesions hospitalization and interventions when these subjects were correspond to the subset of patients most likely to benefit compared with those with severe CHD, justifying our classi- from improved surgical outcomes. The greater survival to fication. Of the original 61 386 patients, samples of 17 474 adulthood may result in a shift in mortality beyond 18 years data files were subject to manual review by 1 or 2 authors of age. Indeed, our data show that the majority of adults in (A.J.M., A.S.M.). Subsets of patients were audited in all 2000 were children in 1985 (Figure 2b). This suggests that categories of subjects, including those excluded, with severe higher mortality rates can be expected in adults with severe and other CHD, with unspecified defects, hospitalized (oper- CHD than in children. Further studies are needed to clarify ated and unoperated), and outpatients. After the application of how improved survival in childhood will affect adult inclusion and exclusion criteria, 71% of the subjects in our mortality. CHD database had a unique ICD-9 diagnostic code used

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Marelli et al Population Prevalence of Congenital Heart Disease 171

throughout the study period. Despite limitations, the ICD-9 in prevalence of adults with severe CHD will level off or has been used to investigate CHD-related births and mortality continue to rise. in the United States.3,22 We were limited to the ICD-9 CHD codes that were Sources of Funding available in the administrative databases. We defined as Dr Marelli was supported by the Heart and Stroke Foundation of severe CHD those lesions with a description in the ICD-9 that Canada (No. G-02-MA-1137). Dr Mackie was supported by the Fonds de la Recherche en Santé du Québec (6498 and 8232). Dr reflected the highest probability of being associated with Rahme was supported by The Arthritis Society (TAS02/0007). Dr cyanosis or requiring surgical intervention early in life. We Pilote was supported by the Canadian Institutes of Health Research thought this approach was reasonable because cyanosis has and is a William Dawson Chair of Medicine at McGill University. been a longstanding component of the classification of CHD lesions,23–25 cyanotic lesions have been linked to determi- Disclosures nants of survival in the first year of life,5 and the surgical None. procedures targeting these lesions from 1955 to 197126–30 are References likely to affect observed CHD prevalence from the 1980s on. 1. Warnes CA, Liberthson R, Danielson GK, Dore A, Harris L, Hoffman JI, Our definition differs from other valuable CHD classifica- Somerville J, Williams RG, Webb GD. Task Force 1: the changing profile tions based on the recommended requirements for expert of congenital heart disease in adult life. J Am Coll Cardiol. 2001;37: care.1 Our purpose was to provide empirical data on the 1170–1175. 2. Hoffman JI, Kaplan S. The incidence of congenital heart disease. JAm observed prevalence of CHD in the general population. Our Coll Cardiol. 2002;39:1890–1900. definition of severe CHD could result in misclassification of 3. Botto LD, Correa A, Erickson JD. Racial and temporal variations in the lesions classified as complex or moderate in the Bethesda prevalence of heart defects. Pediatrics. 2001;107:E32. classification1 such as Eisenmenger’s syndrome, mitral or 4. Ferencz C, Rubin JD, McCarter RJ, Brenner JI, Neill CA, Perry LW, Hepner SI, Downing JW. Congenital heart disease: prevalence at live- tricuspid atresia, and supravalvar and subvalvar aortic steno- birth: the Baltimore-Washington Infant Study. Am J Epidemiol. 1985; ses.1 We also may have misclassified patients with defects 121:31–36. having a broad spectrum of severity depending on the age of 5. Fyler DC. Report of the New England Regional Infant Cardiac Program. Pediatrics. 1980;65:375–461. presentation such as Ebstein’s anomaly, congenital aortic and 6. Perloff JK. Congenital heart disease after childhood: an expanding pop- mitral stenoses, and coarctation of the aorta. We minimized ulation: 22nd Bethesda Conference. J Am Coll Cardiol. 1991;18: misclassification by examining all available data for a given 311–342. subject, including inpatient and outpatient codes and surgical 7. Webb GD, Williams RG. Care of the adult with congenital heart disease: introduction. J Am Coll Cardiol. 2001;37:1166. codes. We established a hierarchy of diagnoses and surgical 8. Hoffman JI, Kaplan S, Liberthson RR. Prevalence of congenital heart interventions imitating what was done for the Report of the disease. Am Heart J. 2004;147:425–439. New England Regional Infant Cardiac Program,5 dividing the 9. Pilote L, Joseph L, Belisle P, Penrod J. Universal health insurance coverage does not eliminate inequities in access to cardiac procedures ICD-9 diagnoses into blocks of decreasing specificity. When after acute myocardial infarction. Am Heart J. 2003;146:1030–1037. specific CHD surgical interventions were performed, we 10. Hennekens CH, Buring JE. Measures of Disease Frequency. In: Mayrent inferred a more specific CHD diagnosis. Even so the propor- SL, ed. Epidemiology in Medicine. 1st ed. Philadelphia, Pa: Lippincott tion of patients with severe CHD may have been Williams & Wilkins; 1987:55–98. 11. Annual demographic statistics 2004. In: Division D, ed. Annual Demo- underestimated. graphic Statistics. Ottawa, Quebec: Statistics Canada; 2005. Catalogue Nos. 91–213-XIB and 91–213-XPB. Conclusions 12. Edmonds LD, Layde PM, James LM, Flynt JW, Erickson JD, Oakley GP Jr. Congenital malformations surveillance: two American systems. Int J This population-based study yielded new information on the Epidemiol. 1981;10:247–252. epidemiology of CHD between 1985 and 2000. The median 13. Division P, ed. Annual Estimates of the Population by Selected Age age of those with severe disease increased. There was a Groups and Sex for the United States. Washington, DC: US Census female predominance among adults with severe CHD. Over- Bureau; 2004. Volume NC-EST2004–02. 14. Samanek M. Boy:girl ratio in children born with different forms of all, in the year 2000, 1 in 84 children had CHD. In the same cardiac malformation: a population-based study. Pediatr Cardiol. 1994; year, a population prevalence of 4.09 per 1000 or 1 per 245 15:53–57. adults with CHD was determined, of whom 9% had severe 15. Nora JJ, Nora AH. Maternal transmission of congenital heart diseases: new recurrence risk figures and the questions of cytoplasmic inheritance lesions. Prevalence rates of severe CHD increased between and vulnerability to teratogens. Am J Cardiol. 1987;59:459–463. 1985 and 2000, but the increase was significantly higher in 16. Burn J, Brennan P, Little J, Holloway S, Coffey R, Somerville J, Dennis adults, so that in the year 2000, the numbers of adults and NR, Allan L, Arnold R, Deanfield JE, Godman M, Houston A, Keeton B, children alive with severe CHD were nearly equal. The Oakley C, Scott O, Silove E, Wilkinson J, Pembrey M, Hunter AS. Recurrence risks in offspring of adults with major heart defects: results results of this study provide empirical data on which recom- from first cohort of British collaborative study. Lancet. 1998;351: mendations for resource allocation in CHD can be anchored. 311–316. These data underscore the published estimates and recom- 17. Botto LD, Correa A. Decreasing the burden of congenital heart anom- alies: an epidemiologic evaluation of risk factors and survival. Prog mendations stating that more specialized care facilities for Pediatr Cardiol. 2003;18:111–121. adults with CHD are necessary to meet the needs of this 18. Cooper WO, Hernandez-Diaz S, Arbogast PG, Dudley JA, Dyer S, growing patient population.31 Although studies using admin- Gideon PS, Hall K, Ray WA. Major congenital malformations after istrative databases have limitations, we have used a unique first-trimester exposure to ACE inhibitors. N Engl J Med. 2006;354: 2443–2451. and contemporary application of prevalence determination. 19. Somerville J. Grown-up congenital heart disease-medical demands look Further studies are needed to determine whether the increase back, look forward 2000. Thorac Cardiovasc Surg. 2001;49:21–26.

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 172 Circulation January 16, 2007

20. Division HS, ed. Causes of Death: Congenital Malformations, Defor- 26. Lillehei CW, Cohen M, Warden HE. Direct vision intracardiac surgical mations and Chromosomal Abnormalities (Q00 to Q99), by Age Group correction of the tetralogy of Fallot, pentalogy of Fallot and pulmonary and Sex, Canada. Ottawa, Quebec: Statistics Canada; 2003:6–12. atresia defects: report of first ten cases. Ann Surg. 1955;142:418–423. 21. Jollis JG, Ancukiewicz M, DeLong ER, Pryor DB, Muhlbaier LH, Mark 27. Mustard WT. Successful two-stage correction of transposition of the great DB. Discordance of databases designed for claims payment versus vessels. Surgery. 1964;55:469–473. clinical information systems: implications for outcomes research. Ann 28. Rastelli GC, Wallace RB, Ongley PA. Complete repair of transposition of Intern Med. 1993;119:844–850. the great arteries with pulmonary stenosis: a review and report of a case 22. Boneva RS, Botto LD, Moore CA, Yang Q, Correa A, Erickson JD. corrected by using a new surgical technique. Circulation. 1969;39:83–95. Mortality associated with congenital heart defects in the United States: 29. Senning A. Surgical correction of transposition of the great vessels. trends and racial disparities, 1979–1997. Circulation. 2001;103: Surgery. 1959;45:966–975. 2376–2381. 30. Fontan F, Baudet E. Surgical repair of tricuspid atresia. Thorax. 1971; 23. Abbott ME. Congenital heart disease. Nelson’s Loose Leaf Medicine: 26:240–248. Volume 5. New York, NY: Nelson and Sons; 1932. 31. Williams RG, Pearson GD, Barst RJ, Child JS, del Nido P, Gersony WM, 24. Bing RJ. Congenital heart disease: an introduction and classification. Kuehl KS, Landzberg MJ, Myerson M, Neish SR, Sahn DJ, Verstappen Am J Med. 1952;12:77–96. A, Warnes CA, Webb CL. Report of the National Heart, Lung, and Blood 25. Lev M, Paul MH, Miller RA. A classification of congenital heart disease Institute Working Group on Research in Adult Congenital Heart Disease. based on the pathologic complex. Am J Cardiol. 1962;10:733–737. J Am Coll Cardiol. 2006;47:701–707.

CLINICAL PERSPECTIVE This is the first study to provide empirical data on the changing epidemiology of congenital heart disease (CHD) in the general population. Advances in pediatric cardiovascular care have resulted in more adults with CHD being followed up in tertiary care centers. Previous estimates of CHD prevalence in adults have been generated using birth prevalence and estimated survival rates. In this study, administrative databases were used to obtain data on children and adults in the province of Quebec, a large geographic region of Canada with universal access to healthcare services. A total of 45 960 patients with International Classification of Disease, ninth revision, diagnoses of severe and other forms of CHD from l985 to 2000 were included. The prevalence of CHD was 4.09 per 1000 adults in 2000; 9% had severe lesions with a significant female predominance. This corresponds to 96 000 patients in Canada and 856 000 patients in the United States with CHD in 2000. The median age of all patients with severe CHD increased from 11 years in 1985 to 17 years in 2000, and the prevalence of severe CHD increased more in adults than in children during the study period. In 2000, there were nearly equal numbers of adults and children with severe CHD. This study provides important data documenting the growing number of adults with CHD in the general population. These findings underscore the need for increasing healthcare resource allocation to this emerging population.

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Coronary Heart Disease

Association of Cystatin C With Mortality, Cardiovascular Events, and Incident Heart Failure Among Persons With Coronary Heart Disease Data From the Heart and Soul Study

Joachim H. Ix, MD; Michael G. Shlipak, MD, MPH; Glenn M. Chertow, MD, MPH; Mary A. Whooley, MD

Background—Serum creatinine and related estimating equations predict cardiovascular events and mortality among persons with coronary heart disease (CHD). Cystatin C is a novel and sensitive endogenous marker of kidney function. Whether cystatin C concentrations are associated with adverse events among ambulatory persons with CHD is unknown. Methods and Results—Nine hundred ninety ambulatory persons with CHD were categorized into quartiles of serum cystatin C at inception, with Յ0.91 mg/L constituting the lowest quartile (I) and Ն1.30 mg/L constituting the highest (IV). Cox proportional hazards models evaluated time to all-cause mortality, cardiovascular events (composite of CHD death, myocardial infarction, and stroke), and incident heart failure. After a median follow-up of 37 months, 132 participants (13%) died, 101 (10%) had cardiovascular events, and 57 (7%) had incident heart failure. Compared with participants in the lowest cystatin C quartile, those in the highest quartile were at increased risk of all-cause mortality (hazard ratio, 3.6; 95% CI, 1.8 to 7.0), cardiovascular events (hazard ratio, 2.0; 95% CI, 1.0 to 3.8), and incident heart failure (hazard ratio, 2.6; 95% CI, 1.0 to 6.9) in analyses adjusted for traditional cardiovascular risk factors. Cystatin C in the highest quartile predicted similar risk for these outcomes among participants with lower (Յ60 mL/min per 1.73 m2) or higher estimated glomerular filtration rate and among participants with or without microalbuminuria. Conclusions—High cystatin C concentrations predict substantial increased risks of all-cause mortality, cardiovascular events, and incident heart failure among ambulatory persons with CHD. This risk is not completely captured by measures of kidney function routinely used in clinical practice. (Circulation. 2007;115:173-179.) Key Words: coronary disease Ⅲ cystatin C Ⅲ heart failure Ⅲ kidney Ⅲ mortality Ⅲ myocardial infarction Ⅲ stroke

ild to moderate chronic kidney disease is common in Clinical Perspective p 179 Mthe US population1 and is an independent risk factor for cardiovascular disease events and all-cause mortality.2 nants, it has been shown to overestimate kidney function Among persons with established coronary heart disease among persons with glomerular filtration rates (GFR) Ͼ60 (CHD), concomitant chronic kidney disease predicts cardio- mL/min per 1.73 m2.7 vascular events as strongly as other established risk factors Cystatin C is a novel endogenous marker of kidney such as diabetes mellitus and hypertension.3 Serum creatinine function that may be more sensitive for detecting mild to concentrations or creatinine-based estimating equations have moderate decrements in GFR.8–10 Most,8–13 but not all,14 served as the primary tool for evaluation of kidney function in studies suggest that it is not affected by age, sex, or muscle clinical practice and in epidemiological research studies. mass. Several recent publications have demonstrated that However, creatinine concentrations are affected by several cystatin C is superior to serum creatinine or creatinine-based nonkidney factors including age, body weight, nutritional estimating equations for prediction of all-cause mortality,15–17 status, ethnicity, and sex4,5 and are insensitive to modest cardiovascular events,16 and incident congestive heart failure decreases in kidney function; for instance, kidney function (CHF)18 in elderly community-based cohorts that were pre- may decline Ն50% before serum creatinine exceeds the dominantly free of cardiovascular disease at inception. normal range.4,6 Although the Modification of Diet and Renal Whether elevated cystatin C concentrations are associated Disease study formula is less affected by nonkidney determi- with adverse clinical outcomes among persons with prevalent

Received June 6, 2006; accepted November 1, 2006. From the Division of Nephrology (J.H.I., G.M.C.), Department of Medicine (J.H.I., M.G.S., G.M.C., M.A.W.), and Department of Epidemiology and Biostatistics (M.G.S., G.M.C., M.A.W.), University of California at San Francisco; and Section of General Internal Medicine, Veterans Affairs Medical Center, San Francisco, Calif (M.G.S., M.A.W.). Correspondence to Joachim H. Ix, MD, Division of Nephrology, Department of Medicine, Box 0532, HSE 672, University of California at San Francisco, San Francisco, CA 94143-0532. E-mail [email protected] © 2007 American Heart Association, Inc. Circulation is available at http://www.circulationaha.org DOI: 10.1161/CIRCULATIONAHA.106.644286

173

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 174 Circulation January 16, 2007

CHD has not been studied extensively. Because persons with nine ratios (mg albumin/g creatinine) were calculated, and microal- CHD represent a group at high risk for cardiovascular events buminuria was defined as Ն30 mg/g.23 and death, risk stratification may be particularly useful in this Other Measurements population. Furthermore, previous studies evaluating the Baseline demographics and medical history were determined by predictive value of cystatin C for cardiovascular morbidity or questionnaire. Participants underwent a complete physical examina- mortality have not determined whether the observed associ- tion that included blood pressure determination by trained study ations and outcomes are robust among persons with and personnel using calibrated sphygmomanometers. Participants were instructed to bring their medication bottles to the study appointment, without microalbuminuria. and study personnel recorded all current medications. High- We evaluated whether serum cystatin C concentrations sensitivity C-reactive protein was measured with the use of the were associated with all-cause mortality, cardiovascular Roche Integra assay and the Beckman Extended Range assay as events, and incident CHF among ambulatory persons with previously described.24,25 Fasting serum samples were used to measure total cholesterol, high-density lipoprotein cholesterol, and CHD enrolled in the Heart and Soul Study. In addition, we triglyceride concentrations. Low-density lipoprotein cholesterol con- evaluate whether the risk associated with elevated cystatin C centrations were calculated by the Friedewald equation.26 differs among persons with or without low estimated GFR or microalbuminuria. Outcomes Annual telephone interviews were conducted with participants or Methods their proxies to ask about interval death or hospitalization. For any reported event, medical records, electrocardiograms, death certifi- Participants cates, autopsy, and coroner’s reports were obtained and reviewed by The Heart and Soul Study is a prospective cohort study designed to 2 independent and blinded adjudicators. If the adjudicators agreed on investigate the influence of psychosocial factors on CHD progres- the outcome classification, their classification was binding. If they sion. Methods have been described previously.19,20 Briefly, partici- disagreed, they conferred, reconsidered their classification, and pants were recruited from several outpatient clinics in the San requested consultation from a third blinded adjudicator as necessary Francisco Bay Area if they met 1 of the following inclusion criteria: to obtain consensus. (1) history of myocardial infarction; (2) angiographic evidence of Cardiovascular events were defined as CHD death, nonfatal Ͼ50% stenosis in 1 or more coronary vessels; (3) evidence of myocardial infarction, or stroke. CHD death was defined as (1) death exercise-induced ischemia by treadmill or nuclear testing; or (4) during the same hospitalization in which an acute myocardial history of coronary revascularization. Participants were excluded if infarction was documented or (2) death not explained by other causes they had had a myocardial infarction within the prior 6 months, were and that occurred within 1 hour of the onset of terminal symptoms. not able to walk 1 block, or were likely to move out of the area Nonfatal myocardial infarction was defined by the American Heart within 3 years. There was no exclusion on the basis of kidney Association diagnostic cretiria.27 Stroke was defined as a new disease. Because the study recruited heavily from a Veterans Affairs neurological deficit not known to be secondary to brain trauma, facility, the resulting cohort was 82% male. The study protocol was tumor, infection, or other cause.28 approved by the appropriate institutional review boards, and all CHF was defined as hospitalization for a clinical syndrome participants provided written informed consent. involving at least 2 of the following: paroxysmal nocturnal dyspnea, Between September 2000 and December 2002, a total of 1024 orthopnea, elevated jugular venous pressure, pulmonary rales, third individuals enrolled and underwent a day-long baseline study ap- heart sound, and cardiomegaly or pulmonary edema on chest pointment that included a medical history interview, a physical radiography.29 These clinical signs and symptoms must have repre- examination, and a comprehensive health status questionnaire. Fast- sented a clear change from the normal clinical status of the ing (12-hour) venous samples were drawn, and serum was frozen at participant and must have been accompanied by either decreased Ϫ70°C. Subjects for whom frozen serum was not available (nϭ34) cardiac output as determined by peripheral hypoperfusion (in the were excluded, resulting in a sample size of 990 participants for this absence of other causes such as sepsis or dehydration) or peripheral analysis. or pulmonary edema requiring intravenous diuretics, inotropic agents, or vasodilator therapy. Supportive documentation of reduced Measurements cardiac output, elevated pulmonary capillary wedge pressure, de- creased oxygen saturation, and end-organ hypoperfusion was as- Kidney Function sessed, when available. Serum cystatin C was measured from frozen samples collected at the All-cause mortality was determined by review of death baseline study visit with the use of a BNII nephelometer (Dade certificates. Behring, Inc, Deerfield, Ill) with a particle-enhanced immunoneph- elometric assay (N Latex Cystatin C, Dade Behring, Inc).21 Mono- Statistical Analysis clonal antibodies to cystatin C were coated on polystyrene particles Because age-, sex-, and race-specific normal ranges for serum that agglutinate to increase the intensity of scattered light in cystatin C have not been established, we categorized cystatin C into proportion to the concentration of cystatin C. The assay range is quartile groups. Differences in baseline characteristics were com- 0.195 to 7.330 mg/L; the reference range for young healthy persons pared with the use of ANOVA or the Kruskal-Wallis test for ranges from 0.53 to 0.95 mg/L. The intra-assay coefficient of continuous variables and the ␹2 test or Fisher exact test for categor- variation ranges from 2.0% to 2.8%, and the interassay coefficient of ical variables, as appropriate. Cox proportional hazards models variation ranges from 2.3% to 3.1%. evaluated whether cystatin C quartile groups were associated with Serum creatinine was measured by the rate Jaffe method. The time to each of the 3 outcome measures. Covariates for adjustment intraindividual coefficient of variation was Ϸ2%. Estimated GFR included important demographics (age, sex, race), traditional cardio- was calculated by the abbreviated (4-variable) Modification of Diet vascular risk factors (age, sex, diabetes mellitus, hypertension, and Renal Disease Study formula, as follows: estimated history of prior myocardial infarction, tobacco use, body mass index, GFRϭ186ϫ(serum creatinineϪ1.154)ϫ(ageϪ0.203)ϫ(0.742 if fe- high-density lipoprotein, and C-reactive protein), and variables male)ϫ(1.21 if black).5 We defined low estimated GFR as Յ60 previously associated with elevated cystatin C concentrations (age, mL/min per 1.73 m2, consistent with stage III or greater severity of sex, body mass index, tobacco use, and C-reactive protein).14 For chronic kidney disease.22 incident CHF, we excluded participants with a self-reported history Urinary albumin and creatinine were measured by nephelometry of heart failure at inception (nϭ174), resulting in a sample size of and the rate Jaffe method, respectively. Urinary albumin-to-creati- 816 participants for this outcome. Multiplicative interaction terms

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Ix et al Cystatin C, Mortality, Events, and Heart Failure 175

TABLE 1. Baseline Characteristics of Participants by Cystatin C Quartiles

Cystatin C Quartiles

I II III IV P Cut points, mg/L Յ0.91 0.92 to 1.05 1.06 to 1.29 Ն1.30 ⅐⅐⅐ N 239 248 262 241 ⅐⅐⅐ Demographics Age, y, meanϮSD 61Ϯ10 65Ϯ969Ϯ10 71Ϯ11 Ͻ0.001 Male sex 171 (72) 203 (82) 226 (86) 206 (85) Ͻ0.001 Race ⅐⅐⅐ ⅐⅐⅐ ⅐⅐⅐ ⅐⅐⅐ 0.01 White 121 (51) 149 (60) 174 (66) 153 (63) ⅐⅐⅐ Black 53 (22) 43 (17) 30 (11) 34 (14) ⅐⅐⅐ Other 65 (27) 54 (22) 58 (22) 54 (22) ⅐⅐⅐ Medical history Diabetes mellitus 56 (24) 62 (25) 54 (21) 88 (37) Ͻ0.001 Hypertension 154 (65) 165 (67) 186 (72) 190 (79) 0.004 Myocardial infarction 115 (49) 132 (54) 133 (51) 149 (62) 0.02 Coronary revascularization 111 (47) 147 (61) 158 (61) 156 (65) Ͻ0.001 CHF 28 (12) 37 (15) 43 (17) 66 (27) Ͻ0.001 Stroke 27 (11) 34 (14) 33 (13) 44 (18) 0.15 Tobacco use 49 (21) 45 (18) 58 (22) 44 (18) 0.62 Measurements Body mass index, kg/m2, meanϮSD 28Ϯ528Ϯ528Ϯ528Ϯ5 0.83 LDL cholesterol, mg/dL, meanϮSD 107Ϯ34 103Ϯ32 106Ϯ32 101Ϯ37 0.21 HDL cholesterol, mg/dL, meanϮSD 49Ϯ15 46Ϯ14 45Ϯ13 43Ϯ13 Ͻ0.001 Triglycerides, mg/dL* 103 (69, 154) 100 (68, 150) 122 (79, 169) 114 (78, 178) 0.002 C-reactive protein, mg/dL* 1.6 (0.6, 3.8) 1.7 (0.7, 3.6) 2.4 (1.1, 5.3) 3.2 (1.5, 7.4) Ͻ0.001 Values are number (%) unless indicated otherwise. LDL indicates low-density lipoprotein; HDL, high-density lipoprotein. *Median (quartile 1, quartile 3).

were created to evaluate the effect modification of the association of lower high-density lipoprotein cholesterol, whereas low- cystatin C with each outcome by the presence or absence of density lipoprotein cholesterol concentrations were not sig- microalbuminuria and low estimated GFR. Stratified analyses were nificantly different across quartiles. Persons with higher conducted to determine whether the association of cystatin with cystatin C also had higher C-reactive protein concentrations. cardiovascular outcomes differed by microalbuminuria or estimated GFR categories. Among the 990 study participants, there were 132 deaths Proportional hazards assumptions were assessed by visually in- (13%) after 40 530 patient-months of follow-up (median, 37 specting log-minus-log plots and plots of Schoenfeld residuals versus months). Six participants (0.6%) were lost to follow-up. survival time for the association of cystatin C with each outcome Among the high cystatin C quartile (IV), there was a 9.4% variable. We found no evidence that the proportionality assumption annual mortality rate compared with 1.5% in the lowest Ͻ was violated. Two-tailed probability values 0.05 were considered quartile (I) (Figure 1). After multivariable adjustment, partic- statistically significant. Analyses were performed with the use of Ͼ Stata statistical software, version 9 (College Station, Tex). ipants in quartile IV had a 3-fold mortality hazard com- The authors had full access to the data and take responsibility for pared with quartile I. The mortality hazards for quartiles II the integrity of the data. All authors have read and agree to the and III were similar to those for quartile I (Table 2). manuscript as written.

Results The mean age of the 990 study participants was 67 years, 82% were male, 61% were white, 51% had a history of myocardial infarction, 24% had diabetes, and 69% had hypertension. The mean cystatin C concentration was 1.20Ϯ0.56 mg/L, and mean estimated GFR was 77Ϯ23 mL/min per 1.73 m2. Associations of baseline characteristics by cystatin C quartiles are shown in Table 1. Participants with higher cystatin C were older and more likely to be male, to be white, and to have comorbid medical conditions. Persons with higher cystatin C also had higher serum triglycerides and Figure 1. Probability value for trend Ͻ0.001 for each outcome.

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 176 Circulation January 16, 2007

TABLE 2. Association of Cystatin C Quartiles With All-Cause Mortality, Cardiovascular Events, and Incident Heart Failure

Cystatin C Quartiles

I II III IV

No. of Events HR 95% CI No. of Events HR 95% CI No. of Events HR 95% CI No. of Events HR 95% CI P for Trend All-cause mortality 12 ⅐⅐⅐ ⅐⅐⅐ 18 ⅐⅐⅐ ⅐⅐⅐ 29 ⅐⅐⅐ ⅐⅐⅐ 73 ⅐⅐⅐ ⅐⅐⅐ Unadjusted ⅐⅐⅐ 1.0 Reference ⅐⅐⅐ 1.3 0.6 to 2.8 ⅐⅐⅐ 1.9 1.0 to 3.7 ⅐⅐⅐ 5.7 3.1 to 10.5 Ͻ0.001 Adjusted* ⅐⅐⅐ 1.0 Reference ⅐⅐⅐ 1.2 0.6 to 2.5 ⅐⅐⅐ 1.4 0.7 to 2.9 ⅐⅐⅐ 3.6 1.8 to 7.0 Ͻ0.001 Cardiovascular events 14 ⅐⅐⅐ ⅐⅐⅐ 17 ⅐⅐⅐ 23 ⅐⅐⅐ ⅐⅐⅐ 47 ⅐⅐⅐ ⅐⅐⅐ Unadjusted ⅐⅐⅐ 1.0 Reference ⅐⅐⅐ 1.2 0.6 to 2.4 ⅐⅐⅐ 1.5 0.8 to 2.9 ⅐⅐⅐ 3.8 2.1 to 6.9 Ͻ0.001 Adjusted* ⅐⅐⅐ 1.0 Reference ⅐⅐⅐ 1.0 0.5 to 2.0 ⅐⅐⅐ 1.1 0.5 to 2.2 ⅐⅐⅐ 2.0 1.0 to 3.8 0.04 Incident heart failure 6 ⅐⅐⅐ ⅐⅐⅐ 7 ⅐⅐⅐ 9 ⅐⅐⅐ ⅐⅐⅐ 35 ⅐⅐⅐ ⅐⅐⅐ Unadjusted ⅐⅐⅐ 1.0 Reference ⅐⅐⅐ 1.1 0.4 to 3.1 ⅐⅐⅐ 1.4 0.5 to 4.0 ⅐⅐⅐ 6.1 2.5 to 14.5 0.001 Adjusted* ⅐⅐⅐ 1.0 Reference ⅐⅐⅐ 0.7 0.2 to 2.2 ⅐⅐⅐ 0.8 0.3 to 2.4 ⅐⅐⅐ 2.6 1.0 to 6.9 0.05 HR indicates hazard ratio. *Adjusted for age, sex, black race, diabetes, hypertension, prior myocardial infarction, tobacco use, body mass index, high-density lipoprotein, and C-reactive protein.

Among the study cohort, 101 cardiovascular events oc- Risk of each adverse clinical outcome associated with a curred. Of these, 13 were CHD deaths, 70 were nonfatal cystatin C concentration Ն1.3 mg/L was similar among myocardial infarctions, and 18 were strokes. The high cysta- persons with or without a low estimated GFR and with or tin C quartile experienced a 7% annual cardiovascular event without microalbuminuria (Table 3). Probability values for rate compared with 2% among the low cystatin C quartile interaction were all Ͼ0.35 with the exception of the associ- (Figure 1). After multivariable adjustment, participants in ation of high cystatin C with cardiovascular events, in which quartile IV had a 2-fold higher cardiovascular event hazard substantially greater risk was observed in the subgroup with compared with quartile I, but no significant differences were low estimated GFR (P for interactionϭ0.04). observed among the lower 3 quartiles (Table 2). When CHD death, nonfatal myocardial infarction, and Discussion stroke were evaluated as individual outcome measures, the In the present study, we demonstrate that higher serum relative hazards for participants in quartile IV compared with cystatin C concentrations predict all-cause mortality, cardio- quartile I were 5.8 (95% CI, 0.6 to 51.3) for CHD death, 2.0 vascular events, and incident heart failure independent of (95% CI, 0.9 to 4.3) for nonfatal myocardial infarction, and traditional cardiovascular risk factors among ambulatory 1.1 (95% CI, 0.4 to 3.6) for stroke. persons with CHD. Moreover, we found that a substantial Among the 816 participants without a prior history of CHF, proportion of persons with cystatin C concentrations in the there were 57 hospitalizations for incident CHF during highest quartile of our study sample did not have an estimated 32 078 patient-months of follow-up (median, 36 months). GFR Յ60 mL/min per 1.73 m2 or microalbuminuria. High Participants in quartile IV experienced a 6% annual incidence cystatin C level had similar predictive value for adverse rate of CHF compared with 1% in quartile I (Figure 1). After clinical outcomes among persons with or without low esti- multivariable adjustment, participants in the highest cystatin mated GFR or microalbuminuria. C quartile had a Ͼ2-fold higher incident CHF hazard com- Observational studies have previously demonstrated that pared with the lowest quartile. There was no statistically mild to moderate kidney disease independently predicts significant difference in CHF hazards among the lowest 3 all-cause mortality and cardiovascular events among persons quartiles (Table 2). with prevalent CHD,3,30,31 and the increased risk may be Serum cystatin C concentrations were highly correlated observed with moderate decrements in GFR.2,16 Detection of with estimated GFR and, to a lesser degree, with albuminuria mild kidney disease in routine clinical practice is problematic, (Spearman rank correlation rϭ0.74 and 0.23, respectively). however. Serum creatinine concentrations4–6 and the Modi- There was substantial overlap, however, such that among fication of Diet and Renal Disease study formula27,32 are both participants with serum cystatin C concentrations in the insensitive to mild decrements in kidney function. Whereas highest quartile (Ն1.30 mg/L), 25% had estimated GFR Ͼ60 measured creatinine clearance by timed urine collection may mL/min per 1.73 m2, and 67% did not have microalbuminuria be more sensitive, this test is fraught with inaccuracy and is (Figure 2). We evaluated whether the event risks associated no longer recommended to routinely assess kidney function with elevated cystatin C differed among participants with low in the ambulatory setting.22 Because cystatin C is sensitive to versus higher estimated GFR and present versus absent mild decrements of GFR and may not be affected by age, sex, microalbuminuria. Because the highest quartile of cystatin C and muscle mass, it may have clinical utility in accurately was associated with substantially higher risk for all outcomes, detecting mild kidney disease among ambulatory per- we dichotomized participants at the highest cystatin C quar- sons.8,9,32 We hypothesized that it might also allow more tile (serum cystatin C Ն1.3 versus Ͻ1.3 mg/L) to enhance accurate cardiovascular risk prediction among persons with power for subgroup analyses. CHD.

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Ix et al Cystatin C, Mortality, Events, and Heart Failure 177

Figure 2. Distribution of estimated GFR (A) and albumin-to-creatinine ratio (B) among the 241 study participants with cystatin C concentrations within the highest quartile of the study sample (Ն1.3 mg/L). Light bars represent partici- pants with estimated GFR Ͼ60 mL/min per 1.73 m2 (A) or without microalbumin- uria (B) despite cystatin C concentrations among the highest quartile of the study sample.

In the present study, we demonstrate that among persons cystatin C concentrations predict not only cardiovascular with cystatin C concentrations in the highest quartile of our events but also mortality and incident heart failure among cohort (Ͼ1.3 mg/L), one fourth had estimated GFR Ͼ60 ambulatory persons with CHD. mL/min per 1.73 m2, and two thirds had no microalbuminuria. Among the strengths of the present study are measure- These participants were demonstrated to have increased risks ments of multiple potential confounding variables and avail- for adverse clinical outcomes, and the relative hazards were ability of albuminuria measurement among an ambulatory similarly increased among persons with or without low cohort with CHD. Several limitations should be considered estimated GFR or microalbuminuria. These results demon- when our results are interpreted, however. Although 2 strate that the risk of adverse events attributable to kidney blinded adjudicators reviewed all medical records, the diag- disease is not completely captured by estimates of kidney nosis of cardiovascular events and incident CHF may be function routinely used in clinical practice. More accurate subject to misclassification. Any misclassification should have biased our results toward the null and should not have characterization of kidney function in persons with CHD been influenced by serum cystatin C concentrations. We could allow targeted therapies that might improve prognosis cannot exclude the possibility of residual confounding. How- and limit use of tests or procedures that have high risk for ever, any unmeasured confounder would need to be highly acute renal failure and associated consequences. prevalent in our cohort and strongly associated with cystatin Although previous studies demonstrate that cystatin C C and each outcome measure to explain the relatively large 15–17 16 predicts all-cause mortality, cardiovascular events, and hazard ratios we observed in the present study. The number of 18 incident CHF among elderly ambulatory cohorts predomi- outcome events within each quartile was relatively small; nantly without CHD, the predictive value of cystatin C among therefore, these findings should be confirmed in future outpatients with prevalent CHD has not been studied exten- studies. For the cardiovascular events and incident heart sively. Jernberg and colleagues33 determined that elevated failure outcomes, the number of covariates included in cystatin C was associated with mortality among persons multivariable models was Ͼ1 for each 10 outcome events. hospitalized with acute coronary syndrome, and Koenig and Therefore, it remains possible that the models were overfit- colleagues34 demonstrated that elevated cystatin C predicted ted. Finally, the present study participants were elderly and second cardiovascular events among a cohort participating in predominantly male. Therefore, our results may not be an in-hospital rehabilitation program shortly after acute myo- generalizable to younger or female patients. cardial infarction or coronary revascularization. Our present In conclusion, we found that higher serum cystatin C study adds to the existing literature by demonstrating that concentrations were associated with all-cause mortality, car-

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 178 Circulation January 16, 2007

TABLE 3. Adjusted Associations of High Cystatin* With References Mortality, Cardiovascular Events, and Heart Failure by Strata of 1. Clase CM, Garg AX, Kiberd BA. Prevalence of low glomerular filtration Estimated GFR and Microalbuminuria rate in nondiabetic Americans: Third National Health and Nutrition Examination Survey (NHANES III). J Am Soc Nephrol. 2002;13: High Cystatin, Hazard Ratio P for 1338–1349. Outcome n (%) (95% CI) Interaction 2. Go AS, Chertow GM, Fan D, McCulloch CE, Hsu CY. Chronic kidney All-cause mortality† disease and the risks of death, cardiovascular events, and hospitalization. N Engl J Med. 2004;351:1296–1305. Low estimated GFR‡ 183 (69) 3.6 (1.5 to 8.7) 0.36 3. Shlipak MG, Simon JA, Grady D, Lin F, Wenger NK, Furberg CD. Renal Higher estimated GFR 58 (8) 2.5 (1.4 to 4.4) ⅐⅐⅐ insufficiency and cardiovascular events in postmenopausal women with coronary heart disease. J Am Coll Cardiol. 2001;38:705–711. Microalbuminuria§ 79 (54) 3.0 (1.1 to 8.0) 0.74 4. Kassirer JP. Clinical evaluation of kidney function–glomerular function. No microalbuminuria 126 (19) 2.4 (1.5 to 3.9) ⅐⅐⅐ N Engl J Med. 1971;285:385–389. 5. Levey AS. Measurement of renal function in chronic renal disease. Cardiovascular events† Kidney Int. 1990;38:167–184. Low estimated GFR‡ 183 (69) 4.8 (1.4 to 15.9) 0.04 6. Seliger SL, Davis C, Stehman-Breen C. Gender and the progression of Higher estimated GFR 58 (8) 1.1 (0.5 to 2.4) ⅐⅐⅐ renal disease. Curr Opin Nephrol Hypertens. 2001;10:219–225. 7. Rule AD, Larson TS, Bergstralh EJ, Slezak JM, Jacobsen SJ, Cosio FG. Microalbuminuria§ 79 (54) 1.1 (0.5 to 2.6) 0.81 Using serum creatinine to estimate glomerular filtration rate: accuracy in No microalbuminuria 126 (19) 1.5 (0.8 to 2.8) ⅐⅐⅐ good health and in chronic kidney disease. Ann Intern Med. 2004;141: 929–937. Incident heart failure† 8. Bokenkamp A, Domanetzki M, Zinck R, Schumann G, Byrd D, Brodehl Low estimated GFR‡ 150 (74) 3.2 (0.9 to 11.3) 0.95 J. Cystatin C: a new marker of glomerular filtration rate in children independent of age and height. Pediatrics. 1998;101:875–881. Higher estimated GFR 51 (8) 1.7 (0.6 to 4.6) ⅐⅐⅐ 9. Fliser D, Ritz E. Serum cystatin C concentration as a marker of renal Microalbuminuria§ 57 (54) 3.2 (0.9 to 11.0) 0.36 dysfunction in the elderly. Am J Kidney Dis. 2001;37:79–83. No microalbuminuria 114 (21) 1.0 (0.4 to 2.6) ⅐⅐⅐ 10. Newman DJ, Thakkar H, Edwards RG, Wilkie M, White T, Grubb AO, Price CP. Serum cystatin C measured by automated immunoassay: a more *Compares highest quartile (Ն1.3 mg/L) with lower 3 quartiles. sensitive marker of changes in GFR than serum creatinine. Kidney Int. †Adjusted for age, sex, black race, diabetes, hypertension, prior myocardial 1995;47:312–318. infarction, tobacco use, body mass index, high-density lipoprotein, and 11. Coll E, Botey A, Alvarez L, Poch E, Quinto L, Saurina A, Vera M, Piera C-reactive protein. C, Darnell A. Serum cystatin C as a new marker for noninvasive esti- ‡Estimated GFR Յ60 mL/min per 1.73 m2. mation of glomerular filtration rate and as a marker for early renal §Albumin-to-creatinine ratio Ն30 mg/g. impairment. Am J Kidney Dis. 2000;36:29–34. 12. Keevil BG, Kilpatrick ES, Nichols SP, Maylor PW. Biological variation of cystatin C: implications for the assessment of glomerular filtration rate. diovascular events, and incident CHF among ambulatory Clin Chem. 1998;44:1535–1539. persons with CHD. Moreover, higher cystatin C concentra- 13. Randers E, Erlandsen EJ. Serum cystatin C as an endogenous marker of the renal function: a review. Clin Chem Lab Med. 1999;37:389–395. tions predicted increased risk of these adverse clinical out- 14. Knight EL, Verhave JC, Spiegelman D, Hillege HL, de Zeeuw D, Curhan comes even among persons without microalbuminuria or low GC, de Jong PE. Factors influencing serum cystatin C levels other than estimated GFR. If confirmed in future studies, serum cystatin renal function and the impact on renal function measurement. Kidney Int. 2004;65:1416–1421. C measurement may prove to be useful for the prediction of 15. Fried LF, Katz R, Sarnak MJ, Shlipak MG, Chaves PH, Jenny NS, cardiovascular events among persons with coronary heart Stehman-Breen C, Gillen D, Bleyer AJ, Hirsch C, Siscovick D, Newman disease. AB. Kidney function as a predictor of noncardiovascular mortality. JAm Soc Nephrol. 2005;16:3728–3735. 16. Shlipak MG, Sarnak MJ, Katz R, Fried LF, Seliger SL, Newman AB, Acknowledgments Siscovick DS, Stehman-Breen C. Cystatin C and the risk of death and The authors thank Eric Vittinghoff, PhD, and David Glidden, PhD, cardiovascular events among elderly persons. N Engl J Med. 2005;352: for their assistance in statistical consultation. 2049–2060. 17. Shlipak MG, Wassel Fyr CL, Chertow GM, Harris TB, Kritchevsky SB, Sources of Funding Tylavsky FA, Satterfield S, Cummings SR, Newman AB, Fried LF. Cystatin C and mortality risk in the elderly: the health, aging, and body Dr Ix was funded by grants from the University of California at San composition study. J Am Soc Nephrol. 2006;17:254–261. Francisco Academic Senate and the American Heart Association 18. Sarnak MJ, Katz R, Stehman-Breen CO, Fried LF, Jenny NS, Psaty BM, Fellow-to-Faculty Transition Award. Dr Shlipak was funded by the Newman AB, Siscovick D, Shlipak MG. Cystatin C concentration as a American Federation for Aging Research and National Institute on risk factor for heart failure in older adults. Ann Intern Med. 2005;142: Aging (Paul Beeson Scholars Program), the Robert Wood Johnson 497–505. Foundation (Generalist Faculty Scholars Program), and National 19. Ix JH, Shlipak MG, Liu HH, Schiller NB, Whooley MA. Association Institutes of Health grant R01 DK066488. Dr Whooley was funded between renal insufficiency and inducible ischemia in patients with by the Department of Veterans Affairs, the American Federation for coronary artery disease: the Heart and Soul Study. J Am Soc Nephrol. Aging Research, the Robert Wood Johnson Foundation, the Ischemia 2003;14:3233–3238. Research and Education Foundation, the Nancy Kimran Heart 20. Ruo B, Rumsfeld JS, Hlatky MA, Liu H, Browner WS, Whooley MA. Research Fund, and Dade Behring, Inc. Dade Behring supported the Depressive symptoms and health-related quality of life: the Heart and cystatin C measurements in the Heart and Soul cohort. Soul Study. JAMA. 2003;290:215–221. 21. Erlandsen EJ, Randers E, Kristensen JH. Evaluation of the Dade Behring N Latex Cystatin C assay on the Dade Behring Nephelometer II system. Disclosures Scand J Clin Lab Invest. 1999;59:1–8. Dr Shlipak has received a modest honorarium from Amgen, Inc. Dr 22. K/DOQI clinical practice guidelines for chronic kidney disease: eval- Chertow has received significant research support from Amgen, Inc. uation, classification, and stratification. Am J Kidney Dis. 2002;39(suppl Dr Whooley has received significant research support from Roche 1):S1–S266. Pharmaceuticals, Amgen, Inc, and Dade Behring, Inc. The other 23. Toto RD. Microalbuminuria: definition, detection, and clinical signif- authors report no conflicts. icance. J Clin Hypertens (Greenwich). 2004;6(suppl 3):2–7.

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Ix et al Cystatin C, Mortality, Events, and Heart Failure 179

24. Beattie MS, Shlipak MG, Liu H, Browner WS, Schiller NB, Whooley 29. Redfield MM, Jacobsen SJ, Burnett JC Jr, Mahoney DW, Bailey KR, MA. C-reactive protein and ischemia in users and nonusers of beta- Rodeheffer RJ. Burden of systolic and diastolic ventricular dysfunction in blockers and statins: data from the Heart and Soul Study. Circulation. the community: appreciating the scope of the heart failure epidemic. 2003;107:245–250. JAMA. 2003;289:194–202. 25. Ix JH, Shlipak MG, Brandenburg VM, Ali S, Ketteler M, Whooley MA. 30. Shlipak MG, Heidenreich PA, Noguchi H, Chertow GM, Browner WS, Association between human fetuin-A and the metabolic syndrome: data McClellan MB. Association of renal insufficiency with treatment and from the Heart and Soul Study. Circulation. 2006;113:1760–1767. outcomes after myocardial infarction in elderly patients. Ann Intern Med. 26. Friedewald WT, Levy RI, Fredrickson DS. Estimation of the concen- 2002;137:555–562. tration of low-density lipoprotein cholesterol in plasma, without use of the 31. Ix JH, Mercado N, Shlipak MG, Lemos PA, Boersma E, Lindeboom W, preparative ultracentrifuge. Clin Chem. 1972;18:499–502. O’Neill WW, Wijns W, Serruys PW. Association of chronic kidney 27. Luepker RV, Apple FS, Christenson RH, Crow RS, Fortmann SP, Goff D, disease with clinical outcomes after coronary revascularization: the Ar- terial Revascularization Therapies Study (ARTS). Am Heart J. 2005;149: Goldberg RJ, Hand MM, Jaffe AS, Julian DG, Levy D, Manolio T, 512–519. Mendis S, Mensah G, Pajak A, Prineas RJ, Reddy KS, Roger VL, 32. Perkins BA, Nelson RG, Ostrander BE, Blouch KL, Krolewski AS, Rosamond WD, Shahar E, Sharrett AR, Sorlie P, Tunstall-Pedoe H. Case Myers BD, Warram JH. Detection of renal function decline in patients definitions for acute coronary heart disease in epidemiology and clinical with diabetes and normal or elevated GFR by serial measurements of research studies: a statement from the AHA Council on Epidemiology serum cystatin C concentration: results of a 4-year follow-up study. JAm and Prevention; AHA Statistics Committee; World Heart Federation Soc Nephrol. 2005;16:1404–1412. Council on Epidemiology and Prevention; the European Society of Car- 33. Jernberg T, Lindahl B, James S, Larsson A, Hansson LO, Wallentin L. diology Working Group on Epidemiology and Prevention; Centers for Cystatin C: a novel predictor of outcome in suspected or confirmed Disease Control and Prevention; and the National Heart, Lung, and Blood non-ST-elevation acute coronary syndrome. Circulation. 2004;110: Institute. Circulation. 2003;108:2543–2549. 2342–2348. 28. Angeja BG, Shlipak MG, Go AS, Johnston SC, Frederick PD, Canto JG, 34. Koenig W, Twardella D, Brenner H, Rothenbacher D. Plasma concen- Barron HV, Grady D. Hormone therapy and the risk of stroke after acute trations of cystatin C in patients with coronary heart disease and risk for myocardial infarction in postmenopausal women. J Am Coll Cardiol. secondary cardiovascular events: more than simply a marker of glomer- 2001;38:1297–1301. ular filtration rate. Clin Chem. 2005;51:321–327.

CLINICAL PERSPECTIVE Previous epidemiological studies have demonstrated associations of moderate kidney disease with mortality, cardiovascular events, and heart failure among diverse populations. However, studies evaluating whether milder kidney disease is associated with these adverse outcomes have been limited by the use of serum creatinine and associated estimating equations because they are insensitive to mild decrements in glomerular filtration rate. Cystatin C is a novel endogenous biomarker of kidney function that may be more sensitive to mild decrements in glomerular filtration rate. Whether cystatin C concentrations are associated with adverse clinical events among persons with coronary heart disease has not been studied extensively. In the present study, we demonstrate that higher cystatin C concentrations are associated with mortality, cardiovascular events, and incident heart failure, independent of traditional cardiovascular risk factors, among 990 persons with coronary heart disease. Higher cystatin C concentrations predicted these outcomes whether or not participants had low estimated glomerular filtration rate or microalbuminuria. If confirmed in future studies, cystatin C may have clinical utility in identifying persons at high risk for adverse outcomes among populations with coronary heart disease.

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Epidemiology

Elevated Homocysteine Is Associated With Reduced Regional Left Ventricular Function The Multi-Ethnic Study of Atherosclerosis

Khurram Nasir, MD, MPH; Michael Tsai, PhD; Boaz D. Rosen, MD; Veronica Fernandes, MD; David A. Bluemke, MD, PhD; Aaron R. Folsom, MD, MPH; João A.C. Lima, MD

Background—An elevated homocysteine (Hcy) level has been reported to be a risk factor for the development of congestive heart failure in individuals free of myocardial infarction. In this study, we aim to investigate the relationship between Hcy levels and regional left ventricular function in an asymptomatic population. Method and Results—Regional peak systolic midwall circumferential strains were calculated from 1178 tagged magnetic resonance imaging studies in participants in the Multi-Ethnic Study of Atherosclerosis (MESA). Left ventricular regions were defined by coronary territories (left anterior descending, left circumflex, right coronary artery). For the 1178 study participants (66Ϯ10 years of age, 58% males), the median (interquartile range) of Hcy was 9.1 (9.0 to 9.3). After adjustment for traditional risk factors, race, height, weight, left ventricular end-diastolic mass/volume, serum creatinine, and measures of atherosclerosis, reduced regional myocardial circumferential shortening across sex-specific quartiles of plasma Hcy in the left anterior descending (Pϭ0.038) and left circumflex (Pϭ0.009) regions persisted, which indicated an important association of reduced function with elevated Hcy. Multiple linear regression analyses confirmed that circumferential systolic dysfunction was associated with log transformed Hcy levels in the left anterior descending (Pϭ0.004) and left circumflex (Pϭ0.0002) regions. In the fully adjusted model, the odds ratio for left ventricular strains below the 10th percentile with 1 SD increases in log-transformed Hcy was 1.33 (95% confidence interval, 1.04 to 1.70; Pϭ0.022) for the left anterior descending, 1.28 (95% confidence interval, 1.00 to 1.64; Pϭ0.046) for the left circumflex, and 1.32 (95% confidence interval, 1.03 to 1.69; Pϭ0.025) for the right coronary artery region. Conclusion—In this asymptomatic population, an elevated Hcy level is associated with reduced regional left ventricular systolic function detected by tagged magnetic resonance imaging. (Circulation. 2007;115:180-187.) Key Words: contractility Ⅲ heart failure Ⅲ homocysteine Ⅲ imaging Ⅲ magnetic resonance imaging

n the past decade, plasma total homocysteine (Hcy) has been Clinical Perspective p 187 Iextensively explored and proposed as an independent cardiovas- cular disease (CVD) risk factor. Elevated total Hcy levels have been This study hypothesized that individuals with Hhe related to greater risk of CVD outcomes, which include myocardial would exhibit alterations in regional LV function. We used infarction,1–3 stroke,4–7 and cardiovascular mortality,6,8,9 in many magnetic resonance imaging (MRI) to test this hypothesis but not all studies.10–13 However, recent clinical trials have not been because, apart from precise measurement of global LV able to demonstrate a reduction of clinical CVD end points with a volumes, LV mass, and LV ejection fraction (LVEF), MRI decrease of Hcy levels.14–17 However, plasma Hcy was reported to permits assessment of regional LV function in a detailed be a risk factor for the development of congestive heart failure and quantitative manner through myocardial tagging.25 (CHF) in the Framingham Heart Study.18 Although experimental Detection of subclinical reduction in LV function poten- evidence strongly suggests that the myocardium may be uniquely tially identifies asymptomatic individuals at high risk for susceptible to Hcy-induced injury,19–24 it is unclear whether hyper- the development of clinical heart failure.26 A detailed homocysteinemia (Hhe) is independently associated with a reduced characterization of regional ventricular function with Hhe intrinsic left ventricular (LV) myocardial function in asymptomatic may also help elucidate the mechanisms that underlie the individuals free of CVD. increased risk of CHF associated with Hhe.

Received June 23, 2006; accepted October 13, 2006. From the Division of Cardiology (K.N., B.D.R., V.F., J.A.C.L.), Department of Medicine, and the Department of Radiology (D.A.B.), Johns Hopkins University, Baltimore, Md, and the Department of Laboratory Medicine and Pathology (M.T.) and the Division of Epidemiology and Community Health (A.R.F.), University of Minnesota, Minneapolis, Minn. Correspondence to Dr João A.C. Lima, Cardiology Division, Blalock 524, Johns Hopkins Hospital, 600 N. Wolfe St, Baltimore, MD 21287–0409. E-mail [email protected] © 2007 American Heart Association, Inc. Circulation is available at http://www.circulationaha.org DOI: 10.1161/CIRCULATIONAHA.106.633750

180

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Nasir et al Regional LV Function and Homocysteine 181

Methods LV apex. Short-axis cine untagged images (series 5) were acquired from the end-diastolic image of the 4-chamber acquisition by Study Population prescribing 10 to 12 slices perpendicular to a line from the middle of The Multi-Ethnic Study of Atherosclerosis (MESA) is a prospective the mitral valve plane to the cardiac apex. A 2-chamber vertical study designed to evaluate mechanisms that underlie the develop- long-axis cine sequence (series 6) was obtained by prescribing a ment and progression of subclinical CVD in asymptomatic individ- single slice along a line extending from the LV apex to the middle of uals. The details of MESA have been previously described.27 In the mitral valve plane as viewed on the 4-chamber view (series 4). short, 6814 men and women, 45 to 85 years of age, from 4 ethnicities The image prescription was designed to minimize variation among (non-Hispanic white, black, Hispanic, and Chinese) were enrolled by the MR field centers. Cine images were obtained with a temporal 6 participating centers (Wake Forest University, Winston Salem, resolution of Ϸ50 ms or less.30 NC; Columbia University, New York, NY; Johns Hopkins Hospital, LV end-diastolic mass, LV-end diastolic volume, and LVEF were Baltimore, Md; University of Minnesota, Minneapolis, Minn; North- determined for each study by a dedicated commercially available western University, Chicago, Ill; and University of California, Los software (MASS, version 4.2, Medis, The Netherlands). LV end-di- Angeles). On entry, all participants underwent extensive evaluation astolic volume and LV end-systolic volume were calculated with that included clinical questionnaires, physical examination, and laboratory tests, which included fasting plasma glucose, triglycer- Simpson’s rule (the summation of areas on each separate slice ides, and total, high-density lipoprotein, and low-density lipoprotein multiplied by the sum of slice thickness and image gap). LV mass cholesterol levels. Individuals with symptomatic CVD were ex- was determined by the sum of the myocardial area (the difference cluded. Total plasma Hcy was measured with high-performance between endocardial and epicardial contour) times slice thickness liquid chromatography with fluorometric detection. The coefficient plus image gap in the end-diastolic phase multiplied by the specific for variation for this assay was 3.8%. gravity of myocardium (1.05 g/mL). LV stroke volume was calcu- The carotid arteries were evaluated with high-resolution B-mode lated as the difference between LV end-diastolic volume and LV ultrasonography. One real-time transverse (short-axis) scanning end-systolic volume. LVEF was calculated as LV stroke volume sequence and 1 longitudinal image of the common carotid artery divided by LV end-diastolic volume multiplied by 100. Concentric were acquired. The maximal intimal medial thickness (IMT) of the remodeling was determined by calculating the ratio of LV end-dia- common carotid artery was defined as the mean of the maximal IMT stolic mass to end-diastolic volume (M/V ratio). of the near and far wall on both the left and right sides, and it was measured at 10 mm proximal to the common carotid bulb. The MRI Protocol for Tagged Images values were expressed as meanϮSE in millimeters. Computed Images were acquired by 1.5T MR scanners (SIGNA [LX and CVi], tomography scanning of the chest was performed either with an General Electric Medical Systems, Waukesha, Wisc, and Siemens ECG-triggered (at 80% of the RR interval) electron beam computed Medical Solutions [Vision and Sonata], Erlangen, Germany). All tomography scanner (Chicago, Los Angeles, and New York field images were obtained by ECG-triggered segmented k-space fast centers: Imatron C-150, GE/Imatron, South San Francisco, Calif) or spoiled gradient-echo (SPGR or FLASH) pulse sequence during with prospectively ECG-triggered scan acquisition at 50% of the RR breath-holding (12 to 18 seconds). Dedicated phase array coils were interval with a multidetector computed tomography system at 4 used for signal reception. After completion of the standard protocol, simultaneously acquired 2.5-mm slices for each cardiac cycle in a 4 tagged short-axis slices were obtained. Parallel stripe tags were sequential or axial scan mode (Baltimore, Md, Forsyth County, NC, prescribed in 2 orthogonal orientations (0° and 90°) with an identical and St. Paul, Minn field centers; Lightspeed [General Electric Health pulse sequence with additional spatial modulation of magnetization– Care, Chalfont St. Giles, United Kingdom] or Volume Zoom encoding gradients. [Siemens, Munich, Germany]).28 The coronary artery calcification Parameters for tagged images were: field of view 40 cm; slice measurements among scanning centers and between participants thickness 8 to 10 mm; repetition time 6 ms (range 3.5 to 7.2 ms); were adjusted with a standard calcium phantom scanned simulta- echo time 3.0 ms (range 2.0 to 4.2 ms); flip angle 12°; matrix size neously with each participant. The mean Agatston score was used in 256ϫ96 to 140; 4 to 9 phase encoding views per segment, and tag all analyses. Concordance for presence of coronary artery calcifica- spacing of 7 mm. With these parameters, 19 to 27 phases per cycle tion between duplicate scans was high and similar for both electron- were acquired, which yielded a temporal resolution of 40 ms (range beam and multidetector row computed tomography (96%, ␬ 0.92).29 20 to 41 ms). The current study examined 1178 MESA participants via tagged MRI studies as part of an ancillary study (from September 2001 to Harmonic Phase Analysis September 2002). The institutional review boards in each of the Tagged short-axis slices were analyzed by harmonic phase imaging participating centers approved the study protocol and informed (HARP, version 2.0, Diagnosoft Inc, Palo Alto, Calif). HARP is a consent was obtained from each participant. new method that enables a fast determination of myocardial strain.31,32 It has been demonstrated that strain measures obtained by Magnetic Resonance Imaging HARP have a good overall agreement when compared with Findtags The detailed protocol has been previously described. In brief, cardiac (open source software), a well-established method of strain analy- MRI examinations consisted of short- and long-axis cine images, sis.33,34 To assess the interobserver and intraobserver agreement for phase contrast images of the aorta, and black blood aorta images. The myocardial MR-tagged image analysis with the HARP technique, 3 protocol was accomplished within 30 to 45 minutes. All images were independent observers performed 2 separate and blinded quantitative acquired during short breath-holding (12 to 15 sec) at resting lung strain analyses of myocardial cine MR tagging images in 24 MESA volume.30 First, 4 midline, ECG-gated, sagittal scout images were acquired trial participants. Intraclass correlation coefficients R for interob- with a fast gradient-echo sequence (series 1). These images were server and intraobserver agreement for peak systolic midwall ECC 35 used to confirm the correct position of the phased-array surface coil. were 0.81 and 0.84, respectively. Second, the same sequence was used to acquire 3 axial scout views To evaluate regional myocardial function, we used peak midwall (series 2); these images were acquired beginning 2 cm above the systolic/diastolic strain. An in-house tool was developed to evaluate diaphragm. Next, a pseudovertical long-axis scout image (series 3) regional strains according to the coronary artery perfusion areas (left was obtained with the largest LV image from series 2. This slice anterior descending [LAD], left circumflex [LCX] and right coro- extended through the middle of the mitral valve plane and the LV nary artery perfusion territories). Assignment of the different regions apex. Four-chamber long-axis cine images (series 4) were next to coronary territories was done according to recently published acquired with a cine ECG-gated fast gradient-echo pulse sequence. standards.36 By convention systolic circumferential strains are neg- The imaging plane for the 4-chamber view was prescribed along ative because they indicate systolic circumferential shortening. points that intersected the middle of the mitral valve plane and the Therefore, an increased negativity denotes enhanced function.

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 182 Circulation January 16, 2007

TABLE 1. Characteristics of the Study Population According to Homocysteine Quartiles

Sex-Specific Quartiles of Homocysteine*

First Quartile Second Quartile Third Quartile Fourth Quartile P for Trend Age, y 62Ϯ10 65Ϯ968Ϯ970Ϯ9 Ͻ0.0001 Men, % 52 54 53 54 0.96 Systolic blood pressure, mm Hg 124Ϯ20 127Ϯ19 129Ϯ20 135Ϯ23 Ͻ0.0001 Hypertension,† % 35 44 47 61 Ͻ0.0001 Cholesterol/HDL ratio 4.1Ϯ1.2 4.1Ϯ1.3 4.0Ϯ1.1 4.1Ϯ1.2 0.9 Serum creatinine, mg/dL 0.88Ϯ0.16 0.93Ϯ0.17 0.98Ϯ0.18 1.12Ϯ0.44 Ͻ0.0001 Diabetes mellitus,‡ % 15 18 15 23 Ͻ0.0001 Current smoker, % 10 13 9 12 0.41 Height, cm 166Ϯ10 166Ϯ10 165Ϯ9 165Ϯ10 0.55 Weight, lb 169Ϯ34 166Ϯ31 167Ϯ30 167Ϯ35 0.89 Coronary artery calcification§ 0 (0-79) 6 (0-139) 19 (0-167) 44 (0-296) Ͻ0.001 Carotid intima media thickness, mm 0.99Ϯ0.55 1.06Ϯ0.65 1.12Ϯ0.60 1.27Ϯ0.78 Ͻ0.001 Values are expressed as meanϮSD or %. HDL indicates high-density lipoprotein. *Homocysteine (␮mol/L) quartiles for women: first, 4.3–7.2; second, 7.3–8.3; third, 8.4–10.1; fourth, 10.2–20.3. Homocysteine (␮mol/L) quartiles for men: first, 3.8–8.3; second, 8.4–9.8; third, 9.9–11.6; fourth, 11.7–41.1. †Hypertension defined as diastolic blood pressure Ն90 mm Hg, systolic blood pressure Ն140 mm Hg, or receiving treatment for hypertension. ‡Diabetes mellitus defined as fasting plasma glucose Ն126 mg/dL or receiving treatment for diabetes. §Data presented as median (interquartile range).

Statistical Analysis Results The associations of Hcy with demographic characteristics and The median (interquartile range [IQR]) and geometric mean cardiovascular risk factors were determined by categorization of the (95% confidence interval) of Hcy for the 1178 study partic- cohort into Hcy quartiles. For comparison of variables with normal Ϯ Ϯ distribution, ANOVA was used, whereas the Kruskal-Wallis test was ipants (mean age SD 66 10 years, 58% males) were 9.1 used for variables with skewed distribution. In case of categorical (9.0 to 9.3) and 9.3 (9.1 to 9.5) ␮mol/L, respectively. The variables, ␹2 testing was used. median (IQR) of Hcy levels among females was 8.3 (7.1 to Multiple linear regression models were used to examine relations 10.1) ␮mol/L compared with 9.8 (8.2 to 11.6) ␮mol/L in of systolic circumferential strains in the LAD, LCX, and right males (PϽ0.001). As a result, the population was categorized coronary artery regions as the dependent variables to plasma Hcy. Three sets of multivariable models were examined in a hierarchical by sex-specific Hcy quartiles. The respective ranges for Hcy fashion: (1) adjusted for age, gender, and race; (2) adjusted for age, (␮mol/L) quartiles for women were: first, 4.3 to 7.2; second, gender, race, height, weight, systolic blood pressure, antihyperten- 7.3 to 8.3; third, 8.4 to 10.1; and fourth, 10.2 to 20.3; and for sive treatment, cigarette smoking status, diabetes, and total choles- men were: first, 3.8 to 8.3; second, 8.4 to 9.8; third, 9.9 to terol/high-density lipoprotein ratio; (3) adjusted for age, gender, race, height, weight, systolic blood pressure, antihypertensive treat- 11.6; and fourth, 11.7 to 41.1. Baseline characteristics of the ment, cigarette smoking status, diabetes status, total cholesterol/ study population by Hcy quartiles are listed in Table 1. high-density lipoprotein ratio, LV mass/volume ratio, serum creati- Higher Hcy levels were associated with greater age, systolic nine, coronary artery calcification, and carotid IMT. blood pressure, serum creatinine, coronary artery calcifica- In all of these models, we summarized the means of the LV Ͻ measurements across quartiles of plasma Hcy. The trend across the tion, and carotid IMT (all P 0.0001). Those with higher Hcy quartiles was tested for statistical significance by scoring each levels were also more likely to be hypertensive and have quartile by its median value and entering the score as a continuous diabetes. term in the regression model. We also report regression coefficients, The least-squares mean values of adjusted regional myo- which represent change in systolic strains, that treat Hcy as a cardial circumferential shortening across sex-specific quar- continuous variable with natural logarithmic transformation as mea- sures of effect, in accordance with the above models. Additionally, tiles of plasma Hcy are shown in Table 2. In models adjusted logistic regression was used to assess the odds of strains below the for age, gender, and race (model 1), regional systolic function 10th percentile for increasing levels of Hcy. The 10th percentile was lower across increasing sex-specific quartiles of plasma cutoff was also used to categorize values of peak systolic strains as Hcy in the LAD and LCX regions (LAD Pϭ0.047; LCX clinically meaningful. We selected this cutoff because the resultant ϭ ϭ values are in the range shown by Kraitchman et al34 to be dysfunc- P 0.023; right coronary artery P 0.075 for trend). The tional. In secondary analyses, we incorporated interaction terms relationship of regional myocardial circumferential strains to individually in multivariate model 3 (log transformed plasma Hcy ϫ plasma Hcy quartiles was robust after adjustment for clinical covariate) to test for the presence of effect modification by age, covariates (model 2). The association was unaltered with gender, race, systolic blood pressure, and creatinine. further adjustment for creatinine levels, left ventricle end-di- The authors had full access to and take full responsibility for the integrity of the data. All authors have read and agree to the astolic mass/volume ratio, coronary artery calcification, and manuscript as written. carotid IMT (model 3).

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Nasir et al Regional LV Function and Homocysteine 183

TABLE 2. Adjusted Mean Left Ventricular Regional Circumferential Strains Across Homocysteine Quartiles

Sex-Specific Quartiles of Homocysteine*

First Quartile Second Quartile Third Quartile Fourth Quartile P for Trend Left anterior descending Model 1† Ϫ15.7Ϯ0.25 Ϫ15.8Ϯ0.25 Ϫ15.6Ϯ0.25 Ϫ15.0Ϯ0.25 0.047 Model 2‡ Ϫ15.8Ϯ0.25 Ϫ15.8Ϯ0.25 Ϫ15.5Ϯ0.26 Ϫ15.0Ϯ0.27 0.044 Model 3§ Ϫ15.9Ϯ0.25 Ϫ15.7Ϯ0.26 Ϫ15.5Ϯ0.25 Ϫ15.0Ϯ0.26 0.038 Left circumflex Model 1† Ϫ17.5Ϯ0.29 Ϫ17.1Ϯ0.30 Ϫ16.9Ϯ0.30 Ϫ16.6Ϯ0.30 0.023 Model 2‡ Ϫ17.6Ϯ0.30 Ϫ17.2Ϯ0.30 Ϫ16.9Ϯ0.30 Ϫ16.5Ϯ0.32 0.011 Model 3§ Ϫ17.7Ϯ0.29 Ϫ17.1Ϯ0.30 Ϫ16.8Ϯ0.30 Ϫ16.4Ϯ0.31 0.009 Right coronary artery Model 1† Ϫ12.4Ϯ0.24 Ϫ12.2Ϯ0.26 Ϫ12.0Ϯ0.25 Ϫ11.8Ϯ0.26 0.075 Model 2‡ Ϫ12.4Ϯ0.25 Ϫ12.1Ϯ0.25 Ϫ12.1Ϯ0.26 Ϫ11.8Ϯ0.27 0.080 Model 3§ Ϫ12.5Ϯ0.25 Ϫ12.1Ϯ0.26 Ϫ12.0Ϯ0.26 Ϫ11.8Ϯ0.27 0.070 *Homocysteine (␮mol/L) quartiles for women: first, 4.3–7.2; second, 7.3–8.3; third, 8.4–10.1; and fourth, 10.2–20.3. Homocysteine (␮mol/L) quartiles for men: first, 3.8–8.3; second, 8.4–9.8; third, 9.9–11.6; and fourth, 11.7–41.1. Quartiles data are adjusted least square means (meanϮSEM). †Model 1: adjusted for age, sex, and race. ‡Model 2: adjusted for age, sex, race, height, weight, cigarette smoking status, diabetes status, systolic blood pressure, antihypertensive medications, and cholesterol/high-density lipoprotein ratio. §Model 3: adjusted for age, sex, race, height, weight, cigarette smoking status, diabetes status, systolic blood pressure, antihypertensive medications, cholesterol/high-density lipoprotein ratio, left ventricular end-diastolic mass/volume ratio, creatinine, carotid IMT, and coronary artery calcification.

When considered as a continuous variable, greater plasma cerned, no differences were observed across increasing sex- Hcy levels (1 SD increase in log transformed Hcy) were specific quartiles of Hcy (Table 4). Overall, the associations significantly associated with regional circumferential systolic of Hcy levels with LV indices (regional and global) did not strains in the LAD (Pϭ0.02 to 0.004) and LCX (0.02 to vary by age, gender, race, systolic blood pressure, or creati- 0.002) regions, respectively (Table 3). Logistic regression nine (P for all interactions Ͼ0.10) (data not shown). analyses demonstrated that in the fully adjusted model (model 3), the odds ratio for LV strains below the 10th percentile with Discussion a 1-SD increase in log-transformed Hcy levels was 1.33 (95% The present study cross-sectionally examined the relationship confidence interval, 1.04 to 1.70; Pϭ0.022) for the LAD, of plasma Hcy to regional LV function in a large community- 1.28 (95% confidence interval, 1.00 to 1.64; Pϭ0.046) for the based sample of asymptomatic individuals. Evidence exists in LCX, and 1.32 (95% confidence interval, 1.03 to 1.69; the literature that Hhe is associated with diverse parameters Pϭ0.025) for the right coronary artery. of subclinical CVD such as endothelial dysfunction,37–40 As far as global LV indices (LVEF, LV end-diastolic mass, vascular stiffness,41,42,12 and atherosclerosis.43,44 Our findings LV end-diastolic volume, LV mass/volume ratio) are con- extend these observations by demonstrating that greater Hcy

TABLE 3. Relations of Plasma Homocysteine to Regional Left Ventricular Circumferential Strains

Left Anterior Descending Left Circumflex Right Coronary Artery

Regression Coefficients* Regression Coefficients Regression Coefficients (95% CI) P (95% CI) P (95% CI) P Model 1† 0.29 (0.04 to 0.55) 0.02 0.34 (0.04 to 0.63) 0.02 0.31 (Ϫ0.11 to 0.38) 0.30 Model 2‡ 0.31 (0.05 to 0.56) 0.02 0.41 (0.11 to 0.71) 0.007 0.16 (Ϫ0.92 to 0.41) 0.21 Model 3§ 0.43 (0.14 to 0.72) 0.004 0.54 (0.20 to 0.89) 0.002 0.25 (Ϫ0.04 to 0.54) 0.086 CI indicates confidence interval. *Linear regression coefficients represent difference in systolic strains (%) for 1-SD increment in log-transformed homocysteine. †Model 1: adjusted for age, sex, and race. ‡Model 2: adjusted for age, sex, race, height, weight, cigarette smoking status, diabetes status, systolic blood pressure, antihypertensive medications, and cholesterol/high-density lipoprotein ratio. §Model 3: adjusted for age, sex, race, height, weight, cigarette smoking status, diabetes status, systolic blood pressure, antihypertensive medications, cholesterol/high-density lipoprotein ratio, left ventricular end-diastolic mass/volume ratio, creatinine, carotid IMT, and coronary artery calcification.

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 184 Circulation January 16, 2007

TABLE 4. Adjusted Global Left Ventricular Indices Across Homocysteine Quartiles

Homocysteine Sex-Specific Quartiles (␮mol/L)*

First Quartile Second Quartile Third Quartile Fourth Quartile P for Trend Left ventricular ejection fraction, % Model 1† 69.5Ϯ0.43 68.9Ϯ0.44 68.0Ϯ0.44 68.9Ϯ0.45 0.10 Model 2‡ 69.8Ϯ0.43 69.0Ϯ0.44 68.1Ϯ0.44 68.8Ϯ0.45 0.09 Model 3§ 69.7Ϯ0.43 68.9Ϯ0.44 68.2Ϯ0.44 68.9Ϯ0.0.46 0.10 Left ventricular end-diastolic mass, g Model 1† 143.7Ϯ1.98 141.5Ϯ2.07 148.6Ϯ2.04 146.1Ϯ2.08 0.21 Model 2‡ 145.7Ϯ1.67 142.6Ϯ1.73 147.8Ϯ1.72 143.0Ϯ1.75 0.72 Model 3§8 145.5Ϯ1.59 142.9Ϯ1.61 147.9Ϯ1.59 142.8Ϯ1.68 0.73 Left ventricular end-diastolic volume, mL Model 1† 121.9Ϯ1.69 121.4Ϯ1.76 127.2Ϯ1.74 122.5Ϯ1.77 0.38 Model 2‡ 123.1Ϯ1.54 122.0Ϯ1.58 126.2Ϯ1.58 121.1Ϯ1.61 0.96 Model 3§ 123.2Ϯ1.33 121.7Ϯ1.36 126.1Ϯ1.34 121.5Ϯ1.42 0.96 Left ventricular end-diastolic mass/volume ratioʈ Model 1† 1.19Ϯ0.01 1.18Ϯ0.01 1.19Ϯ0.01 1.21Ϯ0.01 0.43 Model 2‡ 1.20Ϯ0.01 1.19Ϯ0.01 1.19Ϯ0.01 1.20Ϯ0.01 0.92 Model 3§ 1.20Ϯ0.01 1.18Ϯ0.01 1.19Ϯ0.01 1.19Ϯ0.01 0.81 Values are expressed as meanϮSEM unless otherwise indicated. *Homocysteine (␮mol/L) quartiles for women: first, 4.3–7.2; second, 7.3–8.3; third, 8.4–10.1; and fourth, 10.2–20.3. Homocysteine (␮mol/L) quartiles for men: first, 3.8–8.3; second, 8.4–9.8; third, 9.9–11.6; and fourth, 11.7–41.1. †Model 1: adjusted for age, sex, and race. ‡Model 2: adjusted for age, sex, race, height, weight, cigarette smoking status, diabetes status, systolic blood pressure, antihypertensive medications, and cholesterol/high-density lipoprotein ratio. §Model 3: adjusted for age, sex, race, height, weight, cigarette smoking status, diabetes status, systolic blood pressure, antihypertensive medications, cholesterol/high-density lipoprotein ratio, left ventricular end-diastolic mass/volume ratio, creatinine, carotid IMT, and coronary artery calcification. ʈNot adjusted for left ventricular end-diastolic mass/volume ratio in model 3.

levels are associated with a lower regional LV systolic in a prospective study of 2491 adults, Vasan et al demon- function independent of potential confounders among indi- strated that baseline Hcy concentration was associated posi- viduals free of coronary heart disease. tively with risk of CHF during a follow-up of 8 years.18 Our The association of Hcy with CVD outcomes in the litera- study indirectly supports this concept by demonstrating a ture is inconsistent. In the Physicians’ Health Study, a modest cross-sectional association between elevated serum Hcy and increases in plasma Hcy level conferred a 3-fold increase in reduced myocardial strain in asymptomatic participants of the relative risk for myocardial infarction and mortality from MESA study, most of whom (95%) had LVEF Ͼ55%. cardiovascular disease.2 However, after further follow-up, the association was not statistically significant.45 An elevated Plausible Mechanisms Hcy level was independently associated with incident stroke The mechanisms by which Hhe affects cardiac function as well all-cause and CVD mortality in the Framingham remain an active area of research, with a number of studies Heart Study.5 However, such positive associations are not elucidating the basic pathomechanisms that link Hcy to reported in all studies. In the ARIC study, Folsom et al myocardial dysfunction with appropriate animal models. demonstrated that adjustment for other coronary heart disease Increased Hcy levels can promote endothelial dysfunction of risk factors abolished the association of Hcy with incident coronary resistance vessels.49,50 Investigators have also em- coronary heart disease events.13 The results of the prospective phasized the critical role of Hcy as a source of increased Caperhilly Study also do not support the hypothesis that a oxidative stress, a factor known to promote myocardial high circulating Hcy concentration is a risk factor for acute damage.50 In addition to indirect effects on myocardial coronary events in a middle-aged male population free of performance, Hcy may have direct effect on the myocardium. prior heart disease.12 A number of other prospective epide- Short-term exposure of isolated papillary muscle to Hcy miological studies of subjects initially free of coronary heart decreases the rate of rise and duration of action potentials, disease have shared similar findings. A recent meta-analysis without detectable changes in the resting membrane potential indicated that Hcy level is only weakly related to CVD risk in or the early phase of repolarization.24 In Hhe rats, a depres- healthy populations.46 More recently, evidence for an associ- sion in the ST segment as well an increased duration of the ation of Hcy with prevalent and incident CHF has been QRS complex (cardiac depolarization) has been observed.22 postulated. Among patients with CHF, elevated plasma Hcy Elevated Hcy has also been implicated in induction of cardiac levels have been previously reported.47,48 More importantly, fibrosis, probably by activation of transforming growth

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Nasir et al Regional LV Function and Homocysteine 185

22 factor-␤1. Also, because of its limited ability to metabolize Assignment of LV myocardial regions to major coronary Hcy, the myocardium may be particularly sensitive to ele- territories in our study was performed in accordance with vated Hcy levels. established criteria.36 The importance of this analysis stems from the fact that it is routinely used in other imaging Methodologic Considerations modalities, which include stress echo and nuclear studies. Our study includes 1178 subjects, which makes it one of the Importantly, our study highlights the relationship between largest MRI series of quantitative regional myocardial func- myocardial perfusion in the major coronary territories and tion that we are aware of. HARP is a new tool that is capable regional LV function.34 of a fast and reliable determination of global and regional LV function. Although the limitations of MRI tagging and Conclusion In summary, our study provides evidence that reduced re- regional function assessment by HARP are well known, this gional systolic function is observed in asymptomatic individ- tool permits analysis of a large amount of MRI data within a uals with elevated Hcy levels. If Hhe is a causal factor in reasonable time period. More importantly, however, the progression from segmental to global LV dysfunction, then cross-sectional character of our study does not allow any an association of elevated Hcy with worsening of myocardial conclusion on the temporality of the described relationships. Previous data on the relationship of Hcy with global LV function and development of CHF should be demonstrable in indices are limited and conflicting. Nygard et al demonstrated prospective studies. an inverse association between LVEF and Hcy51 in patients with coronary artery disease, whereas no correlation was Acknowledgments 52 The authors thank the other investigators, the staff, and the partici- observed by Retterstol et al. In a recent study, Cesari et al pants of the MESA study for their valuable contributions. A full list reported Hhe to be strongly associated with the presence of a of participating MESA investigators and institutions can be found at low ejection fraction (Ͻ40%) in individuals with arterial http://www.mesa-nhlbi.org. hypertension but not among noromotensives.53 We did not find any association between LVEF and Hcy, even when Sources of Funding stratified by hypertension status (data not shown). Cesari et al This study was supported by an NHLBI grant (RO1-HL66075-01) studied high-risk individuals referred for coronary artery and MESA study contracts (NO1-HC-9808, NO1-HC-95168, and NO1-HC-95169). Dr Lima is supported by the Reynolds angiography, whereas we studied a low- to intermediate-risk Foundation. cohort without known clinical CVD.53 When one considers Hcy-induced endothelial dysfunction Disclosures of coronary resistance vessels, increased oxidative stress, None. myocardial fibrosis, and direct myocardial effects (which include negative inotropy) as potential mechanisms that may References lead to myocardial dysfunction, one might expect that these 1. Shai I, Stampfer MJ, Ma J, Manson JE, Hankinson SE, Cannuscio C, mechanisms should affect global LV function. However, the Selhub J, Curhan G, Rimm EB. Homocysteine as a risk factor for coronary heart diseases and its association with inflammatory biomarkers, fact that we observed a relationship of increasing Hcy with lipids and dietary factors. Atherosclerosis. 2004;177:375–381. regional myocardial dysfunction in the absence of global 2. Stampfer MJ, Malinow MR, Willett WC, Newcomer LM, Upson B, dysfunction can be explained on the following basis: It is Ullmann D, Tishler PV, Hennekens CH. A prospective study of plasma widely accepted that individuals may progress through an homocyst(e)ine and risk of myocardial infarction in US physicians. JAMA. 1992;268:877–881. asymptomatic phase of LV dysfunction before the develop- 3. Soinio M, Marniemi J, Laakso M, Lehto S, Ronnemaa T. Elevated plasma ment of clinical heart failure.54 Although abnormalities of homocysteine level is an independent predictor of coronary heart disease global LV structure and function represent an advanced stage events in patients with type 2 diabetes mellitus. Ann Intern Med. 2004; in the development of CHF, they likely evolved from more 140:94–100. 4. Tanne D, Haim M, Goldbourt U, Boyko V, Doolman R, Adler Y, Brunner localized alterations, particularly when one considers that the D, Behar S, Sela BA. Prospective study of serum homocysteine and risk most common etiologies of myocardial damage are regional of ischemic stroke among patients with preexisting coronary heart in nature.55,56 Regional alterations of myocardial perfusion, disease. Stroke. 2003;34:632–636. 55,57 5. Bostom AG, Silbershatz H, Rosenberg IH, Selhub J, D’Agostino RB, metabolism, and inflammation, as well as segmental Wolf PA, Jacques PF, Wilson PW. Nonfasting plasma total homocysteine function,58 are well documented in patients with global levels and all-cause and cardiovascular disease mortality in elderly Fra- processes such as cardiomyopathies and LV hypertrophic mingham men and women. Arch Intern Med. 1999;159:1077–1080. disease.59 We also believe that regional function is potentially 6. Bostom AG, Rosenberg IH, Silbershatz H, Jacques PF, Selhub J, D’Agostino RB, Wilson PW, Wolf PA. Nonfasting plasma total homo- more sensitive than global function for detection of early cysteine levels and stroke incidence in elderly persons: the Framingham changes in LV function. Finally, because the MESA study Study. Ann Intern Med. 1999;131:352–355. precluded the inclusion of patients with symptomatic CVD, 7. Schroecksnadel K, Frick B, Fuchs D. Homocysteine is an independent those with global LV dysfunction were likely excluded from risk factor for myocardial infarction, in particular fatal myocardial Ͻ infarction in middle-aged women. Circulation 2004;110:e37–e38; author the study. In our study, only 3 individuals had LVEF 40% reply e37–e38. and 5% had LVEF Ͻ55%. In order to describe a more direct 8. Lange S, Trampisch HJ, Haberl R, Darius H, Pittrow D, Schuster A, von relationship of Hcy with LV dysfunction, follow-up studies Stritzky B, Tepohl G, Allenberg JR, Diehm C. Excess 1-year cardiovas- are required in these patients to assess whether Hcy correlates cular risk in elderly primary care patients with a low ankle-brachial index (ABI) and high homocysteine level. Atherosclerosis. 2005;178:351–357. with further reduction in regional LV function and especially 9. Zylberstein DE, Bengtsson C, Bjorkelund C, Landaas S, Sundh V, Thelle with development of global LV impairment.58–60 D, Lissner L. Serum homocysteine in relation to mortality and morbidity

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 186 Circulation January 16, 2007

from coronary heart disease: a 24-year follow-up of the population study 30. Natori S, Lai S, Finn JP, Gomes AS, Hundley WG, Jerosch-Herold M, of women in Gothenburg. Circulation. 2004;109:601–606. Pearson G, Sinha S, Arai A, Lima JA, Bluemke DA. Cardiovascular 10. de Bree A, Verschuren WM, Blom HJ, Nadeau M, Trijbels FJ, Kromhout function in multi-ethnic study of atherosclerosis: normal values by age, D. Coronary heart disease mortality, plasma homocysteine, and B-vi- sex, and ethnicity. AJR Am J Roentgenol. 2006;186:S357–S365. tamins: a prospective study. Atherosclerosis. 2003;166:369–377. 31. O’Dell WG, McCulloch AD. Imaging three-dimensional cardiac function. 11. Fallon UB, Elwood P, Ben-Shlomo Y, Ubbink JB, Greenwood R, Smith Annu Rev Biomed Eng. 2000;2:431–456. GD. Homocysteine and ischaemic stroke in men: the Caerphilly study. J 32. Osman NF, Prince JL. Visualizing myocardial function using HARP Epidemiol Community Health. 2001;55:91–96. MRI. Phys Med Biol. 2000;45:1665–1682. 12. Fallon UB, Ben-Shlomo Y, Elwood P, Ubbink JB, Smith GD. Homo- 33. Garot J, Bluemke DA, Osman NF, Rochitte CE, McVeigh ER, Zerhouni cysteine and coronary heart disease in the Caerphilly cohort: a 10 year EA, Prince JL, Lima JA. Fast determination of regional myocardial strain follow up. Heart. 2001;85:153–158. fields from tagged cardiac images using harmonic phase MRI. Circu- 13. Folsom AR, Nieto FJ, McGovern PG, Tsai MY, Malinow MR, Eckfeldt lation. 2000;101:981–988. JH, Hess DL, Davis CE. Prospective study of coronary heart disease 34. Kraitchman DL, Sampath S, Castillo E, Derbyshire JA, Boston RC, incidence in relation to fasting total homocysteine, related genetic poly- Bluemke DA, Gerber BL, Prince JL, Osman NF. Quantitative ischemia morphisms, and B vitamins: the Atherosclerosis Risk in Communities detection during cardiac magnetic resonance stress testing by use of (ARIC) study. Circulation. 1998;98:204–210. FastHARP. Circulation. 2003;107:2025–2030. 14. Toole JF, Malinow MR, Chambless LE, Spence JD, Pettigrew LC, 35. Castillo E, Osman NF, Rosen BD, El-Shehaby I, Pan L, Jerosch-Herold Howard VJ, Sides EG, Wang CH, Stampfer M. Lowering homocysteine M, Lai S, Bluemke DA, Lima JA. Quantitative assessment of regional in patients with ischemic stroke to prevent recurrent stroke, myocardial myocardial function with MR-tagging in a multi-center study: interob- infarction, and death: the Vitamin Intervention for Stroke Prevention server and intraobserver agreement of fast strain analysis with Harmonic (VISP) randomized controlled trial. JAMA. 2004;291:565–575. Phase (HARP) MRI. J Cardiovasc Magn Reson. 2005;7:783–791. 15. Lonn E, Yusuf S, Arnold MJ, Sheridan P, Pogue J, Micks M, McQueen MJ, Probstfield J, Fodor G, Held C, Genest J Jr. Homocysteine lowering 36. Cerqueira MD, Weissman NJ, Dilsizian V, Jacobs AK, Kaul S, Laskey with folic acid and B vitamins in vascular disease. N Engl J Med WK, Pennell DJ, Rumberger JA, Ryan T, Verani MS. Standardized 2006;354:1567–1577. myocardial segmentation and nomenclature for tomographic imaging of 16. Bonaa KH, Njolstad I, Ueland PM, Schirmer H, Tverdal A, Steigen T, the heart: a statement for healthcare professionals from the Cardiac Wang H, Nordrehaug JE, Arnesen E, Rasmussen K. Homocysteine Imaging Committee of the Council on Clinical Cardiology of the lowering and cardiovascular events after acute myocardial infarction. American Heart Association. Circulation. 2002;105:539–542. N Engl J Med. 2006;354:1578–1588. 37. Stuhlinger MC, Oka RK, Graf EE, Schmolzer I, Upson BM, Kapoor O, 17. Liem A, Reynierse-Buitenwerf GH, Zwinderman AH, Jukema JW, van Szuba A, Malinow MR, Wascher TC, Pachinger O, Cooke JP. Endothelial Veldhuisen DJ. Secondary prevention with folic acid: effects on clinical dysfunction induced by hyperhomocyst(e)inemia: role of asymmetric outcomes. J Am Coll Cardiol. 2003;41:2105–2113. dimethylarginine. Circulation 2003;108:933–938. 18. Vasan RS, Beiser A, D’Agostino RB, Levy D, Selhub J, Jacques PF, 38. Tawakol A, Omland T, Gerhard M, Wu JT, Creager MA. Hyperhomo- Rosenberg IH, Wilson PW. Plasma homocysteine and risk for congestive cyst(e)inemia is associated with impaired endothelium-dependent vaso- heart failure in adults without prior myocardial infarction. JAMA. 2003; dilation in humans. Circulation. 1997;95:1119–1121. 289:1251–1257. 39. Tawakol A, Forgione MA, Stuehlinger M, Alpert NM, Cooke JP, 19. Kennedy RH, Owings R, Shekhawat N, Joseph J. Acute negative ino- Loscalzo J, Fischman AJ, Creager MA, Gewirtz H. Homocysteine tropic effects of homocysteine are mediated via the endothelium. impairs coronary microvascular dilator function in humans. J Am Coll Am J Physiol Heart Circ Physiol. 2004;287:H812–H817. Cardiol. 2002;40:1051–1058. 20. Joseph J, Joseph L, Shekhawat NS, Devi S, Wang J, Melchert RB, 40. Woo KS, Chook P, Lolin YI, Cheung AS, Chan LT, Sun YY, Sanderson Hauer-Jensen M, Kennedy RH. Hyperhomocysteinemia leads to patho- JE, Metreweli C, Celermajer DS. Hyperhomocyst(e)inemia is a risk factor logical ventricular hypertrophy in normotensive rats. Am J Physiol Heart for arterial endothelial dysfunction in humans. Circulation. 1997;96: Circ Physiol. 2003;285:H679–H686. 2542–2544. 21. Sood HS, Cox MJ, Tyagi SC. Generation of nitrotyrosine precedes acti- 41. Nestel PJ, Chronopoulos A, Cehun M. Arterial stiffness is rapidly induced vation of metalloproteinase in myocardium of hyperhomocysteinemic by raising the plasma homocysteine concentration with methionine. Ath- rats. Antioxid Redox Signal. 2002;4:799–804. erosclerosis. 2003;171:83–86. 22. Miller A, Mujumdar V, Palmer L, Bower JD, Tyagi SC. Reversal of 42. Bortolotto LA, Safar ME, Billaud E, Lacroix C, Asmar R, London GM, endocardial endothelial dysfunction by folic acid in homocysteinemic Blacher J. Plasma homocysteine, aortic stiffness, and renal function in hypertensive rats. Am J Hypertens. 2002;15:157–163. hypertensive patients. Hypertension. 1999;34:837–842. 23. Mujumdar VS, Hayden MR, Tyagi SC. Homocyst(e)ine induces calcium 43. Tsai MY, Arnett DK, Eckfeldt JH, Williams RR, Ellison RC. Plasma second messenger in vascular smooth muscle cells. J Cell Physiol 2000; homocysteine and its association with carotid intimal-medial wall 183:28–36. thickness and prevalent coronary heart disease: NHLBI Family Heart 24. Suarez J, Chagoya de Sanchez V. Inhibition of S-adenosyl-L- Study. Atherosclerosis. 2000;151:519–524. homocysteine hydrolase by adrenaline in isolated guinea-pig papillary 44. Adachi H, Hirai Y, Fujiura Y, Matsuoka H, Satoh A, Imaizumi T. Plasma muscles. Int J Biochem Cell Biol. 1997;29:1279–1284. homocysteine levels and atherosclerosis in Japan: epidemiological study 25. Zerhouni EA, Parish DM, Rogers WJ, Yang A, Shapiro EP. Human heart: by use of carotid ultrasonography. Stroke. 2002;33:2177–2181. tagging with MR imaging: a method for noninvasive assessment of 45. Chasan-Taber L, Selhub J, Rosenberg IH, Malinow MR, Terry P, Tishler myocardial motion. Radiology. 1988;169:59–63. PV, Willett W, Hennekens CH, Stampfer MJ. A prospective study of 26. McMurray JV, McDonagh TA, Davie AP, Cleland JG, Francis CM, folate and vitamin B6 and risk of myocardial infarction in US physicians. Morrison C. Should we screen for asymptomatic left ventricular dys- function to prevent heart failure? Eur Heart J. 1998;19:842–846. J Am Coll Nutr. 1996;15:136–143. 27. Bild DE, Bluemke DA, Burke GL, Detrano R, Diez Roux AV, Folsom 46. Homocysteine Studies Collaboration. Homocysteine and risk of ischemic AR, Greenland P, Jacob DR Jr, Kronmal R, Liu K, Nelson JC, O’Leary heart disease and stroke: a meta-analysis. JAMA 2002;288:2015–2022. D, Saad MF, Shea S, Szklo M, Tracy RP. Multi-ethnic study of athero- 47. Cooke GE, Eaton GM, Whitby G, Kennedy RA, Binkley PF, Moesch- sclerosis: objectives and design. Am J Epidemiol. 2002;156:871–881. berger ML, Leier CV. Plasma atherogenic markers in congestive heart 28. Detrano RC, Anderson M, Nelson J, Wong ND, Carr JJ, McNitt-Gray M, failure and posttransplant (heart) patients. J Am Coll Cardiol. 2000;36: Bild DE. Coronary calcium measurements: effect of CT scanner type and 509–516. calcium measure on rescan reproducibility–MESA study. Radiology. 48. Ventura P, Panini R, Verlato C, Scarpetta G, Salvioli G. Hyperhomocys- 2005;236:477–484. teinemia and related factors in 600 hospitalized elderly subjects. Metab- 29. Carr JJ, Nelson JC, Wong ND, McNitt-Gray M, Arad Y, Jacobs DR Jr, olism. 2001;50:1466–1471. Sidney S, Bild DE, Williams OD, Detrano RC. Calcified coronary artery 49. Lentz SR. Homocysteine and vascular dysfunction. Life Sci. 1997;61: plaque measurement with cardiac CT in population-based studies: stan- 1205–1215. dardized protocol of Multi-Ethnic Study of Atherosclerosis (MESA) and 50. Aguilar B, Rojas JC, Collados MT. Metabolism of homocysteine and its Coronary Artery Risk Development in Young Adults (CARDIA) study. relationship with cardiovascular disease. J Thromb Thrombolysis. 2004; Radiology. 2005;234:35–43. 18:75–87.

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Nasir et al Regional LV Function and Homocysteine 187

51. Nygard O, Nordrehaug JE, Refsum H, Ueland PM, Farstad M, Vollset coronary artery disease, or acute myocardial infarction. Am Heart J. SE. Plasma homocysteine levels and mortality in patients with coronary 1986;111:760–767. artery disease. N Engl J Med. 1997;337:230–236. 56. Mody FV, Brunken RC, Stevenson LW, Nienaber CA, Phelps ME, 52. Retterstol L, Paus B, Bohn M, Bakken A, Erikssen J, Malinow MR, Berg Schelbert HR. Differentiating cardiomyopathy of coronary artery disease K. Plasma total homocysteine levels and prognosis in patients with from nonischemic dilated cardiomyopathy utilizing positron emission previous premature myocardial infarction: a 10-year follow-up study. tomography. J Am Coll Cardiol. 1991;17:373–383. J Intern Med. 2003;253:284–292. 57. Eisenberg JD, Sobel BE, Geltman EM. Differentiation of ischemic from 53. Cesari M, Zanchetta M, Burlina A, Pedon L, Maiolino G, Sticchi D, nonischemic cardiomyopathy with positron emission tomography. Am J Cardiol. 1987;59:1410–1414. Pessina AC, Rossi GP. Hyperhomocysteinemia is inversely related with 58. Sunnerhagen KS, Bhargava V, Shabetai R. Regional left ventricular wall left ventricular ejection fraction and predicts cardiovascular mortality in motion abnormalities in idiopathic dilated cardiomyopathy. Am J Cardiol. high-risk coronary artery disease hypertensives. Arterioscler Thromb 1990;65:364–370. Vasc Biol. 2005;25:115–121. 59. Kramer CM, Reichek N, Ferrari VA, Theobald T, Dawson J, Axel L. 54. Wang TJ, Evans JC, Benjamin EJ, Levy D, LeRoy EC, Vasan RS. Natural Regional heterogeneity of function in hypertrophic cardiomyopathy. Cir- history of asymptomatic left ventricular systolic dysfunction in the com- culation. 1994;90:186–194. munity. Circulation. 2003;108:977–982. 60. O’Gara PT, Bonow RO, Maron BJ, Damske BA, Van Lingen A, Bacharach 55. Iskandrian AS, Hakki AH, Kane S. Resting thallium-201 myocardial SL, Larson SM, Epstein SE. Myocardial perfusion abnormalities in patients perfusion patterns in patients with severe left ventricular dysfunction: with hypertrophic cardiomyopathy: assessment with thallium-201 emission differences between patients with primary cardiomyopathy, chronic computed tomography. Circulation 1987;76:1214–1223.

CLINICAL PERSPECTIVE Increased homocysteine has been shown to be associated with incident congestive heart failure in the Framingham Heart Study. Although experimental evidence strongly suggests that the left ventricle is susceptible to homocysteine-induced injury, the relationship between elevated homocysteine levels and regional function has not been studied in individuals without a history of coronary artery disease. In our study, 1078 participants of the Multi-Ethnic Study of Atherosclerosis (MESA), without a known history of cardiovascular disease, underwent myocardial strain with magnetic resonance imaging. Higher levels of homocysteine were associated with reduced regional peak systolic circumferential strain. The relationship persisted after adjustment for traditional risk factors, left ventricle size, renal function, and measures of atherosclerosis. On the other hand, no such relationship existed with global left ventricular functions. Future studies are required to assess the relationship of elevated homocysteine levels with reduction in left ventricular function to establish causality.

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Greater Fish, Fruit, and Vegetable Intakes Are Related to Lower Incidence of Venous Thromboembolism The Longitudinal Investigation of Thromboembolism Etiology

Lyn M. Steffen, PhD; Aaron R. Folsom, MD; Mary Cushman, MD; David R. Jacobs, Jr, PhD; Wayne D. Rosamond, PhD

Background—Little is known about the role of dietary intake in the development of deep vein thrombosis or pulmonary embolus (venous thromboembolism [VTE]). Homocysteine, factor VIII, and von Willebrand factor levels, risk factors for VTE, are influenced by dietary intake. We tested the hypothesis that foods rich in B vitamins and ␻-3 fatty acids are negatively associated and meat intake is positively associated with incidence of VTE. Methods and Results—In a prospective study over 12 years, 14 962 middle-aged adults participating in the Atherosclerosis Risk in Communities study were followed up for incident VTE. All hospitalizations were identified, and 196 VTEs were validated by chart review. A food frequency questionnaire assessed dietary intake at baseline and year 6. In separate proportional hazards regression analyses, risk of developing VTE was computed across quintiles of selected nutrients, major food groups, and the Western diet pattern, with adjustment for demographic and lifestyle factors, body mass index, and diabetes. Hazard ratios and 95% confidence intervals of VTE incidence across quintiles of fruit and vegetable intake were 1.0 (reference), 0.73 (0.48 to 1.11), 0.57 (0.37 to 0.90), 0.47 (0.29 to 0.77), and 0.59 (0.36 to 0.99) ϭ (Ptrend 0.03). Eating fish 1 or more times per week was associated with 30% to 45% lower incidence of VTE for quintiles 2 to 5 compared with quintile 1, suggestive of a threshold effect. Hazard ratios of VTE across quintiles of red and processed meat intake were 1.0, 1.24 (0.78 to 1.98), 1.21 (0.74 to 1.98), 1.09 (0.64 to 1.87), and 2.01 (1.15 to 3.53) ϭ (Ptrend 0.02). Hazard ratios were attenuated only slightly after adjustment for factors VIIc and VIIIc and von Willebrand factor. Conclusions—A diet including more plant food and fish and less red and processed meat is associated with a lower incidence of VTE. (Circulation. 2007;115:188-195.) Key Words: diet Ⅲ epidemiology Ⅲ fish Ⅲ folate Ⅲ fruit and vegetables Ⅲ venous thromboembolism Ⅲ vitamin B6

ittle is known about the role of dietary intake in the Clinical Perspective p 195 Ldevelopment of venous thromboembolism (VTE); thus, there are no dietary recommendations for VTE. Homocys- In Norway, the rate of postoperative thrombosis emboli teine,1 factor VIII coagulant activity (FVIIIc), and von decreased considerably during World War II (1940–1944), Willebrand factor (vWF), putative risk factors for VTE,1–3 are although the rate increased again after 1944.12 During this affected by dietary intake.4–9 Supplemental folic acid, alone same period, food rationing changed food consumption pat- 13 or in combination with vitamin B12 and vitamin B6, as well as terns. With food rationing, consumption of cholesterol, total foods rich in these vitamins, including fruit, vegetables, and fat, and calories decreased because of a reduced intake of cereal, reduce homocysteine levels.4–6 Levels of FVIIIc7 and meat, whole milk, cream, margarine, cheese, eggs, and fruit, vWF,8 coagulation factors that also relate to blood viscosity whereas intake increased for fish, cod liver oil, skimmed and vascular function, may be influenced by dietary intake.7–9 milk, whole grain bread, potatoes, and fresh vegetables, thus 3 13 Recently, plasma levels of coagulation factor VII (FVIIc) increasing intakes of ␻-3 fatty acids, vitamin B6, and folate. and fibrin fragment D-dimer10 have been suggested as risk These ecological data provide a hypothesis that changes in factors for VTE and may be responsive to dietary fat7 or dietary pattern beneficially influenced coagulation balance, vitamin B intake.11 resulting in a lower VTE rate.

Received May 23, 2006; accepted October 27, 2006. From the Division of Epidemiology and Community Health, University of Minnesota School of Public Health, Minneapolis (L.M.S., A.R.F., D.R.J.); Department of Medicine, College of Medicine, University of Vermont, Burlington (M.C.); Department of Nutrition, University of Oslo, Oslo, Norway (D.R.J.); and Department of Epidemiology, University of North Carolina at Chapel Hill (W.D.R.). Correspondence to Lyn M. Steffen, PhD, MPH, Division of Epidemiology and Community Health, University of Minnesota School of Public Health, 1300 S Second St, Suite 300, Minneapolis, MN 55454. E-mail [email protected] © 2007 American Heart Association, Inc. Circulation is available at http://www.circulationaha.org DOI: 10.1161/CIRCULATIONAHA.106.641688

Downloaded from circ.ahajournals.org 188at Mohammed Mahboob on June 10, 2007 Steffen et al Dietary Intake and Risk of VTE 189

Although folate and vitamins B6 and B12 reduce homocys- fruit) and 12 vegetable categories (green beans; broccoli; cabbage, teine levels and ␻-3 fatty acids improve vascular function, it cauliflower, or brussel sprouts; carrots; corn; spinach, collards, or is unknown whether dietary intake of these nutrients is related other greens; peas or lima beans; dark yellow or winter squash; sweet potatoes; beans or lentils; tomatoes; and potatoes, not including to risk of developing VTE. Furthermore, the relations of french fries), which were listed on the food frequency questionnaire. foods and dietary patterns with risk of developing VTE are Small amounts of vegetables included in mixed dishes were not even less well understood. We therefore examined the rela- recorded.

tions of (1) nutrients (folate, vitamins B6 and B12, and ␻-3 and saturated fatty acids); (2) foods rich in these nutrients; and (3) Case Ascertainment dietary patterns with incidence of VTE in a prospective study Study participants were followed for VTE (deep vein thrombosis [DVT] or pulmonary embolism) end points through December 31, of middle-aged black and white men and women. We 2001, via annual telephone calls and surveillance of community hypothesized a priori that foods rich in B vitamins and ␻-3 hospitals.3,14 For all hospitalizations, International Classification of fatty acids are negatively associated and intakes of saturated Diseases, Ninth Revision, Clinical Modification discharge codes fatty acid and meat are positively associated with incidence of were recorded and used to abstract medical records for possible VTE. Medical records were reviewed by 2 physicians, VTE events VTE. were classified independently, and differences were resolved through discussion. VTE required objective evidence from imaging or Methods autopsy. DVT was nearly always defined as a positive duplex ultrasound or venogram or rarely, in the earliest years, by a positive Study Design and Population Doppler ultrasound or impedance plethysmography. Pulmonary The Longitudinal Investigation of Thromboembolism Etiology embolism nearly always was defined by a ventilation-perfusion scan (LITE) is a prospective study of VTE occurrence in 2 population- with multiple segmental or subsegmental mismatched defects or a based cohorts: the Atherosclerosis Risk in Communities (ARIC) positive pulmonary angiogram or computed tomographic scan. study and the Cardiovascular Health Study.3,14 Similar methods were used in data collection for both studies; however, different instru- Statistical Analysis ments were used to assess dietary intake, and thus it was not feasible Of the 15 792 ARIC participants, we excluded from the analyses to pool the diet data from the 2 studies. For this report, we used data those with prevalent VTE at baseline (nϭ236), use of warfarin at from the ARIC cohort only. ARIC examined 15 792 45- to 64-year- baseline (nϭ73), or cancer-related incident VTE (nϭ72). We further old black and white male and female residents recruited in between excluded 26 individuals with missing dietary information and 375 1987 and 1989 from Forsyth County, North Carolina; the city of individuals with energy intake Ͻ500 and Ͻ700 kcal for women and Jackson, Miss; selected suburbs of Minneapolis, Minn; and Wash- Ͼ Ͼ 15 men, respectively, or 3500 and 4500 kcal for women and men, ington County, Maryland. respectively. These cut points approximate the lower and upper 1% distribution of energy intake. Forty-eight individuals were excluded Data Collection because they were not white or black, leaving 14 962, including 2482 The institutional review boards of the 4 participating centers ap- black and 5771 white women and 1531 black and 5178 white men. proved this study. Study participants underwent a comprehensive All analyses were conducted with the use of the statistical software baseline examination for cardiovascular disease risk factors and up to package SAS, version 8.0 (SAS Institute, Cary, NC). Follow-up time 3 triennial reexaminations.15 Height and weight were measured. was calculated as time from baseline to incident VTE, death, last Fasting blood specimens were collected, centrifuged at 4°C, and follow-up contact, or through December 31, 2001, whichever oc- frozen at Ϫ70°C until analysis in a central laboratory. FVIIc, FVIIIc, curred first. For participants with no VTE before examination 3 (year and vWF levels were measured at baseline.16 Body mass index 6 of follow-up), we averaged diet data from examination 1 and (BMI) was calculated as weight in kilograms divided by the square examination 3.21 Those with VTE by examination 3 were censored. of standing height in meters. Diabetes status (yes or no) was defined Nutrient and food intakes between examinations 1 and 3 were at baseline as fasting glucose Ն126 mg/dL,17 nonfasting glucose moderately correlated with Spearman r values ranging from 0.49 to Ն200 mg/dL, or a history of or treatment for diabetes. Information 0.56 (PϽ0.001). When well-known within-person variation in re- was obtained about vitamin supplement intake. sponse to diet questionnaires of the food frequency questionnaire Usual dietary intake was assessed with the use of the Willett type is considered, the correlations suggest considerable tracking 66-item semiquantitative food frequency questionnaire, interviewer- (lack of change over time), although they could also represent some administered at baseline and 6 years later (examination 3).18 For each change in diet over 6 years.21 From 32 food subgroups, principal food, participants were asked to report the frequency of consumption components analysis was performed, which yielded scores for 2 diet in 9 categories, ranging from never or less than once per month to patterns, which we labeled as prudent (healthy) and Western.22 The Ն6 times per day. Interviewers obtained additional information, Western diet pattern was characterized by a diet rich in red and including the brand name breakfast cereal usually consumed. Data processed meat, fast food, and high-fat dairy products and low in were analyzed for selected nutrient and food intakes. Nutrients in fish, fruit, and vegetables; the opposite characterized the prudent diet these analyses included folate, vitamins B6 and B12, fiber, and pattern. saturated and ␻-3 fatty acids. The selected nutrients, food groups, and diet pattern scores were Foods were grouped into whole grain, refined grain, fruit and categorized into quintiles of intake. Means or proportions were vegetables, dairy, fish, and red and processed meat. The whole and computed to describe baseline characteristics of participants. Cox refined grain groups were formed according to previously developed proportional hazards regression analyses were used to estimate the procedures.19,20 Food items classified as whole grain were dark bread hazard ratio for developing VTE across quintiles of intake of (1) ␻ and whole grain cold breakfast cereal, in which whole grain cold nutrients (folate, vitamins B6 and B12, and saturated and -3 fatty cereals contained at least 25% whole grain or bran by weight. Food acids); (2) food groups (whole grains, refined grains, fruit and items classified as refined grain included cold breakfast cereal with vegetables, dairy, red and processed meat, and fish); and (3) diet Ͻ25% whole grain or bran, cooked cereal (oatmeal, cream of wheat, pattern scores (prudent or healthy and Western patterns). To consider and cream of rice, which were queried in a single item, could not be whether collinearity might be a problem in the statistical models, separated), white bread, bagels, doughnuts, pastry, muffins, biscuits, Spearman correlations were calculated. Spearman correlations be- cookies, cake, brownies, pasta, and rice (brown rice and wild rice tween the food groups ranged from Ϫ0.28 (meat versus fruit/ were not separately queried). The fruit and vegetable food group vegetables) and 0.27 (fish versus fruit/vegetables) and between Ϫ comprised 6 fruit categories (fresh apples or pears; oranges; orange 0.32 (saturated fat and folate) and 0.58 (folate and vitamin B6) for or grapefruit juice; peaches, apricots, or plums; bananas; and other nutrients. We concluded that collinearity was not an issue.

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 190 Circulation January 16, 2007

TABLE 1. Mean (SE) or Prevalence of Baseline Characteristics Across Quintiles of Fruit and Vegetable (962 14؍Intake Among ARIC Study Participants, 1987–1989 (N

Quintiles of Fruit and Vegetable Intake

Characteristic 12345P Demographic factors Age, y 53.3 (0.11) 53.9 (0.11) 54.2 (0.10) 54.4 (0.10) 54.8 (0.11) Ͻ0.001 Women, % 46 52 56 59 62 Ͻ0.001 Blacks, % 32 26 25 23 28 0.17 Education, Ͼhigh school, % 37 43 45 47 48 Ͻ0.001 Current smokers, % 36 29 24 21 21 Ͻ0.001 Clinical characteristics* BMI, kg/m2 27.5 (0.10) 27.5 (0.10) 27.6 (0.09) 27.9 (0.09) 27.9 (0.10) Ͻ0.001 Overweight/obese (BMI Ն25), % 65 65 67 69 67 Ͻ0.001 Waist circumference, cm 96.8 (0.25) 96.6 (0.25) 96.6 (0.25) 97.0 (0.25) 97.4 (0.25) Ͻ0.001 Diabetes status, % 9 10 12 14 14 Ͻ0.001 FVIIc, % standard 120.1 (0.53) 117.9 (0.53) 119.0 (0.53) 119.5 (0.53) 118.9 (0.53) 0.62 FVIIIc, % standard 132.1 (0.70) 130.1 (0.69) 131.0 (0.69) 132.0 (0.69) 130.3 (0.70) 0.36 vWF, % standard 117.6 (0.86) 117.8 (0.84) 116.9 (0.84) 118.5 (0.84) 117.6 (0.85) 0.40 Daily nutrient intake† Energy intake, kcal 1318 (10) 1496 (10) 1595 (10) 1747 (10) 1982 (10) Ͻ0.001 Total fat, g 63.9 (0.23) 62.6 (0.23) 61.1 (0.22) 58.6 (0.23) 53.6 (0.24) Ͻ0.001 Saturated fat, g 23.7 (0.10) 23.2 (010) 22.4 (0.10) 21.3 (0.10) 19.3 (0.10) Ͻ0.001 ␻-3 fatty acids, g 0.18 (0.004) 0.21 (0.005) 0.25 (0.005) 0.27 (0.005) 0.35 (0.005) Ͻ0.001 Protein, g 66.6 (0.31) 69.0 (0.30) 71.6 (0.30) 72.8 (0.30) 76.5 (0.31) Ͻ0.001 Fiber, g 11.7 (0.10) 14.3 (0.09) 16.6 (0.09) 19.1 (0.09) 24.3 (0.10) Ͻ0.001 Folate, ␮g 162.1 (1.33) 199.4 (1.29) 222.7 (1.28) 252.9 (1.29) 307.2 (1.34) Ͻ0.001

Vitamin B6,mg 1.35 (0.008) 1.54 (0.008) 1.69 (0.008) 1.85 (0.008) 2.13 (0.008) Ͻ0.001

Vitamin B12, ␮g 7.07 (0.07) 7.45 (0.07) 7.67 (0.07) 7.87 (0.07) 8.26 (0.07) Ͻ0.001 Food intake (servings per day)† Whole grain 0.95 (0.02) 1.11 (0.02) 1.24 (0.02) 1.4 (0.02) 1.5 (0.02) Ͻ0.001 Refined grain 2.8 (0.03) 2.7 (0.02) 2.4 (0.02) 2.3 (0.02) 1.8 (0.03) Ͻ0.001 Fruit and vegetables 2.0 (0.03) 3.0 (0.02) 3.8 (0.02) 4.8 (0.02) 6.7 (0.03) Ͻ0.001 Fish 0.21 (0.005) 0.24 (0.005) 0.28 (0.005) 0.31 (0.005) 0.40 (0.005) Ͻ0.001 Dairy 1.6 (0.02) 1.6 (0.02) 1.6 (0.02) 1.7 (0.02) 1.7 (0.02) Ͻ0.001 Red and processed meat 1.22 (0.01) 1.15 (0.01) 1.08 (0.01) 1.00 (0.01) 0.84 (0.01) Ͻ0.001 *Adjusted for age, gender, race, and field center. †Adjusted for age, gender, race, field center, and energy intake.

Regression models were adjusted for age (continuous), race study participants was 54 years. The average BMI was 27.7, (black, white), gender (male, female), energy intake (continuous), 67.7% were overweight or obese, and 11.7% had diabetes. vitamin supplement use (any, none), BMI (continuous), diabetes Mean intake of folate was below the recommended daily (yes, no), and other dietary factors (continuous; see table footnotes ␮ for details). To determine whether associations of dietary factors intake of 400 g for US adults, but intakes of vitamins B6 and with incident VTE might be mediated by FVIIc, FVIIIc, or vWF, we B12 were above the recommended levels of 1.5 to 1.7 mg and further adjusted for these potential explanatory factors in a supple- 2.4 ␮g, respectively. Men and women consumed an average mental model. Test for linear trend across increasing quintiles of of 4.3 servings per day of fruit and vegetables, which is below dietary intake was performed with the use of the continuous variable the recommended number of servings per day. Mean intake of of nutrient, food, or diet score as the level of exposure. red and processed meat was 1.1 servings per day, and daily All authors had full access to and take full responsibility for the integrity of the data. All authors have read and agree to the fish intake was 0.25 servings. manuscript as written. Relations Between Dietary Intake and Results Incident VTE Over an average 12.5 years of follow-up, 197 incident Nutrient Intake non–cancer-related VTE events were validated among ARIC Incidence of VTE was 34% to 51% lower among individuals study participants. At baseline (Table 1), the average age of consuming Ն160 ␮g of folate per day (quintiles 2 to 5) than

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Steffen et al Dietary Intake and Risk of VTE 191

TABLE 2. Hazard Ratios of VTE Across Quintiles of Nutrient Intake Among ARIC Study Participants, (962 14؍N) 2001–1987

Quintiles of Daily Nutrient Intake

12345Ptrend Folate, ␮g/d Quintile ranges Ͻ160 160–206 206–249 249–310 Ͼ310 No. of cases 57 34 40 32 34 HR (95% CI) 1.0 0.58 (0.37–0.91) 0.66 (0.41–1.05) 0.49 (0.28–0.86) 0.49 (0.24–1.02) 0.06

Vitamin B6, mg/d Quintile ranges Ͻ1.25 1.25–1.57 1.57–1.86 1.86–2.26 Ͼ2.26 No. of cases 54 40 38 33 32 HR (95% CI) 1.0 0.66 (0.43–1.02) 0.57 (0.35–0.93) 0.44 (0.25–0.78) 0.37 (0.17–0.80) 0.007 ␻-3 Fatty acids, g/d Quintile ranges Ͻ0.1 0.1–0.16 0.16–0.24 0.24–0.39 Ͼ0.39 No. of cases 57 32 38 31 39 HR (95% CI) 1.0 0.56 (0.36–0.87) 0.64 (0.42–0.99) 0.54 (0.34–0.85) 0.70 (0.43–1.13) 0.37 Saturated fatty acids, g/d Quintile ranges Ͻ13.5 13.5–17.7 17.7–22.0 22.0–28.1 Ͼ28.1 No. of cases 38 48 30 46 35 HR (95% CI) 1.0 1.29 (0.81–2.06) 0.84 (0.46–1.51) 1.40 (0.72–2.73) 1.14 (0.43–3.00) 0.78 Adjusted for age, race, gender, field center, energy intake, vitamin supplement use, BMI, diabetes, and the other nutrients in the table. HR indicates hazard ratio.

among those consuming Ͻ160 ␮g per day (Ptrendϭ0.06) highest quintile of Western diet score had a 60% higher risk

(Table 2). There was a decreasing dose-response relation of VTE than those in the bottom quintile (Ptrendϭ0.04).

between vitamin B6 and VTE incidence (Ptrendϭ0.007), with the hazard ratio for the highest versus lowest quintile being Other Analyses Adjustment for additional standard cardiovascular risk fac- 0.37 (95% CI, 0.17 to 0.80). For ␻-3 fatty acid intake, the tors, including smoking, physical activity, and alcohol intake pattern suggests a threshold of 30% to 46% lower risk of VTE for all participants and hormone replacement therapy for in quintiles 2 to 5 compared with quintile 1, with the 95% CIs women, did not change the associations between dietary excluding 1 in quintiles 2 to 4 but including 1 in quintile 5. A intake and risk of VTE.23 Further adjustment for FVIIc, linear trend was tested but was not significant (Pϭ0.37). VTE FVIIIc, and vWF in the models in Tables 2 to 4 only slightly was not related to intakes of fatty acids (Table 2) or vitamin attenuated associations (data not shown). Results were also B (data not shown). 12 similar after we excluded from analysis VTE events with Food Intake obvious precipitants (eg, surgery, trauma, recent hospitaliza- Compared with eating Ͻ2.5 servings per day of fruit and tion, or severe immobility14) to isolate idiopathic VTE vegetables, eating Ն2.5 servings per day was associated with (nϭ111) (data not shown).

a 27% to 53% lower risk of VTE (Ptrendϭ0.03) (Table 3). Eating Ն0.1 serving of fish per day (or Ն1 serving per week Discussion for quintiles 2 to 5) was associated with 30% to 45% lower In this prospective study of black and white middle-aged Ն risk of VTE than eating Ͻ0.1 servings of fish per day, adults, consumption of 4 servings of fruit and vegetables suggestive of a threshold pattern effect but not a linear trend per day or at least 1 serving of fish per week was associated with lower incidence of VTE. In a comparison of the highest (Ptrendϭ0.30). Intakes of dairy and refined grain (data not shown) and whole grain were not related to VTE risk. quintile of intake with the lowest, red and processed meat and Individuals consuming Ͼ1.5 servings of red and processed a Western diet pattern were positively associated with inci- meat per day had 2 times higher risk of developing VTE than dent VTE. To support these food findings, nutrient intakes of vitamin B6, folate, and ␻-3 fatty acids were inversely related those consuming Ͻ0.5 servings per day (Ptrendϭ0.02). to VTE, although the shape of the relation for folate and ␻-3 Dietary Patterns fatty acids was that of a threshold pattern, whereas a dose-

With prudent diet scores above the first quintile, there was a response relation was observed for vitamin B6. Vitamin B12 nonsignificant 28% to 38% lower risk of VTE across quin- and saturated fatty acids were not related to VTE risk. HRs tiles 2 to 5, again suggestive of a threshold pattern effect but were only slightly attenuated after adjustment for FVIIc,

not a linear trend (Ptrendϭ0.12) (Table 4). Participants in the FVIIIc, and vWF.

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 192 Circulation January 16, 2007

TABLE 3. Hazard Ratios of VTE Across Quintiles of Fruit and Vegetable, Whole Grain, Fish, and Meat (962 14؍Intake Among ARIC Study Participants, 1987–2001 (N

Quintiles of Daily Food Intake

12345Ptrend Fruit and vegetables, servings/d Quintile ranges Ͻ2.5 2.5–3.5 3.5–4.5 4.5–5.8 Ͼ5.8 No. of cases 56 41 35 28 37 HR (95% CI) 1.0 0.73 (0.48–1.11) 0.57 (0.37–0.90) 0.47 (0.29–0.77) 0.59 (0.36–0.99) 0.03 Whole grain, servings/d Quintile ranges Ͻ0.4 0.4–0.8 0.8–1.25 1.25–2.0 Ͼ2.0 No. of cases 37 46 33 48 33 HR (95% CI) 1.0 1.13 (0.73–1.75) 0.87 (0.54–1.41) 1.23 (0.79–1.93) 0.89 (0.54–1.46) 0.69 Fish, servings/d Quintile ranges Ͻ0.1 0.1–0.14 0.15–0.25 0.25–0.43 Ͼ0.43 No. of cases 57 32 38 31 39 HR (95% CI) 1.0 0.58 (0.37–0.90) 0.60 (0.39–0.92) 0.55 (0.35–0.88) 0.70 (0.44–1.10) 0.30 Red and processed meat, servings/d Quintile ranges Ͻ0.5 0.5–0.75 0.75–1.0 1.0–1.5 Ͼ1.5 No. of cases 33 41 39 34 50 HR (95% CI) 1.0 1.24 (0.78–1.98) 1.21 (0.74–1.98) 1.09 (0.64–1.87) 2.01 (1.15–3.53) 0.02 Adjusted for age, race, gender, field center, energy intake, vitamin supplement use, BMI, diabetes, and the other food groups in the table. HR indicates hazard ratio.

Because we are not aware of studies of dietary intake with homocysteine as a risk factor for VTE.26,27 Surpris- and VTE risk, we draw on “Virchow’s triad”24 to explain ingly, whole grain foods, good sources of folate and

our study results. In 1856, Virchow hypothesized that vitamin B6, were not related to VTE risk. Homocysteine-

venous thrombosis was the result of an increase in blood mia may result from low levels of folic acid, vitamin B6, 24 4–6,26,28–30 coagulability, stasis, and damage to the wall of the vein. and vitamin B12. In a case-control study in which Although genetics may play a particular role in coagula- plasma nutrients were measured after DVT and compared tion, all 3 factors may also be influenced by environment, with controls, folate, and pyridoxal-5Ј-phosphate, the co-

such as dietary intake. We address the role of diet in enzyme form of vitamin B6, were significantly and in- coagulation and venous stasis, but not vessel wall injury, versely associated with DVT.27 The relation between folate because few published data are available. and DVT was attenuated when homocysteine was added to Most attention on nutrient factors and VTE has related the model; pyridoxal-5Ј-phosphate, however, remained to homocysteine as a risk factor.1,25,26 Our findings of inversely associated with risk of DVT independent of lower VTE risk with increasing dietary intake of vitamin homocysteine and folate.27 In clinical trials, levels of 4–6,30 B6, folate, and foods rich in these nutrients are consistent homocysteine among healthy adults and VTE pa-

TABLE 4. Hazard Ratios of VTE Across Quintiles of Prudent and Western Dietary Pattern Scores Among (962 14؍ARIC Study Participants, 1987–2001 (N

Quintiles of Diet Pattern Score

12345Ptrend Prudent Diet Pattern Score Quintile ranges ϽϪ0.7 Ϫ0.7– Ϫ0.3 Ϫ0.3–0.03 0.03–0.5 Ͼ0.5 No. of cases 49 37 39 35 37 HR (95% CI) 1.0 0.72 (0.47–1.11) 0.69 (0.45–1.07) 0.62 (0.39–0.97) 0.69 (0.44–1.09) 0.12 Western Diet Pattern Score Quartile ranges ϽϪ0.8 Ϫ0.8–Ϫ0.4 Ϫ0.4– Ϫ0.1 0.1–0.7 Ͼ0.7 No. of cases 40 35 36 37 49 HR (95% CI) 1.0 0.84 (0.53–1.33) 0.93 (0.59–1.48) 1.00 (0.62–1.62) 1.60 (0.97–2.66) 0.04 Adjusted for age, race, gender, field center, energy intake, vitamin supplement use, BMI, and diabetes. HR indicates hazard ratio.

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Steffen et al Dietary Intake and Risk of VTE 193

tients30 declined with vitamin supplementation of folate An important limitation to consider when our results are

alone, folate in combination with vitamins B6 and B12,or interpreted is the use of a food frequency questionnaire foods rich in these nutrients. Randomized clinical trials containing only 66 items, thus restricting the number of using supplementation with folic acid or folic acid in food categories to characterize usual dietary intake, which

combination with vitamins B6 and B12 also reduced plasma likely results in underestimated energy intake. Dietary D-dimer,31 a strong risk marker for incident VTE.10 How- intake may be misclassified by this questionnaire, contrib- ever, in a recent randomized trial, vitamin supplementation uting to measurement error in the point estimates that may

of folate with vitamins B6 and B12 was not effective in potentially result in large biases either toward or away reducing the risk of recurrent VTE.32 Unlike vitamin from the null.41 Furthermore, the completeness of reported supplements as used in the trial,32 consuming a dietary food intake, and therefore energy intake, may differ according to level of BMI; however, the statistical models pattern of foods rich in folate and vitamin B6, such as fish, chicken, whole grains, potatoes, carrots, legumes, and were adjusted for BMI. An additional limitation to our bananas, also provides other nutrients and food compounds study was that homocysteine was not measured on the that may promote vascular health and prevent the occur- entire cohort. Finally, although our data support reduced ␻ rence of VTE events. There likely is a synergistic effect risk with greater intakes of folate, vitamin B6, -3 fatty among foods and their nutritive and nonnutritive compo- acids, fruit and vegetables, and fish, the reader should nents, in which the sum of the diet’s constituent parts has interpret the shape of the relationships with caution be- greater health effects than the individual effects of single cause our study lacked precision to establish whether the foods or nutrients.33 relations have threshold or dose-response shapes. Elevated FVIIc, FVIIIc, and vWF, which are related to A major strength of this investigation is that it is the first prospective study on diet and VTE. It had a large number increased risk of VTE incidence,3,34 and elevated FVIIIc, of white and black men and women enrolled and 12.5 years which is related to recurrence,34 may be influenced by of follow-up. Another strength is the second diet interview dietary intake.7,8,35–40 In a randomized clinical trial com- that updated dietary information in the regression models, paring the effectiveness of a Mediterranean-type diet, a because food choices and frequency may change over diet rich in olive oil, fish, and fruit and vegetables, with a time.21 The use of repeated measurements of diet may high-saturated-fat diet in lowering hemostatic factors, reduce measurement error due to intraindividual variation levels of FVIIc and FVIIIc were reduced in men consum- in dietary intake. We adjusted for several VTE risk factors, ing the Mediterranean-type diet but not in those consuming but it is possible that there was residual confounding by the high-saturated-fat diet.7 Plasma vWF was significantly unaccounted for factors that are associated with both diet reduced in men with atherosclerosis who consumed a and VTE. lipid-lowering diet (low in total, saturated, and monoun- In conclusion, before the present study, there has been saturated fat but high in polyunsaturated fat) compared no study of food in relation to prevalent, incident, or with men who had been following their normal diet for 3 recurrent VTE. One randomized clinical trial, currently 8 9 years. Shahar et al reported an inverse cross-sectional reported only in abstract form,32 found no association ␻ relation of -3 fatty acids and fish intake with blood levels between B vitamin supplementation and recurrent VTE. of FVIIIc and vWF in ARIC. Links between dietary fat and Nevertheless, a causal relationship between diet and VTE coagulation factors tend to support our study findings that is suggested by 2 separate ecological studies from Norway, ␻ greater intakes of -3 fatty acids and fish are inversely one showing a lower rate of VTE occurrence during World related to incident VTE, and meat intake is positively War II,12 the other showing that dietary patterns changed related to VTE risk. However, because adjustment for with food rationing during this same time period.13 FVIIc, FVIIIc, and vWF did not affect the association Our hypothesis of a relationship between diet and VTE is between diet and VTE, we could not confirm their role as also consistent with results from several randomized intermediaries between dietary intake and VTE. clinical trials and feeding studies demonstrating lower In addition to the B vitamins and ␻-3 fatty acids found levels of putative risk factors for VTE with dietary in foods, it is feasible that the biologically active nutrients intervention.5–8,35–40 Our findings provide evidence that a and food compounds in fruit, vegetables, and fish act diet including abundant plant food and fish and little meat singularly or synergistically to enhance some aspects of is associated with lower risk of incident VTE. Potential health.33 When we examined the relations of 2 diet prevention strategies for VTE may include eating a healthy patterns, representing combinations of foods, with risk of diet consistent with the US Department of Agriculture incident VTE, we found that a high score (quintile 5) on Dietary Guidelines for Americans.42 the Western diet pattern, characterized by high intake of red and processed meat, fast food, and refined grain and Acknowledgment low intakes of fish, fruit, and vegetables, had a 60% greater The authors thank the staff and participants of the ARIC study for risk of incident VTE than a lower score (quintile 1). their important contributions. Similarly, prudent diet pattern scores in quintiles 2 to 5 Sources of Funding were associated with lower risk of VTE, suggestive of a The ARIC study is performed as a collaborative study supported threshold effect, even though the linear trend was not by National Heart, Lung, and Blood Institute contracts N01-HC- statistically significant. 55015, N01-HC-55016, N01-HC-55018, N01-HC-55019, N01-

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 194 Circulation January 16, 2007

HC-55020, N01-HC-55021, and N01-HC-55022. The LITE study women: the Iowa Women’s Health Study. Am J Clin Nutr. 1998;68: was funded by grant HL59367 from the National Heart, Lung, and 248–257. Blood Institute. 20. Steffen LM, Jacobs DR, Stevens J, Shahar E, Carithers T, Folsom AR. Associations of whole grain, refined grain, and fruit and vegetable consumption with all-cause mortality, incident coronary heart disease Disclosures and ischemic stroke: the ARIC study. Am J Clin Nutr. 2003;78: None. 383–390. 21. Hu FB, Stampfer MJ, Rimm E, Ascherio A, Rosner BA, Spiegelman References D, Willett WC. Dietary fat and coronary heart disease: a comparison 1. den Heijer M, Lewington S, Clarke R. Homocysteine, MTHFR and risk of approaches for adjusting for total energy intake and modeling of venous thrombosis: a meta-analysis of published epidemiological repeated dietary measurements. Am J Epidemiol. 1999;149:531–540. studies. J Thromb Haemost. 2005;3:292–299. 22. Dunteman GH. Principal Components Analysis. Newbury Park, Calif: 2. Martinelli I. von Willebrand factor and factor VIII as risk factors for Sage Publications; 1989. arterial and venous thrombosis. Sem Hematol. 2005;42:49–55. 23. Tsai AW, Cushman M, Rosamond WD, Heckbert SR, Polak JF, 3. Tsai AW, Cushman M, Rosamond WD, Heckbert SR, Tracy RP, Folsom AR. Cardiovascular risk factors and venous thromboembolism Aleksic N. Folsom AR. Coagulation factors, inflammation markers, incidence: the Longitudinal Investigation of Thromboembolism and venous thromboembolism: the Longitudinal Investigation of Etiology. Arch Intern Med. 2002;162:1182–1189. Thromboembolism Etiology (LITE). Am J Med. 2002;113:636–642. 24. Virchow R. Neuer fall von todlichen: emboli der lungenarterie. Arch 4. Homocysteine Lowering Trialists’ Collaboration. Lowering blood Pathol Anat. 1856;10:225–228. homocysteine with folic acid based supplements: meta-analysis of 25. Tsai AW, Cushman M, Tsai MY, Heckbert SR, Rosamond WD, randomized trials. BMJ. 1998;316:894–898. Aleksic N, Yanez ND, Psaty BM, Folsom AR. Serum homocysteine, 5. Tucker KL, Olson B, Bakun P, Dallal GE, Selhub J, Rosenberg IH. thermolabile variant of methylene tetrahydrofolate reductase (MTHFR), and venous thromboembolism: Longitudinal Investigation Breakfast cereal fortified with folic acid, vitamin B-6, and vitamin of Thromboembolism Etiology (LITE). Am J Hematol. 2003;72: B-12 increases vitamin concentrations and reduces homocysteine con- 192–200. centrations: a randomized trial. Am J Clin Nutr. 2004;79:805–811. 26. Gemmati D, Previati M, Serino ML, Moratelli S, Guerra S, Capitani 6. Jacques PF, Selhub J, Bostom AG, Wilson PW, Rosenberg IH. The S, Forini E, Ballerini G, Scapoli GL. Low folate levels and thermo- effect of folic acid fortification on plasma folate and total homo- labile methylenetetrahydrofolate reductase as primary determinant of cysteine concentrations. N Engl J Med. 1999;340:1449–1454. mild hyperhomocystinemia in normal and thromboembolic subjects. 7. Mezzano D, Leighton F, Martinez C, Marshall G, Cuevas A, Castillo Arterioscler Thromb Vasc Biol. 1999;19:1761–1767. O, Panes O, Munoz B, Perez DD, Mizon C, Rozowski J, San Martin 27. Cattaneo M, Lombardi R, Lecchi A, Bucciarelli P, Mannucci PM. A, Pereira J. Complementary effects of Mediterranean diet and Low plasma levels of vitamin B6 are independently associated with a moderate red wine intake on haemostatic cardiovascular risk factors. heightened risk of deep-vein thrombosis. Circulation. 2001;104: Eur J Clin Nutr. 2001;55:444–451. 2442–2446. 8. Blann AD, Jackson P, Bath PM, Watts GF. von Willebrand factor, a 28. Alfthan G, Laurinen MS, Valsta LM, Pastinen T, Aro A. Folate intake, possible indicator of endothelial cell damage, decreases during plasma folate and homocysteine status in a random Finnish popu- long-term compliance with a lipid-lowering diet. J Intern Med. 1995; lation. Eur J Clin Nutr. 2003;57:81–88. 237:557–561. 29. Quinlivan EP, Gregory JF III. Effect of food fortification on folic acid 9. Shahar E, Folsom AR, Wu KK, Dennis BH, Shimakawa T, Conlan intake in the United States. Am J Clin Nutr. 2003;77:221–225. MG, Davis CE, Williams OD. Associations of fish intake and dietary 30. den Heijer M, Brouwer IA, Bos GM, Blom HJ, van der Put NM, n-3 polyunsaturated fatty acids with a hypocoagulable profile: the Spaans AP, Rosendaal FR, Thomas CM, Haak HL, Wijermans PW, Atherosclerosis Risk in Communities (ARIC) Study. Arterioscler Gerrits WB. Vitamin supplementation reduces blood homocysteine Thromb. 1993;13:1205–1212. levels: a controlled trial in patients with venous thrombosis and 10. Cushman M, Folsom AR, Wang L, Aleksic N, Rosamond WD, Tracy healthy volunteers. Arterioscler Thromb Vasc Biol. 1998;18:356–361. RP, Heckbert SR. Fibrin fragment D-dimer and the risk of future 31. Klerk M, Verhoef P, Verbruggen B, Schouten EG, Blom HJ, Bos GM, venous thrombosis. Blood. 2003;101:1243–1248. den Heijer M. Effect of homocysteine reduction by B-vitamin sup- 11. Mangoni AA, Arya R, Ford E, Asonganyi B, Sherwood RA, Ouldred plementation on markers of clotting activation. Thromb Haemost. E, Swift CG, Jackson SH. Effects of folic acid supplementation on 2002;88:230–235. inflammatory and thrombogenic markers in chronic smokers: a ran- 32. Bos GMJ, Heijer den M, Willems H, Blom H, Gerrits W, Cattaneo M, domised controlled trial. Thromb Res. 2003;110:13–17. Eichinger S, Rosendaal F. Homocysteine lowering by B vitamins and 12. Jensen RA. Postoperative thrombosis-emboli. Acta Chir Scand. 1952; the secondary prevention of deep-vein thrombosis and pulmonary 103:263–278. embolism: a first randomised, placebo-controlled, double-blind trial. 13. Strom A. Examination into the diet of Norwegian families during the Blood. 2004;104:A489. war-years 1942–45. Acta Med Scand Suppl. 1948;214:1–47. 33. Jacobs DR, Steffen LM. Nutrients, foods, and dietary patterns as 14. Cushman M, Tsai AW, White RH, Heckbert SR, Rosamond WD, exposures in research: issues and challenges. Am J Clin Nutr. 2003; Enright P, Folsom AR. Deep vein thrombosis and pulmonary 78(suppl):508S–513S. embolism in two cohorts: the Longitudinal Investigation of Throm- 34. Kraaijenhagen RA, in’t Anker PS, Koopman MM, Reitsma PH, Prins boembolism Etiology. Am J Med. 2004;117:19–25. MH, van den Ende A, Buller HR. High plasma concentration of factor 15. ARIC Investigators. The Atherosclerosis Risk in Communities VIIIc is a major risk for venous thromboembolism. Thromb Haemost. (ARIC) study: design and objectives. Am J Epidemiol. 1989;129: 2000;83:5–9. 687–702. 35. Sandstrom B, Marckmann P, Jespersen J. Fat consumption and factor 16. Folsom AR, Wu KK, Rosamond WD, Sharrett AR, Chambless LE. VII coagulant activity. In: Somogyi JC, Biró GY, Hötzel D, eds. Prospective study of hemostatic factors and incidence of coronary Nutrition and Cardiovascular Risks: 29th Symposium of the Group of heart disease: the Atherosclerosis Risk in Communities (ARIC) study. European Nutritionists, Balatonföldvár, April 24-26, 1991 (Forum of Circulation. 1997;96:1102–1108. Nutrition/Bibliotheca Nutritio Et Dieta). Basel, Switzerland: Karger; 17. Alberti K, Zimmet P. Definition, diagnosis and classification of 1992:93–101. diabetes mellitus and its complications, part 1: diagnosis and classi- 36. Mitropoulos KA, Miller GJ, Martin JC, Reeves BEA, Cooper J. fication of diabetes mellitus provisional report of a WHO consul- Dietary fat induces changes in factor VII coagulant activity through tation. Diabet Med. 1998;15:539–553. effects on plasma free stearic acid concentration. Arterioscler 18. Willett WC, Sampson L, Stampfer MJ, Rosner B, Bain C, Witschi J, Thromb. 1994;14:214–222. Hennekens CH, Speizer FE. Reproducibility and validity of a semi- 37. Miller GJ. Dietary fatty acids and the haemostatic system. Athero- quantitative food frequency questionnaire. Am J Epidemiol. 1985;122: sclerosis. 2005;179:213–227. 51–65. 38. Marckmann P, Sandstrom B, Jespersen J. Effects of total fat content 19. Jacobs DR, Meyer KA, Kushi LH, Folsom AR. Whole-grain intake and fatty acid composition in diet on factor VII coagulant activity and may reduce the risk of ischemic heart disease death in postmenopausal blood lipids. Atherosclerosis. 1990;80:227–233.

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Steffen et al Dietary Intake and Risk of VTE 195

39. Miller GJ, Martin JC, Mitropoulos KA, Reeves BE, Thompson RL, 41. Kipnis V, Subar AF, Midthune D, Freedman LS, Ballard-Barbash R, Meade TW, Cooper JA, Cruickshank JK. Plasma factor VII is acti- Troiano RP, Bingham S, Schoeller DA, Schatzkin A, Carroll RJ. vated by postprandial triglyceridaemia, irrespective of dietary fat Structure of dietary measurement error: results of the OPEN composition. Atherosclerosis. 1991;86:163–171. biomarker study. Am J Epidemiol. 2003;158:14–21. 40. Mennen LI, Witteman JC, den Breeijen JH, Schouten EG, de Jong PT, 42. US Department of Agriculture, Center for Nutrition Policy and Pro- Hofman A, Grobbee DE. The association of dietary fat and fiber with motion. 2005 Dietary Guidelines for Americans. 6th ed. Available at: coagulation factor VII in the elderly: the Rotterdam Study. Am J Clin http://www.usda.gov/cnpp/DG2005/index.html. Accessed May 16, Nutr. 1997;65:732–736. 2006.

CLINICAL PERSPECTIVE Little is known about the role of dietary intake on the development of deep vein thrombosis or pulmonary embolus (venous thromboembolism [VTE]). Suspicion of a role is raised because elevated homocysteine, factor VIII, and von Willebrand factor are risk factors for VTE and are influenced by dietary intake. In a prospective study of almost 15 000 middle-aged adults, we observed a 41% lower risk of incident VTE among those eating Ͼ5.5 servings of fruit and vegetables per day and a 30% to 45% lower risk with 1 or more servings per week of fish. Adults who ate Ͼ1.5 servings per day of red and processed meat had twice the risk of developing VTE of those who ate Ͻ0.5 serving per day. The dietary pattern associated with lower risk of VTE in this study is similar to that suggested for reduced arterial disease by the American Heart Association 2006 Diet Recommendations. Such a diet may reduce the risk of VTE.

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Health Services and Outcomes Research

Regional Differences in Process of Care and Outcomes for Older Acute Myocardial Infarction Patients in the United States and Ontario, Canada

Dennis T. Ko, MD, MSc; Harlan M. Krumholz, MD, SM; Yongfei Wang, MS; JoAnne M. Foody, MD; Fredrick A. Masoudi, MD, MSPH; Edward P. Havranek, MD; John J. You, MD; David A. Alter, MD, PhD; Therese A. Stukel, PhD; Alice M. Newman, MS; Jack V. Tu, MD, PhD

Background—Previous comparisons of acute myocardial infarction (AMI) treatment between the United States and Canada are limited because they compared selected patients from randomized trials, used administrative data that lacked clinical detail, or did not consider regional differences in AMI treatment. Methods and Results—We compared medication use, invasive cardiac procedure use, and 30-day risk-standardized mortality rates of 38 886 fee-for-service Medicare beneficiaries hospitalized with AMI in the United States and 5634 similarly aged patients in Ontario, Canada, from 1998 and 2001. Baseline characteristics and illness severity across the US regions and Ontario were not substantially different. Cardiac catheterization use in AMI patients was significantly higher in the United States compared with Ontario (38.7% versus 16.8%, PϽ0.001), but significant regional variations existed, in which the northeastern United States had significantly lower utilization rates (25.6%) compared with other US regions. ␤-Blocker use among ideal candidates was highest in the northeastern United States (77.6% versus 69.7% in the United States as a whole, PϽ0.001) and angiotensin-converting enzyme inhibitor use was highest in Ontario (69.1% versus 58.2% in the United States, PϽ0.001). Risk-standardized mortality rates at 30 days were not substantially different across the regions. Conclusions—Previous studies have suggested a clear divergence in invasive cardiac therapy for AMI patients between the United States and Canada on the basis of health care financing and structural differences. Our findings of similar treatment patterns in the northeastern United States and Ontario suggest that regional practices may have a greater impact on treatment patterns than the respective health care delivery systems. (Circulation. 2007;115:196-203.) Key Words: outcomes Ⅲ myocardial infarction Ⅲ heart catheterization

he market-oriented health care system of the United Editorial p 158 States, with its limited governmental control, is sharply T Clinical Perspective p 203 different from Canada’s single-payer system with govern- mental reimbursement for most health care services.1–3 Na- term mortality rates were not substantially different, which tional comparative studies have therefore generated great suggests that higher rates of cardiac invasive therapy after interest because they are natural experiments that evaluate AMI may not improve outcomes.4–8 However, many United different health care systems that cannot easily be studied by States–Canada comparison studies have used data from ran- randomized trials. Previous studies evaluating acute myocar- domized trials,5–8 and therefore may not accurately reflect dial infarction (AMI) treatment between the United States how AMI patients are typically treated in clinical practice.9–12 and Canada, for example, have demonstrated substantially Others have used population-based administrative data, but higher rates of coronary invasive procedures in the United have lacked clinical details on processes of care.4 Finally, all States.4–8 These studies found that short- and intermediate- of these studies examined AMI patients hospitalized in the

Received August 23, 2006; accepted October 19, 2006. From the Division of Cardiology, Schulich Heart Centre, Sunnybrook Health Sciences Centre (D.T.K., D.A.A., J.V.T.), Institute for Clinical Evaluative Sciences (D.T.K., J.J.Y., D.A.A., T.A.S., A.M.N., J.V.T.), Toronto, Ontario, Canada; Section of Cardiovascular Medicine, Department of Medicine, Yale University School of Medicine (H.M.K., Y.W.), Section of Health Policy and Administration, Department of Epidemiology and Public Health (H.M.K.), Robert Wood Johnson Clinical Scholars Program, Yale University School of Medicine (H.M.K.), and the Center for Outcomes Research and Evaluation, Yale–New Haven Hospital (H.M.K.), New Haven, Conn; Yale University School of Medicine, Section of Cardiovascular Medicine, West Haven Veteran’s Administration Medical Center, West Haven, CT Qualidigm, Middletown, Conn (J.M.F.); Department of Medicine, Denver Health Medical Center (F.A.M., E.P.H.), Department of Medicine, University of Colorado Health Sciences Center (F.A.M., E.P.H.), Denver, Col; Center for the Evaluative Clinical Sciences, Dartmouth Medical School, Hanover, NH (T.A.S.); and Division of General Internal Medicine, Sunnybrook Health Sciences Centre, University of Toronto, Toronto, Ontario, Canada (J.V.T.). Guest Editor for this article was Gregg C. Fonarow, MD. Correspondence to Dr Jack V. Tu, Institute for Clinical Evaluative Sciences, G-106, 2075 Bayview Ave, Toronto, Ontario, Canada M4N, 3M5. E-mail [email protected] © 2007 American Heart Association, Inc. Circulation is available at http://www.circulationaha.org DOI: 10.1161/CIRCULATIONAHA.106.657601

196

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Ko et al Regional Differences in US and Canadian AMI Care 197

early 1990s, which predates many recent advances in medical or (3) 2 of the following: presence of chest pain, a 2-fold elevation and interventional therapies.4–7 As a result, differences in of creatine kinase, or ST-segment elevation on ECG. processes of care and outcomes for AMI patients in the United States and Canada remain incompletely understood. Census Regions Implicit in these cross-country comparisons is the assump- We divided the US cohort into 4 census regions (northeastern, southern, midwestern, and western United States) based on the state tion that the observed differences in processes of care are in which the patients were hospitalized.22 These regions were chosen driven by organizational differences between the 2 health a priori such that the sample sizes of AMI patients in each region care systems. None of these national comparison studies have would be comparable to the Ontario cohort. considered whether local regional variations in AMI treat- ment patterns may play a role, however. Indeed, evidence has Process of Care and Mortality demonstrated tremendous regional variation in the utilization Patients were identified as eligible for discharge medications if they had survived to hospital discharge and were not transferred out to of medical therapy and cardiac invasive procedures within the another acute care hospital. We identified ideal candidates for 13–17 United States. Therefore, it is possible that the discrep- aspirin, ␤-blockers, and angiotensin-converting enzyme inhibitors at ancy in cardiac procedure rates between the United States and discharge by exclusion of eligible patients who had contraindications Canada may reflect regional differences in practice patterns to therapy. We harmonized the definition of “ideal” so that data rather than national differences in the respective healthcare elements common to both databases could be used in an identical 19,23 delivery systems. fashion. We assessed mortality after hospital admission by linking the Our first objective was to compare contemporary AMI clinical data sets to the Medicare enrollment database and billing treatment patterns and outcomes between the United States records in the United States, and the Ontario Registered Persons and Canada with clinically abstracted data from cohorts that Database in Canada, which contains information on the vital status of are representative of routine clinical practice. Furthermore, all Ontario residents.24,25 Risk-standardized mortality rates at 30 we aimed to explore whether differences in AMI treatment days were chosen as our primary mortality outcome and risk- standardized mortality rates at 1 year and 3 years were chosen as between the United States and Canada were consistent across secondary outcomes. each US region. Statistical Analysis Methods Because of national privacy laws that restrict the combination of National Heart Care Project individual patient information, analyses of US and Canadian data were performed separately. We compared demographics, clinical The National Heart Care Project is an initiative of the Centers for characteristics, and therapies between the entire US and Ontario Medicare and Medicaid Services designed to improve the quality of cohorts. We also compared these characteristics across geographic care for Medicare beneficiaries hospitalized with AMI in the United regions with ␹2 tests for categorical variables and F tests with 4 States.18 A random sample of Medicare beneficiaries hospitalized degrees of freedom for continuous variables. We considered proba- from March 1998 to April 1999 and from March 2000 to April 2001 bility values of Ͻ0.001 as statistically significant in order to reduce with a principal discharge diagnosis of AMI (International Classi- type I error due to multiple comparisons performed in the present fication of Diseases, 9th Revision, Clinical Modification, code 410) study. We acknowledge that because of the large number of patients was identified.19 in our cohorts, it is possible that small differences may be detected as statistically significant, and we therefore carefully considered Enhanced Feedback for Effective Cardiac clinically important results. We also examined the relationship Treatment Project between regional cardiac catheterization rates and availability of The Enhanced Feedback for Effective Cardiac Treatment (EFFECT) cardiac facilities by linear regression. An R2 statistic was calculated project is an ongoing initiative to improve the care of cardiovascular to assess the strength of association and Studentized residuals were patients in Ontario, Canada.20,21 Patients hospitalized with a most used to assess potential outliers. responsible diagnosis of AMI from April 1, 1999, to March 31, 2001, To account for differences in baseline risk in comparisons of were identified randomly from each of the 103 participating hospitals mortality rates, we first calculated mortality risk scores for each with the Canadian Institute for Health Information hospital discharge national cohort from the in-hospital prediction risk score from the database (International Classification of Diseases, 9th Revision, Global Registry of Acute Coronary Events (GRACE) study.26 We Clinical Modification, code 410). A “most responsible diagnosis” is then calculated direct standardized mortality rates of patients hospi- the condition most responsible for the length of a patient’s hospital talized in the United States (or different regions in the United States) stay. This term has been demonstrated to be essentially equivalent from the risk score distribution in the Canadian cohort as the for AMI patients to “principal diagnosis,” commonly used in the standard. This method allowed estimation of mortality rates that United States to describe the discharge diagnosis responsible for the would be expected in the US cohort if it were to have the same risk hospital admission.4 distribution as the Canadian cohort, which thus enabled us to adjust for differences in baseline risk without combining our data sets. Study Sample Statistical analyses were conducted with SAS version 9.1 software Because the data sets in the United States and Ontario were (SAS Institute Inc, Cary, NC), and direct standardized mortality constructed independently, we restricted both AMI cohorts in an calculation was performed with STATA statistical software (Version identical fashion to ensure comparability. We excluded patients who 8.0; STATA Corp, College Station, TX). were Ͻ65 years or Ͼ105 years of age, who had invalid social The authors had full access to the data and take full responsibility security numbers (or Ontario health card numbers), who were for the integrity of the data. All authors have read and agree to the transferred into a care center from another acute care institution, who manuscript as written. had been hospitalized with an AMI in the previous year, or had left bundle-branch block on ECG. We only included patients who had a confirmed AMI defined as Results either (1) creatine kinase–MB fraction Ͼ0.05, (2) elevation of The US study cohort included 38 886 patients and the troponin I or T above the upper limit of normal at a given hospital, Canadian cohort included 5634 patients who were hospital-

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 198 Circulation January 16, 2007

TABLE 1. Definition of the AMI Study Sample ischemia noninvasively after AMI and had the highest utili- zation rates of stress tests during hospitalization (Table 3). Description United States Canada Table 4 shows the demographic and admission character- Initial sample 71 120 10 634 istics of AMI patients who received cardiac catheterization Exclusions during hospitalization. AMI patients hospitalized in different Age Ͻ65 years or Ͼ105 years 8203 4514 regions had similar age, clinical characteristics, and mean Hospitalized with an AMI in the previous year 5724 * GRACE risk score. In addition, patients who received cardiac Transferred from other institutions 11 878 * catheterization in the northeastern United States and Ontario, Repeat hospitalizations during study period 3653 * areas of lower utilization of cardiac catheterization, had Unable to link with mortality database 522 0 similar demographic and admission characteristics compared with other regions with higher utilization rates. Unable to link with administrative database 3012 3 There was a strong association between the use of cardiac Left bundle-branch block 4419 483 catheterization and the availability of cardiac-invasive facil- In Puerto Rico or Virgin Islands 1200 NA ities (R2ϭ0.81) (Figure 1). Observed rates of cardiac cathe- No. of AMI patients 38 886 5634 terization were below the best-fit line of expected rates in the Some AMI patients satisfied multiple exclusion criteria. NA indicates not northeastern United States and higher than expected in the applicable. western United States. None of the regions were considered *Patients had been excluded prior to clinical chart abstraction and therefore outliers as all regions had Studentized residuals within the the exact number was unknown. expected range of Ϫ2to2.

ized in Ontario after inclusion and exclusion criteria were Mortality Outcomes applied (Table 1). The area under the receiver-operating characteristic curve of the GRACE mortality model was 0.76 for the US cohort Clinical, Physician, and Hospital Characteristics and 0.80 for the Canadian cohort, which indicated good Table 2 provides the clinical, physician, and hospital charac- discriminative ability of the GRACE risk score within each teristics of AMI patients treated in each region. Mean age of cohort. Risk-standardized mortality rates at 30 days were hospitalized AMI patients ranged from 76.9 years in Ontario not significantly different between the United States to 78.7 years in the northeastern United States. Although (17.3%; 95% CI, 17.0% to 17.7%) and Ontario (16.5%; many admission characteristics of AMI patients were not 95% CI, 15.6% to 17.4%; Pϭ0.1). Furthermore, 30-day substantially different across all regions, Ontario patients had standardized rates were not substantially different when the lowest rates of hypertension, diabetes, prior myocardial compared across all geographic regions: northeastern infarction, and prior coronary interventions (Table 2). Mean United States (15.3%; 95% CI, 14.9% to 16.5%), southern GRACE risk scores of AMI patients also did not differ United States (19.0%; 95% CI, 18.4% to 19.6%), midwest- significantly across the US regions and Ontario (Table 2). ern United States (17.1%; 95% CI, 16.4% to 17.8%), AMI patients in the northeastern United States and Ontario western United States (16.4%; 95% CI, 15.7% to 17.1%), ϭ were the most likely to have been admitted to hospitals and Ontario (16.5%; 95% CI, 15.6% to 17.4%; P 0.06 without cardiac-invasive facilities compared with other US across all 5 regions). At 1 year after AMI hospitalization, regions (Table 2). risk-standardized mortality rates were significantly lower in Ontario (27.7%; 95% CI, 26.7% to 28.8%) compared Regional Variations in Medical Therapy and with the United States (31.9%; 95% CI, 31.5% to 32.3%; Ͻ Cardiac Procedure Use P 0.001). Similarly, risk-standardized mortality rates Among ideal candidates, ␤-blocker use was highest in the were also significantly lower at 3 years after AMI hospi- northeastern United States, angiotensin-converting enzyme talization (40.3%; 95% CI, 39.1% to 41.5% in Ontario versus 45.9%; 95% CI, 45.5% to 46.4% in the United inhibitor use was highest in Ontario, and early reperfusion States; PϽ0.001). therapy use was highest in the western United States (Table 3). Cardiac catheterization rates were significantly higher in Discussion the United States compared with Ontario (Table 3). We To the best of our knowledge, the present study is the first observed substantial regional differences in cardiac catheter- United States–Canada comparison study that uses clinical ization use in the United States, however. When cardiac data from population-representative samples to compare AMI catheterization rates in each US region were compared with treatment patterns and outcomes. Although we observed rates in Ontario, the rates in the northeastern United States significantly higher overall utilization rates of cardiac inva- were more similar to Ontario than other regions of the United sive procedures in the United States compared with Canada, States. Similar utilization patterns of percutaneous coronary these differences varied substantially according to geographic intervention and coronary artery bypass surgery were ob- regions in the United States. In fact, AMI treatment patterns served, with the northeastern United States being more were quite similar in the northeastern United States and similar to Ontario than other regions of the United States Ontario, as both regions had relatively low rates of cardiac- (Table 3). On the other hand, physicians in the northeastern invasive procedures but among the highest rates of medica- United States and Ontario were more likely to evaluate tion use compared with other US regions. Less intensive use

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Ko et al Regional Differences in US and Canadian AMI Care 199

TABLE 2. Characteristics of AMI patients in the United States and Ontario

Northeastern Southern Midwestern Western Ontario, United States United States United States United States United States Canada P Characteristics (Nϭ38 886) (Nϭ6655) (Nϭ13 187) (Nϭ10 213) (Nϭ8831) (Nϭ5634) Values* Demographics Age, y, mean (SD) 77.9 (7.4) 78.7 (7.4) 77.5 (7.5) 78.3 (7.5) 77.4 (7.3) 76.9 (7.1) Ͻ0.001 65 to 74, n (%) 14 380 (37.0) 2154 (32.4) 5184 (39.3) 3575 (35.0) 3467 (39.3) 2322 (41.2) 75 to 84, n (%) 16 400 (42.2) 2928 (44.0) 5442 (41.3) 4358 (42.7) 3672 (41.6) 2394 (42.5) Ն85, n (%) 8106 (20.9) 1573 (23.6) 2561 (19.4) 2280 (22.3) 1692 (19.2) 918 (16.3) Female, n (%) 19 276 (49.6) 3483 (52.3) 6669 (50.6) 5108 (50.0) 4016 (45.5) 2565 (45.5) Ͻ0.001 Admission characteristics SBP, mean (SD) 143.8 (32.9) 144.2 (33.1) 144.3 (33.5) 143.7 (32.6) 143.0 (32.3) 147.3 (33.5) Ͻ0.001 DBP, mean (SD) 77.5 (20.0) 77.2 (20.3) 77.4 (19.9) 77.6 (19.9) 77.7 (20.1) 80.6 (18.7) Ͻ0.001 Heart rate, mean (SD) 87.9 (24.8) 88.7 (24.7) 88.8 (24.9) 87.7 (24.6) 86.2 (24.7) 85.5 (25.0) Ͻ0.001 Mean creatinine,† mean (SD) 1.4 (1.2) 1.4 (2.0) 1.4 (1.0) 1.4 (0.9) 1.4 (1.0) 1.3 (0.8) Ͻ0.001 Cardiogenic shock, n (%) 533 (1.4) 112 (1.7) 157 (1.2) 97 (1.0) 167 (1.9) 133 (2.4) Ͻ0.001 Cardiac arrest, n (%) 1180 (3.0) 184 (2.8) 398 (3.0) 267 (2.6) 331 (3.7) 125 (2.2) Ͻ0.001 GRACE score, mean (SD) 165 (37) 168 (37) 164 (37) 165 (36) 165 (37) 165 (35) 0.17 Cardiac risk factors and comorbidities, n (%) Hypertension 26 840 (69.0) 4701 (70.6) 9392 (71.2) 6939 (67.9) 5808 (65.8) 2858 (50.7) Ͻ0.001 Diabetes 11 933 (30.7) 2152 (32.3) 4246 (32.2) 3067 (30.0) 2468 (27.9) 1499 (26.6) Ͻ0.001 Current smoker 6135 (15.8) 948 (14.2) 2265 (17.2) 1556 (15.2) 1366 (15.5) 1095 (19.4) Ͻ0.001 Prior myocardial infarction 11 894 (30.6) 2171 (32.6) 3979 (30.2) 3182 (31.2) 2562 (29.0) 1471 (26.1) Ͻ0.001 Prior CABG 6045 (15.5) 920 (13.8) 2118 (16.1) 1675 (16.4) 1332 (15.1) 410 (7.3) Ͻ0.001 Prior PCI 4375 (11.3) 622 (9.3) 1470 (11.2) 1236 (12.1) 1047 (11.9) 164 (2.9) Ͻ0.001 Prior stroke or TIA 6939 (17.8) 1116 (16.8) 2569 (19.5) 1814 (17.8) 1440 (16.3) 794 (14.1) Ͻ0.001 Hospital characteristics Ͻ0.001 Cardiac surgery suite, n (%) 15 657 (40.3) 1704 (25.6) 5711 (43.3) 4842 (47.4) 3400 (38.5) 467 (8.3) Cardiac catheterization lab, n (%) 7761 (20.0) 1600 (24.0) 3075 (23.3) 1549 (15.2) 1537 (17.4) 464 (8.2) No invasive facilities, n (%) 10 372 (26.7) 2605 (39.1) 2876 (21.8) 2883 (28.2) 2008 (22.7) 4703 (83.5) Unknown, n (%) 5096 (13.1) 746 (11.2) 1525 (11.6) 939 (9.2) 1886 (21.4) 0 (0.0) Nonteaching hospital, n (%) 24 923 (64.1) 4006 (60.2) 8926 (67.7) 5663 (55.5) 6328 (71.7) 4768 (84.6) Ͻ0.001 Hospital beds, mean (SD) 286.8 (215.7) 276.6 (221.4) 321.7 (235.1) 302.2 (225.3) 223.6 (142.7) 176.1 (136.5) Ͻ0.001 Cardiologist, n (%) 10 425 (26.8) 1581 (23.8) 3546 (26.9) 2358 (23.1) 2940 (33.3) 1910 (33.9) Ͻ0.001 LOS, d, mean (SD) 6.7 (6.4) 7.1 (6.2) 6.8 (5.4) 6.7 (5.7) 6.2 (8.3) 8.3 (8.5) Ͻ0.001 SBP indicates systolic blood pressure; DBP, diastolic blood pressure; CABG, coronary artery bypass grafting; PCI, percutaneous coronary intervention; TIA, transient ischemic attack; and LOS, length of stay. *P values in the table compared characteristics between the entire US and Ontario cohorts. Tests of significance (not shown in table) across all 5 regions had P values Ͻ0.001 for all variables. †To convert creatinine from mg/dL to ␮mol/L, multiply by 88.4.

of cardiac-invasive procedures in the northeastern United observe that lower utilization rates in the northeastern United States and Ontario was not associated with poorer outcomes States and Ontario were associated with a lower supply of in the short and longer term after hospitalization compared cardiac-invasive facilities. It is well known that Canadian with other US regions. This observation may imply that federal budget deficits during the 1990s created a strain on regions can deliver quality of care to AMI patients with specialized services.3 Although Ontario continues to have the similar mortality outcomes at lower cost. Our findings also lowest utilization of cardiac catheterization, our observed rate suggest that, in addition to rules and regulations in each of 16.8% during hospitalization is actually Ͼ2-fold higher country, local regional practice patterns may also substan- than the rate of 6.3% at 30 days observed in 1991 to 1992.4 tially influence treatment patterns. Future United States– It has also been suggested that Canadian patients do not Canada comparisons should perhaps include regional analy- receive timely access to quality care.30,31 Our observation that ses, as it is possible that simple national comparisons may Ontario had the fewest number of catheterization facilities obscure important differences that exist between regions. that treat patients with similar illness severity lends support to Cardiac-invasive procedures are highly dependent on the recent initiatives to improve timely access and quality health availability of resources.16,27–29 It is therefore not surprising to care for Canadians.32

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 200 Circulation January 16, 2007

TABLE 3. Process of Care of AMI Patients in the United States and Ontario

Northeastern Southern Midwestern Western Ontario, United States United States United States United States United States Canada P Characteristics (nϭ38 886) (nϭ6655) (nϭ13 187) (nϭ10 213) (nϭ8831) (nϭ5634) Values* Procedures during hospitalization Cardiac catheterization 15 044 (38.7) 1702 (25.6) 5378 (40.8) 4202 (41.1) 3762 (42.6) 948 (16.8) Ͻ0.001 PCI 7099 (18.3) 700 (10.5) 2260 (17.1) 2199 (21.5) 1940 (22.0) 368 (6.5) Ͻ0.001 CABG 2489 (6.4) 231 (3.5) 904 (6.9) 784 (7.7) 570 (6.5) 223 (4.0) Ͻ0.001 Echocardiography 21 882 (56.3) 4450 (66.9) 7366 (55.9) 5548 (54.3) 4518 (51.2) 2462 (43.7) Ͻ0.001 Stress test 3718 (9.6) 1037 (15.6) 1025 (7.8) 833 (8.2) 823 (9.3) 669 (11.9) Ͻ0.001 Medications at hospital discharge for eligible patients Aspirin 19 668 (76.2) 3128 (74.9) 6350 (74.1) 5487 (77.8) 4703 (78.2) 3268 (77.7) 0.032 ␤-Blockers 16 193 (62.7) 2969 (71.1) 5155 (60.1) 4521 (64.1) 3548 (59.0) 2788 (66.3) Ͻ0.001 ACE inhibitors 8042 (52.6) 1452 (54.4) 2639 (50.6) 2201 (53.4) 1750 (53.1) 1281 (67.6) Ͻ0.001 Angiotensin receptor blockers 899 (5.9) 151 (5.7) 298 (5.7) 236 (5.7) 214 (6.5) 65 (3.4) Ͻ0.001 Lipid-lowering therapy 8192 (31.7) 1327 (31.8) 2571 (30.0) 2778 (32.3) 2016 (33.5) 1330 (31.6) 0.88 Medications for ideal candidates Thrombolysis 1326 (47.1) 271 (53.2) 410 (47.1) 254 (40.8) 391 (48.0) 859 (62.9) Ͻ0.001 Primary PCI 540 (19.2) 43 (8.4) 137 (15.7) 157 (25.2) 203 (24.9) 18 (1.3) Ͻ0.001 Reperfusion therapy 1781 (63.2) 310 (60.9) 519 (59.6) 391 (62.8) 561 (68.9) 870 (63.7) 0.77 Aspirin at discharge 10 457 (76.2) 1932 (78.1) 3333 (74.0) 2760 (76.5) 2432 (77.7) 3003 (79.5) Ͻ0.001 ␤-Blockers at discharge 7965 (69.7) 1324 (77.6) 2512 (65.8) 2323 (71.9) 1806 (67.5) 1811 (68.6) 0.28 ACE inhibitors at discharge 6381 (58.2) 1136 (60.9) 2091 (55.5) 1778 (59.5) 1376 (58.5) 1016 (69.1) Ͻ0.001 Values are expressed as n (%). ACE indicates angiotensin-converting enzyme; PCI, percutaneous coronary intervention. *P values compared characteristics between the entire US and Ontario cohorts. Tests of significance (not shown in table) across all 5 regions had P values Ͻ0.001 for all variables.

Because the northeastern United States does not face population is predominantly dependent on a patient’s baseline similar resource constraints compared to Ontario, it is more risk of future cardiovascular events (ie, greater benefits in difficult to understand why the northeastern United States higher-risk patients);34 a “paradox” has been observed, how- consistently performed fewer cardiac invasive procedures, ever, in which aggressive medical and interventional thera- had a smaller number of invasive facilities, and performed pies are often applied to patients at relatively low risk of lower proportion of cardiac invasive procedures per facility adverse outcomes.35,36 Finally, data suggest that invasive than other US regions.16,29 We observed that patients who therapies need to be used in a timely fashion to optimize their underwent cardiac catheterization had admission characteris- benefits.37,38 Therefore, delays in the use of invasive therapies tics and similar GRACE risk scores, which indicated that may have diminished the potential benefits of cardiac inva- illness severity was unlikely a major reason to explain sive procedures in clinical practice.39 variations in invasive procedures. Other possible reasons to We chose risk-standardized mortality at 30 days as our explain procedural variations that we were unable to explore primary outcome of interest because it is likely most reflec- include differences in physician attitudes toward invasive tive of in-hospital processes of care. Risk-standardized mor- treatment, patient preferences, and unmeasured regional tality rates were similar between the United States and characteristics.13,16 Ontario at 30 days. However, we found that longer-term Although we cannot determine the optimal rate of cardiac mortality at 1 year and at 3 years were lower in Ontario in catheterization after AMI, the less aggressive invasive ap- secondary analyses. A phenomenon of better short-term proach observed in the northeastern United States and On- survival rates in the United States followed by better tario was not associated with worse short- or long-term intermediate-term survival rates in Canada has also been survival after adjustment for patient and hospital factors. previously observed in other studies.4,40 We could not assess Findings from our study mirror those of other population- the possible reasons for this finding in our study, but a recent based studies, where increased invasive procedure utilization study that surveyed both countries found that US residents had no significant impact on mortality after AMI.4,13,29,33 had more chronic conditions but were less likely to have a Several reasons may explain why benefits shown in random- regular physician, more likely to have unmet health needs, ized trials do not translate to the population level. First, and more likely to forgo needed medications because of evidence has suggested that there may only be marginal coast.41 It is also possible that long-term survival after an additional benefits of coronary interventions in patients who AMI is more dependent on the general health status of already receive optimal medical therapy.29 Second, the im- patients and/or the quality of outpatient care (eg, primary pact of any evidence-based cardiovascular therapy in the care, medication adherence) than on the quality of in-hospital

circ.aha at Mohammed Mahboob on June 10, 2007

Downloaded from journals.org Ko et al Regional Differences in US and Canadian AMI Care 201

TABLE 4. Characteristics of AMI Patients Who Received Cardiac Catheterization During Hospitalization in the United States and Ontario

Northeastern Southern Midwestern Western Ontario, United States United States United States United States United States Canada P Characteristics (Nϭ15 044) (Nϭ1702) (Nϭ5378) (Nϭ4202) (Nϭ3762) (Nϭ948) Values* Demographics Age, y, mean (SD) 74.9 (6.0) 75.4 (5.9) 74.6 (5.9) 75.1 (6.0) 74.9 (6.0) 73.4 (5.3) Ͻ0.001 65 to 74, n (%) 7678 (51.0) 800 (47.0) 2855 (53.1) 2108 (50.2) 1915 (50.9) 584 (61.6) 75 to 84, n (%) 6295 (41.8) 776 (45.6) 2192 (40.8) 1774 (42.2) 1553 (41.3) 336 (35.4) Ն85 1071 (7.1) 126 (7.4) 331 (6.2) 320 (7.6) 294 (7.8) 28 (3.0) Female, n (%) 6443 (42.8) 767 (45.1) 2322 (43.2) 1864 (44.4) 1490 (39.6) 368 (38.8) 0.016 Admission characteristics SBP, mean (SD) 145.0 (31.4) 146.7 (32.0) 145.7 (31.4) 144.5 (31.6) 143.8 (30.8) 149.5 (32.0) Ͻ0.001 DBP, mean (SD) 78.1 (19.1) 78.1 (20.1) 78.3 (18.9) 77.7 (19.1) 78.2 (18.9) 82.8 (18.2) Ͻ0.001 Heart rate, mean (SD) 83.0 (22.9) 83.8 (22.8) 83.6 (22.9) 82.8 (22.6) 82.1 (23.0) 80.0 (21.6) Ͻ0.001 Mean creatinine,† mean (SD) 1.2 (1.4) 1.4 (3.6) 1.2 (0.8) 1.2 (0.7) 1.2 (0.8) 1.2 (0.7) 0.39 Cardiogenic shock, n (%) 190 (1.3) 30 (1.8) 56 (1.0) 38 (0.9) 66 (1.8) 17 (1.8) 0.16 Cardiac arrest, n (%) 365 (2.4) 47 (2.8) 113 (2.1) 87 (2.1) 118 (3.1) 17 (1.8) 0.22 GRACE risk, mean (SD) 154.7 (33.6) 156.6 (33.8) 152.8 (33.0) 154.5 (33.0) 157.0 (35.0) 153.9 (33.4) 0.45 Diabetes, n (%) 4167 (27.7) 503 (29.6) 1568 (29.2) 1140 (27.1) 956 (25.4) 235 (24.8) 0.052 Current smoker, n (%) 2744 (18.2) 278 (16.3) 1074 (20.0) 781 (18.6) 611 (16.2) 198 (20.9) 0.041 Prior myocardial infarction, n (%) 4292 (28.5) 521 (30.6) 1501 (27.9) 1241 (29.5) 1029 (27.4) 226 (23.8) 0.002 Prior CABG, n (%) 2423 (16.1) 249 (14.6) 918 (17.1) 708 (16.8) 548 (14.6) 82 (8.6) Ͻ0.001 Prior PCI, n (%) 2295 (15.3) 231 (13.6) 803 (14.9) 704 (16.8) 557 (14.8) 54 (5.7) Ͻ0.001 Prior stroke or TIA, n (%) 1954 (13.0) 215 (12.6) 767 (14.3) 547 (13.0) 425 (11.3) 84 (8.9) Ͻ0.001 SBP indicates systolic blood pressure; DBP, diastolic blood pressure; CABG, coronary artery bypass grafting; PCI, percutaneous coronary intervention; and TIA, transient ischemic attack. *P values in the table compared characteristics between the entire US and Ontario cohorts. Most characteristics compared across all 5 regions had P values Ͻ0.001 (not shown in table) except for mean creatinine (Pϭ0.008) and cardiac arrest (Pϭ0.004). †To convert creatinine from mg/dL to ␮mol/L, multiply by 88.4.

care. Further research studies are needed to identify the pies can be prescribed at a low cost. Previous data suggest reasons for the important long-term outcome differences that regional differences in the use of medical therapy persist observed in this study. even after adjustment for other patient, physician, and hospi- Although the use of evidence-based medical therapies tal characteristics.13 Understanding of the possible factors that varied significantly less across regions, we still observed a underlie these regional variations will be crucial to ongoing 12% absolute difference for ␤-blockers and 6% for angioten- improvements to the care and outcomes of patients with AMI. sin-converting enzyme inhibitors among the highest to lowest Several limitations of our study merit consideration. First, utilization regions among ideal candidates. There are no privacy laws in the United States and Canada restricted obvious reasons to explain why regional variations in medical transfer of patient level data outside each country and treatment exist, as published evidence that demonstrates its prevented the combination of data sets. However, both data effectiveness is widely available and evidence-based thera- sets contained detailed information on demographics, admis- sion characteristics, medication use, and invasive procedures

) 80

% that provided comprehensive insights into regional variations (

e 2 t R = 0.8139 in processes of care. We did not examine regional practice a r

60 n variations in Canada because similar clinically abstracted o i

t West Midwest a

z data sets were not available from other provinces. However, i South r 40 e

t it has been demonstrated that population utilization rates of e h t Northeast

a cardiac catheterization in Ontario (509.6 per 100 000 popu-

c 20

c lation) are highly representative of the overall rates in Canada a

i Ontario

d 41 r (471.5 per 100 000). Third, although we did not have a 0 C 0 1020304050607080 information on outpatient use of medication after AMI Proportion of AMI patients hospitalized to cardiac invasive facility (%) discharge, prescription at hospital discharge is one of the strongest predictors of long-term use.42–44 Fourth, the US The use of cardiac catheterization in acute myocardial infarction and the proportion of patients hospitalized to a cardiac invasive cohort was a random sample of an equal number of AMI facility in each region of the United States and Ontario, Canada. patients hospitalized in each state and the Ontario cohort was

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 202 Circulation January 16, 2007

a random sample of AMI patients representative of the entire province. However, we performed additional analyses by management patterns after acute myocardial infarction in Canada and the United States. The SAVE investigators. N Engl J Med. 1993;328: weighting the US sample based on the population in each 779–784. state and found similar mortality results. Finally, we were 7. Fu Y, Chang WC, Mark D, Califf RM, Mackenzie B, Granger CB, Topol unable to compare other nonfatal outcomes after AMI such as EJ, Hlatky M, Armstrong PW. Canadian-American differences in the quality of life, left ventricular ejection fraction, or heart management of acute coronary syndromes in the GUSTO IIb trial: failure hospitalization because these data were not available. one-year follow-up of patients without ST-segment elevation. Circu- lation. 2000;102:1375–1381. In conclusion, despite marked differences between the 8. Kaul P, Armstrong PW, Chang W-C, Naylor CD, Granger CB, Lee KL, American and the Canadian health care systems, the similar- Peterson ED, Califf RM, Topol EJ, Mark DB. Long-term mortality of ity of AMI treatment between the northeastern United States patients with acute myocardial infarction in the United States and Canada. and Ontario suggests that local regional practice patterns play Comparison of patients enrolled in global utilization of streptokinase and t-PA for occluded coronary arteries (GUSTO)-I. Circulation. 2004;110: a substantial role in the influence of the delivery of care. 1754–1760. Further exploration of factors that underlie these regional 9. Gurwitz JH, Col NF, Avorn J. The exclusion of the elderly and women variations may provide insights to improve the quality of care from clinical trials in acute myocardial infarction. JAMA. 1992;268: for patients with AMI. 1417–1422. 10. Lee PY, Alexander KP, Hammill BG, Pasquali SK, Peterson ED. Rep- resentation of elderly persons and women in published randomized trials Sources of Funding of acute coronary syndromes. JAMA. 2001;286:708–713. The analyses on which this publication is based were performed 11. Hutchins LF, Unger JM, Crowley JJ, Coltman CA Jr, Albain KS. Under- under Contract Number 500-02-CO-01 entitled Utilization and representation of patients 65 years of age or older in cancer-treatment Quality Control Peer Review Organization for the State of Colorado, trials. N Engl J Med. 1999;341:2061–2067. which is sponsored by the Centers for Medicare and Medicaid 12. Jha P, Deboer D, Sykora K, Naylor CD. Characteristics and mortality Services (CMS, formerly the Health Care Financing Administra- outcomes of thrombolysis trial participants and nonparticipants: a tion), Department of Health and Human Services. The content of the population-based comparison. J Am Coll Cardiol. 1996;27:1335–1342. publication does not necessarily reflect the views or policies of the 13. Krumholz HM, Chen J, Rathore SS, Wang Y, Radford MJ. Regional Department of Health and Human Services, nor does mention of variation in the treatment and outcomes of myocardial infarction: inves- trade names, commercial products, or organizations imply endorse- tigating New England’s advantage. Am Heart J. 2003;146:242–249. ment by the US government. The authors assume full responsibility 14. O’Connor GT, Quinton HB, Traven ND, Ramunno LD, Dodds TA, for the accuracy and completeness of the ideas presented. This article Marciniak TA, Wennberg JE. Geographic variation in the treatment of is a direct result of the Health Care Quality Improvement Program acute myocardial infarction: the Cooperative Cardiovascular Project. initiated by CMS, which has encouraged identification of quality JAMA. 1999;281:627–633. improvement projects from analysis of patterns of care, and therefore 15. Krumholz HM, Radford MJ, Wang Y, Chen J, Heiat A, Marciniak TA. required no special funding on the part of this contractor. Ideas and National use and effectiveness of beta-blockers for the treatment of contributions to the author that concern experiences that engage the elderly patients after acute myocardial infarction: National Cooperative issues presented are welcome. CMS did not play a role in the design Cardiovascular Project. JAMA. 1998;280:623–629. and conduct of the present study, or in the analysis and interpretation 16. Pilote L, Califf RM, Sapp S, Miller DP, Mark DB, Weaver WD, Gore JM, of the data (Tracking Number: HIQIOSCHEARTCARE-001-OK-0506). Armstrong PW, Ohman EM, Topol EJ. Regional variation across the United States in the management of acute myocardial infarction; The EFFECT study was supported by a Canadian Institutes of GUSTO-1 Investigators. Global utilization of streptokinase and tissue Health Research Team Grant in Cardiovascular Outcomes Research. plasminogen activator for occluded coronary arteries N Engl J Med. Dr Ko is supported by a Heart and Stroke Foundation of Ontario 1995;333:565–572. Clinician Scientist Award. Dr Alter is supported by a Heart and 17. Wennberg DE, Dartmouth Atlas of Cardiovascular Health Care Working Stroke Foundation of Ontario Career Investigator Award. Dr Tu is Group. The Dartmouth Atlas of Cardiovascular Health Care. Chicago, Ill: supported by a Canadian Research Chair in Health Service Research AHA Press; 1999. and a Career Investigator Award from the Heart and Stroke Foun- 18. Marciniak TA, Ellerbeck EF, Radford MJ, Kresowik TF, Gold JA, dation of Ontario. Krumholz HM, Kiefe CI, Allman RM, Vogel RA, Jencks SF. Improving the quality of care for Medicare patients with acute myocardial infarction: Disclosures results from the Cooperative Cardiovascular Project. JAMA. 1998;279: None. 1351–1357. 19. Burwen DR, Galusha DH, Lewis JM, Bedinger MR, Radford MJ, References Krumholz HM, Foody JM. National and state trends in quality of care for acute myocardial infarction between 1994–1995 and 1998–1999: the 1. Iglehart JK. The American health care system–Medicaid. N Engl J Med. medicare health care quality improvement program. Arch Intern Med. 1999;340:403–408. 2003;163:1430–1439. 2. Iglehart JK. The American health care system–Medicare. N Engl J Med. 20. Tu JV, Donovan LR, Lee DS, Austin PC, Ko DT, Wang JT, Newman 1999;340:327–332. 3. Detsky AS, Naylor CD. Canada’s health care system–reform delayed. AM. Quality of Cardiac Care in Ontario. Toronto, Ontario: ICES; 2004. N Engl J Med. 2003;349:804–810. 21. Lee DS, Austin PC, Rouleau JL, Liu PP, Naimark D, Tu JV. Predicting 4. Tu JV, Pashos CL, Naylor CD, Chen E, Normand SL, Newhouse JP, mortality among patients hospitalized for heart failure: derivation and McNeil BJ. Use of cardiac procedures and outcomes in elderly patients validation of a clinical model. JAMA. 2003;290:2581–2587. with myocardial infarction in the United States and Canada. N Engl 22. Census Regions and Divisions. Washington, DC: US Department of J Med. 1997;336:1500–1505. Commerce; 2002. 5. Mark DB, Naylor CD, Hlatky MA, Califf RM, Topol EJ, Granger CB, 23. Tran CT, Lee DS, Flintoft VF, Higginson L, Grant FC, Tu JV, Cox J, Knight JD, Nelson CL, Lee KL, Clapp-Channing NE, Sutherland W, Holder D, Jackevicius C, Pilote L, Tanser P, Thompson C, Tsoi E, Pilote L, Armstrong PW. Use of medical resources and quality of life Warnica W, Wielgosz A, Canadian Cardiovascular Outcomes Research after acute myocardial infarction in Canada and the United States. N Engl Team/Canadian Cardiovascular S, Acute Myocardial Infarction Quality J Med. 1994;331:1130–1135. Indicator P. CCORT/CCS quality indicators for acute myocardial 6. Rouleau JL, Moye LA, Pfeffer MA, Arnold JM, Bernstein V, Cuddy TE, infarction care. Can J Cardiol. 2003;19:38–45. Dagenais GR, Geltman EM, Goldman S, Gordon D, Hamm P, Klein M, 24. Fleming C, Fisher ES, Chang CH, Bubolz TA, Malenka DJ. Studying Lamas GA, McCans J, McEwan P, Menapace FJ, Parker JO, Sestier F, outcomes and hospital utilization in the elderly. The advantages of a merged database for Medicare and Veterans Affairs hospitals. Med Care. Sussex B, Braunwald E, for the SAVE Investigators. A comparison of 1992;30:377–391.

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Ko et al Regional Differences in US and Canadian AMI Care 203

25. Tu JV, Austin PC, Walld R, Roos L, Agras J, McDonald KM. Devel- 36. Ko DT, Mamdani M, Alter DA. Lipid-lowering therapy with statins in opment and validation of the Ontario acute myocardial infarction mor- high-risk elderly patients: the treatment-risk paradox. JAMA. 2004;291: tality prediction rules. J Am Coll Cardiol. 2001;37:992–997. 1864–1870. 26. Granger CB, Goldberg RJ, Dabbous O, Pieper KS, Eagle KA, Cannon 37. Braunwald E, Antman EM, Beasley JW, Califf RM, Cheitlin MD, CP, Van De Werf F, Avezum A, Goodman SG, Flather MD, Fox KA, Hochman JS, Jones RH, Kereiakes D, Kupersmith J, Levin TN, Pepine Global Registry of Acute Coronary Events I. Predictors of hospital mor- CJ, Schaeffer JW, Smith EE 3rd, Steward DE, Theroux P, Gibbons RJ, tality in the global registry of acute coronary events. Arch Intern Med. Alpert JS, Faxon DP, Fuster V, Gregoratos G, Hiratzka LF, Jacobs 2003;163:2345–2353. AK, Smith SC Jr. ACC/AHA 2002 guideline update for the man- agement of patients with unstable angina and non-ST-segment ele- 27. Alter DA, Tu JV, Austin PC, Naylor CD. Waiting times, revascularization vation myocardial infarction–summary article: a report of the modality, and outcomes after acute myocardial infarction at hospitals with American College of Cardiology/American Heart Association task and without on-site revascularization facilities in Canada. J Am Coll force on practice guidelines (Committee on the Management of Cardiol. 2003;42:410–419. Patients With Unstable Angina). J Am Coll Cardiol. 2002;40: 28. Guadagnoli E, Hauptman PJ, Ayanian JZ, Pashos CL, McNeil BJ, Cleary 1366–1374. PD. Variation in the use of cardiac procedures after acute myocardial 38. Cannon CP, Gibson CM, Lambrew CT, Shoultz DA, Levy D, French WJ, infarction. N Engl J Med. 1995;333:573–578. Gore JM, Weaver WD, Rogers WJ, Tiefenbrunn AJ. Relationship of 29. Stukel TA, Lucas FL, Wennberg DE. Long-term outcomes of regional symptom-onset-to-balloon time and door-to-balloon time with mortality variations in intensity of invasive vs medical management of medicare in patients undergoing angioplasty for acute myocardial infarction. patients with acute myocardial infarction. JAMA. 2005;293:1329–1337. JAMA. 2000;283:2941–2947. 30. Ezekowitz JA, Armstrong PW. The waiting game: facing the conse- 39. Nallamothu BK, Bates ER, Herrin J, Wang Y, Bradley EH, Krumholz quences. CMAJ. 2002;167:1247–1248. HM, for the NRMI Investigators. Times to treatment in transfer patients 31. Sanmartin C, Shortt SED, Barer ML, Sheps S, Lewis S, McDonald PW. undergoing primary percutaneous coronary intervention in the United Waiting for medical services in Canada: lots of heat, but little light. States: National Registry of Myocardial Infarction (NRMI)-3/4 analysis. CMAJ. 2000;162:1305–1310. Circulation. 2005;111:761–767. 32. Tu JV, Pinfold SP, McColgan P, Laupacis A. Access to Health Services in 40. Ko DT, Tu JV, Masoudi FA, Wang Y, Havranek EP, Rathore SS, Ontario: ICES Atlas. Toronto: Institute for Clinical Evaluative Sciences; 2005. Newman AM, Donovan LR, Lee DS, Foody JM, Krumholz HM. Quality 33. Van de Werf F, Gore JM, Avezum A, Gulba DC, Goodman SG, Budaj A, of care and outcomes of older patients with heart failure hospitalized in Brieger D, White K, Fox KA, Eagle KA, Kennelly BM, GRACE Inves- the United States and Canada. Arch Intern Med. 2005;165:2486–2492. tigators. Access to catheterisation facilities in patients admitted with acute 41. Faris PD, Grant FC, Galbraith PD, Gong Y, Ghali WA, Canadian Car- diovascular Outcomes Research Team. Diagnostic cardiac catheterization coronary syndrome: multinational registry study. BMJ. 2005;330:441. and revascularization rates for coronary heart disease. Can J Cardiol. 34. Alter DA, Manuel DG, Gunraj N, Anderson G, Naylor CD, Laupacis A. 2004;20:391–397. Age, risk-benefit trade-offs, and the projected effects of evidence-based 42. Butler J, Arbogast PG, BeLue R, Daugherty J, Jain MK, Ray WA, Griffin therapies. Am J Med. 2004;116:540–545. MR. Outpatient adherence to beta-blocker therapy after acute myocardial 35. Bhatt DL, Roe MT, Peterson ED, Li Y, Chen AY, Harrington RA, infarction. J Am Coll Cardiol. 2002;40:1589–1595. Greenbaum AB, Berger PB, Cannon CP, Cohen DJ, Gibson CM, Saucedo 43. Butler J, Arbogast PG, Daugherty J, Jain MK, Ray WA, Griffin MR. JF, Kleiman NS, Hochman JS, Boden WE, Brindis RG, Peacock WF, Outpatient utilization of angiotensin-converting enzyme inhibitors among Smith SC Jr, Pollack CV Jr, Gibler WB, Ohman EM, CRUSADE Inves- heart failure patients after hospital discharge. J Am Coll Cardiol. 2004; tigators. Utilization of early invasive management strategies for high-risk 43:2036–2043. patients with non-ST-segment elevation acute coronary syndromes: 44. Fonarow GC, Gheorghiade M, Abraham WT. Importance of in-hospital results from the CRUSADE Quality Improvement Initiative. JAMA. initiation of evidence-based medical therapies for heart failure—a review. 2004;292:2096–2104. Am J Cardiol. 2004;94:1155–1160.

CLINICAL PERSPECTIVE Previous studies have suggested a clear divergence in cardiac invasive therapy for acute myocardial infarction patients between the United States and Canada on the basis of health care finance and structural differences. With clinical data from population-representative samples in the United States and Canada, we observed that acute myocardial infarction treatment patterns were actually quite similar in the northeastern United States and Ontario, as both regions had relatively low rates of cardiac-invasive procedures, whereas the rates of medication use compared with other US regions were among the highest. Less intensive use of cardiac-invasive procedures in the northeastern United States and Ontario was not associated with poorer mortality outcomes at short- and longer-term after hospitalization compared with other US regions. This observation may imply that regions can deliver quality of care to acute myocardial infraction patients with similar mortality outcomes at lower cost. Our findings also suggest that, in addition to rules and regulations in each country, local regional practice patterns may also play a substantial role in influencing how care is delivered to patients.

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Heart Failure

Effects of Cardiac Resynchronization Therapy With or Without a Defibrillator on Survival and Hospitalizations in Patients With New York Heart Association Class IV Heart Failure

JoAnn Lindenfeld, MD; Arthur M. Feldman, MD, PhD; Leslie Saxon, MD; John Boehmer, MD; Peter Carson, MD; Jalal K. Ghali, MD; Inder Anand, MD; Steve Singh, MD; Jonathan S. Steinberg, MD; Brian Jaski, MD; Teresa DeMarco, MD; David Mann, MD; Patrick Yong, MSEE; Elizabeth Galle, MS; Fred Ecklund, MA; Michael Bristow, MD, PhD

Background—Cardiac resynchronization therapy (CRT) alone or combined with an implantable defibrillator (CRT-D) has been shown to improve exercise capacity and quality of life and to reduce heart failure (HF) hospitalizations and mortality in patients with New York Heart Association (NYHA) class III and IV HF. There is concern that the device procedure may destabilize these very ill class IV patients. We sought to examine the outcomes of NYHA class IV patients enrolled in the Comparison of Medical Therapy, Pacing, and Defibrillation in Heart Failure (COMPANION) trial to assess the potential benefits of CRT and CRT-D. Methods and Results—The COMPANION trial randomized 1520 patients with NYHA class III and IV HF to optimal medical therapy, CRT, or CRT-D. In the class IV patients (nϭ217), the primary end point of time to death or hospitalization for any cause was significantly improved by both CRT (hazard ratio [HR], 0.64; 95% CI, 0.43 to 0.94; Pϭ0.02) and CRT-D (HR, 0.62; 95% CI, 0.42 to 0.90; Pϭ0.01). Time to all-cause death and HF hospitalization was also significantly improved in both CRT (HR, 0.57; 95% CI, 0.37 to 0.87; Pϭ0.01) and CRT-D (HR, 0.49; 95% CI, 0.32 to 0.75; Pϭ0.001) Time to all-cause death trended to an improvement in both CRT (HR, 0.67; 95% CI, 0.41 to 1.10; Pϭ0.11) and CRT-D (HR, 0.63; 95% CI, 0.39 to 1.03; Pϭ0.06). Time to sudden death appeared to be significantly reduced in the CRT-D group (HR, 0.27; 95% CI, 0.08 to 0.90; Pϭ0.03). There was a nonsignificant reduction in time to HF deaths for both CRT (HR, 0.68; 95% CI, 0.34 to 1.37; Pϭ0.28) and CRT-D (HR, 0.79; 95% CI, 0.41 to 1.52; Pϭ0.48). Conclusions—CRT and CRT-D significantly improve time to all-cause mortality and hospitalizations in NYHA class IV patients, with a trend for improved mortality. These devices should be considered in ambulatory NYHA class IV HF patients similar to those enrolled in COMPANION. (Circulation. 2007;115:204-212.) Key Words: defibrillation Ⅲ heart failure Ⅲ pacemakers Ⅲ survival

ardiac resynchronization therapy (CRT) alone or com- Editorial p 161 bined with an implantable defibrillator (CRT-D) has C Clinical Perspective p 212 been shown to improve exercise capacity and quality of life and to reduce heart failure (HF) hospitalizations and mortality the implant procedure may destabilize the HF and worsen short- in patients with New York Heart Association (NYHA) class term outcomes before the benefits of device therapy can be realized. III and IV HF.1–7 Only small numbers of patients enrolled in Therefore, we sought to examine the outcomes of NYHA class IV these trials have been classified as NYHA class IV, however. HF patients enrolled in the Comparison of Medical Therapy, Class IV patients typically have limited myocardial reserve Pacing, and Defibrillation in Heart Failure (COMPANION) trial, a and a poor survival. It has been suggested that class IV trial that randomized patients to 1 of 3 treatment groups: optimal patients may not benefit from CRT or indeed even CRT-D, as medical therapy (OPT), OPT ϩ CRT, and OPT ϩ CRT-D.

Received April 7, 2006; accepted October 11, 2006. From the Department of Medicine, University of Colorado Health Sciences Center (J.L., M.R.B.), and Women’s Health Research Center (J.L.), Denver, Colo; Thomas Jefferson Medical College (A.M.F.), Philadelphia, Pa; University of Southern California (Keck School of Medicine) (L.A.S.), Los Angeles; Milton Hershey Medical Center, Penn State School of Medicine (J.B.), Hershey, Pa; Washington DC Veterans Administration (S.S., P.C.), Washington, DC; Wayne State School of Medicine (J.K.G.), Detroit, Mich; Minneapolis Veterans Administration (I.A.), Minneapolis, Minn; San Diego Cardiac Center (B.J.), San Diego, Calif; University of California (T.M.), San Francisco; Cardiovascular Associates and St Mary and Elizabeth Hospital (D.M.), Louisville, Ky; and Boston Scientific (P.Y., E.G., F.E.), St. Paul, Minn. Clinical trial registration information—URL:http://www.clinicaltrials.gov. Unique identifier: NCT00180258. Correspondence to Dr JoAnn Lindenfeld, Division of Cardiology, University of Colorado Health Sciences Center, 4200 E. Ninth Ave, B-130, Denver, CO 80262. E-mail [email protected] © 2007 American Heart Association, Inc. Circulation is available at http://www.circulationaha.org DOI: 10.1161/CIRCULATIONAHA.106.629261

204

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Lindenfeld et al Effects of CRT on Class IV Heart Failure Patients 205

Methods 0.30, stay 0.05) and included the following baseline variables: The COMPANION trial was performed in 128 US centers. The treatment, systolic blood pressure, diastolic blood pressure, heart complete protocol as well as the results for the entire cohort are rate, LVEF, QRS, NYHA class, ischemic status, diabetes status, left reported elsewhere.3,8 Criteria for enrollment included NYHA class bundle-branch block, right bundle-branch block, atrial fibrillation, III or IV HF caused by either ischemic or nonischemic cardiomy- renal disease, hypertension, peripheral vascular disease, body mass opathy, an electrocardiographic QRS interval of at least 120 milli- index, age, gender, and medication (␤-blocker, angiotensin seconds with a PR interval of Ͻ150 milliseconds, sinus rhythm, no converting-enzyme or angiotensin receptor blocker, digoxin, diuret- clinical indication for pacemaker or implantable defibrillator, a left ic). A log(-log) plot was used for each survival analysis to validate ventricular ejection fraction (LVEF) of Յ0.35, and a hospitalization the proportionality assumption. To analyze the end points of mor- for the treatment of HF or the equivalent in the 12 months preceding tality and hospitalization, data on patients who withdrew before enrollment.8 Exclusion criteria included expectation of a heart reaching an end point, who were not known to have died and for transplantation in the subsequent 6 months, medically refractory whom complete postwithdrawal information on hospitalization could atrial arrhythmias, unexplained syncope, myocardial infarction not be obtained, were censored at the time of elective hospitalization within 60 days, surgically uncorrected primary valvular heart dis- for device implantation or on the date of the last contact. For the ease, coronary artery intervention (catheter or surgical) within 60 NYHA class IV patients, some were censored before the end of the Ͻ days, progressive or unstable angina, and life expectancy 6 months present study. That is, for the primary end point of all-cause death or for non-HF conditions.8 In addition, no patient could have a hospitalization, 6 (11%), 0 (0%), and 0 (0%) of the OPT, CRT, and scheduled or unscheduled admission for HF or intravenous inotropic CRT-D patients, respectively, were censored; for the mortality end 8 or vasoactive therapy in excess of 4 hours in the previous month. points, 2 (4%), 0 (0%), and 5 (6%) of the OPT, CRT, and CRT-D The steering committee, clinical endpoints committee, and sponsor patients were censored. Survival curves were plotted based on the were blinded to treatment assignments. Physicians, patients, inde- Kaplan-Meier method. Two-year rates for mode of death adjusted for pendent statisticians, and members of the data-management group follow-up time were calculated based on life table survival estimates and the data and safety monitoring board were not blinded to from SAS software’s LIFETEST procedure. Continuous functional treatment assignments. Enrolled patients were randomly assigned in a 1:2:2 ratio to capacity variables were tested with a Wilcoxon 2-sample test. ␹2 treatment with OPT, CRT, or CRT-D. Randomization was strat- NYHA class change was evaluated with the Mantel-Haenszel for ified by clinical site and ␤-blocker use, but not by NYHA class. NYHA class change and implant success by the Fisher exact test. For All patients were taking diuretics (unless not needed), angioten- the functional capacity end points (6-minute walk, quality of life, sin-converting enzyme inhibitors or angiotensin receptor blockers NYHA class), CRT and CRT-D were combined in the original (unless not tolerated), ␤-blockers (unless not tolerated), and protocol design because there was no expectation that CRT-D would spironolactone (unless not tolerated). Successful implantation of influence exercise capacity in a manner dissimilar to CRT. The CRT or CRT-D was defined as successful implantation of all duration of implant hospitalization days was based on the sum of all leads. implants (ie, some patients had Ͼ1 attempt or successful implant) Once randomized, patients received CRT with implantation of and tested with a Wilcoxon 2-sample test. All probability values are a biventricular pacemaker (Contak TR model 1241, Guidant, 2-sided. Indianapolis, Ind) or a CRT-defibrillator (Contak CD model 1823, The authors had full access to the data and take full responsibility Guidant) with the use of commercially available leads for right for the integrity of the data. All authors have read and agreed to the atrial pacing and right ventricular pacing or for pacing with manuscript as written. defibrillation (Endotak models 0125, 0154, and 0155, Guidant). Implantation techniques were described elsewhere.3 The primary end point of the present study, as well as the main trial, was a composite of death from any cause and hospitalization Results from any cause analyzed from the time of randomization to the time of first event. Unscheduled administration of intravenous Baseline Characteristics vasoactive or inotropic agents for Ͼ4 hours in the emergency Fourteen percent (nϭ217) of the 1520 COMPANION pa- department was counted as a hospitalization. The implant hospi- talization was not included in the primary end point. All-cause tients were NYHA class IV, whereas the remaining 86% were mortality was a secondary end point. An additional end point of NYHA class III. Table 1 compares baseline demographics of interest was death from any cause or hospitalization for heart participants with NYHA class III and class IV HF. NYHA failure. Components of the primary end point were adjudicated by class IV patients had lower LVEF, a larger left ventricular a blinded 7-member endpoints committee. Sudden death was defined as observed or unobserved sudden death in the absence of end-diastolic diameter, higher resting heart rate, and lower progressive HF. This category included patients who experienced systolic and diastolic blood pressure than NYHA class III sudden cardiac death and survived in a postresuscitative course subjects. Ischemic heart disease was more often the cause of but never regained consciousness and did not leave the hospital. HF in patients with NYHA class IV compared with class III HF death was defined as progressively worsening HF manifested patients, and diabetes was more common in NYHA class IV. by increased symptoms that required increases in medications, which included intravenous medications. HF hospitalization was Six-minute walk distance was significantly lower in the defined as a hospitalization for the treatment of HF that required NYHA class IV subjects. Conduction disorders and QRS increases in medications, which included intravenous width were not significantly different. Patients with NYHA medications. class IV HF were less likely to be taking angiotensin Statistical Analyses converting-enzyme inhibitors or angiotensin receptor block- All analyses were performed according to intention to treat. Com- ers and ␤-blockers. The use of diuretics was slightly and parisons between baseline characteristics were conducted by significantly more common in class IV (100%) participants ANOVA for continuous variables and a ␹2 test for categorical than in class III participants (95%). Spironolactone and variables. Hazard ratios (HRs) and associated probability values digoxin use was similar in both groups. were calculated based on time to first event with a Cox proportional hazard model. All hazard ratios are unadjusted except where indi- Table 2 presents the baseline characteristics of the NYHA cated. Hazard ratios were adjusted with stepwise selection (entry class IV patients by treatment group. There were no signifi-

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 206 Circulation January 16, 2007

TABLE 1. Baseline Demographics for NYHA Class III and NYHA Class IV Patients

NYHA Class III NYHA Class IV Characteristic (Nϭ1303) (Nϭ217) P Age, y, meanϮSD 65.6Ϯ11.5 66.7Ϯ10.4 0.19 Gender, n (%) Male 890 (68) 137 (63) 0.13 Female 413 (32) 80 (37) ⅐⅐⅐ Duration of HF, y, meanϮSD 4.5Ϯ4.1 5.4Ϯ4.9 Ͻ0.01 LVEF, %, meanϮSD 22.6Ϯ6.9 20.8Ϯ6.6 Ͻ0.01 LVEDD, meanϮSD 68.8Ϯ8.1 70.9Ϯ10.0 Ͻ0.01 Resting heart rate, bpm, meanϮSD 72.5Ϯ12.6 76.6Ϯ13.1 Ͻ0.01 Systolic blood pressure, mm Hg, meanϮSD 114.9Ϯ16.8 110.5Ϯ18.0 Ͻ0.01 Diastolic blood pressure, mm Hg, meanϮSD 67.2Ϯ9.9 65.6Ϯ10.2 0.02 6-Minute walk distance, m, meanϮSD 265.3Ϯ106.5 167.0Ϯ104.9 Ͻ0.01 QRS width, ms, meanϮSD 158.1Ϯ24.0 161.0Ϯ26.6 0.10 Ischemic, n (%) Yes 704 (54) 134 (62) 0.03 No 599 (46) 83 (38) ⅐⅐⅐ History of diabetes, n (%) No 785 (60) 112 (52) 0.02 Yes 517 (40) 105 (48) ⅐⅐⅐ Conduction disorder, n (%) Left bundle-branch block 926 (71) 148 (68) 0.54 Nonspecific intraventricular 235 (18) 46 (21) ⅐⅐⅐ Right bundle-branch block 142 (11) 23 (11) ⅐⅐⅐ ACE inhibitor or angiotensin II, n (%) Yes 1176 (90) 177 (82) Ͻ0.01 No 127 (10) 40 (18) ⅐⅐⅐ ␤-Blocker, n (%) Yes 914 (70) 112 (52) Ͻ0.01 No 389 (30) 105 (48) ⅐⅐⅐ ACE inhibitor/angiotensin II ϩ ␤-blocker, n (%) Yes 824 (63) 96 (44) Ͻ0.01 No 479 (37) 121 (56) ⅐⅐⅐ Loop diuretics, n (%) Yes 1240 (95) 216 (100) Ͻ0.01 No 63 (5) 1 (0) ⅐⅐⅐ Digoxin, n (%) Yes 944 (72) 156 (72) 0.86 No 359 (28) 61 (28) ⅐⅐⅐ Spironolactone, n (%) Yes 704 (54) 122 (56) 0.55 No 599 (46) 95 (44) ⅐⅐⅐ LVEDD indicates left ventricular end-diastolic diameter; ACE, angiotensin-converting enzyme.

cant differences by treatment group in any of the baseline tional end point of the study in NYHA class IV subjects. demographics. The primary end point of time to death or hospitalization for any cause was significantly prolonged by both CRT Efficacy of Device Therapy (HR, 0.64; Pϭ0.02) and by CRT-D (HR, 0.62; Pϭ0.01) For NYHA class IV patients, the median duration of compared with optimal medical therapy (OPT). Despite follow-up for the primary end point was 7.2 months for favorable trends, time to death from any cause was not OPT, 14.2 months for CRT, and 14.1 month for CRT-D. significantly different when CRT (HR, 0.67; Pϭ0.11) or Figure 1 demonstrates the primary, secondary and addi- CRT-D (HR, 0.63; Pϭ0.06) was compared with OPT.

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Lindenfeld et al Effects of CRT on Class IV Heart Failure Patients 207

TABLE 2. Baseline Characteristics of NYHA Class IV Patients by Treatment Group

Characteristic CRT-D (Nϭ83) CRT (Nϭ79) OPT (Nϭ55) P Age, y, meanϮSD 64.8Ϯ11.4 66.9Ϯ10.4 69.0Ϯ8.4 0.07 Gender, n (%) Male 50 (60) 56 (71) 31 (56) 0.18 Female 33 (40) 23 (29) 24 (44) ⅐⅐⅐ Duration of HF, y, meanϮSD 5.1Ϯ4.7 5.6Ϯ5.3 5.4Ϯ4.4 0.79 LVEF, %, meanϮSD 21.5Ϯ6.8 20.2Ϯ5.8 20.7Ϯ7.3 0.42 LVEDD, meanϮSD 70.1Ϯ9.5 71.2Ϯ8.9 71.7Ϯ12.4 0.68 Resting heart rate, bpm, meanϮSD 77.7Ϯ13.4 75.9Ϯ13.1 76.0Ϯ12.9 0.62 Systolic blood pressure, mm Hg, meanϮSD 109.1Ϯ15.9 112.2Ϯ18.8 110.0Ϯ19.9 0.54 Diastolic blood pressure, mm Hg, meanϮSD 66.8Ϯ10.4 65.5Ϯ10.3 63.8Ϯ9.5 0.23 6-Minute walk distance, m, meanϮSD 177.9Ϯ112.6 161.2Ϯ98.2 158.5Ϯ102.6 0.50 QRS width, ms, meanϮSD 161.4Ϯ27.2 160.1Ϯ26.2 161.9Ϯ26.9 0.93 Ischemic, n (%) Yes 47 (57) 51 (65) 36 (65) 0.47 No 36 (43) 28 (35) 19 (35) ⅐⅐⅐ History of diabetes, n (%) No 45 (54) 43 (54) 24 (44) 0.39 Yes 38 (46) 36 (46) 31 (56) ⅐⅐⅐ Conduction disorder, n (%) Left bundle-branch block 59 (71) 49 (62) 40 (73) 0.46 Nonspecific intraventricular 18 (22) 18 (23) 10 (18) ⅐⅐⅐ Right bundle-branch block 6 (7) 12 (15) 5 (9) ⅐⅐⅐ ACE inhibitor or angiotensin II, n (%) Yes 69 (83) 63 (80) 45 (82) 0.86 No 14 (17) 16 (20) 10 (18) ⅐⅐⅐ ␤-Blocker, n (%) Yes 47 (57) 40 (51) 25 (45) 0.43 No 36 (43) 39 (49) 30 (55) ⅐⅐⅐ ACE inhibitor/angiotensin II ϩ ␤-blocker, n (%) Yes 44 (53) 43 (54) 34 (62) 0.57 No 39 (47) 36 (46) 21 (38) ⅐⅐⅐ Loop diuretics, n (%) Yes 83 (100) 78 (99) 55 (100) 0.42 No ⅐⅐⅐ 1 (1) ⅐⅐⅐ ⅐⅐⅐ Digoxin, n (%) Yes 56 (67) 61 (77) 39 (71) 0.38 No 27 (33) 18 (23) 16 (29) ⅐⅐⅐ Spironolactone, n (%) Yes 52 (63) 39 (49) 31 (56) 0.23 No 31 (37) 40 (51) 24 (44) ⅐⅐⅐ LVEDD indicates left ventricular end-diastolic diameter; ACE, angiotensin-converting enzyme.

Time to mortality or HF hospitalization was significantly Figure 2 depicts the time to sudden death and time to HF death improved by both CRT (HR, 0.57; Pϭ0.01) and CRT-D in each of the 3 treatment groups of NYHA class IV patients. The (HR, 0.49; Pϭ0.001) compared with OPT. Analysis of time to sudden death was significantly prolonged by CRT-D time to HF death or HF hospitalization demonstrated a compared with OPT (HR, 0.27; Pϭ0.03). The time to sudden death significant benefit of CRT-D (CRT-D versus OPT: HR, was not affected by CRT compared with OPT (HR, 0.81; Pϭ0.64). 0.58; Pϭ0.03). There was a strong trend for a benefit of The time to HF death was not significantly altered by either CRT CRT versus OPT in time to HF death or HF hospitalization (HR, 0.68; Pϭ0.28) or CRT-D (HR, 0.79; Pϭ0.48) compared with (CRT versus OPT: HR, 0.64; Pϭ0.07). CRT did not differ OPT. CRT was not significantly different from CRT-D for either from CRT-D for any of these endpoints. time to sudden death or HF death.

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 208 Circulation January 16, 2007

Figure 2. Time to mode of death by treatment arm for NYHA class IV. A, Time to sudden death. B, Time to HF death. FU indicates follow-up.

Although baseline covariates were not significantly different among the 3 treatment groups, an analysis ad- justed for covariates was performed for all the end points. The adjusted analysis differed from the unadjusted analy- sis in only 1 comparison, which became significant in the adjusted analysis: time to all-cause death in CRT versus OPT (HR, 0.58; 95% CI, 0.35 to 0.96; Pϭ0.04). The proportionality assumption appeared to be met in all analyses except the sudden death and the HF death analyses. At 1 year in the NYHA class IV patients, 44% of OPT patients died, compared with 36% of CRT and 30% of CRT-D patients. Table 3 shows mode of death over 2 years, adjusted for follow-up time: 62% of OPT subjects Figure 1. NYHA class IV endpoints by treatment arm. A, Pri- died compared with 45% of CRT and 55% of CRT-D mary time to all-cause death or hospitalization. B, Secondary time to all-cause death. C, Time to all-cause death or HF hospi- subjects. Total mortality, HF mortality, and sudden cardiac talization. FU indicates follow-up. death, as expected, were higher in NYHA class IV subjects than in class III subjects: 29%, 26%, and 41% of NYHA

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Lindenfeld et al Effects of CRT on Class IV Heart Failure Patients 209

TABLE 3. Mode of Death at 2 Years TABLE 5. Device Implantation Attempts and Success

OPT CRT CRT-D Successful Unsuccessful Mode of Death (n/Nϭ55/253) (n/Nϭ79/538) (n/Nϭ83/512) Successful on After Ͼ1 After Ͼ1 No All death Treatment First Attempt Attempt Attempt Attempt ϭ NYHA class IV 29 (62) 34 (45) 34 (55) Class IV (N 162) ϭ NYHA class III 46 (26) 92 (23) 68 (20) CRT-D (N 83) 64 (77) 6 (7) 10 (12) 3 (4) ϭ HF CRT (N 79) 63 (80) 2 (3) 14 (18) 0 (0) ϭ NYHA class IV 14 (29) 17 (26) 21 (41) Class III (N 1050) ϭ NYHA class III 18 (10) 34 (10) 29 (8) CRT-D (N 512) 440 (86) 31 (6) 37 (7) 4 (1) ϭ Sudden cardiac death CRT (N 538) 447 (83) 27 (5) 57 (11) 7 (1) NYHA class IV 8 (25) 11 (16) 4 (9) Values expressed as n (%). NYHA class III 10 (5) 36 (9) 13 (5) Other walk test did not improve significantly when CRT/CRT-D NYHA class IV 7 (28) 6 (12) 9 (18) was compared with OPT, but there were only 12 patients in NYHA class III 18 (14) 22 (6) 26 (8) the OPT group. However, quality-of-life score and percent Values are expressed as n (%). of patients who improved at least 1 NYHA class improved significantly in the CRT/CRT-D patients compared with OPT. class IV OPT, CRT, and CRT-D patients died of HF over No NYHA class IV patients died during the implantation 2 years, but time to HF death was not significantly hospitalization. The duration of implantation hospitaliza- different (Figure 2B). In the OPT, CRT, and CRT-D tion did not differ between NYHA class IV and III groups, 25%, 16%, and 9% of NYHA class IV patients patients, both of whom averaged Ϸ4 days. Implantation died of sudden cardiac death. Time to sudden death was success rates by NYHA class and treatment group are prolonged by CRT-D compared with OPT, but CRT and shown in Table 5. Success rates were slightly higher in CRT-D were not significantly different (Figure 2A). In NYHA class III subjects compared with class IV subjects OPT patients, HF was the cause of death in 48% and 39% in the CRT patients (88% versus 83%, respectively), and of all deaths in the NYHA class IV and III subjects significantly higher in the CRT-D patients (92% versus respectively. Also in OPT patients, sudden death ac- 84%, Pϭ0.04, respectively). counted for 28% and 22% of all deaths in the NYHA class IV and III groups, respectively. Adverse Events By the time of the 1-month visit, 67% of NYHA class IV Overall, 40% (nϭ87) of NYHA class IV patients had a severe CRT-D and CRT patients improved at least 1 NYHA class, adverse event within 1 year of randomization compared with compared with only 31% of OPT patients. Of those patients 21% (nϭ275) of NYHA class III patients (PϽ0.0001). For who improved, there was a significant treatment benefit with NYHA class III patients, the percent of patients with a severe regard to mortality (CRT-D versus OPT: HR, 0.43; PՅ0.01; adverse event was consistent across all 3 treatment arms; CRT versus OPT: HR, 0.41; Pϭ0.02). No mortality benefit however, for NYHA class IV patients, the percent varied was seen in those who did not improve in NYHA class status across OPT, CRT, and CRT-D arms (47%, 43%, and 33%, by 1 month. respectively; all PϽ0.015). Table 4 shows the baseline to 6-month changes in 6-minute walk, quality of life, and NYHA class for Crossovers combined CRT-D/CRT versus OPT patients. The 6-minute There were 19 patients (35%) in the OPT class IV group that “crossed over” to device therapy during the course of the TABLE 4. Functional Capacity in NYHA Class IV Patients: study. Seven (9%) of the CRT and 4 (5%) of CRT-D class IV Change From Baseline to 6 Months patients crossed over. In class III, the crossovers were 37%, 11%, and 3% for OPT, CRT, and CRT-D, respectively. There Indicator Number Median (Q1, Q3) % Improved P were no significant differences in crossovers between NYHA 6-Minute walk class III and IV. CRT/CRT-D 69 45.6 (Ϫ15.2, 106.4) ⅐⅐⅐ 0.55 OPT 12 45.6 (Ϫ22.3, 60.9) ⅐⅐⅐ ⅐⅐⅐ Discussion Quality of life Our present analysis of the COMPANION trial demonstrates CRT/CRT-D 109 Ϫ25.0 (Ϫ44.0, Ϫ8.0) ⅐⅐⅐ Ͻ0.01 that both CRT and CRT-D improve the time to all-cause mortality or first hospitalization in patients with NYHA class OPT 29 Ϫ4.0 (Ϫ20, 9.0) ⅐⅐⅐ ⅐⅐⅐ IV HF. Both CRT and CRT-D also improved time to NYHA all-cause mortality or HF hospitalization. There was a non- Ͻ CRT/CRT-D 119 ⅐⅐⅐ 78 0.01 significant trend toward improved time to all-cause mortality OPT 27 ⅐⅐⅐ 52 ⅐⅐⅐ in both groups, and, in an analysis adjusted for covariates, Q indicates quartile. CRT did demonstrate a significant benefit for time to all-

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 210 Circulation January 16, 2007

TABLE 6. Demographics and Mortality in Studies of Patients With Advanced HF

REMATCH17 COMPANION FIRST18 PROMISE17 RALES17 BEST (IV)16 COPERNICUS17 ORAL IV (OPT) NYHA class III/IV III/IV IV IV IV IV IV IV No. 471 527 841 ⅐⅐⅐ 1133 15 46 217 Age, y 65 64 65 63 64 68 68 67 LVEF, % 18 21 25 21 20 17 17 21 Systolic blood pressure, mm Hg 105 115 122 113 125 107 100 110 ACE inhibitor or ARB, % 84 ⅐⅐⅐ 95 88 97 66 53 82 ␤-Blocker, % ⅐⅐⅐ ⅐⅐⅐ 11 ⅐⅐⅐ randomized 1:1 to ␤-blocker 34 13 48 Mortality, y, % All 62 ⅐⅐⅐ ⅐⅐⅐ 22 ⅐⅐⅐ 51 76 44 Treatment ⅐⅐⅐ 37 18 ⅐⅐⅐ 11.5 ⅐⅐⅐ ⅐⅐⅐ ⅐⅐⅐ Placebo ⅐⅐⅐ 30 26 ⅐⅐⅐ 18.5 ⅐⅐⅐ ⅐⅐⅐ ⅐⅐⅐ FIRST indicates the Flolan International Randomized Survival Trial; PROMISE, Prospective Randomized Milrinone Survival Evaluation trial; RALES, Randomized Aldactone Evaluation Study; BEST, Beta-Blocker Evaluation of Survival Trial; COPERNICUS, Carvedilol Prospective Randomized Cumulative Survival Study; REMATCH, Randomized Evaluation of Mechanical Assistance in Treatment if Chronic Heart Failure; ACE, angiotensin-converting enzyme; and ARB, angiotensin receptor blocker.

cause mortality. CRT-D was associated with a significant improvement in time to HF deaths and HF hospitalization None of the large trials of prophylactic implantable with a strong trend for improvement in the same end point cardiac defibrillators alone have included patients with with CRT. A significant survival benefit was demonstrated NYHA class IV HF.10–14 Our data demonstrate that CRT-D for both the CRT and CRT-D patients who had symptomatic decreases sudden death (Pϭ0.03) in class IV patients. improvement by at least 1 NYHA class at 1 month. Improve- Similar to the entire cohort data, CRT-D improved sudden ment in the primary end point was demonstrated for both death, although there was no significant difference be- CRT and CRT-D despite an 18% and 12% rate of failed tween CRT and CRT-D.3 Salutary effects of CRT-D on implantation for CRT and CRT-D, respectively. Although all-cause or HF hospitalizations as measured in the com- only CRT-D prolonged the time to sudden death, our data do bined end points, which include mortality, was similar to not demonstrate a significant difference between CRT and results with CRT alone. CRT-D in any of the end points examined. These data Although COMPANION NYHA class IV subjects were represent the first demonstration that NYHA class IV patients ambulatory at the time of randomization, they clearly had benefit from either CRT or CRT-D. These results are partic- severe HF. Our NYHA class IV OPT subjects were very ularly impressive given the high rate of crossovers in the OPT comparable to subjects in other studies with ambulatory group. but advanced HF (Table 6). LVEF was 21% in the present Only a few trials of CRT have included patients with study, which was equivalent to the LVEF of the NYHA NYHA class IV HF.1,5,7 The number of NYHA class IV class IV subjects in the Beta-Blocker Evaluation of Sur- patients was Ͻ200 in all these trials and only 1 other trial vival Trial (BEST) and Carvedilol Prospective Random- randomized patients before device implantation was at- ized Cumulative Survival Trial (COPERNICUS), a trial of tempted.5 None of the trials have reported NYHA class IV carvedilol in subjects with advanced HF.15,16 COMPAN- patients separately, although 1 trial of CRT alone reported ION class IV subjects had a systolic blood pressure of no significant difference in outcomes between class III and 110 mm Hg, lower than the 123 mm Hg in COPERNICUS IV patients.5 There were only 50 NYHA class IV subjects, and 117 mm Hg in the NYHA class IV subjects in however, and wide confidence intervals. BEST.15,16 Indeed, the 1-year mortality of 44% in the Progressive HF was the cause of death in 39% of class NYHA class IV OPT group was exceeded only by the III OPT patients and 48% of class IV OPT patients in the 1-year mortality of 62% for patients in the Flolan Interna- present study at 2 years. This compares to 26% and 56% of tional Randomized Survival Trial (FIRST) and the 1-year deaths caused by progressive HF in NYHA class III and IV mortality of 51% and 76% of noninotrope- and inotrope- in Metoprolol Controlled-Release Randomized Interven- dependent patients in the Randomized Evaluation of Me- tion Trial in Heart Failure (MERIT-HF).9 It is likely that chanical Assistance in Treatment of Chronic Heart Failure progressive HF was a more common cause of death in (REMATCH).17,18 Thus, COMPANION NYHA class IV NYHA class III patients in COMPANION than in patients were more ill than patients in previous studies of MERIT-HF because of the requirement for a HF hospital- NYHA class IV ambulatory patients, but slightly less ill ization in the preceding 12 months in COMPANION, and perhaps more stable than patients in FIRST, more than which led to a more advanced degree of HF in COMPAN- half of whom required intravenous inotropes, or those in the noninotrope group of REMATCH–NYHA class IV ION class III patients versus those in MERIT-HF. patients who were felt to need cardiac transplantation but

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Lindenfeld et al Effects of CRT on Class IV Heart Failure Patients 211

had an absolute contraindication. It is important to recog- Underwood J, Pickering F, Truex C, McAtee P, Messenger J; MIRACLE nize that the class IV patients in COMPANION were Study Group. Multicenter InSync randomized clinical evaluation. Cardiac resynchronization in chronic heart failure. N Engl J Med. 2002;346: ambulatory outpatients who had had at least 1 HF hospi- 1845–1853. talization or equivalent in the 12 months prior to random- 2. Auricchio A, Kloss M, Trautmann SI, Rodner S, Klein H. Exercise ization. They could not be enrolled, however, if they had performance following cardiac resynchronization therapy in patients with had a hospitalization for HF or intravenous inotropic or heart failure and ventricular conduction delay. Am J Cardiol. 2002;89: 198–203. vasoactive therapy in excess of 4 hours in the 30 days 3. Bristow MR, Saxon LA, Boehmer J, Krueger S, Kass DA, DeMarco T, before randomization. They also did not have a recent Carson P, DiCarlo L, DeMets D, White BG, DeVries DW, Feldman AM; coronary intervention, refractory atrial arrhythmias, or a Comparison of Medical Therapy, Pacing, and Debibrillation in Heart recent myocardial infarction. Failure (COMPANION) Investigators. Cardiac-resynchronization therapy To date, it has been uncertain whether patients with with or without an implantable defibrillator in advanced chronic heart failure. N Engl J Med. 2004;350:2140–2150. NYHA class IV CHF benefit from either CRT or CRT-D. 4. Cazeau S, Leclercq C, Lavergne T, Walker S, Varma C, Linde C, Arguments have been made that a procedure in these very Garrigue S, Kappenberger L, Haywood GA, Santini M, Bailleul C, ill patients may destabilize the HF and thus cause pro- Daubert JC; Multisite Stimulation in Cardiomopathies (MUSTIC) Study longed hospitalization and increased mortality. Further- Investigators. Effects of multisite biventricular pacing in patients with heart failure and intraventricular conduction delay. N Engl J Med. 2001; more, it has been argued that implantable cardioverter- 344:873–880. defibrillator therapy may not be warranted as these patients 5. Cleland JG, Daubert JC, Erdmann E, Freemantle N, Gras D, Kappen- primarily die of progressive HF. However, the absolute berger L, Tavazzi L; Cardiac Resynchronization-Heart Failure number of lives saved and hospitalizations prevented were (CARE-HF) Study Investigators. The effect of cardiac resynchronization on morbidity and mortality in heart failure. N Engl J Med. 2005;352: much greater than in a less-ill population. Our data argue 1539–1549. strongly that CRT devices can be employed with an 6. Higgins SL, Hummel JD, Niazi IK, Giudici MC, Worley SJ, Saxon LA, excellent risk-benefit ratio in class IV HF patients who did Boehmer JP, Higginbotham MB, DeMarco T, Foster E, Yong PG. not require a hospitalization for HF in the preceding Cardiac resynchronization therapy for the treatment of heart failure in patients with intraventricular conduction delay and malignant ventricular month. Furthermore, our data demonstrate that an implant- tachyarrhythmias. J Am Coll Cardiol. 2003;42:1454–1459. able cardioverter-defibrillator added to CRT may benefit 7. Young JB. Cardiac transplantation 2003. Curr Opin Cardiol. 2003;18: class IV patients in that it may produce the same incre- 127–128. mental reduction in sudden death noted in the entire 8. Bristow MR, Feldman AM, Saxon LA. Heart failure management using implantable devices for ventricular resynchronization: Comparison of 19 COMPANION cohort. Thus, CRT and CRT-D are both Medical Therapy, Pacing, and Defibrillation in Chronic Heart Failure beneficial in altering mortality and morbidity in this very (COMPANION) trial. COMPANION Steering Committee and COM- ill population of NYHA class IV patients. PANION Clinical Investigators. J Card Fail. 2000;6:276–285. 9. Merit-HF Study Investigators. Effect of metoprolol CR/XL in chronic heart failure: Metoprolol CR/XL Randomised Intervention Trial Limitations in-Congestive Heart Failure (MERIT-HF). Lancet. 1999;353:2001–2007. The present study was not stratified by NYHA class, but the 10. Moss AJ, Hall WJ, Cannom DS, Daubert JP, Higgins SL, Klein H, Levine treatment groups were similar in all baseline demographics. A JH, Saksena S, Waldo AL, Wilber D, Brown MW, Heo M. Improved proportionality assumption was violated in the time to sudden survival with an implanted defibrillator in patients with coronary disease death and time to HF death analyses. The present study is at high risk for ventricular arrhythmia. Multicenter Automatic Defi- brillator Implantation Trial Investigators. N Engl J Med. 1996;335: limited by the retrospective nature of the evaluation. The 1933–1940. present study, however, is the only evaluation of CRT or 11. Moss AJ, Zareba W, Hall WJ, Klein H, Wilber DJ, Cannom DS, Daubert CRT-D in NYHA class IV patients alone and is the largest JP, Higgins SL, Brown MW, Andrews ML; Multicenter Automatic group yet studied. Debibrillator Implantation Trial II Investigators. Prophylactic implan- tation of a defibrillator in patients with myocardial infarction and reduced ejection fraction. N Engl J Med 2002;346:877–883. Disclosures 12. Bardy GH, Lee KL, Mark DB, Poole JE, Packer DL, Boineau R, Dr Lindenfeld has received speaking honoraria and consulting fees Domanski M, Troutman C, Anderson J, Johnson G, McNulty SE, Clapp- from Boston Scientific and Medtronic, as well as consulting fees Channing N, Davidson-Ray LD, Fraulo ES, Fishbein DP, Luceri RM, Ip from St. Jude. Dr Feldman is a consultant to Boston Scientific. Dr JH; Sudden Cardiac Death in Heart Failure Trial (SCD-HeFT) Investi- Saxon has received consultant fees and research support from Boston gators. Amiodarone or an implantable cardioverter-defibrillator for con- Scientific and Medtronic. Dr Boehmer has received consulting fees gestive heart failure. N Engl J Med 2005;352:225–237. from Boston Scientific and research funding from Boston Scientific 13. Kadish A, Dyer A, Daubert JP, Quigg R, Estes NA, Anderson KP, and Medtronic. Dr Ghali has received research funding from Calkins H, Hoch D, Goldberger J, Shalaby A, Sanders WE, Schaechter A, Guidant. Dr Anand has received consulting fees and research funds Levine JH; Defibrillators in Non-Ischemic Cardiomyopathy Treatment from Boston Scientific. Dr Steinberg has received speaking hono- Evaluation (DEFINITE) Investigators. Prophylactic defibrillator implan- raria, consulting fees, and research support from Boston Scientific; tation in patients with nonischemic dilated cardiomyopathy. N Engl J Med speaking honoraria and research support from Medtronic; and 2004;350:2151–2158. research support from St. Jude. Dr Jaski has received speaking 14. Buxton AE, Lee KL, Fisher JD, Josephson ME, Prystowsky EN, Hafley G. A randomized study of the prevention of sudden death in patients with honoraria and consulting fees from Boston Scientific. Dr DeMarco is coronary artery disease. Multicenter Unsustained Tachycardia Trial a consultant for Boston Scientific. P. Yong, E. Galle, and F. Ecklund Investigators. N Engl J Med. 1999;341:1882–1890. are employees of Boston Scientific. Dr Bristow has received con- 15. Packer M, Fowler MB, Roecker EB, Coats AJ, Katus HA, Krum H, sulting fees from Boston Scientific. Drs Carson and Singh report no Mohacsi P, Rouleau JL, Tendera M, Staiger C, Holcslaw TL, conflicts. Amann-Zalan I, DeMets DL; Carvedilol Prospective Randomized Cumu- lative urvival (COPERNICUS Study Group. Effect of carvedilol on the References morbidity of patients with severe chronic heart failure: results of the 1. Abraham WT, Fisher WG, Smith AL, Delurgio DB, Leon AR, Loh E, carvedilol prospective randomized cumulative survival (COPERNICUS) Kocovic DZ, Packer M, Clavell AL, Hayes DL, Ellestad M, Trupp RJ, study. Circulation. 2002;106:2194–2199.

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 212 Circulation January 16, 2007

16. Anderson JL, Krause-Steinrauf H, Goldman S, Clemson BS, Domanski 18. Califf RM, Adams KF, McKenna WJ, Gheorghiade M, Uretsky BF, MJ, Hager WD, Murray DR, Mann DL, Massie BM, McNamara DM, McNulty SE, Darius H, Schulman K, Zannad F, Handberg-Thurmond E, Oren R, Rogers WJ; Beta-Blocker Evaluation of Survival Trial (BEST) Harrell FE Jr., Wheeler W, Soler-Soler J, Swedberg K. A randomized Investigators. Failure of benefit and early hazard of bucindolol for class controlled trial of epoprostenol therapy for severe congestive heart IV heart failure. J Card Fail. 2003;9:266–277. failure: The Flolan International Randomized Survival Trial (FIRST). Am 17. Stevenson LW, Miller LW, Desvigne-Nickens P, Ascheim DD, Parides Heart J. 1997;134:44–54. MK, Renlund DB, Oren RM, Krueger SK, Costanzo MR, Wann LS, 19. Carson P, Anand I, O’Conner C, Jaski B, Steinberg J, Lwin A, Lindenfeld Levitan RG, Mancini D; REMATCH Investigators. Left ventricular assist device as destination for patients undergoing intravenous inotropic J, Ghali J, Barnet JH, Feldman AM, Bristow MR. Mode of death in therapy: a subset analysis from REMATCH (Randomized Evaluation of advanced heart failure: the Comparison of Medical, Pacing, and Defibril- Mechanical Assistance in Treatment of Chronic Heart Failure). Circu- lation Therapies in Heart Failure (COMPANION) Trial. J Am Coll lation. 2004;110:975–981. Cardiol. 2005;46:2329–2334.

CLINICAL PERSPECTIVE Cardiac resynchronization therapy (CRT) alone or in combination with an implantable defibrillator (CRT-D) improves quality of life, exercise capacity, and survival in patients with New York Heart Association (NYHA) class III to IV heart failure (HF) and systolic dysfunction. However, very few patients with NYHA class IV HF have been studied, and there are concerns that the implantation procedure might destabilize the HF and that either CRT or CRT-D may not significantly prolong life in these very ill patients who most often die of progressive HF. In the COMPANION trial, patients with systolic dysfunction and NYHA class III and IV HF were randomized to receive optimal medical therapy, optimal medical therapy ϩ CRT, or optimal medical therapy ϩ CRT-D. There were 217 NYHA class IV patients randomized. The patients were relatively clinically stable in that they could not be randomized if they had had a hospital admission in the previous 30 days. In these NYHA class IV patients, both CRT and CRT-D prolonged the time to all-cause death and hospitalization. There was also a strong trend for an improvement in all-cause death by both CRT and CRT-D. Both CRT and CRT-D improved the time to all-cause death and HF hospitalizations. The duration of the implantation hospitalization was not significantly different between NYHA class III and class IV subjects. Thus, in ambulatory, clinically stable NYHA class IV patients with systolic dysfunction, both CRT and CRT-D may provide a clinical benefit.

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Hypertension

Promotion of Faster Weight Gain in Infants Born Small for Gestational Age Is There an Adverse Effect on Later Blood Pressure?

Atul Singhal, MD, MRCP; Tim J. Cole, PhD, ScD; Mary Fewtrell, MD, MRCP; Kathy Kennedy, MSc; Terence Stephenson, DM, FRCP, FRPCH; Alun Elias-Jones, FRCP, FRCPCH; Alan Lucas, MD, FRCP, FMed Sci

Background—Being born small for gestational age is associated with later risk factors for cardiovascular disease, such as high blood pressure. Promotion of postnatal growth has been proposed to ameliorate these effects. There is evidence in animals and infants born prematurely, however, that promotion of growth by increased postnatal nutrition increases rather than decreases later cardiovascular risk. We report the long-term impact of growth promotion in term infants born small for gestational age (birth weight Ͻ10th percentile). Methods and Results—Blood pressure was measured at 6 to 8 years in 153 of 299 (51%) of a cohort of children born small for gestational age and randomly assigned at birth to receive either a standard or a nutrient-enriched formula. The enriched formula contained 28% more protein than standard formula and promoted weight gain. Diastolic and mean (but not systolic) blood pressure was significantly lower in children assigned to standard compared with nutrient-enriched formula (unadjusted mean difference for diastolic blood pressure, Ϫ3.2 mm Hg; 95% CI, Ϫ5.8 to Ϫ0.5; Pϭ0.02) independent of potential confounding factors (adjusted difference, Ϫ3.5 mm Hg; Pϭ0.01). In observational analyses, faster weight gain in infancy was associated with higher later blood pressure. Conclusions—In the present randomized study targeted to investigate the effect of early nutrition on long-term cardiovascular health, we found that a nutrient-enriched diet increased later blood pressure. These findings support an adverse effect of relative “overnutrition” in infancy on long-term cardiovascular disease risk, have implications for the early origins of cardiovascular disease hypothesis, and do not support the promotion of faster weight gain in infants born small for gestational age. (Circulation. 2007;115:213-220.) Key Words: blood pressure Ⅲ follow-up studies Ⅲ gestational age Ⅲ nutrition

nfants born small for gestational age (SGA) are at long- Clinical Perspective p 220 Iterm risk of atherosclerotic cardiovascular disease (CVD)1 and its risk factors (such as high blood pressure2), but whether percentile crossing for weight was associated with higher manipulation of nutrition and growth in infancy can influence later blood pressure,13–15 insulin resistance,16 and, in 2 sys- these effects is controversial. tematic reviews, an increased risk of later obesity.17,18 Pro- Previously, epidemiological associations between greater motion of growth in infancy may also be detrimental for later weight at 1 year of age and a lower risk of later CVD have led CVD risk. For instance, nutritional supplementation with to suggestions that promotion of infant growth may be dried milk powder in infancy has been shown to have an beneficial for later CVD.3,4 Paradoxically, however, interven- adverse long-term effect on, or program, higher blood pres- tion studies in animals indicate that promotion of early sure in adulthood.19 However, most previous studies in growth by a higher nutrient intake increases later cardiovas- humans are observational, and there are few randomized cular risk.5–8 Recent data in humans are also conflicting. For studies with long-term prospective follow-up that can inform instance, faster growth in early life was not associated with the optimal management of infants born SGA. mortality from CVD in a Finnish cohort9 but was associated In the present study, we investigated the role of early with later cardiovascular risk factors such as endothelial nutrition and weight gain on later CVD risk using an dysfunction,10 insulin resistance,11 and dislipidemia12 in pre- experimental approach. We studied term infants born SGA, term infants. Similarly, in infants born at term, upward whom it was ethical at the time to assign randomly to a

Received February 8, 2006; accepted October 20, 2006. From the Medical Research Council Childhood Nutrition Research Centre (A.S., M.F., K.K., A.L.) and Centre for Pediatric Epidemiology and Biostatistics (T.J.C.), Institute of Child Health, London; Academic Division of Child Health, University Hospital, Nottingham (T.S.); and Leicester General Hospital National Health Service Trust, Gwendolen Road, Leicester (A.E.-J.), UK. Correspondence to Dr Atul Singhal, Institute of Child Health, 30 Guilford St, London, UK. E-mail [email protected] © 2007 American Heart Association, Inc. Circulation is available at http://www.circulationaha.org DOI: 10.1161/CIRCULATIONAHA.106.617811

213

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 214 Circulation January 16, 2007

nutrient-enriched formula that promoted growth (which was TABLE 1. Composition of the Trial Diets* thought to be beneficial)20 or to a standard formula. Our Term Formula Nutrient-Enriched Formula principal and a priori outcome was later blood pressure, an Energy, kJ 284 301 established risk factor for CVD, known to be programmed by Energy, kcal 68 72 21 22 early nutrition and to track into adult life. Protein, g 1.4 1.8 Casein 0.6 0.7 Methods Whey 0.9 1.1 Carbohydrate, g 7.0 7.2 Participants Lactose 7.0 6.2 Research nurses recruited infants from 5 hospitals in Cambridge, Notting- Maltodextrin 0.0 1.0 ham, and Leicester in the United Kingdom between 1993 and 1995.20 Infants meeting the eligibility criteria (Ն37 weeks of gestation, with birth Fat, g 3.8 4.0 weight below the 10th percentile for gestation and sex according to UK Minerals growth charts, and free of congenital abnormalities) were enrolled as soon Calcium, mg 39 70 as possible after birth.20 Gestational age was determined from the last Phosphorus, mg 27 35 menstrual period or first trimester ultrasound scan. Infants of mothers who had unequivocally decided to formula-feed were randomly assigned (at a Chloride, mg 45 45 mean age of 4 days; nϭ299) to receive a standard, term formula (Farley’s Sodium, mg 17 22 Ostermilk, Farley’s Health Products, a division of HJ Heinz Company Ltd, Potassium, mg 57 78 Stockley Park Uxbridge, UK) or a nutrient-enriched formula (Farley’s Zinc, mg 0.3 0.6 PremCare). The randomization schedule, generated by random permuted blocks and prepared by a member of the team not involved in data Iron, mg 0.6 0.6 collection, was assigned with the use of sealed envelopes.20 A reference Copper, ␮g4257 group of breast-fed infants (Ն37 weeks of gestation and birth weight Ͻ10th Iodine, ␮g 4.5 4.5 ϭ 20 percentile) was also recruited (n 175). Informed written consent was Manganese, ␮g 3.4 5.0 obtained from the parent, and the study was approved by the research ethics Vitamins committees of each participating center. Retinol, ␮g 100 100 Study Design Thiamine, ␮g4295 Formulas were assigned until age 9 months. Anthropometric and demo- Riboflavin, ␮g 55 100 20 graphic information was obtained as described. Social class was based on B6, ␮g 35 80 the occupation of the parent providing the main financial support for the B12, ␮g 0.1 0.2 family according to the registrar general’s classification. Folate, ␮g 3.4 25 C, mg 6.9 15 Composition of Formulas D, ␮g 1.0 1.3 Both formulas fulfilled the European Community directive for the composition of formulas for term infants. Nutrient-enriched formula E, mg 0.5 1.5 contained 28% more protein than standard formula and also more K, ␮g 2.7 6.0 minerals, trace elements, and vitamins to support the projected Biotin, ␮g 1.0 1.1 increase in weigh and length gain (Table 1). Pantothenic acid, mg 0.2 0.4 Carnitine, mg 0.0 1.1 Follow-Up Taurine, mg 5.0 5.1 All children who were alive, traceable, and willing to participate were reviewed in their homes between 1999 and 2002. Ethical Choline, mg 4.8 5.1 approval for the follow-up was obtained from research ethics Osmolality, mOsm/L 300 280 committees at each center, and written consent was obtained from *Values are per 100 mL. children’s parents or guardians. The derivation of the sample followed up at 6 to 8 years of age is given in the Figure. All researchers and study participants were blinded to the original Statistical Analysis dietary assignments. Diastolic and systolic blood pressures were Sample size was based on our previous study21 and estimated to measured with an automated oscillometric device (Accutorr, Data- detect a half SD difference in mean diastolic blood pressure and scope Corp, NJ), which has been recommended by the European MAP between randomized formula-fed groups at 80% power and 5% Society of Hypertension and validated previously with protocols significance21 (requiring Ϸ128 subjects). A subset of 153 formula- from the US Association for the Advancement of Medical Instru- fed children agreed to participate. mentation and the British Hypertension Society.23,24 Mean arterial Our outcomes were diastolic blood pressure and MAP, which have blood pressure (MAP) was derived automatically from the shape of been shown previously to be programmed by early nutrition.21 the blood pressure curve according to the following formula: Secondary outcomes were systolic blood pressure and size at follow-up. Comparisons between (randomized) groups were made systolic pressureϩ(2 ϫ diastolic pressure) with Student t test for continuous variables and ␹2 test for categorical MAPϭ 3 variables. Multiple linear regression analyses were used to adjust differences in outcome between randomized groups for possible Measurements were made, after 10 minutes of rest, on 2 occasions 5 baseline differences. Initial analyses were on an intention-to-treat minutes apart in the left arm with the use of an appropriately sized cuff,21 basis and included all children followed up irrespective of whether and the average value was obtained. Height and weight were measured by they complied with the initial diet. Secondary analyses were con- trained observers using a portable stadiometer accurate to 1 mm (Holtain fined to those who complied with the dietary intervention. Instruments Ltd, Crymych, UK) and electronic scales accurate to 0.1 kg Linear regression analyses were used to assess the influence of (Seca, CMS Weight Equipment Ltd, London, UK), respectively. growth rate on later blood pressure. In 6 separate regression models,

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Singhal et al Early Nutrition and Later Blood Pressure 215

Infants with birth weight <10th centile, gestation ≥ 37 wk; no congenital malformations or chromosomal abnormality

Mother formula-feeding infant

299 infants randomly assigned to Mother breast-feeding infant

Standard term formula Nutrient-enriched formula Breast-fed reference group (n = 147) (n = 152) (n = 175)

Completed 9 months Completed 9 months Withdrawn by 9 mo: n = 28 of trial diet n=126 of trial diet n=121 Withdrawn by 9 mo n = 21 Withdrawn by 9 mo n = 31 1Related to formula n = 8 1Related to formula n = 14 Not related to formula n = 13 Not related to formula n = 17

Seen at 6-8 yrs Seen at 6-8 yrs Seen at 6-8 yrs n = 83 n = 70 n = 97

2Not traced n = 46 2Not traced n = 49 2Not traced n = 40 Refused n = 15 Refused n = 27 Refused n = 23 Moved abroad n = 3 Moved abroad n = 3 Moved abroad n = 13 3Withdrawn n = 0 3Withdrawn n = 2 3Withdrawn n = 1 Died n = 0 Died n = 1 Died n = 1

1Withdrawn from the study by the mothers for reasons considered by research nurse to be related to formula (e.g. constipation, vomiting, or feeling 'unsettled' or 'not satisfied' with formula). 2Not traced or did not reply to initial letter; 3withdrawn by family doctor (reasons not given).

Derivation of sample.

growth was expressed as the change in weight or length SD score (z 2). A greater percentage of boys was reviewed in the score) (calculated with the use of percentiles for British infants) from nutrient-enriched formula group compared with the standard- birth to age 9 months, 9 to 18 months, and 18 months to 6 to 8 years. formula group (Table 3); this would not be expected to Regression analyses were adjusted for potential confounding factors (age, sex, z score for weight and height at 6 to 8 years, and social class). confound our findings because diastolic blood pressure, Assumptions of linearity were checked in exploratory analyses with the MAP, and systolic blood pressure did not differ by sex (eg, use of nonparametric statistics and visual inspection of scatterplots. mean blood pressureϭ78 [SD 8] mm Hg in boys versus 78 The authors had full access to the data and take full responsibility [SD 8] mm Hg in girls; Pϭ0.6). There were no other for the integrity of the data. All authors have read and agree to the significant differences between randomly assigned dietary manuscript as written. groups in anthropometric, socioeconomic, or demographic Results factors at baseline or at follow-up (Table 3) or in maternal age, anthropometry, education, and incidence of smoking or Main Effect: Comparison of Groups Randomly high blood pressure during pregnancy (data not shown). Assigned to Different Formulas There were no adverse events reported (reasons for partici- Children reviewed at 6 to 8 years were representative of the pants withdrawing from the trial are given in the Figure).20 original study population; there were no differences in demo- At age 6 to 8 years, diastolic blood pressure and MAP were graphic, anthropometric, or socioeconomic factors in infancy lower in children randomly assigned to standard formula between children who were or were not followed up (Table compared with nutrient-enriched formula (for diastolic blood

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 216 Circulation January 16, 2007

TABLE 2. Characteristics of Children Who Were and Were Not Followed up at Ages 6 to 8 Years

Standard Formula vs Nutrient-Enriched Formula Trial Breast-Fed Reference Group

Followed up (nϭ153) Not Followed up (nϭ146) Followed up (nϭ97) Not Followed up (nϭ78) Growth At birth Gestation, wk 39.2 (1.3) 39.2 (1.4) 39.3 (1.5) 39.1 (1.3) Birth weight, kg 2.6 (0.3) 2.6 (0.3) 2.6 (0.3) 2.5 (0.3) Birth weight z score Ϫ1.7 (0.5) Ϫ1.7 (0.5) Ϫ1.7 (0.5) Ϫ1.7 (0.5) At enrollment Weight, kg 2.5 (0.3) 2.5 (0.3) 2.5 (0.3) 2.5 (0.3) Weight z score Ϫ1.79 (0.6) Ϫ1.75 (0.6) Ϫ1.8 (0.5) Ϫ1.8 (0.5) Length, cm 47.2 (1.9) 47.1 (2.2) 47.9 (1.8) 47.2 (1.8)† Length z score Ϫ1.43 (0.9) Ϫ1.47 (1.0) Ϫ1.1 (0.8) Ϫ1.4 (0.9)‡ Change in weight z score enrollment to 9 mo 0.92 (1.0) 0.82 (1.0) 0.88 (0.9) 0.74 (1.0) Change in length z score enrollment to 9 mo 0.81 (1.0) 0.79 (1.1) 0.45 (1.0) 0.40 (0.9) Demographic % Male (n)* 44 (68) 48 (70) 54 (52) 52 (41) Maternal age, y 26.8 (4.9) 26.4 (5.4) 30.0 (4.4) 28.8 (5.0) Social class: % nonmanual (n)* 20 (31) 24 (35) 64 (62) 68 (53) Maternal qualifications* % None (n) 30 (46) 39 (57) 3 (3) 4 (3) % Degree or higher (n) 3 (5) 7 (10) 34 (33) 37 (29) Data are mean (SD) except *percentage (n) analyzed by ␹2 test. Small loss of n (Ͻ10%) for some variables. Significant differences between those followed up and not followed up: †Pϭ0.01; ‡Pϭ0.02.

pressure: mean difference, Ϫ3.2 mm Hg; Pϭ0.02) (Table 4). score) (mean difference, Ϫ3.2 mm Hg; 95% CI, Ϫ6.0 to Similar findings were obtained after adjustment for potential Ϫ0.6; Pϭ0.02). There was a trend for nutrient-enriched confounding factors (age, sex, z score for weight and height formula to have a greater adverse effect on later blood at 6 to 8 years, and social class) (Table 4). pressure in those with lower birth weight for gestation; the There was little evidence that the relation between formula interaction between nutrient-enriched formula and birth type and later blood pressure differed according to sex weight z score was statistically significant for MAP (Pϭ0.04) (PϾ0.4 for formulaϫsex interactions for diastolic blood but not for systolic (Pϭ0.08) or diastolic blood pressure ϭ pressure, MAP, or systolic blood pressure). When stratified (P 0.4). by sex, blood pressure tended to be lower in both boys and girls randomized to standard formula versus nutrient-enriched Nonrandomized Analyses In a secondary analysis, we tested the hypothesis that upward formula (mean difference [95% CI] in boys: diastolic: Ϫ2.1 percentile crossing for weight programmed later blood pres- [Ϫ6.2 to 2.0] mm Hg, Pϭ0.3; MAP: Ϫ2.5 [Ϫ6.2 to sure. Infant weight gain was assessed as both a dichotomous 1.3] mm Hg, Pϭ0.2; systolic: Ϫ1.7 [Ϫ6.4 to 3.0] mm Hg, and continuous variable. Pϭ Ϫ Ϫ Ϫ 0.5; in girls: diastolic: 4.6 [ 8.4 to 0.8] mm Hg, When infant weight gain was analyzed as a dichotomous ϭ Ϫ Ϫ ϭ P 0.02; MAP: 3.5 [ 7.4 to 0.4] mm Hg, P 0.08; systolic: variable, diastolic blood pressure was greater in children Ϫ Ϫ ϭ 2.6 [ 7.3 to 2.2] mm Hg, P 0.3). whose weight z score increased between birth and age 9 The effect of early diet on later diastolic blood pressure months (mean [SD]: 63.5 [8.0] mm Hg; nϭ104) than in those was also evident after the analysis was confined to those who without such acceleration in weight (59.9 [7.1] mm Hg; completed the initial diet (137 of 153) (per protocol) and nϭ29) (unadjusted mean difference, 3.6 mm Hg; 95% CI, 0.3 adjustment for potential confounding factors (as above) to 6.8; Pϭ0.03). Similar findings were obtained after adjust- (mean difference, Ϫ3.7 mm Hg; 95% CI, Ϫ6.5 to Ϫ0.9; ment for potential confounding factors (adjusted mean differ- Pϭ0.009). Similar findings were obtained for MAP (mean ence, 3.7 mm Hg; 95% CI, 0.3 to 7.2; Pϭ0.03); for MAP difference, Ϫ3.2 mm Hg; 95% CI, Ϫ5.9 to Ϫ0.5; Pϭ0.02) (unadjusted mean difference, 4.6 mm Hg; 95% CI, 1.4 to 7.7; but not for systolic blood pressure (mean difference, Pϭ0.005; adjusted mean difference, 4.7 mm Hg; 95% CI, 1.5 Ϫ2.2 mm Hg; 95% CI, Ϫ5.5 to 1.0; Pϭ0.2). to 8.0; Pϭ0.005); and for systolic blood pressure (unadjusted mean difference, 5.6 mm Hg; 95% CI, 1.8 to 9.4; Pϭ0.004; Association With Birth Weight adjusted mean difference, 5.1 mm Hg; 95% CI, 1.3 to 9.0; Diastolic blood pressure was lower in infants randomized to Pϭ0.01). standard versus nutrient-enriched formula after adjustment As a continuous variable, the change in weight z score for birth weight for gestational age z score (birth weight z between birth and 9 months was not associated with higher

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Singhal et al Early Nutrition and Later Blood Pressure 217

TABLE 3. Anthropometric and Demographic Characteristics of Children

Randomized Formula-Fed Groups

Standard Formula Nutrient-Enriched Formula Breast-Fed Group (nϭ83) (nϭ70) (nϭ97) In infancy Gestation, wk 39.3 (1.4) 39.1 (1.2) 39.3 (1.5) Birth weight, kg 2.6 (0.3) 2.5 (0.3) 2.6 (0.3) Birth weight z score Ϫ1.65 (0.5) Ϫ1.75 (0.6) Ϫ1.65 (0.5) Apgar score at 5 min 9.5 (0.7) 9.7 (0.6) 9.0 (1.1) At enrollment Weight, kg 2.5 (0.3) 2.5 (0.3) 2.5 (0.3) Weight z score Ϫ1.77 (0.5) Ϫ1.82 (0.6) Ϫ1.83 (0.5) Length, cm 47.3 (1.7) 47.1 (2.1) 47.9 (1.8) Length z score Ϫ1.39 (0.8) Ϫ1.45 (1.0) Ϫ1.08 (0.8) Social class: % nonmanual (n)* 22 (18) 19 (13) 64 (62) Age9mo Completed trial diet to 9 mo, % (n)* 88 (73) 91 (64) 93 (90) Weight, kg 8.2 (1.0) 8.5 (0.9)† 8.3 (0.8) Weight z score Ϫ0.95 (1.1) Ϫ0.67 (1.0) Ϫ0.93 (0.9) Length, cm 69.6 (2.8) 70.6 (2.7)† 70.1 (2.1) Length z score Ϫ0.66 (1.1) Ϫ0.53 (1.1) Ϫ0.60 (0.8) Change in weight z score enrollment to 9 mo 0.76 (1.1) 1.15 (0.9)† 0.90 (0.9) Change in length z score enrollment to 9 mo 0.73 (1.0) 0.92 (1.1) 0.44 (0.9) 18 mo Weight, kg 10.1 (1.3) 10.4 (1.1) 10.6 (1.2) Weight z score Ϫ1.04 (1.1) Ϫ0.81 (1.0) Ϫ0.74 (1.0) Height, cm 79.6 (3.4) 80.3 (3.0) 81.2 (2.7) Height z score Ϫ0.60 (1.1) Ϫ0.46 (1.1) Ϫ0.29 (0.9) Change in weight z score 9 to 18 mo Ϫ0.06 (0.6) Ϫ0.13 (0.6) 0.20 (0.6) Change in height z score 9 to 18 mo 0.06 (0.7) 0.02 (0.8) 0.28 (0.5) 6–8 y % (Proportion) followed up* 56 (83/147) 46 (70/152) 55 (97/175) % Male (n)* 33 (27) 59 (41)‡ 54 (52) Age, y 6.8 (0.6) 6.7 (0.4) 6.9 (0.8) Weight, kg 22.0 (4.2) 22.7 (6.1) 22.7 (4.0) Weight z score Ϫ0.27 (1.1) Ϫ0.13 (1.2) Ϫ0.16 (1.2) Height, cm 117.3 (6.9) 118.1 (6.4) 119.8 (6.6) Height z score Ϫ0.51 (1.1) Ϫ0.31 (1.1) Ϫ0.22 (1.1) Body mass index 16.0 (1.8) 16.3 (2.9) 16.0 (16.0) Change in weight z score 18 mo to 6–8 y 0.77 (1.0) 0.64 (1.0) 0.62 (0.9) Change in height z score 18 mo to 6–8 y 0.15 (0.8) 0.17 (0.7) 0.16 (0.8) Data are mean (SD) except *% (n). Comparison of randomized formula-fed groups all not significant except †Pϭ0.03, ‡Pϭ0.001. Small loss of n (Ͻ15%) for some variables.

later diastolic blood pressure (0.7 mm Hg increase per z score justment for confounding factors (as above) together with change; 95% CI, Ϫ0.8 to 2.3; Pϭ0.3) but was associated with randomized diet (for diastolic blood pressure: 0.8 mm Hg higher MAP (1.5 mm Hg increase per z score change; 95% increase per z score change in weight between birth and 9 CI, 0.1 to 3.0; Pϭ0.04) and systolic blood pressure months; 95% CI, Ϫ0.7 to 2.3; Pϭ0.3; MAP: 1.5 mm Hg (1.7 mm Hg increase per z score change; 95% CI, Ϫ0.01 to increase per z score change; 95% CI, 0.04 to 3.0; Pϭ0.04; 3.4; Pϭ0.05) independent of potential confounding factors and systolic blood pressure: 1.7 mm Hg increase per z (as above). Similar observations were obtained after ad- score change; 95% CI, Ϫ0.08 to 3.4; Pϭ0.06).

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 218 Circulation January 16, 2007

TABLE 4. Blood Pressure in Randomized Formula-Fed Groups at Age 6–8 Years

Unadjusted Blood Pressure, mm Hg Adjusted Blood Pressure, mm Hg*

Standard Nutrient Enriched Mean Mean (nϭ83) (nϭ70) Difference 95% CI P Difference 95% CI P Diastolic 61.3 (8.2) 64.5 (8.3) Ϫ3.2 Ϫ5.8 to Ϫ0.5 0.02 Ϫ3.5 Ϫ6.2 to Ϫ0.7 0.01 MAP 76.9 (8.3) 79.5 (7.8) Ϫ2.5 Ϫ5.1 to 0.1 0.06 Ϫ3.0 Ϫ5.6 to Ϫ0.3 0.03 Systolic 100.5 (10.2) 102.2 (9.8) Ϫ1.7 Ϫ4.9 to 1.5 0.3 Ϫ2.0 Ϫ5.3 to 1.3 0.2 Data are mean (SD) in each formula group. *Adjusted for age, sex, z score for weight and height at 6–8 years, and social class.

Blood pressure at age 6 to 8 years was not associated with slower weight and length gain in infancy,27 but, at least in acceleration in length between birth and age 9 months or with term infants born SGA, discourage the promotion of faster the change in weight or length z score between age 9 and 18 weight gain. months or between 18 months and 6 to 8 years (data not shown). The effect of a nutrient-enriched early diet on later blood pressure appears to be mediated via an effect on the rate of Breast-Fed Reference Group early weight gain. This finding is consistent with detrimental Blood pressure in breast-fed infants (mean [SD]: diastolic, effects of faster growth in infancy on cardiovascular risk 62.5 [8.6]; MAP, 78.2 [7.8]; systolic, 101.7 [9.4] mm Hg) factors5–8 and life span28 in animals, endothelial dysfuc- was similar to infants fed a standard or nutrient-enriched ntion10 and insulin resistance11 in infants born prematurely, formula. However, exploratory analyses suggested that faster and a tendency to obesity in healthy term infants.17,18,29 Both early weight gain also programmed higher later blood pres- faster linear growth and faster weight gain have been impli- sure in breast-fed children. Thus, breast-fed infants whose cated previously.10,16,29 Faster weight gain2,30 and length gain2 ϭ weight z score increased in the first 9 months (n 76) had in older children have also been consistently associated with higher diastolic blood pressure (mean [SD], 63.1 higher later blood pressure. However, these studies have [8.5] mm Hg) than those whose weight did not (55.6 usually estimated growth from the difference in weight z ϭ [6.4] mm Hg; n 11) (mean difference, 7.5 [95% CI, 2.2 to score between birth and a point later in childhood or adult life 12.8] mm Hg; Pϭ0.006). Similar findings were made for and therefore have not defined the period of postnatal weight mean and systolic blood pressure (PϽ0.02 for both). gain that has the greatest negative impact on later blood Discussion pressure. Surprisingly, there are relatively few data on the programming In the present prospective experimental study, we found that effects of weight gain in infancy, which, as the period of most consumption of a nutrient-enriched formula programmed rapid growth, could be particularly influential. A recent Brazilian higher blood pressure in SGA infants. We postulate that these study showed that faster weight gain in infancy, childhood, or effects, seen up to 8 years after random dietary assignment, were the result of faster weight gain in infancy promoted by after puberty had exactly the same negative implications for later the nutrient-enriched diet. Our findings support the concept systolic blood pressure (0.37 mm Hg increase in systolic blood 13 that early nutrition has a major impact on long-term cardio- pressure per z score increase in weight gain per year). Simi- vascular health in infants born at term, as shown previously in larly, faster weight gain in infancy was associated with later experimental studies in infants born preterm,11,21 and these systolic blood pressure in prospective studies of adults from 14 15 findings provide further experimental evidence to support the Finland and children from the United States. In contrast, a hypothesis that faster growth in infancy has adverse effects follow-up of a British cohort did not find any effect of infant 30 for later blood pressure.13–15 Therefore, together with evi- weight gain on adult blood pressure, whereas a greater fall in dence from animal studies5–8 and from studies in preterm ponderal index between 6 and 18 months of age predicted higher 31 infants10–12 and both SGA16 and healthy term infants13–15,17,18 systolic blood pressure in adults living in Hong Kong. These (including 2 systematic reviews for effects on later obesi- previous conflicting findings might be explained by differences ty17,18), our findings argue against the promotion of faster in study populations (eg, from a developed or developing weight gain in infants born SGA. country), different ages at follow-up, and the impact of con- The size of the effect of early diet on later blood pressure founding factors associated with an observational study design. (Ϸ3 mm Hg) may amplify with age25 and for populations is In contrast, our experimental study, in which random dietary substantial in terms of its potential impact on public health, assignment produced differences in weight and length growth, although the 95% CIs around the point estimate raise the shows that faster weight gain in infancy programs both higher possibility that the effect size could be smaller. Lowering systolic and diastolic blood pressure later in childhood. population-wide diastolic blood pressure by only 2 mm Hg, We considered several potential limitations to the interpre- however, would be expected to prevent Ϸ100 000 myocardial tation of our data. First, only 51% of the original population and cerebrovascular events per year in the United States was followed up. However, this is unlikely to introduce alone.26 The present study therefore has important implica- systematic bias because our sample was representative of tions for infant nutrition policy. Our data indirectly support those recruited at birth, and subject characteristics of children the promotion of breast-feeding, which is associated with who were reviewed at age 6 to 8 years did not differ between

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Singhal et al Early Nutrition and Later Blood Pressure 219

randomized groups at birth or at follow-up. Moreover, there growth) that is the mechanistic link between nutrient-enriched is no prior reason why the epidemiological association formula and higher later blood pressure. Thus, the high protein between acceleration in weight in infancy and later blood content of the nutrient-enriched formula was designed to fuel pressure should differ between children reviewed and those faster weight and length gain, and the higher content of other not reviewed. Similarly, the excess of boys in the nutrient- nutrients was calculated to meet the increased nutrient needs for enriched compared with the standard formula group at infants growing faster. However, we cannot exclude the possi- follow-up did not affect the findings because there were no bility that differences between formulas in specific nutrients sex differences in blood pressure at age 8 years and because such as calcium or phosphorus would have been influential for the effects of nutrient-enriched formula on later blood pres- later blood pressure, although there are no convincing studies to sure were independent of sex. suggest this explanation. Nonetheless, the primary role of Second, the applicability of our findings to infants not born growth in programming is supported by our preliminary obser- SGA is uncertain. Previous studies in animals,5–8 our data in vation that faster weight gain programmed higher blood pressure preterm infants,10,11 and studies in term infants17,18 suggest that even in breast-fed infants. Similarly, faster weight gain has been programming effects of a nutrient-enriched diet and faster shown to program later obesity independent of protein intake29 weight and length gain on later CVD risk factors, including and the method of infant feeding.36 blood pressure,13–15 are independent of size at birth. Whether Thus, we suggest that, as in animal models,5–8 the cost of similar effects are seen in infants with appropriate birth weight faster growth in infancy is a later increase in cardiovascular needs further research, although nutrition intervention studies risk. Although faster weight gain has short-term advantages are more difficult to justify in such groups. Nonetheless, the for morbidity in infants with low birth weight from low- adverse effect of a nutrient-enriched early diet on later outcome income countries,37 whether risk-benefit analysis will lead to in our study appeared to be greater in babies with lower birth similar conclusions in such infants from high-income coun- weight, consistent with data from animal studies.28 Previously, tries remains uncertain. Importantly, unlike the case for blood pressure has been shown to be lower in breast-fed than infants born prematurely, a nutrient-enriched diet in infancy formula-fed individuals.21 That this was not shown here may be did not improve cognitive development at age 18 months in because breast-fed infants, when born SGA, also show deleteri- our SGA infants.38 Clearly, further data are required to inform ous (spontaneous) faster postnatal weight gain. In fact, we found the debate on the effects of early growth for later cardiovas- a 7.5 mm Hg difference in diastolic blood pressure between 78% cular outcomes and in particular to explain the discrepancies of breast-fed infants who showed faster weight gain and those in findings from some historical cohorts3,4,9 and more recent who did not. observational13–15 and experimental studies.11,21 Our findings, Third, we studied children and did not have prospective however, support the hypothesis that the promotion of faster follow-up into adulthood. Therefore, it is theoretically possible weight gain has adverse programming effects on later blood that the positive association between early weight gain and later pressure, at least in term infants born SGA. blood pressure reverses in adult life, although a previous asso- ciation between faster weight gain in infancy and higher blood Acknowledgments pressure in adulthood14 argues against this hypothesis. We thank Wendy Jenkins, Dee Beresford, Penny Lucas, and Emma Finally, the present study did not identify mechanisms for Sutton for technical assistance. the observed effects. Nevertheless, differences in blood pres- sure between randomized dietary groups were independent of Sources of Funding This study was funded by the Medical Research Council of the body weight adjusted for height (in regression analyses), a United Kingdom. We thank Farley’s Health Products (a division of major determinant of blood pressure also strongly influenced HJ Heinz Company Ltd, Stockley Park Uxbridge, UK) for supplying by faster weight gain in infancy.17,18 In fact, as in infants born the trial formulas. prematurely,32 a nutrient-enriched diet did not have long-term effects on height or weight in SGA infants. Furthermore, in Disclosures accordance with a previous report,33 only acceleration in None. weight and not length was associated with higher later blood pressure. This is consistent with our previous findings of a References 1. Barker DJP, Gluckman PD, Godfrey KM, Harding JE, Owen JA, smaller effect of faster early length gain compared with Robinson JS. Fetal nutrition and cardiovascular disease in adult life. weight gain on later endothelial function.10 Lancet. 1993;341:938–941. Previously, we postulated that faster growth programs the 2. Huxley RR, Shiell AW, Law CM. The role of size at birth and postnatal metabolic syndrome possibly by an effect on hormonal systems catch-up growth in determining systolic blood pressure: a systematic review of the literature. J Hypertens. 2000;18:815–831. 34 such as insulin-like growth factor-1. Our present findings were 3. Fall CHD, Vijayakumar B, Barker DJP, Osmond C, Duggleby S. Weight consistent with this hypothesis and support an evolutionary in infancy and prevalence of coronary heart disease in adult life. BMJ. conserved mechanism, which suggests that faster early growth 1995;310:17–20. 4. Eriksson JG, Forsen T, Tuomilehto J, Osmond C, Barker DJP. Early can have negative effects on long-term health and, in many growth and coronary heart disease in later life: longitudinal study. BMJ. species, there is a tradeoff within individuals between acceler- 2001;322:949–953. ated early growth and shorter life span.5,35 Because the present 5. Metcalfe NB, Monaghan P. Compensation for a bad start: grow now, pay randomized study addressed infant feeding and not growth, later? Trends Ecol Evol. 2001;16:254–260. 6. Plagemann A, Heidrich I, Gotz F, Rohde W, Dorner G. Obesity and however, the present study cannot determine whether it is faster enhanced diabetes and cardiovascular risk in adult rats due to early weight gain itself or infant nutrition (1 determinant of early postnatal overfeeding. Exp Clin Endocrinol. 1992;99:154–158.

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 220 Circulation January 16, 2007

7. Plagemann A, Harder T, Rake A, Voits M, Fink H, Rohde W, Dorner G. 22. Sanchez-Bayle M, Munoz-Fernandez MT, Gonzalez-Requejo A. A lon- Perinatal elevation of hypothalamic insulin, acquired malformations of gitudinal study of blood pressure in Spanish schoolchildren. Arch Dis hypothalamic galaninergic neurons, and syndrome X-like alterations in Child. 1999;81:169–171. adulthood of neonatally overfed rats. Brain Res. 1999;836:146–155. 23. Anwar YA, Tendler BE, McCabe EJ, Mansoor GA, White WB. Eval- 8. Lewis DS, Bertrand HA, McMahan CA, McGill HC, Carey KD, Masoro uation of the Datascope Accutorr Plus according to the recommendations EJ. Preweaning food intake influences the adiposity of young adult of the association for the advancement of medical instrumentation. Blood baboons. J Clin Invest. 1986;78:899–905. Press Monit. 1997;2:105–110. 9. Barker DJP, Osmond C, Forsen TJ, Kajantie E, Eriksson JG. Trajectories 24. O’Brien E, Waeber B, Parati G, Staessen J, Myers MG. Blood pressure of growth among children who have coronary events as adults. N Engl measuring devices: recommendations of the European Society of Hyper- J Med. 2005;353:1802–1809. tension. BMJ. 2001;322:531–536. 10. Singhal A, Cole TJ, Fewtrell MS, Deanfield J, Lucas A. Is slower early 25. Law CM, De Swiet M, Osmond C, Fayers PM, Barker DJP, Cruddas AM, Fall CHD. Initiation of hypertension in utero and its amplification growth beneficial for long-term cardiovascular health? Circulation. 2004; throughout life. BMJ. 1993;306:24–27. 109:1108–1113. 26. Cook NR, Cohen J, Herbert PR, Taylor JO, Hennekens CH. Implications 11. Singhal A, Fewtrell M, Cole TJ, Lucas A. Low nutrient intake and early of small reductions in diastolic blood pressure for primary prevention. Lancet growth for later insulin resistance in adolescents born preterm. . Arch Intern Med. 1995;155:701–709. 2003;361:1089–1097. 27. Ong KKL, Preece MA, Emmett PM, Ahmed ML, Dunger DB. Size at 12. Finken MJJ, Keijzer-Veen MG, Van Montfoort AG, Dekker FH, Frolich birth and early childhood growth in relation to maternal smoking, parity M, Romijn JA, Hille ETM, Van der Heijden AJ, Wit JM. Early catch-up and infant breast-feeding: longitudinal birth cohort study and analysis. growth in weight of very preterm low birthweight infants is associated Pediatr Res. 2002;52:863–867. with higher levels of LDL-cholesterol and apo-B at age 19. Pediatr Res. 28. Ozanne SE, Hales CN. Catch-up growth and obesity in male mice. 2003;53(suppl):32. Abstract. Nature. 2004;427:411–412. 13. Horta BL, Barros FC, Victora CG, Cole TJ. Early and late growth and 29. Gunnarsdottir I, Thorsdottir I. Relationship between growth and feeding blood pressure in adolescence. J Epidemiol Community Health. 2003;57: in infancy and body mass index at the age of 6 years. Int J Obes. 226–230. 2003;27:1523–1527. 14. Jarvelin MR, Sovio U, King V, Lauren L, Xu B, McCarthy MI, Harti- 30. Law CM, Shiell AW, Newsome CA, Syddall HE, Shinebourne EA, kainen AL, Laitinen J, Zitting P, Rantakallio P, Elliot P. Early life factors Fayers PM, Martyn CN, DeSwiet MF. Infant and childhood growth and and blood pressure at age 31 years in the 1966 northern Finland birth adult blood pressure: a longitudinal study from birth to 22 years of age. cohort. Hypertension. 2004;44:838–846. Circulation. 2002;105:1088–1092. 15. Hemachandra AH, Howards PP, Schisterman EF, Klebanoff MA. Inter- 31. Cheung YB, Low L, Osmond C, Barker D, Karlberg J. Fetal growth and action between birth weight and postnatal growth does not increase risk early postnatal growth are related to blood pressure in adults. Hyper- for high blood pressure at age 7 years: results from the United States tension. 2000;36:795–800. Collaborative Perinatal Project. Pediatr Res. 2005;58:1102. Abstract. 32. Morley R, Lucas A. Randomized diet in the neonatal period and growth 16. Soto N, Bazaes RA, Pena V, Salazar T, Avila A, Iniguez G, Ong KK, performance until 7.5–8 y of age in preterm children. Am J Clin Nutr. Dunger DB, Mericq MV. Insulin sensitivity and secretion are related to 2000;71:822–828. catch-up in small for gestational age infants at age 1 year: results from a 33. Lundgren EM, Cnattingius HMS, Jonsson GB, Tuvemo TH. Linear prospective cohort. J Clin Endocrinol Metab. 2003;88:3645–3650. catch-up growth does not increase the risk of elevated blood pressure and reduces the risk of overweight in males. J Hypertens. 2001;19: 17. Monteiro POA, Victora CG. Rapid growth in infancy and childhood and 1533–1538. obesity in later life: a systematic review. Obes Rev. 2005;6:143–154. 34. Singhal A, Lucas A. Early origins of cardiovascular disease: is there a 18. Baird J, Fisher D, Lucas P, Kleijnen J, Roberts H, Law C. Being big or unifying hypothesis? Lancet. 2004;363:1642–1645. growing fast: systematic review of size and growth in infancy and later 35. Metcalfe NB, Monaghan P. Growth versus lifespan: perspectives from obesity. BMJ. 2005;331:929–934. evolutionary ecology. Exp Gerontol. 2003;38:935–940. 19. Martin RM, McCarthy A, Smith GD, Davies DP, Ben-Shlomo Y. Infant 36. Ong KKL, Ahmed ML, Emmett PM, Preece MA, Dunger DB. Asso- nutrition and blood pressure in early adulthood: the Barry Caerphilly ciation between postnatal catch-up growth and obesity in childhood: Growth Study. Am J Clin Nutr. 2003;77:1489–1497. prospective cohort study. BMJ. 2000;320:967–971. 20. Fewtrell MS, Morley R, Abbott RA, Singhal A, Stephenson T, Mac- 37. Victora CG, Barros FC, Horta BL, Martorell R. Short-term benefits of Fadyen UM, Clements H, Lucas A. Catch-up growth in small for gesta- catch-up growth for small for gestational-age infants. Int J Epidemiol. tional age term infants: a randomized trial. Am J Clin Nutr. 2001;74: 2001;30:1325–1330. 516–523. 38. Morley R, Fewtrell MS, Abbott RA, Stephenson T, MacFadyen UM, 21. Singhal A, Cole TJ, Lucas A. Early nutrition in preterm infants and later Lucas A. Neurodevelopment in children born small for gestational age: a blood pressure: two cohorts after randomized trials. Lancet. 2001;357: randomized trial of nutrient enriched versus standard-formula, and com- 413–419. parison with a reference breast-fed group. Pediatrics. 2004;113:515–521.

CLINICAL PERSPECTIVE Being born small for gestational age is associated with later risk factors for cardiovascular disease, such as high blood pressure. Promotion of postnatal growth has been proposed to ameliorate these effects. There is evidence in animals and infants born prematurely, however, that promotion of growth by increased postnatal nutrition increases rather than decreases later cardiovascular risk. We tested this hypothesis by measuring blood pressure at 6 to 8 years of age in a cohort of children born small for gestational age (birth weight Ͻ10th percentile) and randomly assigned at birth to receive either a nutrient-enriched formula, which promoted growth, or a standard, control formula. Diastolic and mean arterial blood pressures were found to be significantly lower in children assigned to standard compared with nutrient-enriched formula (unadjusted mean difference for diastolic blood pressure, Ϫ3.2 mm Hg; 95% CI, Ϫ5.8 to Ϫ0.5; Pϭ0.02), independent of potential confounding factors. In observational analyses, faster weight gain in infancy was associated with higher later blood pressure. The present study showed that a nutrient-enriched diet in infancy increased later blood pressure with an effect size that is important for the health of populations rather than individuals. These findings are consistent with data from animal models that suggest an adverse effect of relative “overnutrition” in infancy on long-term cardiovascular disease risk, indirectly support the promotion of breast-feeding, which is associated with slower weight gain in infancy, and argue against the promotion of faster weight gain, at least in term infants born small for gestational age.

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Cardiac and Systemic Hemodynamic Characteristics of Hypertension and Prehypertension in Adolescents and Young Adults The Strong Heart Study

Jennifer S. Drukteinis, MD; Mary J. Roman, MD; Richard R. Fabsitz, PhD; Elisa T. Lee, PhD; Lyle G. Best, MD; Marie Russell, MD; Richard B. Devereux, MD

Background—The epidemic of overweight is increasing the prevalence of both prehypertension and early-onset hypertension, but few population-based data exist on their impact on cardiac structure and function in adolescents and young adults. Methods and Results—We analyzed clinical characteristics, hemodynamic parameters, and left ventricular structure and function in 1940 participants, 14 to 39 years of age, in the Strong Heart Study. Hypertension occurred in 294 participants (15%), who were more often men (70% versus 30%), older (age, 31Ϯ7 versus 25Ϯ8 years), and more commonly diabetic (23% versus 4.5%; all PϽ0.001) than their normotensive counterparts. Prehypertension occurred in 675 (35%) of participants with similar trends in gender, age, and diabetes status. After adjustment for covariates, both hypertensive and prehypertensive participants had higher left ventricular wall thickness (0.83 and 0.78 versus 0.72 cm), left ventricular mass (182 and 161 versus 137 g), and relative wall thickness (0.30 and 0.29 versus 0.28 cm) and 3- and 2-fold-higher prevalences of left ventricular hypertrophy than their normotensive counterparts (all PϽ0.001). Hypertension and prehypertension also were associated with higher mean pulse pressure/stroke volume index (1.24 and 1.15 versus 1.02 mm Hg/mL · m2) and total peripheral resistance index (3027 and 2805 versus 2566 dynes·s·cmϪ5 ·m2; all PϽ0.001). Conclusions—In a population with high prevalences of obesity and diabetes, hypertension and prehypertension are associated with increases in both cardiac output and peripheral resistance index. Despite the young age of participants with hypertension and prehypertension, they had prognostically adverse preclinical cardiovascular disease, including left ventricular hypertrophy and evidence of increased arterial stiffness. (Circulation. 2007;115:221-227.) Key Words: echocardiography Ⅲ hemodynamics Ⅲ hypertension Ⅲ hypertrophy Ⅲ prehypertension

n the United States and other industrialized countries, Approximately 60% of American adults have prehyperten- Iphysical inactivity and high-calorie diets are leading to sion or hypertension, with hypertension increasing by Ϸ10% increasing prevalences of obesity and diabetes.1 Concor- in the past decade.6 Recent studies in population-based dantly, hypertension also has become increasingly prevalent.2 samples with mean ages from 45 to 55 years at baseline have Obesity and higher blood pressure have been shown to track shown increased cardiovascular event rates in adults with from childhood and adolescence to adulthood and to predict “high-normal” BP (130 to 139/85 to 89 mm Hg)7 or prehy- adult cardiovascular risk.3,4 According to the Seventh Report pertension (BP, 120 to 139/80 to 89 mm Hg without antihy- of the Joint National Committee (JNC) on Prevention, De- pertensive medication).8 Although the increased prevalence tection, Evaluation, and Treatment of High Blood Pressure of hypertension and the common occurrence of prehyperten- (JNC-7),5 a new category based on blood pressure (BP) level, sion are affecting cardiovascular mortality in middle and called prehypertension, requires attention and health- older age, their earlier impact on cardiac structure and promoting lifestyle modifications at an even earlier stage to function in adolescents and young adults has not been prevent the progressive rise in BP and cardiovascular disease. extensively characterized in large, population-based samples.

Continuing medical education (CME) credit is available for this article. Go to http://cme.ahajournals.org to take the quiz. Received February 24, 2006; accepted October 11, 2006. From the Weill Medical College of Cornell University, New York, NY (J.S.D., M.J.R., R.B.D.); National Heart, Lung, and Blood Institute, Bethesda, Md (R.R.F.); University of Oklahoma Health Sciences Center, Oklahoma City (E.T.L.); Missouri Breaks Industries Research, Inc, Timber Lake, SD (L.G.B.); and MedStar Research Institute, Washington, DC (M.R.). This manuscript presents the views of the authors and not necessarily those of the Indian Health Service. Correspondence to Richard B. Devereux, MD, Division of Cardiology, Box 222, New York Presbyterian Hospital, 525 E 68th St, New York, NY 10021. E-mail [email protected] © 2007 American Heart Association, Inc. Circulation is available at http://www.circulationaha.org DOI: 10.1161/CIRCULATIONAHA.106.668921

Downloaded from circ.ahajournals.org 221at Mohammed Mahboob on June 10, 2007 222 Circulation January 16, 2007

Previous studies have examined cardiac structure and func- ventricular (LV) internal dimensions were measured at end diastole tion in white, black, or Hispanic populations with modest and end systole by American Society of Echocardiography recom- mendations.22,23 Aortic annular diameter was measured as previously prevalences of obesity and diabetes.9,10 Accordingly, the described.24 Doppler transaortic flow was assessed by identifying the present study was undertaken to evaluate the associations of apical view in which peak flow velocity was maximal and, after hypertension and prehypertension with clinical characteris- calibration, tracing the black-white interface outlining the Doppler tics, systemic hemodynamics, and cardiac structure and flow envelope.25 Heart rate was measured simultaneously. Prolonged Ն function in adolescent and young adult American Indian isovolumic relaxation time was recognized as 92.4 ms in this population, representing the 97.5th percentile of values in 497 participants (age, 14 to 39 years) in the Strong Heart Study nondiabetic, nonobese, nonhypertensive participants in the fourth (SHS),11–13 a population in which obesity, diabetes, and SHS examination. hypertension are highly prevalent. Calculation of Derived Variables Methods End-diastolic LV dimensions were used to calculate LV mass by an The SHS is a population-based survey of cardiovascular risk factors anatomically validated formula.26 Concentric LV geometry was and cardiovascular disease in American Indian communities in identified by taking into account age effects on LV relative wall Arizona, Oklahoma, and South and North Dakota. As previously thickness according to the method of de Simone et al.27 Systolic described,12 members of 13 communities in Arizona, Oklahoma, and fractional shortening of the LV internal dimension and end-systolic South and North Dakota (45 to 74 years of age) were recruited from stress were calculated by standard methods.21 Aortic annular cross- ␲ϫ 2 reservations or (in Oklahoma) in defined geographic areas (overall sectional area was calculated as follows: (diameter/2). Doppler participation rate, 62%) for an initial examination in 1989 to 1992. stroke volume was calculated as annular cross-sectional area times Two additional examinations of the initial SHS cohort to assess the time-velocity integral.26 Arterial stiffness was estimated by the change over time were performed during 1993 to 1995 and during ratio of pulse pressure to stroke volume index.28 1997 to 1999. The fourth SHS examination in July 2001 to September 2003 included 520 SHS cohort members and 3138 of Measures of Myocardial Performance their relatives in 95 large 3-generation families (Strong Heart Family The primary approach to assess myocardial contractile efficiency Study [SHFS]).14 Standardized measurements of seated brachial BP was to examine LV midwall shortening in relation to circumferential and aspects of body habitus, including body mass index, waist-to-hip end-systolic stress measured at the level of the LV midwall using ratio, body composition by bioelectric impedance, fasting glucose, previously described methods.20–30 Midwall shortening calculated insulin and lipid concentrations, and 2-hour glucose tolerance test from echocardiographic measurements was expressed as a percent- and glycosylated hemoglobin levels, were obtained. age of the value predicted from circumferential end-systolic stress Brachial artery BP (first and fifth Korotkoff sounds) was measured using equations derived from previously studied healthy subjects30; 3 consecutive times on seated participants after they had rested 5 this variable is called stress-corrected midwall shortening.31 minutes with the use of a mercury sphygmomanometer (W.A. Baum Co, Copiague, NY). An appropriately sized cuff was placed on the Statistical Analysis right arm; pulse occlusion pressure was determined; and the cuff was Data, expressed as meanϮSD for continuous variables and propor- inflated to 20 mm Hg above that pressure. The mean of the last 2 of tions for categorical variables, were analyzed by SPSS 12.0 software these measurements was used to estimate BP. Hypertension was (SPSS, Chicago, Ill). Differences between 2 groups for continuous defined by systolic BP Ն140 mm Hg, diastolic pressure variables were assessed by t tests, with log transformation when Ն90 mm Hg, or use of antihypertensive medications.15 Participants needed to satisfy the assumption of normality, and the ␹2 statistic was were asked to bring their medications to the clinic and to recall (with used to determine differences of categorical variables. Relations assistance from an adult for minors) additional medications. Prehy- between clinical and echocardiographic variables were assessed with pertension was defined by systolic BP of 120 to 139 mm Hg or consideration of appropriate covariates. Independence of differences diastolic BP of 80 to 89 mm Hg5 in the absence of antihypertensive from effects of covariates (age, gender, diabetes, and center location) treatment. Fat-free mass and adipose mass were estimated with an was assessed by ANCOVA in the general linear model with the RJL impedance meter (model B14101, RJL Equipment Co, Detroit, Sidak post hoc test or by logistic regression analysis for categorical Mich) and equations based on total body water12 that have been variables. Two-tailed values of PϽ0.05 indicated statistical validated in the American Indian population.16,17 Obesity was significance. defined by a body mass index Ն30 kg/m2 and central adiposity by The authors had full access to and take full responsibility for the waist circumference Ͼ102 cm for men and Ͼ88 cm for women.18 integrity of the data. All authors have read and agree to the Diabetes was diagnosed by 1997 American Diabetes Association manuscript as written. diagnostic criteria.19 Participants (or their parent or guardian in the case of minors) gave written informed consent under protocols Results approved by institutional review boards of the clinical centers. Characteristics Echocardiographic Methods Of the 3658 participants in the fourth SHS examination, 1940 Imaging and Doppler echocardiograms, performed as previously relatives of members of the original SHS cohort were Ͻ40 described,14,20,21 were recorded on videotape using phased-array years of age. Most participants (57.5%) were women; the echocardiographs with fundamental and harmonic M-mode and mean age was 26.8Ϯ7.7 years; 294 (15%) and 675 (35%) met 2-dimensional imaging, as well as pulsed, continuous-wave, and color-flow Doppler capabilities (Sequoia, Siemens Inc, Mountain JNC-7 criteria for hypertension and prehypertension, respec- View, Calif). Participants were examined in a partial decubitus tively. Among hypertensive participants, 45 (15%) had sys- position with the head of table tilted at Ϸ30°. tolic hypertension, defined as BP Ն140 mm Hg; 138 (47%) had diastolic hypertension, defined as BP Ն90 mm Hg; 57 Echocardiographic Measurements (19%) had both diastolic and systolic hypertension; and 54 Correct orientation of planes for imaging and Doppler recordings (18%) had normal BP on antihypertensive medication. Hy- was verified as previously described.21,22 Measurements were made with a computerized review station equipped with digitizing tablet pertension was more prevalent in men than women and and monitor screen overlay for calibration and measurement perfor- increased with age in both genders (the Figure), rising from mance. Interventricular and posterior wall thicknesses and left 1.7% in women and 9.3% in men at 14 to 19 years of age

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Drukteinis et al Hypertension and Prehypertension in Young Adults 223

older (30.9Ϯ6.8 and 27.2Ϯ7.5 versus 25.2Ϯ7.7 years; both PϽ0.001). In analyses that adjusted for age, gender, diabetes, and center, hypertension and prehypertension were associated with higher weight, body mass index, percent body fat, body surface area, waist circumference, and adipose mass (Table 1). However, no difference was seen in ankle/brachial index between groups. According to National Institutes of Health guidelines,19 79% and 69% of hypertensive and prehyperten- sive participants, respectively, had central adiposity by waist circumference criteria compared with 55% in the nonhyper- tensive group (PϽ0.001). The hypertensive participants had Prevalence of hypertension (vertical axis) is greater in male higher mean urine albumin-to-creatinine ratio and higher (solid bars) than female (striped bars) adolescent and young adult SHS participants and increases with age (horizontal axis) hemoglobin A1c levels, whereas prehypertensive participants in both genders. did not differ statistically from their normotensive counter- parts. Both the hypertensive and prehypertensive groups had (PϽ0.001) to 6.2% of women and 16.5% of men at 20 to 24 higher total cholesterol, insulin, and low-density lipoprotein years of age (Pϭ0.007), 5.1% of women and 16.4% of men cholesterol than their normotensive counterparts, with no at 25 to 29 years of age (PϽ0.001), 4.5% of women and difference between groups in plasma creatinine, triglycerides, 21.0% of men at 30 to 34 years of age (PϽ0.001), and 18.0% or high-density lipoprotein cholesterol after adjustment for of women and 41.6% of men at 35 to 39 years of age covariates. (PϽ0.001). Compared with the nonhypertensive group, hy- pertensive and prehypertensive participants were more likely Systemic Hemodynamics to be men (70% and 52% versus 38%, respectively), to be Heart rate and cardiac output were higher in both hyperten- obese (75% and 65% versus 53%, respectively), and to have sive and prehypertensive participants than in normotensive diabetes (23% and 9% versus 6%, respectively) and impaired participants in absolute terms and after adjustment for age, fasting glucose (10% and 6% versus 3%, respectively; all gender, diabetes, and center location (Table 2). Total periph- intergroup differences, PϽ0.05). Hypertension was more eral resistance was higher in both the hypertensive and prevalent in Oklahoma or Arizona than in North and South prehypertensive groups; however, after adjustment for covari- Dakota (18% and 19% versus 11%, respectively; PϽ0.001). ates, the difference remained significant in only the hyper- Hypertensive and prehypertensive participants were also tensive group. Peripheral resistance index and pulse pressure/

TABLE 1. Clinical Characteristics of Normotensive, Prehypertensive, and Hypertensive Adolescent and Young Adult SHS Participants

BP P *

NT (Nϭ971) PH (Nϭ675) HT (Nϭ294) HT vs PH PH vs HT NT vs HT Systolic BP, mm Hg 108.0Ϯ7.0 123.2Ϯ7.5 134.9Ϯ14.7 ⅐⅐⅐ ⅐⅐⅐ ⅐⅐⅐ Diastolic BP, mm Hg 68.1Ϯ7.2 78.7Ϯ7.7 89.6Ϯ10.8 ⅐⅐⅐ ⅐⅐⅐ ⅐⅐⅐ Body mass index, kg/m2 30.0Ϯ8.2 33.8Ϯ8.3 36.6Ϯ9.2 Ͻ0.001 0.003 Ͻ0.001 Weight, kg 82.5Ϯ22.6 97.5Ϯ24.6 108.8Ϯ28.2 Ͻ0.001 Ͻ0.001 Ͻ0.001 Body surface area, m2 1.89Ϯ0.24 2.07Ϯ0.25 2.19Ϯ0.27 Ͻ0.001 0.001 Ͻ0.001 Body fat, % 36.8Ϯ11.6 37.9Ϯ10.8 38.2Ϯ9.5 Ͻ0.001 0.087 Ͻ0.001 Adipose mass, kg 32.3Ϯ17.5 38.1Ϯ17.9 43.2Ϯ20.0 Ͻ0.001 0.007 Ͻ0.001 Waist circumference, cm 96.5Ϯ18.5 106.2Ϯ18.1 114.3Ϯ19.1 Ͻ0.001 0.001 Ͻ0.001 Ankle/arm index 1.10Ϯ0.11 1.11Ϯ0.11 1.11Ϯ0.11 0.757 0.494 0.170 Plasma creatinine, mg/dL 0.77Ϯ0.16 0.81Ϯ0.16 0.83Ϯ0.28 0.995 0.973 0.931 Log urinary albumin/creatinine 2.08Ϯ0.10 2.18Ϯ1.1 2.53Ϯ1.5 0.932 0.010 0.003 Total cholesterol, mg/dL 165Ϯ33.1 181Ϯ38.3 190Ϯ42.7 Ͻ0.001 0.986 0.003 HDL cholesterol, mg/dL 50.3Ϯ13.0 48.3Ϯ12.7 50.0Ϯ15.0 0.074 0.101 0.954 LDL cholesterol, mg/dL 89.5Ϯ25.0 101Ϯ29.6 105Ϯ33.0 Ͻ0.001 0.623 0.030 Triglycerides, mg/dL 130Ϯ121 172.5Ϯ236 211.4Ϯ305 0.121 0.990 0.555 Insulin, ␮U/mL 15.6Ϯ14.7 19.5Ϯ20.5 22.4Ϯ26.6 Ͻ0.001 0.043 Ͻ0.001

Hemoglobin A1C,% 5.9Ϯ1.9 6.3Ϯ2.1 7.0Ϯ2.3 0.992 0.949 0.882 NT indicates normotensive; PH, prehypertensive; HT, hypertensive; HDL, high-density lipoprotein; and LDL, low-density lipoprotein. Values are meanϮSD unless otherwise indicated. *Adjusted for age, gender, diabetes, and center location.

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 224 Circulation January 16, 2007

TABLE 2. Systemic Hemodynamics of Normotensive and Hypertensive Adolescent and Young Adult SHS Participants

BP P *

NT PH HT NT vs PH PH vs HT NT vs HT Heart rate, bpm 65Ϯ10 68Ϯ11 71Ϯ12 Ͻ0.001 0.020 Ͻ0.001 Cardiac output, mL/min 5037Ϯ987 5624Ϯ1175 6111Ϯ1390 Ͻ0.001 0.014 Ͻ0.001 Cardiac output/body surface area, mL ⅐ min ⅐ mϪ2 2683Ϯ511 2736Ϯ544 2790Ϯ541 0.172 0.793 0.095 Total peripheral resistance, dyne ⅐ cm ⅐ sϪ5 1365Ϯ278 1369Ϯ283 1398Ϯ311 0.866 0.257 0.097 Pulse pressure/stroke index, mm Hg/mL ⅐ m2 1.02Ϯ0.26 1.15Ϯ0.30 1.24Ϯ0.35 0.034 0.074 Ͻ0.001 Peripheral resistanceϫbody surface area, dyne ⅐ cm ⅐ sϪ5 ⅐ m2 2566Ϯ552 2805Ϯ569 3027Ϯ631 Ͻ0.001 Ͻ0.001 Ͻ0.001 Abbreviations as in Table 1. Values are meanϮSD unless otherwise indicated. *Adjusted for age, gender, diabetes status, and center location.

stroke index were higher in both the hypertensive and after covariate adjustment. Mitral deceleration time was prehypertensive groups. slightly higher in hypertensive and prehypertensive partici- pants, although this difference did not remain significant after LV Systolic Function covariate adjustment. LV endocardial fractional shortening and ejection fraction from linear LV dimension or 2-dimensional wall motion LV Geometry scores were not statistically different among groups (Table 3). Hypertensive and prehypertensive patients had, on average, Hypertensive and prehypertensive participants had lower thicker interventricular septal and LV posterior walls than midwall shortening than the nonhypertensive group. Stress- normotensive participants (Table 5). Both LV chamber diam- corrected midwall shortening was not statistically different eter and relative wall thickness were increased in the hyper- among groups. Circumferential end-systolic stress was sig- tensive and prehypertensive groups. As a result, LV mass in nificantly elevated in both the hypertensive and prehyperten- absolute terms or indexed for body surface area and height to sive groups. The circumferential end-systolic stress/end-sys- its allometric power was higher in both the hypertensive and tolic volume index, a load-adjusted measure of chamber prehypertensive groups. The prevalence of LV hypertrophy contractility, did not differ among groups after adjustment for was Ͼ3-fold higher in hypertensive and 2-fold higher in age, gender, diabetes, and center. prehypertensive participants than in their normotensive coun- terparts. Classification of LV geometry, using age-adjusted LV Diastolic Filling relative wall thickness27 and LV mass/height2.7 partition Isovolumic relaxation time was longer and the mean E values of 46.9 g/m2.7 in women and 49.2 g/m2.7 in men, shows velocity was slightly lower in hypertensive and prehyperten- minimally higher prevalence of concentric remodeling and sive participants, but the differences were not significant after 3-fold- and 2-fold-higher prevalences of eccentric LV hyper- adjustment for covariates (Table 4). More participants had trophy in the hypertensive and prehypertensive groups, re- prolonged isovolumic relaxation time in both the hyperten- spectively. Concentric LV hypertrophy was rare (nϭ6, 0.4%) sive and prehypertensive groups; however, after adjustment in the present study population. for covariates, the difference remained significant only in the hypertensive group. Mean mitral A velocity and atrial filling Discussion fraction were higher and the mean mitral E/A ratio was lower The present study documents high prevalences of hyperten- in both hypertensive and prehypertensive participants, even sion (15%) and prehypertension (35%) in a large population-

TABLE 3. Left Ventricular Systolic Function of Normotensive and Hypertensive Adolescent and Young Adult SHFS Participants

BP P *

NT PH HT NT vs PH PH vs HT NT vs HT Ejection fraction, % 60.4Ϯ5.0 59.6Ϯ5.0 59.5Ϯ5.2 0.150 0.863 0.867 2-Dimensional ejection fraction, % 62.5Ϯ2.5 62.4Ϯ2.7 62.3Ϯ3.3 1.00 1.00 1.00 Endocardial fractional shortening, % 32.5Ϯ3.4 32.1Ϯ3.5 32.1Ϯ3.6 0.306 0.809 0.981 Midwall shortening, % 18.7Ϯ1.6 18.1Ϯ1.6 17.8Ϯ1.6 Ͻ0.001 0.871 Ͻ0.001 Stress-corrected midwall shortening, % 1139Ϯ10 112Ϯ10 112Ϯ10 0.852 0.371 0.736 Circumferential end-systolic stress, kdyne/cm2 158Ϯ25 175Ϯ28 190Ϯ35 Ͻ0.001 Ͻ0.001 Ͻ0.001 Circumferential end-systolic stress/end-systolic 1.12Ϯ0.31 1.05Ϯ0.27 1.04Ϯ0.27 0.053 0.903 0.634 volume index, ϫ10 kdynes ⅐ cmϪ2 ⅐ mϪ2 Abbreviations as in Table 1. Values are meanϮSD unless otherwise indicated. *Adjusted for age, gender, diabetes status, and center location.

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Drukteinis et al Hypertension and Prehypertension in Young Adults 225

TABLE 4. Diastolic Function of Normotensive and Hypertensive Adolescent and Young Adult SHFS Participants

BP P *

NT PH HT NT vs PH PH vs HT NT vs HT IVRT, ms 75.3Ϯ10.4 77.3Ϯ10.4 80.0Ϯ11.8 0.631 0.923 0.463 IVRT Ͼ92 ms, n (%) 44 (4.6) 43 (6.5) 38 (13.3) 0.375 0.001 Ͻ0.001 Mitral E velocity, cm/s 73.3Ϯ13.2 70.3Ϯ13.8 68.4Ϯ16.3 0.943 0.893 0.992 Mitral A velocity, cm/s 48.9Ϯ11.3 52.1Ϯ11.7 58.5Ϯ14.5 Ͻ0.001 0.004 Ͻ0.001 Mitral E/A ratio 1.84Ϯ0.55 1.69Ϯ0.53 1.47Ϯ0.38 Ͻ0.001 0.001 Ͻ0.001 Mitral deceleration time, ms 166.2Ϯ36.1 169.5Ϯ35.4 177.0Ϯ37.4 0.840 0.799 0.455 Atrial filling fraction, % 25 28 32 Ͻ0.001 0.003 Ͻ0.001 IVRT indicates isovolumic relaxation time. Other abbreviations as in Table 1. Values are meanϮSD unless otherwise indicated. *Adjusted for age, gender, diabetes status, and center location.

based sample of adolescent and young adult American than without diabetes (24% versus 6%), consistent with Indians (mean age, 29 years) who are particularly affected by evidence implicating these factors in precipitating hyperten- the epidemic of obesity that is engulfing industrialized sion in middle-aged to elderly adults.34 nations. Twice as many participants in this population and Previous data in hypertensive adults indicated that lower age range had diastolic hypertension than systolic hyperten- stress-corrected midwall shortening predicts a higher rate of sion. The higher prevalence of diastolic hypertension in this cardiovascular events.35 Our population did not exhibit a younger age group is consistent with previous data that difference in circumferential end-systolic stress/end-systolic arterial elasticity decreases with age, consequently increasing volume index or stress-corrected midwall shortening between systolic hypertension, and is an independent risk factor for groups. The lack of difference in these contractility indexes cardiovascular disease.31,32 The presence of systolic hyper- between young normotensive and hypertensive SHS partici- tension in 102 participants and increased pulse pressure/ pants indicates that myocardial function has not yet been stroke index in the entire group of hypertensive participants compromised despite LV geometric abnormalities at this identifies an early prevalence of increased arterial stiffness early and relatively mild stage of hypertension. with multiple features related to insulin resistance, as sug- The percentage of patients with prolonged isovolumic relax- gested by the Bogalusa Heart Study.33 In the present study ation time was higher in the hypertensive and prehypertensive population, hypertension was more prevalent in obese than groups, which also had higher peak mitral A velocity and atrial nonobese participants (21% versus 9%), particularly those filling fractions. Hypertensive and prehypertensive participants with central adiposity (19% versus 9%), and in those with had significantly lower mean mitral E/A ratios, suggesting

TABLE 5. LV Geometry of Normotensive and Hypertensive Adolescent and Young Adult SHFS Participants

BP P *

NT PH HT NT vs PH PH vs HT NT vs HT IVSd, cm 0.79Ϯ0.10 0.86Ϯ0.11 0.91Ϯ0.11 Ͻ0.001 0.005 Ͻ0.001 LVIDd, cm 5.2Ϯ0.38 5.4Ϯ0.44 5.5Ϯ0.47 Ͻ0.001 0.485 Ͻ0.001 PWTd, cm 0.72Ϯ0.10 0.78Ϯ0.10 0.83Ϯ0.10 Ͻ0.001 0.018 Ͻ0.001 LVIDs, cm 3.5Ϯ0.35 3.7Ϯ0.40 3.8Ϯ0.45 Ͻ0.001 0.974 Ͻ0.001 LV mass, g 136.5Ϯ31.8 161.1Ϯ37.4 181.9Ϯ42.4 Ͻ0.001 0.014 Ͻ0.001 LV mass/BSA, g/m2 71.9Ϯ12.9 77.7Ϯ14.3 82.9Ϯ15.2 Ͻ0.001 0.346 Ͻ0.001 LV mass/height2.7, g/m2.7 34.7Ϯ7.7 38.4Ϯ7.9 41.4Ϯ8.6 Ͻ0.001 0.100 Ͻ0.001 RWTd, cm 0.28Ϯ0.04 0.29Ϯ0.04 0.30Ϯ0.04 Ͻ0.001 0.145 Ͻ0.001 LV hypertrophy, n (%) 62 (6.5) 76 (11.4) 56 (19.9) Ͻ0.000† Ͻ0.000† Ͻ0.000† Normal LV geometry, n (%) 889 (92.8) 579 (87.2) 220 (78.3) Ͻ0.000† Ͻ0.000† Ͻ0.000† Concentric remodeling, n (%) 7 (0.7) 9 (1.4) 5 (1.8) NS† NS† NS† Concentric LV hypertrophy, n (%) 4 (0.4) 2 (0.3) 0 (0) NS† NS† NS† Eccentric LV hypertrophy, n (%) 58 (6.1) 74 (11.1) 56 (19.9) Ͻ0.000† Ͻ0.000† Ͻ0.000† IVSd indicates interventricular septal thickness in diastole; LVIDd, LV internal dimension diastole; LVIDs, LVID in systole; PWTd, posterior wall thickness in diastole; BSA, body surface area; and RWTd, relative wall thickness, diastole. Other abbreviations as in Table 1. Values are meanϮSD unless otherwise indicated. *Adjusted for age. †Unadjusted for age.

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 226 Circulation January 16, 2007

slightly impaired LV relaxation that was offset by a greater atrial adult men and normotensive control subjects. From our data, it contribution to ventricular filling in a young, mildly hyperten- appears that prehypertension represents an intermediate point sive population. Our results are more marked than the marginal between hypertension and normal BP. In ANCOVAs, prehyper- differences in LV diastolic function reported by Palatini et al36 in tensive participants differ similarly from hypertensive and the the Hypertension and Ambulatory Recording Venetia Study now more stringently defined normotensive participants with (HARVEST) in young adults 18 to 45 years of age with mild, regard to changes in cardiac structure and function. The differ- stage 1 hypertension, a difference that may be due to the greater ences seen between prehypertensive and normotensive partici- prevalence and severity of obesity in our large, population-based pants may be due in part to a “supranormalization” of normo- sample of young participants. tensive control subjects. This suggests that new partition values The importance of systemic hypertension in the pathogen- for “normal” cardiac structural and functional measures are esis of LV hypertrophy is well established, and previous needed if prehypertension is documented to be a robust predictor population-based studies have shown strong associations of increased cardiovascular risk. Although our data suggest between LV mass and BP within the normal to slightly parallels between cardiovascular effects of prehypertension in elevated range.37–40 Even after adjustment for age, both the adolescents and young adults and the well-known detrimental hypertensive and prehypertensive groups exhibited cardiac effects of hypertension, further data from longitudinal observa- structural features associated with increased cardiovascular tional and therapeutic studies are needed before we can draw risk, including increased interventricular septal, posterior conclusions about the clinical implications of prehypertension. wall, and relative wall thicknesses and higher LV mass, LV mass/body surface area, and LV mass/height.2.7 In the Coro- Study Limitations nary Artery Risk Development in Young Adults (CARDIA) The present study assessed a population-based sample of study,37,38 systolic BP was responsible for much of the American Indians with higher prevalences of overweight and increase in LV mass over a 10-year period in young, healthy diabetes than the general US population. Although this may black and white men and women with low prevalences of limit the generalizability of the present results to some hypertension (1.6% to 3.8%). The present study differs from populations, our findings are relevant to the increasing previous ones in population-based samples in grouping ado- proportion of young adults in developed countries who suffer lescents and young adults by the JNC-7 categories of normal from overweight and/or diabetes. In addition, despite exten- BP, prehypertension, and hypertension and by documenting sive efforts to standardize measurements, with the use of the substantial prevalences of prognostically important measures same echocardiographic method as in other clinical and of preclinical cardiovascular disease in prehypertensive and population-based studies, subtle differences in performance especially in hypertensive participants despite their young technique among sites also could have influenced results. age. Future studies are needed to determine whether the abnormalities of LV structure that we have identified at this Conclusions stage contribute to subsequent cardiovascular events. Our results show that even small elevations in BP, as seen The prevalence of LV hypertrophy was elevated in hyper- with prehypertension, can have detrimental effects on hemo- tensive and prehypertensive SHFS participants and was dynamics and cardiovascular structure and function in an similar to findings in hypertensive children and adolescents adolescent and young adult American Indian population with by Daniels et al.41 Approximately 20% of hypertensive many of the same risk factors plaguing the United States and participants in the present study had eccentric LV hypertro- industrialized countries around the world. phy, but few had either concentric LV hypertrophy or concentric LV remodeling. This is in contrast to 16% of Acknowledgments hypertensive children having concentric hypertrophy in a We thank the Indian Health Service, SHS participants, and partici- study by Hanevold et al,10 with especially higher prevalences pating tribal communities for their extraordinary cooperation and among Hispanic or black subjects. The absence of concentric involvement that made this study possible; Betty Jarvis, RN, Tauqeer Ali, MD, and Marcia O’Leary for their coordination of 3 study LVH in young SHS participants may reflect their generally centers; Dawn Fishman, BA, for her data coordination and manage- mild hypertension, whereas the high prevalence of eccentric ment of the database; Tauqeer Ali, MD, Rosina Briones, RDMS, LVH may be related to the high prevalence of obesity in our Joanne Carter, RDMS, for their technical assistance; and Virginia M. population, with a larger volume of circulating plasma.42 Burns for her assistance in the preparation of this manuscript. Our study is the first large population-based study to examine specifically the cardiac and systemic hemodynamic status of the Sources of Funding new JNC-7 category of prehypertension. A recent study by This work was supported in part by grants HL41642, HL41652, Zhang et al8 provides evidence for an increase in incident HL41654, HL65521, and M10RR0047–34 (GCRC) from the Na- tional Institutes of Health, Bethesda, Md. cardiovascular disease among prehypertensive adults 45 to 74 years of age at baseline and an even more striking incident among those with both prehypertension and diabetes in the SHS, Disclosures None. suggesting a need for more vigilant monitoring of prehyperten- 43 sive adults. A previous report by Toikka et al used borderline References hypertension as a model for prehypertension before the new 1. Mokdad AJ, Ford ES, Bowman BA, Dietz WH, Vinivor F, Bales VS, JNC-7 criteria and found no difference in LV mass but differ- Marks JS. Prevalence of obesity, diabetes, and obesity-related health risk ences in LV geometry between borderline hypertensive young factors, 2001. JAMA. 2003;289:76–79.

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Drukteinis et al Hypertension and Prehypertension in Young Adults 227

2. Fields LE, Burt VL, Cutler JA, Hughes J, Roccella EJ, Sorlie P. The burden of 23. Schiller NB, Shah PM, Crawford M, De Maria A, Devereux R, Feigenbaum H, adult hypertension in the United States 1999 to 2000: a rising tide. Hypertension. Gutgesell H, Reichek N, Sahn D, Schnittger I, Silverman NH, Tajik AJ. The 2004;44:398–404. American Society of Echocardiography Committee on Standards, Subcommittee 3. Gidding SS, Bao W, Srinivasan SR, Berenson GS. Effects of secular trends in on Quantitation of Two-Dimensional Echocardiograms: recommendations for obesity on coronary risk factors in children: the Bogalusa Heart Study. J Pediatr. quantitation of the left ventricle by two-dimensional echocardiography. JAmSoc 1995;127:868–874. Echocardiogr. 1989;2:358–367. 4. Lauer RM, Clarke WR, Mahoney LT, Witt J. Childhood predictors of adult blood 24. Roman MJ, Devereux RB, Kramer-Fox R, O’Loughlin J. Two-dimensional pressure: the Muscatine Study. Pediatrics. 1993;40:23–40. echocardiographic aortic root dimensions in children and adults: biologic deter- 5. Chobanian AV, Bakris GL, Black HR, Cushman WC, Green LA, Izzo JL Jr, minants and normal limits. Am J Cardiol. 1989;64:507–512. Jones DW, Materson BJ, Oparil S, Wright JT Jr, Roccella EJ. The Seventh 25. Dubin J, Wallerson DC, Cody RJ, Devereux RB. Comparative accuracy of Report of the Joint National Committee on Prevention, Detection, Evaluation and Doppler echocardiographic methods for clinical stroke volume determinations. Treatment of High Blood Pressure: the JNC 7 Report. JAMA. 2003;289: Am Heart J. 1990;120:116–123. 2560–2572. 26. Devereux RB, Alonso Dr, Lutas EM, Gottlieb GJ, Campo E, Sachs I, Reichek N. 6. Wang Y, Wand QJ. The prevalence of prehypertension and hypertension among Echocardiographic assessment of left ventricular hypertrophy: comparison to US adults according to the new Joint National Committee guidelines: new necropsy findings. Am J Cardiol. 1986;57:450–458. challenges of the old problem. Arch Intern Med. 2004;164:2126–2134. 27. de Simone G, Daniels SR, Kimball TR, Roman MJ, Romano C, Chinali M, 7. Vasan RS, Larson MG, Leip EP, Evans JC, O’Donnell CJ, Kannel WB, Levy D. Galderisi M, Devereux RB. Evaluation of concentric left ventricular geometry in Impact of high-normal blood pressure on the risk of cardiovascular disease. humans, evidence for age-related systematic underestimation. Hypertension. N Engl J Med. 2001;345:1291–1297. 2005;45:64–68. 8. Zhang Y, Lee ET, Devereux RB, Yeh J, Best LG, Fabsitz RR, Howard BV. 28. de Simone G, Roman MJ, Daniels SR, Mureddu G, Kimball TR, Greco R, Prehypertension, diabetes, and cardiovascular disease risk in a population-based Devereux RB. Age-related changes in total arterial capacitance from birth to sample: the Strong Heart Study. Hypertension. 2006;47:410–414. maturity in a normotensive population. Hypertension. 1997;29:1213–1217. 9. Gardin JM, Wagenknecht LE, Anton-Culver H, Flack J, Gidding S, Kurosaki T, 29. de Simone G, Devereux RB, Roman MJ, Ganau A, Saba PS, Alderman MH, Wong ND, Manolio TA. Relationship of cardiovascular risk factors to echocar- Laragh JH. Assessment of left ventricular function by the mid-wall fractional diographic left ventricular mass in healthy young black and white adult men and shortening-end-systolic stress relation in human hypertension. J Am Coll Cardiol. women: the CARDIA Study. Circulation. 1995;92:380–387. 1994;23–1444–1451. 10. Hanevold C, Waller J, Daniels S, Portman R, Sorof J. The effects of obesity, 30. Devereux RB, de Simone G, Pickering TG, Schwartz JE, Roman MJ. Relation of gender and ethnic group on left ventricular hypertrophy and geometry in hyper- left ventricular midwall function to cardiovascular risk factors and arterial tensive children: a collaborative study of the international pediatric hypertension structure and function. Hypertension. 1998;31:929–936. association. Pediatrics. 2004;113:328–333. 31. Gratzka CD, Cameron JD, Kingwell BA, Dart AM. Relation between coronary 11. Lee ET, Welty TK, Fabsitz R, Cowan LD, Le NA, Oopik AJ, Cucciara AJ, artery diseases, aortic stiffness, and left ventricular structure in a population Savage PJ, Howard BV. The Strong Heart Study: a study of cardiovascular sample. Hypertension. 1998;32:575–578. disease in American Indians: design and methods. Am J Epidemiol. 1990;132: 32. Arnett DK, Evans GW, Riley WA. Arterial stiffness: a new cardiovascular risk 1141–1155. factor? Am J Epidemiol. 1994;140:669–682. 12. Howard BV, Welty TK, Fabsitz RR, Cowan LD, Oopik AJ, Le NA, Yeh J, 33. Urbina EM, Srinivasan SR, Kielyka RL, Tang R, Bond MG, Chen W, Berenson Savage PJ, Lee ET. Risk factors for coronary heart disease in diabetic and GS. Correlates of carotid stiffness in young adults: the Bogalusa Heart Study. non-diabetic Native Americans. Diabetes. 1992;41(suppl 2):4–11. Atheroslcerosis. 2004;176:157–164. 13. Welty TK, Lee ET, Yeh JL, Cowan LD, Go O, Fabsitz RR, Le NA, Oopik AJ, 34. de Simone G, Devereux RB, Chinali M, Roman MJ, Best LG, Welty TK, Lee Robbins DC, Howard BV. Cardiovascular disease risk factors among American ET, Howard BV, for the Strong Heart Study Investigators. Risk factors for arterial Indians: the Strong Heart Study. Am J Epidemiol. 1995;142:269–287. hypertension in adults with initial optimal blood pressure: the Strong Heart Study. 14. Chinali M, de Simone G, Roman MJ, Lee ET, Best LG, Howard BV, Devereux Hypertension. 2006;47:1–7. RB. Impact of obesity on cardiac geometry and function in a population of 35. de Simone G, Devereux RB, Koren MJ, Mensah GA, Casale PN, Laragh JH. adolescents: the Strong Heart Study. J Am Coll Cardiol. 2006;47:2267–2273. Midwall left ventricular mechanics: an independent predictor of cardiovascular 15. National Heart, Lung, and Blood Institute. The Seventh Report of the Joint risk in arterial hypertension. Circulation. 1996;93:259–265. National Committee on Detection, Evaluation, and Treatment of High Blood 36. Palatini P, Visentin P, Mormino P, Pietra M, Piccolo D, Cozzutti E, Mione V, Pressure. Bethesda, Md: National Institutes of Health; December 2003. NIH Bocca P, Perissinotto F, Pessina AC. Left ventricular performance in the early publication 03–5233. stages of systemic hypertension. Am Heart J. 2001;142:1016–1023. 16. Stolarczyk LM, Heyward VH, Hicks VL, Baumgartner RN. Predictive accuracy 37. Gardin JM, Brunner D, Schreiner PJ, Xie X, Reid CL, Ruth K, Bild DE, Gidding of bioelectric impedance in estimating body composition of American women. S. Demographics and correlates of five-year change in echocardiographic left Am J Clin Nutr. 1994;59:964–970. ventricular mass in young black and white adult men and women: the Coronary 17. Rising R, Swinburn B, Larson K, Ravussin E. Body composition in Pima Indians: Artery Risk Development in Young Adults (CARDIA) study. J Am Coll Cardiol. validation of bioelectric resistance. Am J Clin Nutr. 1991;53:594–598. 2002;40:529–535. 18. National Institutes of Health. Clinical Guidelines on the Identification, Eval- 38. Lorber R, Gidding SS, Daviglus ML, Colangelo LA, Liu K, Gardin JM. Influence uation, and Treatment of Overweight and Obesity in Adults: The Evidence of systolic blood pressure and body mass index on left ventricular structure in Report. Bethesda, Md: National Institutes of Health; September 1998. NIH healthy African-American and white young adults: the CARDIA study. JAm publication 98–4083. Coll Cardiol. 2003;41:955–960. 19. Expert Committee on the Diagnosis and Classification of Diabetes Mellitus. 39. Burke GL, Arcilla RA, Culpepper WS, Webber LS, Chiang YK, Berenson GS. Report of the Expert Committee on the Diagnosis and Classification of Diabetes Blood pressure and echocardiographic measures in children: the Bogalusa Heart Mellitus. Diabetes Care. 1997;20:1183–1197. Study. Circulation. 1987;75:106–114. 20. Devereux RB, Roman JR, Paranicas M, O’Grady MJ, Wood EA, Howard BV, 40. Urbina EM, Gidding SS, Bao W, Pickoff AS, Berdusis K, Berenson GS. Welty TK, Lee ET, Fabsitz RR. Relations of Doppler stroke distance and aortic Effect of body size, ponderosity, and blood pressure on left ventricular growth annular diameter to left ventricular stroke volume in normotensive and hyper- in children and young adults in the Bogalusa Heart Study. Circulation. tensive American Indians: the Strong Heart Study. Am J Hypertens. 1997;10: 1995;91:2400–2406. 619–628. 41. Daniels SR, Loggie JM, Khoury P, Kimball TR. Left ventricular geometry and 21. Devereux RB, Roman MJ, de Simone G, O’Grady MJ, Paranicas M, Yeh J-L, severe left ventricular hypertrophy in children and adolescents with essential Fabsitz RR, Howard BV, for the Strong Heart Study Investigators. Relations of hypertension. Circulation. 1998;97:1907–1911. left ventricular mass to demographic and hemodynamic variables in American 42. Ganau A, Arru A, Saba PS, Piga G, Glorioso N, Tonolo G, Mardeddu G, Bianchi Indians: the Strong Heart Study. Circulation. 1997;96:1416–1423. G. Stroke volume and left heart anatomy in relation to plasma volume in essential 22. Sahn DJ, De Maria A, Kisslo J, Weyman AE. The Committee on M-mode hypertension. J Hypertens. 1991;9(suppl 6):S150–S151. Standardization of the American Society of Echocardiography: recommendations 43. Toikka JO, Laine H, Ahotupa M, Haapanen A, Viikari JS, Hartiala JJ, Raitakari regarding quantitation in M-mode echocardiography: results of a survey of echo- OT. Increased arterial intima-medial thickness and in vivo LDL oxidation in cardiographic measurements. Circulation. 1978;58:1072–1083. young men with borderline hypertension. Hypertension. 2000;36:929–933.

Go to http://cme.ahajournals.org to take the CME quiz for this article.

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Imaging

Subclinical Coronary and Aortic Atherosclerosis Detected by Magnetic Resonance Imaging in Type 1 Diabetes With and Without Diabetic Nephropathy

Won Yong Kim, MD; Anne Sofie Astrup, MD; Matthias Stuber, PhD; Lise Tarnow, MD; Erling Falk, MD; René M. Botnar, PhD; Cheryl Simonsen, RT; Lotte Pietraszek; Peter R. Hansen, MD; Warren J. Manning, MD; Niels T. Andersen, PhD; Hans-Henrik Parving, MD

Background—Patients with type 1 diabetes and nephropathy maintain an excess cardiovascular mortality compared with diabetic patients with normoalbuminuria. We sought to evaluate coronary and aortic atherosclerosis in a cohort of asymptomatic type 1 diabetic patients with and without diabetic nephropathy using cardiovascular magnetic resonance imaging. Methods and Results—In a cross-sectional study, 136 subjects with long-standing type 1 diabetes without symptoms or history of cardiovascular disease, including 63 patients (46%) with nephropathy and 73 patients with normoalbuminuria, underwent cardiovascular magnetic resonance imaging. All subjects underwent cardiac exercise testing and noninvasive tests for peripheral artery disease and autonomic neuropathy. Coronary artery stenoses were identified in 10% of subjects with nephropathy (versus 0% with normoalbuminuria; Pϭ0.007). Coronary plaque burden, expressed as right coronary artery mean wall thickness (1.7Ϯ0.3 versus 1.3Ϯ0.2 mm; PϽ0.001) and maximum right coronary artery wall thickness (2.2Ϯ0.5 versus 1.6Ϯ0.3 mm; PϽ0.001), was greater in subjects with nephropathy. The prevalence of thoracic (3% versus 0%; Pϭ0.28) and abdominal aortic plaque (22% versus 16%; Pϭ0.7) was similar in both groups. Subjects with and without abdominal aortic plaques had similar coronary plaque burden. Conclusions—In asymptomatic type 1 diabetes, cardiovascular magnetic resonance imaging reveals greater coronary plaque burden in subjects with nephropathy compared with those with normoalbuminuria. (Circulation. 2007;115:228- 235.) Key Words: aorta Ⅲ atherosclerosis Ⅲ coronary disease Ⅲ diabetes mellitus Ⅲ magnetic resonance imaging

therosclerotic cardiovascular disease, in particular cor- Clinical Perspective p 235 Aonary heart disease (CHD), is a major cause of mortality and morbidity in both type 1 and type 2 diabetes.1 Risk imaging of atherothrombosis in the aorta6,7 and carotid8 assessment in asymptomatic patients with type 1 diabetes is and coronary arteries.9–11 We sought to evaluate the effect of less stringent than for type 2 diabetes. The US guidelines for diabetic nephropathy on atherosclerosis in type 1 diabetes the prevention of CHD consider diabetes a CHD risk equiv- using CMRI to evaluate subclinical coronary and aortic alent, regardless of diabetes type, age, and kidney function.2 atherosclerosis in asymptomatic type 1 diabetic patients with The European guidelines do not consider type 1 diabetes to be and without nephropathy. a high-risk state unless microalbuminuria is present.3 In a recent prospective follow-up study, we found that the in- Methods creased cardiovascular mortality in type 1 diabetes was due Patient Cohort and Clinical Measurements predominantly to a poor prognosis in patients with diabetic The study cohort consisted of 136 subjects with long-standing type 4 nephropathy. 1 diabetes without cardiovascular symptoms or disease who were Cardiovascular magnetic resonance imaging (CMRI) al- randomly selected from a larger cohort of Ϸ3000 patients seen in the lows noninvasive detection of coronary artery stenoses5 and Steno Diabetes Center between July 2003 and February 2005. There

Received April 20, 2006; accepted November 2, 2006. From the Department of Cardiology (W.Y.K., E.F.) and MR Center (W.Y.K., C.S.), Skejby Hospital, Aarhus University Hospital, Aarhus, Denmark; Steno Diabetes Center, Gentofte, Denmark (A.S.A., L.T., L.P., H.-H.P.); Russell H. Morgan Department of Radiology and Radiological Science, and Department of Electrical and Computer Engineering, Baltimore, Md (M.S.); Department of Nuclear Medicine, Technical University Munich, Munich, Germany (R.M.B.); Department of Cardiology, Gentofte University Hospital, Gentofte, Denmark (P.R.H.); Department of Medicine, Cardiovascular Division and Department of Radiology, Beth Israel Deaconess Medical Center and Harvard Medical School, Boston, Mass (W.J.M.); and Department of Biostatistics, Institute of Public Health, Aarhus University, Aarhus, Denmark (N.T.A.). Guest Editor for this article was Edgardo Escobar, MD. Correspondence to Won Yong Kim, MD, Department of Cardiology, Skejby Hospital, Brendstrupgaardsvej, 8200 Aarhus N, Denmark. E-mail [email protected] © 2007 American Heart Association, Inc. Circulation is available at http://www.circulationaha.org DOI: 10.1161/CIRCULATIONAHA.106.633339

228

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Kim et al Subclinical Atherosclerosis in Type 1 Diabetes 229

were 400 patients with diabetic nephropathy in this larger cohort, which included 46 subjects from a prior 10-year prognostic follow-up study (31 with nephropathy, 15 with albuminuria).4 Of the 136 subjects in the study cohort, 63 (43%) had diabetic nephropathy, defined as previously persistent albuminuria (urinary albumin excre- tion rate Ͼ300 mg/24 h in 2 of 3 consecutive 24-hour urine collections), presence of retinopathy, and no evidence of other renal or urinary tract disease.12 Blood samples were obtained after 10 minutes of rest. Serum creatinine concentration was assessed by a kinetic Jaffé method. Urinary albumin excretion rate was measured in two 24-hour urine collections by an enzyme immunoassay.13 The most recent measure- ment of glomerular filtration rate in the patient file was noted because all patients with diabetic nephropathy have their glomerular filtration rate measured every year by a 51Cr-EDTA-clearance tech- nique.14 The 24-hour ambulatory blood pressure was measured with a Takeda (Osaka, Japan) TM 2421 device. Values were averaged for each hour before the average 24-hour blood pressure was calculated. Ankle-brachial pressure index and systolic blood pressures in the big toe were measured by strain-gauge technique.15 Tests for autonomic neuropathy, heart rate variability, and orthostatic blood pressure were performed. In addition, somatic nerve function (vibratory perception threshold) was evaluated by biothesiometry. Bicycle exercise ECG was carried out in accordance with consensus guide- lines.16 Test results were analyzed by a masked cardiologist (P.R.H.) and classified as either pathological or normal. If the subject failed to Figure 1. Three-dimensional reformatted black-blood vessel wall reach 85% of the estimated maximum heart rate, the test was imaging of the proximal RCA showing focal coronary vessel wall classified as inconclusive. The present study was approved by the thickening (dotted line represents maximum wall thickness of local ethics committee, and all patients gave written informed Ϸ2.6 mm). Average vessel wall thickness (Ϸ1.8 mm) is calcu- consent. lated as the mean of the anterior and posterior wall thicknesses along the proximal RCA as indicated by the black lines. CMRI Studies All subjects underwent CMRI examination with a Philips Intera 1.5 by consensus of 2 reviewers (W.Y.K., C.S.) blinded to all clinical T MR whole-body scanner (Philips Medical Systems, Best, The data. Aortic vessel wall and plaque data were classified as thoracic or Netherlands) equipped with a 5-element cardiac phased-array coil abdominal according to their location above or below the diaphragm, and a cardiac software package (R9.1.1). All coronary CMRI respectively. Atherosclerotic plaque was defined as characteristic angiograms were acquired according to a previously validated luminal protrusions of Ն1 mm in radial thickness.7 The cross- protocol.5,17,18 sectional vessel wall area in each slice was computed. The total Right coronary artery (RCA) vessel wall scanning was done in a thoracic and abdominal aortic vessel wall volumes were then subset of subjects when total scan time did not exceed 1.5 hours (24 calculated as the sum of each slice area multiplied by the slice gap. with nephropathy, 37 patients without nephropathy) with To quantify the magnitude of aortic atherosclerosis in each subject, 3-dimensional black-blood imaging according to a previously vali- 2 estimates of plaque burden, slice plaque burden and area plaque dated protocol.11,19 burden, were determined.7 To evaluate aortic atherosclerosis, subjects underwent thoracoab- dominal aortic CMRI with acquisition of 24 transverse slices Statistical Analysis spanning the aortic arch to the aortoiliac bifurcation using a Variables were compared between groups by a 2-sample Student t ECG-gated, T2-weighted, and fat-suppressed black-blood turbo test or Wilcoxon rank-sum test. Frequencies were compared by use spin-echo sequence.7 of a ␹2 test or Fisher exact test. Multiple linear regression analysis was performed to investigate Coronary Arteries: Stenosis and Plaque Burden the associations between CPB and the clinical variables (including The 3-dimensional coronary data sets were reformatted along the tests for peripheral artery disease and autonomic neuropathy) listed entire visualized course of the coronaries and analyzed by an in Tables 1 and 2. Each of the clinical variables was included in the experienced reviewer (W.Y.K.) blinded to all clinical data. The analysis separately with and without diabetic nephropathy. The analysis was performed twice with and without the assumption that image quality of all coronary images was visually graded as follows: the association is different for subjects with and without nephropa- 1ϭpoor/uninterpretable, 2ϭgood, 3ϭvery good, and 4ϭexcellent.5 thy. This was done by including an interaction term for clinical Coronary angiograms of grades 2 through 4 were classified on the variables and nephropathy for the model assuming different slopes. basis of the visual assessment of the coronary artery lumen: none or The multiple linear regression analysis also allows us to compare minimal luminal narrowing or with significant disease (marked CPB between subjects with and without nephropathy adjusted for the 5 attenuation of coronary lumen signal). RCA vessel wall scans of clinical variables. Two-tailed values of PϽ0.05 were considered grades 2 through 4 were classified by visual interpretation of vessel significant. All calculations were made with SPSS version 13.0 walls: plaque or no plaque, depending on the presence or absence of (SPSS, Chicago, Ill) and STATA version 9.2 (College Station, Tex). focal coronary vessel wall thickening in any of the visualized segments. The authors had full access to the data and take full responsibility Coronary plaque burden (CPB) was quantified as the mean and for the integrity of the data. All authors have read and agree to the maximum RCA vessel wall thicknesses along the entire proximal course manuscript as written. of the vessel by semiautomated segmentation of the 3-dimensional reformatted black-blood vessel wall scans (Figure 1).11,20 Results Aorta: Vessel Wall Volume and Plaque Burden Subject Characteristics and Nonimaging tests Aortic CMRI data were transferred to a commercial EasyVision The clinical characteristics of the study cohort are summa- Work Station (version 3.0, Philips Medical Systems) and analyzed rized in Table 1. Subjects with nephropathy were on average

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 230 Circulation January 16, 2007

TABLE 1. Clinical Characteristics in 136 Type 1 Diabetic Patients With and Without Diabetic Nephropathy

Patients With Patients With Normoalbuminuria Diabetic Nephropathy (Nϭ73) (Nϭ63) P Clinical variables Male sex male, n (%) 43 (59) 36 (57) 0.84 Age, y 52Ϯ948Ϯ9 0.01 Duration of diabetes, y 31Ϯ734Ϯ9 0.02 Duration of nephropathy, y ⅐⅐⅐ 15Ϯ7 ⅐⅐⅐ BMI, kg/m2 24.7Ϯ2.7 24.4Ϯ2.9 0.62

HbA1c,% 8.0Ϯ0.9 8.8Ϯ1.1 Ͻ0.001 Glomerular filtration rate, mL ⅐ minϪ2 ⅐ 1.73 mϪ2 ⅐⅐⅐ 66Ϯ34 ⅐⅐⅐ Urinary albumin excretion rate, mg/24 h 8 (2–29) 209 (3–6375)* Ͻ0.001 S-creatinine, ␮mol/L 89 (61–116) 119 (68–825) Ͻ0.001 Systolic blood pressure, mm Hg 129Ϯ15 135Ϯ18 0.065 Diastolic blood pressure, mm Hg 71Ϯ671Ϯ10 0.58 Total cholesterol, mmol/L 4.8Ϯ0.8 4.7Ϯ0.8 0.94 HDL cholesterol, mmol/L 1.9Ϯ0.5 1.8Ϯ0.5 0.2 LDL cholesterol, mmol/L 2.4Ϯ0.6 2.3Ϯ0.8 0.46 VLDL cholesterol, mmol/L 0.4Ϯ0.2 0.6Ϯ0.3 0.007 Triglycerides, mmol/L 0.8 (0.3–2.6) 1.1 (0.5–4.8) 0.002 Hemoglobin, mmol/L 8.5Ϯ0.7 7.9Ϯ0.8 Ͻ0.001 Current smoker, n (%) 17 (23) 15 (24) 0.94 Renoprotective and cardioprotective treatment Statins, n (%) 20 (27) 36 (57) Ͻ0.001 Aspirin, n (%) 19 (26) 45 (71) Ͻ0.001 Patients receiving AHT, n (%) 33 (45) 59 (93) Ͻ0.001 Average AHT drugs per patient, n 0.8Ϯ1.2 2.3Ϯ1.2 Ͻ0.001 BMI indicates body mass index; HDL, high-density lipoprotein; LDL, low-density lipoprotein; VLDL, very-low-density lipoprotein; and AHT, antihypertensive treatment. Data are meanϮSD or median (range) as appropriate. *Some patients with previously persistent albuminuria had a urinary albumin excretion rate Ͻ300 mg/24 h after receiving AHT.

4 years younger than normoalbuminuric patients. Glycemic pharmacological treatments in the 2 groups were different control was poorer in subjects with nephropathy (PϽ0.001), (Table 1). None of the subjects (0%) with normoalbuminuria and the data suggest a higher systolic blood pressure had ischemia on exercise testing, whereas 2 subjects (3%) (Pϭ0.065). Total cholesterol, low-density lipoprotein choles- with nephropathy had inducible ischemia (Pϭ0.05). Subjects terol, and high-density lipoprotein cholesterol levels were with nephropathy more often had inconclusive exercise tests similar, whereas triglycerides (PϽ0.03) and very low-density (47% versus 14%; PϽ0.001). The 2 groups had markedly lipoprotein cholesterol (PϽ0.007) were higher in subjects different autonomic and somatic nerve function (Table 2). with nephropathy. The renoprotective and cardioprotective Subjects with nephropathy also had reduced ankle-brachial

TABLE 2. Tests for Autonomic Neuropathy and Peripheral Arterial Disease

Subjects With Subjects With Diabetic Normoalbuminuria (Nϭ68) Nephropathy (Nϭ63) P Vibratory perception threshold, mV 22Ϯ10 30Ϯ14 Ͻ0.001 Heart rate at rest, bpm 69Ϯ15 77Ϯ10 Ͻ0.001 Heart rate variation during deep breathing, bpm 15Ϯ97Ϯ6 Ͻ0.001 Postural blood pressure decrease, mm Hg 2.2Ϯ10 5.4Ϯ16 0.17 Ankle-brachial index 1.1Ϯ0.1 1.0Ϯ0.1 0.03 Blood pressure, big toe, mm Hg 113Ϯ27 96Ϯ37 0.001 Postural blood pressure decrease shows a reduction in systolic blood pressure after 7 minutes in the upright position vs lying down.

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Kim et al Subclinical Atherosclerosis in Type 1 Diabetes 231

TABLE 3. Coronary Plaque Burden

Subjects With Subjects With Diabetic Normoalbuminuria (Nϭ37) Nephropathy (Nϭ24) P RCA interpretable, n (%) 33 (89) 21 (88) 0.8 RCA VW image quality 3.2Ϯ0.9 2.9Ϯ0.7 0.2 RCA VW mean thickness, mm 1.3Ϯ0.2 1.7Ϯ0.3 Ͻ0.001 RCA VW maximum thickness, mm 1.6Ϯ0.3 2.2Ϯ0.5 Ͻ0.001 RCA plaque detected, n (%) 5 (15) 16 (76) Ͻ0.001 VW indicates vessel wall. Data are meanϮSD when appropriate.

pressure index (PϽ0.03) and big toe blood pressure minuria, all clinical characteristics were similar between the (PϽ0.001; Table 2). subset of subjects who had CPB assessment and those who did not. In subjects with nephropathy, systolic (127Ϯ18 CMRI Measurements versus 139Ϯ17 mm Hg; Pϭ0.01) and diastolic (67Ϯ8 versus Coronary MRI was completed in 131 of the 136 subjects 74Ϯ10 mm Hg; Pϭ0.01) blood pressures were lower in those (96%). Ten percent of patients with nephropathy versus 0% subjects who underwent CPB scanning (nϭ21) compared of those with normoalbuminuria had CMRI evidence of with those who did not (nϭ42). Heart rate (73Ϯ11 versus coronary stenoses (1 vessel, 8%; 2 vessel, 2%) (PϽ0.007 97Ϯ8 bpm; Pϭ0.02) and statin use (33% versus 69%; versus normoalbuminuric). Both subjects with ischemia on Pϭ0.007) also were lower in subjects who underwent CPB exercise testing showed evidence of coronary stenosis by assessment. CMRI. The CPB (mean and maximum RCA vessel wall CMRI of aortic atherosclerosis was completed in 128 of thicknesses) was increased in patients with nephropathy the subjects (94%). In 8 subjects, imaging was not completed compared with patients with normoalbuminuria (Table 3 and or the image quality was inadequate for assessment. The Figure 2). The prevalence of visually identifiable coronary prevalence of thoracic and abdominal plaques was similar in plaque also was higher in subjects with nephropathy (76% both patient groups (PϭNS; Table 4). Subjects with (nϭ8) versus 15%; PϽ0.001; Table 3). In patients with normoalbu- and without (nϭ44) abdominal aortic plaques had similar CPB (mean vessel wall thickness, 1.6Ϯ0.3 versus 1.5Ϯ0.3 mm; Pϭ0.15; maximum vessel wall thickness, 2.0Ϯ0.4 versus 1.8Ϯ0.5 mm; Pϭ0.22). The results of the multiple regression analysis of CPB for the clinical variables with at least 1 value of PϽ0.10 are summarized in Table 5. There was a significant association between CPB and the duration of diabetes (Figure 3) and total cholesterol level. In subjects with normoalbuminuria, CPB was significantly associated with systolic (significantly dif- ferent from subjects with nephropathy; Figure 4) and diastolic blood pressure, whereas in patients with nephropathy, CPB showed an association with low-density lipoprotein choles- terol levels. After the effects of the clinical variables were accounted for, there was still a significant effect of nephrop- athy on CPB, with CPB being significantly higher for subjects with nephropathy compared with subjects with normoalbuminuria. Thus, for all clinical variables (indepen- dent variables), regression analysis showed significantly greater intercepts (ie, CPB) for subjects with nephropathy compared with subjects with normoalbuminuria (eg, Figure 3).

Discussion Figure 2. Three-dimensional reformatted coronary MRI of the To the best of our knowledge, the present study is the first to proximal RCA in 2 subjects without coronary luminal stenosis: a evaluate CPB and aortic plaque burden noninvasively in an 58-year-old man with long-standing type 1 diabetes and nor- asymptomatic cohort with long-standing type 1 diabetes. moalbuminuria (A) and a 44-year-old man with long-standing type 1 diabetes and diabetic nephropathy (C). The correspond- In this cross-sectional study of 136 asymptomatic type 1 ing 3-dimensional black-blood vessel wall scans show no CMRI diabetic patients, CMRI revealed greater CPB and higher evidence of atherosclerotic plaque (B; average and maximum prevalence of coronary artery stenoses in patients with vessel wall thickness, 1.1 and 1.3 mm, respectively) and an diabetic nephropathy compared with those with normoalbu- increased atherosclerotic plaque burden (D; average and maxi- mum vessel wall thickness, 2.3 and 3.0 mm, respectively). The minuria. None of our type 1 diabetic patients with normoal- anterior and posterior RCA walls are indicated by arrows. buminuria had coronary artery stenoses or ischemia on

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 232 Circulation January 16, 2007

TABLE 4. Aortic Plaque Burden

Subjects With Subjects With Diabetic Normoalbuminuria (Nϭ68) Nephropathy (Nϭ60) P Thoracic plaque frequency, n (%) 0 (0) 2 (3) 0.28 Thoracic VW volume, mm3 6718Ϯ1631 7130Ϯ1594 0.14 Thoracic APB, % 0Ϯ0 0.08Ϯ0.5 0.15 Thoracic SPB, % 0Ϯ0 0.4Ϯ2 0.12 Abdominal plaque frequency, n (%) 11 (16) 13 (22) 0.7 Abdominal VW volume, mm3 5781Ϯ2725 5330Ϯ2229 0.31 Abdominal APB, % 0.68Ϯ1.9 0.64Ϯ1.6 0.90 Abdominal SPB, % 0.76Ϯ2.1 0.72Ϯ2.0 0.91 VW indicates vessel wall; APB, area plaque burden; and SPB, slice plaque burden. Data are meanϮSD when appropriate.

exercise testing, and the prevalence of subclinical coronary different risk factors. Nephropathy was an independent risk plaque was low (15%). These data suggest that asymptomatic factor for CHD but not for peripheral artery disease.25 patients with type 1 diabetes and normoalbuminuria are not Existing CHD risk prediction models have been developed necessarily at high risk of atherosclerotic cardiovascular with data from the general population26 or from patients with disease, which is in accordance with the most recent prospec- type 2 diabetes27 and therefore do not accurately predict tive follow-up study4 and current European guidelines for the cardiovascular events in type 1 diabetes.28 Furthermore, prevention of CHD.3 intensive contemporary cardioprotective and renoprotective In comparison, the prevalence of asymptomatic coronary medication has frequently been implemented, making risk stenosis investigated with invasive coronary angiography was stratification more challenging because the modifiable risk 47% among 110 pretransplant patients with type 1 diabetes21 factors have already been favorably changed from baseline. and 34% in a cohort of 29 asymptomatic diabetics.22 These Our data based on regression analysis showed that CPB was differences may be explained by clinical differences in the correlated to diabetes duration and present total cholesterol study groups and methodology and more recent intensified levels, whereas blood pressure showed a positive association interventions toward reducing CHD risk factors. with CPB only in patients with normoalbuminuria. The lack Aortic plaque burden was similar between groups, and the of correlation between blood pressure and CPB in patients prevalence of abdominal and thoracic plaques was compara- with nephropathy most likely reflects the more intensive use ble to a previously reported group of asymptomatic subjects of antihypertensive therapy. The cross-sectional design of our from the Framingham Heart Study offspring cohort.7 study precludes assessment of cardiovascular risk profile Prior studies have shown an association between angio- during the entire follow-up period; therefore, the long-term graphically determined extent of coronary artery stenosis and effect of modifiable risk factors on atherosclerosis could not atherosclerosis in the thoracic aorta in patients with CHD.23,24 be evaluated. In our population of asymptomatic type 1 diabetics, we found Cardiovascular autonomic neuropathy has been identified similar magnitudes of CPB in subjects with and without as an independent risk factor for cardiovascular morbidity aortic atherosclerosis. Our data suggest that nephropathy is and mortality in type 1 diabetic patients with nephropathy.29 associated with a differential impact on aortic and coronary Cardiovascular autonomic neuropathy results from damage to atherosclerosis and that coronary atherosclerosis is preferen- the autonomic nerve fibers that innervate the heart and blood tially accelerated in subjects with nephropathy. A similar vessels, resulting in abnormalities in heart rate control and finding was reported from a prospective study, showing that vascular dynamics.30 The exact mechanism by which auto- peripheral artery disease and CHD in type 1 diabetes had nomic neuropathy increases cardiovascular risk in type 1

TABLE 5. Factors Influencing CPB in Multiple Linear Regression Models

Dependent Variable: Mean RCA Vessel Wall Thickness Dependent Variable: Maximum RCA Vessel Wall Thickness

ϪDN and ϩDN ϪDN ϩDN ϪDN Coefficient ϩDN Coefficient Coefficient Coefficient Coefficient ϪDN and ϩDN Variable (SE)/P (Nϭ33) (SE)/P (Nϭ21) (SE)/P (SE)/P (SE)/P Coefficient (SE)/P Duration of diabetes (10-y increase) 0.10 (0.05)/0.07 0.11 (0.07)/0.2 0.10 (0.04)/0.02* 0.20 (0.08)/0.02* 0.09 (0.11)/0.5 0.14 (0.07)/0.04* Systolic blood pressure (10-mm Hg increase) 0.1 (0.02)/0.001* Ϫ0.03 (0.04)/0.5 0.03 (0.02)/0.2† 0.15 (0.04)/0.001* Ϫ0.05 (0.06)/0.5 0.05 (0.04)/0.2† Diastolic blood pressure (10-mm Hg increase) 0.11 (0.06)/0.07 Ϫ0.05 (0.08)/0.6 0.03 (0.05)/0.5 0.22 (0.09)/0.02* Ϫ0.08 (0.13)/0.6 0.07 (0.08)/0.4 Cholesterol (1-mmol/L increase) 0.09 (0.05)/0.07 0.11 (0.08)/0.2 0.1 (0.04)/0.03* 0.08 (0.08)/0.3 0.24 (0.12)/0.07 0.15 (0.07)/0.04* LDL (1-mmol/L increase) Ϫ0.03 (0.08)/0.7 0.16 (0.09)/0.1 0.07 (0.06)/0.26 Ϫ0.06 (0.12)/0.6 0.31 (0.14)/0.04* 0.13 (0.09)/0.2† LDL indicates low-density lipoprotein; DN, diabetic nephropathy. *PϽ0.05, significant regression; †PϽ0.05, significant difference in regression slope for ϩDN vs ϪDN.

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Kim et al Subclinical Atherosclerosis in Type 1 Diabetes 233

Figure 3. Regression of diabetes duration and RCA mean ves- Figure 4. Regression of systolic blood pressure and RCA mean sel wall thickness. Separate regression lines plotted for subjects vessel wall thickness. Separate regression lines plotted for sub- with normoalbuminuria (solid line) and nephropathy (dotted line) jects with normoalbuminuria (solid line) and nephropathy (dotted show similar regression slopes but with a significant difference line) show significantly different regression slopes. in intercepts corresponding to 0.34-mm vessel wall thickness. of type 1 diabetes patients with nephropathy (versus 10% diabetes is unknown. Our data did not show significant associ- with normoalbuminuria) develop cardiovascular events over ations between CPB and cardiovascular autonomic neuropathy. a decade of follow-up.4 The development of atherothrombosis in the presence of At baseline in 1993, the cohort of patients with diabetic nephropathy has been attributed to several mechanisms, nephropathy had elevated lipid levels and higher blood including hypertension, dyslipidemia, and abnormalities of pressure compared with patients with normoalbuminuria.4 fibrinolysis and coagulation.31 Diabetic nephropathy is asso- The high prevalence (76%) of coronary plaques compared ciated with an atherogenic lipoprotein profile, including elevated low-density lipoprotein, very low-density lipopro- with the relatively low prevalence (10%) of significant coro- tein, and lipoprotein(a) and decreased high-density lipopro- nary stenosis in our patients with nephropathy confirms the tein.32,33 Furthermore, a hypercoagulable state characterized concept of outward (positive or Glagov type) remodeling by increased plasminogen activator inhibitor-I, factor VII, with preservation of lumen size despite the progression of and plasma fibrinogen levels has been described.34 Reduced atherosclerotic plaques. Outward arterial remodeling is more renal function can lead to the accumulation of advanced often associated with morphological predictors of plaque 39 glycosylation end products in the circulation and tissue,35 instability. Acute coronary syndromes frequently result which may accelerate atherosclerosis. from rupture of an atherosclerotic plaque in an area of only Noninvasive surrogate measures for cardiovascular disease mild to moderate luminal narrowing.40,41 These data suggest events have the potential to increase the efficiency of clinical that coronary atherothrombosis may play an important role in trials and to improve risk stratification. A leading candidate the overall high cardiovascular mortality in patients with for such a surrogate is the carotid artery intima-media nephropathy. Given the low number of ischemic exercise thickness. In the Epidemiology of Diabetes Interventions and tests (3%) and coronary stenosis (10%) in subjects with Complications study, the carotid intima-media thickness of nephropathy, screening this population with either cardiac type 1 diabetics was related to conventional risk factors, exercise testing or noninvasive coronary angiography would including hypertension, dyslipidemia, and smoking, as well not be appropriate for risk stratification. Furthermore, a large as urinary albumin excretion.36 Furthermore, intensive ther- proportion (47%) of subjects with nephropathy were unable apy presumably mediated through improved glycemic control to achieve 80% of their maximum predicted heart rate with resulted in decreased progression of intima-media thickness exercise, resulting in inconclusive test results. Current guide- after 6 years.37 Whether the decrease in the progression of lines recommend stress testing in asymptomatic diabetic intima-media thickness translates into a reduction in cardio- patients only if Ն2 risk factors are present or if the resting vascular events remains unknown. ECG suggests ischemia or infarction.1 The excess cardiovascular mortality in type 1 diabetes is The assessment of coronary artery stenosis was based on present predominantly in patients with nephropathy.4,38 De- visual inspection of the coronary angiograms using a previ- spite favorable alteration in cardiovascular risk factors, 40% ously validated approach with a sensitivity for detection of

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 234 Circulation January 16, 2007

significant coronary stenosis (compared with quantitative 6. Fayad ZA, Nahar T, Fallon JT, Goldman M, Aguinaldo JG, Badimon JJ, x-ray angiography) of 88% to 93%.5 Therefore, it is unlikely Shinnar M, Chesebro JH, Fuster V. In vivo magnetic resonance eval- uation of atherosclerotic plaques in the human thoracic aorta: a com- that the low prevalence of coronary stenoses in our study parison with transesophageal echocardiography. Circulation. 2000;101: group was a false-negative finding. Furthermore, the repro- 2503–2509. ducibility of CMRI for evaluating coronary and aortic ath- 7. Jaffer FA, O’Donnell CJ, Larson MG, Chan SK, Kissinger KV, Kupka erosclerosis has been validated.19,42 Because of time con- MJ, Salton C, Botnar RM, Levy D, Manning WJ. Age and sex distribution of subclinical aortic atherosclerosis: a magnetic resonance imaging exam- straints, only 54 subjects were able to have CPB assessments. ination of the Framingham Heart Study. Arterioscler Thromb Vasc Biol. Because CPB was not positively associated with blood 2002;22:849–854. pressures in subjects with nephropathy, it is unlikely that the 8. Cai J, Hatsukami TS, Ferguson MS, Kerwin WS, Saam T, Chu B, Takaya N, Polissar NL, Yuan C. In vivo quantitative measurement of intact lower blood pressures in those subjects with nephropathy fibrous cap and lipid-rich necrotic core size in atherosclerotic carotid who had CPB measured compared with those who did not plaque: comparison of high-resolution, contrast-enhanced magnetic res- have it measured would affect the observed difference in CPB onance imaging and histology. Circulation. 2005;112:3437–3444. between subjects with and without nephropathy. Other limi- 9. Fayad ZA, Fuster V, Fallon JT, Jayasundera T, Worthley SG, Helft G, Aguinaldo JG, Badimon JJ, Sharma SK. Noninvasive in vivo human tations include spatial resolution, analysis of CPB restricted coronary artery lumen and wall imaging using black-blood magnetic to the RCA, and lack of plaque characterization. resonance imaging. Circulation. 2000;102:506–510. 10. Botnar RM, Stuber M, Kissinger KV, Kim WY, Spuentrup E, Manning WJ. Noninvasive coronary vessel wall and plaque imaging with magnetic Conclusions resonance imaging. Circulation. 2000;102:2582–2587. A low prevalence of coronary stenoses was found in asymp- 11. Kim WY, Stuber M, Bornert P, Kissinger KV, Manning WJ, Botnar RM. tomatic middle-aged patients with long-term type 1 diabetes Three-dimensional black-blood cardiac magnetic resonance coronary on contemporary cardioprotective and renoprotective medi- vessel wall imaging detects positive arterial remodeling in patients with nonsignificant coronary artery disease. Circulation. 2002;106:296–299. cation. However, CMRI identified a greater CPB in asymp- 12. Parving H-H, Andersen AR, Smidt UM, Svendsen PA. Early aggressive tomatic type 1 diabetic patients with nephropathy compared antihypertensive treatment reduces rate of decline in kidney function in with those with normoalbuminuria. Given the well-known diabetic nephropathy. Lancet. 1983;1:1175–1179. high risk of cardiovascular disease in patients with type 1 13. Feldt-Rasmussen B, Dinesen B, Deckert M. Enzyme immunoassay: an improved determination of urinary albumin in diabetics with incipient diabetes and nephropathy, the aortic plaque prevalence and nephropathy. Scand J Clin Lab Invest. 1985;45:539–544. plaque burden were unexpectedly low in this population. The 14. Brochner-Mortensen J, Rodbro P. Selection of routine method for deter- prognostic value of CMRI awaits follow-up studies in sub- mination of glomerular filtration rate in adult patients. Scand J Clin Lab Invest. 1976;36:35–43. groups of medium- to high-risk populations such as patients 15. Lassen NA, Tvedegaard E, Jeppesen FI, Nielsen PE, Bell G, Gundersen with type 1 diabetes. J. Distal blood pressure measurement in occlusive arterial disease: strain gauge compared to xenon-133. Angiology. 1972;23:211–217. 16. Gibbons RJ, Balady GJ, Beasley JW, Bricker JT, Duvernoy WF, Disclosures Froelicher VF, Mark DB, Marwick TH, McCallister BD, Thompson PD, Dr Parving is a current consultant to or has had a past relationship Winters WL Jr, Yanowitz FG, Ritchie JL, Cheitlin MD, Eagle KA, with the following companies: Merck & Co, Inc; Sanofi-Aventis; Gardner TJ, Garson A Jr, Lewis RP, O’Rourke RA, Ryan TJ. ACC/AHA Novartis; Novo Nordisk A/S; and Pfizer, Inc. Dr Stuber was guidelines for exercise testing: executive summary: a report of the supported by a Biomedical Engineering Grant from the Whitaker American College of Cardiology/American Heart Association Task Force Foundation (RG-02-0745) and a grant from the Donald W. Reynolds on Practice Guidelines (Committee on Exercise Testing). Circulation. Foundation. Dr Stuber is compensated as a consultant by Philips 1997;96:345–354. Medical Systems NL, the manufacturer of equipment described in 17. Stuber M, Botnar RM, Danias PG, Sodickson DK, Kissinger KV, Van this presentation. The terms of this arrangement have been approved CM, De BJ, Manning WJ. Double-oblique free-breathing high resolution by the Johns Hopkins University in accordance with its conflict of three-dimensional coronary magnetic resonance angiography. J Am Coll interest policies. The other authors report no conflicts. Cardiol. 1999;34:524–531. 18. Spuentrup E, Katoh M, Buecker A, Manning WJ, Schaeffter T, Nguyen TH, Kuhl HP, Stuber M, Botnar RM, Gunther RW. Free-breathing 3D References steady-state free precession coronary MR angiography with radial 1. Consensus development conference on the diagnosis of coronary heart k-space sampling: comparison with cartesian k-space sampling and car- disease in people with diabetes: 10–11 February 1998, Miami, Florida: tesian gradient-echo coronary MR angiography: pilot study. Radiology. American Diabetes Association. Diabetes Care. 1998;21:1551–1559. 2004;231:581–586. 2. Executive Summary of the Third Report of the National Cholesterol 19. Desai MY, Lai S, Barmet C, Weiss RG, Stuber M. Reproducibility of 3D Education Program (NCEP) Expert Panel on Detection, Evaluation, and free-breathing magnetic resonance coronary vessel wall imaging. Eur Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel Heart J. 2005;26:2320–2324. III). JAMA. 2001;285:2486–2497. 20. Etienne A, Botnar RM, Van Muiswinkel AM, Boesiger P, Manning WJ, 3. De BG, Ambrosioni E, Borch-Johnsen K, Brotons C, Cifkova R, Stuber M. “Soap-bubble” visualization and quantitative analysis of 3D Dallongeville J, Ebrahim S, Faergeman O, Graham I, Mancia G, Manger coronary magnetic resonance angiograms. Magn Reson Med. 2002;48: Cats V, Orth-Gomer K, Perk J, Pyorala K, Rodicio JL, Sans S, Sansoy V, 658–666. Sechtem U, Silber S, Thomsen T, Wood D. European guidelines on 21. Manske CL, Wilson RF, Wang Y, Thomas W. Prevalence of, and risk cardiovascular disease prevention in clinical practice: Third Joint Task factors for, angiographically determined coronary artery disease in type Force of European and Other Societies on Cardiovascular Disease Pre- I-diabetic patients with nephropathy. Arch Intern Med. 1992;152: vention in Clinical Practice. Eur Heart J. 2003;24:1601–1610. 2450–2455. 4. Astrup AS, Tarnow L, Rossing P, Pietraszek L, Hansen PR, Parving HH. 22. Larsen J, Brekke M, Sandvik L, Arnesen H, Hanssen KF, Dahl-Jorgensen Improved prognosis in type 1 diabetic patients with nephropathy: a K. Silent coronary atheromatosis in type 1 diabetic patients and its prospective follow-up study. Kidney Int. 2005;68:1250–1257. relation to long-term glycemic control. Diabetes. 2002;51:2637–2641. 5. Kim WY, Danias PG, Stuber M, Flamm SD, Plein S, Nagel E, Langerak 23. Rohani M, Jogestrand T, Ekberg M, van der LJ, Kallner G, Jussila R, SE, Weber OM, Pedersen EM, Schmidt M, Botnar RM, Manning WJ. Agewall S. Interrelation between the extent of atherosclerosis in the Coronary magnetic resonance angiography for the detection of coronary thoracic aorta, carotid intima-media thickness and the extent of coronary stenoses. N Engl J Med. 2001;345:1863–1869. artery disease. Atherosclerosis. 2005;179:311–316.

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Kim et al Subclinical Atherosclerosis in Type 1 Diabetes 235

24. Taniguchi H, Momiyama Y, Fayad ZA, Ohmori R, Ashida K, Kihara T, lipoprotein (VLDL), intermediate density lipoprotein (IDL), and low Hara A, Arakawa K, Kameyama A, Noya K, Nagata M, Nakamura H, density lipoprotein (LDL) composition. Atherosclerosis. 1991;89:49–57. Ohsuzu F. In vivo magnetic resonance evaluation of associations between 34. Gruden G, Cavallo-Perin P, Bazzan M, Stella S, Vuolo A, Pagano G. aortic atherosclerosis and both risk factors and coronary artery disease in PAI-1 and factor VII activity are higher in IDDM patients with microal- patients referred for coronary angiography. Am Heart J. 2004;148: buminuria. Diabetes. 1994;43:426–429. 137–143. 35. Makita Z, Bucala R, Rayfield EJ, Friedman EA, Kaufman AM, Korbet 25. Forrest KY, Becker DJ, Kuller LH, Wolfson SK, Orchard TJ. Are pre- SM, Barth RH, Winston JA, Fuh H, Manogue KR. Reactive glycosylation dictors of coronary heart disease and lower-extremity arterial disease in endproducts in diabetic uraemia and treatment of renal failure. Lancet. type 1 diabetes the same? A prospective study. Atherosclerosis. 2000; 1994;343:1519–1522. 148:159–169. 36. Effect of intensive diabetes treatment on carotid artery wall thickness in 26. Wilson PW, D’Agostino RB, Levy D, Belanger AM, Silbershatz H, the epidemiology of diabetes interventions and complications: Epidemi- Kannel WB. Prediction of coronary heart disease using risk factor cate- ology of Diabetes Interventions and Complications (EDIC) Research gories. Circulation. 1998;97:1837–1847. Group. Diabetes. 1999;48:383–390. 27. Stevens RJ, Kothari V, Adler AI, Stratton IM. The UKPDS risk engine: 37. Nathan DM, Lachin J, Cleary P, Orchard T, Brillon DJ, Backlund JY, a model for the risk of coronary heart disease in type II diabetes (UKPDS O’Leary DH, Genuth S. Intensive diabetes therapy and carotid 56). Clin Sci. 2001;101:671–679. intima-media thickness in type 1 diabetes mellitus. N Engl J Med. 2003; 28. Zgibor JC, Piatt GA, Ruppert K, Orchard TJ, Roberts MS. Deficiencies of 348:2294–2303. cardiovascular risk prediction models for type 1 diabetes. Diabetes Care. 38. Borch-Johnsen K, Kreiner S. Proteinuria: value as predictor of cardio- 2006;29:1860–1865. vascular mortality in insulin dependent diabetes mellitus. BMJ. 1987;294: 29. Astrup AS, Hansen BV, Hilsted J, Parving HH, Rossing P, Tarnow L. Cardiac autonomic neuropathy predicts cardiovascular morbidity and 1651–1654. mortality in type 1 diabetic patients with diabetic nephropathy. Diabetes 39. Burke AP, Kolodgie FD, Farb A, Weber D, Virmani R. Morphological Care. 2006;29:334–339. predictors of arterial remodeling in coronary atherosclerosis. Circulation. 30. Vinik AI, Maser RE, Mitchell BD, Freeman R. Diabetic autonomic 2002;105:297–303. neuropathy. Diabetes Care. 2003;26:1553–1579. 40. Little WC, Constantinescu M, Applegate RJ, Kutcher MA, Burrows MT, 31. Deckert T, Kofoed-Enevoldsen A, Norgaard K, Borch-Johnsen K, Feldt- Kahl FR, Santamore WP. Can coronary angiography predict the site of a Rasmussen B, Jensen T. Microalbuminuria: implications for micro- and subsequent myocardial infarction in patients with mild-to-moderate macrovascular disease. Diabetes Care. 1992;15:1181–1191. coronary artery disease? Circulation. 1988;78:1157–1166. 32. Jensen T, Stender S, Deckert T. Abnormalities in plasmas concentrations 41. Falk E, Shah PK, Fuster V. Coronary plaque disruption. Circulation. of lipoproteins and fibrinogen in type 1 (insulin-dependent) diabetic 1995;92:657–671. patients with increased urinary albumin excretion. Diabetologia. 1988; 42. Chan SK, Jaffer FA, Botnar RM, Kissinger KV, Goepfert L, Chuang ML, 31:142–145. O’Donnell CJ, Levy D, Manning WJ. Scan reproducibility of magnetic 33. Winocour PH, Durrington PN, Bhatnagar D, Ishola M, Mackness M, resonance imaging assessment of aortic atherosclerosis burden. J Car- Arrol S. Influence of early diabetic nephropathy on very low density diovasc Magn Reson. 2001;3:331–338.

CLINICAL PERSPECTIVE Patients with type 1 diabetes and nephropathy maintain an excess cardiovascular mortality compared with diabetic patients with normoalbuminuria. Risk assessment in asymptomatic patients with type 1 diabetes is less stringent than for type 2 diabetes, and the effect of diabetic nephropathy on atherosclerosis in type 1 diabetes is not well described. In this cross-sectional study of 136 asymptomatic type 1 diabetic patients with long-standing diabetes, cardiovascular magnetic resonance imaging revealed greater coronary plaque burden and higher prevalence of coronary artery stenoses in patients with diabetic nephropathy compared with those with normoalbuminuria. These data suggest that coronary atherothrombosis may play an important role in the overall high cardiovascular mortality in patients with type 1 diabetes and nephropathy.

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Rapid Detection of Myocardial Infarction by Subsecond, Free-Breathing Delayed Contrast-Enhancement Cardiovascular Magnetic Resonance

Burkhard Sievers, MD; Michael D. Elliott, MD; Lynne M. Hurwitz, MD; Timothy S.E. Albert, MD; Igor Klem, MD; Wolfgang G. Rehwald, PhD; Michele A. Parker, MS; Robert M. Judd, PhD; Raymond J. Kim, MD

Background—An ultrafast, delayed contrast-enhancement cardiovascular magnetic resonance technique that can acquire subsecond, “snapshot” images during free breathing (subsecond) is becoming widely available. This technique provides myocardial infarction (MI) imaging with complete left ventricular coverage in Ͻ30 seconds. However, the accuracy of this technique is unknown. Methods and Results—We prospectively compared subsecond imaging with routine breath-hold delayed contrast- enhancement cardiovascular magnetic resonance (standard) in consecutive patients. Two cohorts with unambiguous standards of truth were prespecified: (1) patients with documented prior MI (nϭ135) and (2) patients without MI and with low likelihood of coronary disease (lowest Framingham risk category; nϭ103). Scans were scored masked to identity and clinical information. Sensitivity, specificity, and accuracy of subsecond imaging for MI diagnosis were 87%, 96%, and 91%, respectively. Compared with the standard technique (98%, 100%, 99%), the subsecond technique had modestly reduced sensitivity (Pϭ0.0001), but specificity was excellent. Missed infarcts were generally small or subendocardial (87%). Overall, regional transmural extent of infarction scores were highly concordant (2083/2294; 91%); however, 51 of 337 regions (15%) considered predominantly infarcted (Ͼ50% transmural extent of infarction) by the standard technique were considered viable (Յ25% transmural extent of infarction) by the subsecond technique. Quantitative analysis demonstrated moderately reduced contrast-to-noise ratios for subsecond imaging between infarct and remote myocardium (12.0Ϯ7.2 versus 20.1Ϯ6.6; PϽ0.0001) and infarct and left ventricular cavity (Ϫ2.5Ϯ2.7 versus 3.6Ϯ3.7; PϽ0.0001). Conclusions—MI can be rapidly detected by subsecond delayed contrast-enhancement cardiovascular magnetic resonance during free breathing with high accuracy. This technique could be considered the preferred approach in patients who are more acutely ill or unable to hold their breath. However, compared with standard imaging, sensitivity is mildly reduced, and the transmural extent of infarction may be underestimated. (Circulation. 2007;115:236-244.) Key Words: imaging Ⅲ magnetic resonance imaging Ⅲ myocardial infarction

yocardial infarction (MI) occurs in almost a million Clinical Perspective p 244 Mpeople in the United States each year, and coronary heart disease is the leading cause of hospital admissions.1 According Delayed enhancement cardiovascular magnetic resonance to the joint American College of Cardiology and European (DE-CMR) may overcome some of these limitations. Studies Society of Cardiology consensus document concerning the in animal models suggest a nearly perfect match of infarcted redefinition of MI, the diagnosis of MI is substantially based on regions by DE-CMR to histology.4 In patients, DE-CMR can cardiac biomarkers and ECG changes.2 However, biomarkers provide accurate and reproducible diagnosis of both acute and are only elevated for 4 to 10 days after an acute event2; thus, chronic MI.5,6 Furthermore, even small subendocardial in- biomarkers are not useful for the diagnosis of subacute or chronic farcts can be detected reliably in the absence of Q waves.6–8 MI. The ECG also has limitations: Q waves that form the funda- CMR is a rapidly progressing field, and new DE-CMR mental basis of the diagnosis of chronic MI may be absent or, if techniques are becoming widely available. One of these new initially present, may disappear at a later time point.2,3 techniques, which is highly promising, uses a single-shot

Received April 24, 2006; accepted October 19, 2006. From Duke Cardiovascular Magnetic Resonance Center (B.S., M.D.E., L.M.H., T.S.E.A., I.K., W.G.R., M.A.P., R.M.J., R.J.K.), Department of Medicine (B.S., M.D.E., T.S.E.A., I.K., M.A.P., R.M.J., R.J.K.), and Department of Radiology (L.M.H., R.M.J., R.J.K.), Duke University Medical Center, Durham, NC; and Siemens Medical Solutions, Chicago, Ill (W.G.R.). Correspondence to Raymond Kim, MD, Duke Cardiovascular Magnetic Resonance Center, Duke DUMC-3934, Durham, NC 27710. E-mail [email protected] © 2007 American Heart Association, Inc. Circulation is available at http://www.circulationaha.org DOI: 10.1161/CIRCULATIONAHA.106.635409

Downloaded from circ.ahajournals.org 236at Mohammed Mahboob on June 10, 2007 Sievers et al MI Detection by Subsecond, Free-Breathing CMR 237

acquisition that can acquire subsecond (“snapshot”) images TABLE 1. Pulse Sequence Parameters during free breathing rather than acquiring data for 1 image Standard Subsecond over several heartbeats during a prolonged breath-hold as Temporal resolution, ms 207 207† required by the standard technique. This rapid technique can Field of view, read, mm 320–400 320–400 potentially speed up imaging 10- to 20-fold and enable image Field of view, phase, % 81.3 83.3 acquisition of the entire heart during free breathing in Ͻ30 Image matrix 166ϫ256 104ϫ192 seconds. This advantage has the potential to fundamentally Voxel size, mm3 1.8ϫ1.4ϫ6.0ϭ15.1 2.9ϫ1.9ϫ6.0ϭ33.1 change current CMR practice. Either clinical throughput Interslice gap, mm 4 4 could be increased multifold, with several scans completed Repetition time, ms 9.8 2.5 per hour, or the improved efficiency could be utilized for a Echo time, ms 3.8 1.1 comprehensive examination: Cardiac morphology, function, viability, stress-rest perfusion, and coronary imaging could be Flip angle, degree 20 50 performed in under an hour. Therefore, not surprisingly, Trigger pulse* 2 3 many centers are considering switching or have already Bandwidth, Hz/pixel 130 1130 switched to this new subsecond method as the preferred Breath-hold Yes No delayed-enhancement technique. Standard indicates segmented inversion-recovery, gradient-echo sequence; However, there are limited data regarding the utility of subsecond, single-shot inversion-recovery, steady state free precession. *Data acquisition every second (standard) or every third (subsecond) RR subsecond imaging in clinical practice.9 The purpose of the interval. present study was to systematically evaluate the accuracy of †With parallel imaging acceleration (factorϭ2). subsecond imaging compared with standard segmented inversion-recovery DE-CMR for the diagnosis of MI and the administration (gadoversetamide, 0.15 mmol/kg). This time- determination of infarct size. window/dose combination for infarct imaging was chosen on the basis of previous results that have demonstrated accurate and Methods reproducible infarct size measurements.5,13 Patients Subsecond DE-CMR Consecutive patients scheduled to undergo standard DE-CMR at the Duke Cardiovascular Magnetic Resonance Center starting February Rapid imaging was performed with a single-shot, inversion- 2003 and ending October 2004 were prospectively evaluated. Only 2 recovery, steady state free-precession sequence. Similar to standard cohorts with prespecified unambiguous standards of truth for either imaging, an inversion recovery prepulse was used to accentuate the the presence or absence of MI were considered for enrollment. conspicuity of infarcted regions, and the inversion time was set to Patients with nonischemic cardiac disorders such as idiopathic null normal myocardium (typically 20 to 50 ms longer than for the dilated cardiomyopathy, hypertrophic cardiomyopathy, and myocar- standard method). Dissimilar to standard imaging, parallel image ditis were excluded. acquisition14 was used to speed up imaging; however, this was at the The infarct group (nϭ135) consisted of patients with documented cost of the signal-to-noise ratio. Importantly, the entire data set for 1 MI in their medical history according to the recent joint European image was acquired in a single contiguous stream (single-shot) rather Society of Cardiology/American College of Cardiology consensus than in a segmented manner over several heartbeats. The patient was document for the redefinition of MI.2 Thus, patients had typical rise instructed to take slow and shallow breaths through the 25 to 30 and fall of cardiac biomarkers with either appropriate ECG changes, seconds of image acquisition. A potential limitation is that transla- ischemic symptoms, or both at the time of the acute event. The tional motion of the heart during free breathing may result in no-infarct group (nϭ103) consisted of subjects without known asymmetrical coverage of the LV during image acquisition. To test coronary heart disease and with a very low probability of developing the significance of this limitation, we performed subsecond imaging coronary disease over the next 10 years10 (lowest Framingham risk twice, immediately before and after standard imaging, in a subgroup score category: 1% for women, 2% for men). The study was of MI patients. approved by the Duke University institutional review board; all patients gave informed written consent. Parameter Matching Even with parallel image acquisition, it is not possible for subsecond Data Acquisition imaging to match both the high temporal and high spatial resolution All available cardiac enzyme data were collected. Patients with peak of standard imaging. We chose to match the temporal resolution at enzyme levels occurring Ͻ14 days before CMR were considered the cost of a lower spatial resolution (Table 1); therefore, both acute MI patients. All ECGs performed within a short time period of standard and subsecond data acquisition occurs during a 207-ms time CMR (2 days for acute MI patients and 10 weeks for chronic MI or window during mid-diastole, to take advantage of the fact that the no-infarct patients) were collected. Patients were imaged in a 1.5-T heart is relatively motionless during this time. We did not match the scanner (Siemens, Sonata) with the use of an 8-element phased-array spatial resolution or pursue an intermediate compromise because any system. Multiple short-axis views that covered the left ventricle (LV) increase in data readout period would result in increased image from base to apex (every 1 cm) were obtained by cine imaging and blurring and partial volume effects due to substantial cardiac motion both delayed enhancement techniques. All images were ECG gated. outside the diastolic standstill period, thus negating any potential Standard DE-CMR improvement in spatial resolution. Segmented inversion-recovery gradient-echo imaging was per- formed in the standard fashion.11,12 Briefly, data for 1 image were Timing of DE-CMR Techniques acquired over multiple heartbeats; thus, images were obtained during Subsecond imaging was performed immediately before standard repetitive 6- to 10-second breath-holds. Inversion times were typi- imaging. Because subsecond imaging took Ͻ30 seconds to complete, cally 280 to 380 ms to null normal myocardium.12 Specific sequence there was little time delay between the 2 techniques, and the potential parameters are displayed in Table 1. Importantly, images were that time after contrast injection could account for possible differ- acquired between 10 and 30 minutes after intravenous contrast ences in the results between the 2 techniques was minimized.

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 238 Circulation January 16, 2007

Analysis TABLE 2. Clinical Characteristics Clinical Data Infarct Group No-Infarct Group ECGs were analyzed for Q waves according to Minnesota codes Parameter (nϭ135) (nϭ103) 1-1-1 to 1-2-7.15 To determine Framingham risk score category, Age, y 60.2Ϯ13.5 43.1Ϯ16.5 standard score sheets based on total cholesterol level were used (see Figures 3 and 4 of Wilson et al10). Gender Male 93 (69) 41 (40) CMR Scans were placed in random order and scored visually by the CHD risk factors consensus of 2 observers who were masked to patient identity and Diabetes 41 (30) 6 (6) associated clinical information. Standard and subsecond images were Hypertension 85 (63) 22 (21) read independently during separate sessions. Hyperenhancement was scored on a 17-segment model16 with a 5-point scale (0, no Hyperlipidemia 59 (44) 16 (16) hyperenhancement; 1, 1% to 25%; 2, 26% to 50%; 3, 51% to 75%; Smoking 51 (38) 19 (18) 4, 76% to 100%).8 Infarct size as a percentage of LV myocardium Heart rate, bpm 71.3Ϯ14.0 71.8Ϯ13.3 was calculated by summing the regional scores (each weighted by the midpoint of the range of hyperenhancement) and dividing by the Atrial fibrillation 5 (4) 0 (0) total number of regions.8 Hypoenhanced regions within hyperen- LV ejection fraction, %* 41.9Ϯ12.7 64.7Ϯ5.6§ hanced myocardium were interpreted as regions of microvascular ࿣ obstruction (“no-reflow”) and included in infarct size calculations.12 12-lead ECG (Q wave†) 59/135 (46) 0/64 (0) To assess interobserver variability, images from 68 patients were Infarct characteristic randomly selected and rescored independently by 2 readers. Cine Cardiac enzymes on admission‡ images were evaluated separately from the DE-CMR images and scored for LV ejection fraction by visual inspection.17 CK, U/L 891 (394, 2331) ⅐⅐⅐ Quantitative analysis was performed to evaluate potential mech- CK-MB, ng/mL 99 (49, 204) ⅐⅐⅐ anisms for differences in the accuracy of subsecond and standard Troponin T, ng/mL 3 (1.2, 6.1) ⅐⅐⅐ imaging. Twenty-one patients from the infarct group who had hyperenhancement on both techniques were randomly selected. A Acute (Ͻ14 d) 61 (45) ⅐⅐⅐ single short-axis image with the largest hyperenhanced region was Chronic (Ͼ14 d) 74 (55) ⅐⅐⅐ selected for each patient. Regions of interest were placed within Ϯ infarct, remote myocardium, and LV cavity. Signal intensity (SI) Values are expressed as mean SD, n (%), or median (25th, 75th percen- ratios between the various compartments were calculated as follows: tile). CHD indicates coronary heart disease; CK, creatine kinase. ϭ *Determined by cine CMR. ratioinfarct/remote mean SI of infarct/mean SI of remote; ratioinfarct/ †According to Minnesota codes 1-1-1 to 1-2-7. cavityϭmean SI infarct/mean SI cavity. Additionally, a region of interest was placed outside the body to ‡Cardiac enzyme values were available in 95 patients. measure the SD of background noise and to generate signal-to-noise §All with normal LV ejection fraction and no regional wall motion abnormal- ratios (SNR) and contrast-to-noise ratios (CNR) as follows: ities. ϭ ϫ ࿣ SNRinfarct mean SI of infarct/(1.43 SD of background); CNRinfarct- ECGs were obtained in 64 of the no-infarct group subjects within 10 weeks to-remoteϭ(mean SI of infarctϪmean SI of remote)/(1.43ϫSD of of CMR. background). The correction factor of 1.43 accounts for the under- estimation of noise that occurs when noise is measured from patients had acute infarcts at the time of imaging. Twelve- magnitude images18 after adjustment for an 8-element coil array.19 This method of measuring noise has limitations in the setting of lead ECGs were acquired in all patients from the infarct group parallel imaging. On accelerated images such as subsecond images, and in 64 from the no-infarct group. The interval between the noise background varies with spatial position caused by the ECG and CMR was 1Ϯ1, 5Ϯ14, and 6Ϯ11 days in acute MI, spatial dependence of noise amplification characterized by coil chronic MI, and no-infarct patients, respectively, and no 20 geometry. Thus, SNR and CNR values from subsecond images patient had any cardiac events during this interval. Q waves should be considered only an approximation. were present in 46% of the infarct group and 0% of the Statistical Analysis no-infarct group. Cardiac enzyme values were available in Continuous data are expressed as meanϮSD, except where noted. 70% (95/135) of the infarct group. The remainder were Comparisons between patient groups were made with the use of the referred from outside hospitals, and although the medical ␹2 2-sample t test for continuous data and test for discrete data. records of these patients clearly documented prior MI accord- Comparisons between imaging sequences were made with the paired t test for continuous data and McNemar test for discrete data. ing to European Society of Cardiology/American College of Bland-Altman21 and linear regression analyses were performed to Cardiology consensus criteria,2 including reference to the fact assess the relationship between the 2 DE-CMR techniques for infarct that enzyme levels were abnormal, the specific values were size. Interobserver agreement for the presence of hyperenhancement not available. was calculated with the ␬ statistic. All statistical tests were 2-tailed; PϽ0.05 was considered significant. Statistical analysis was per- formed with S-PLUS 6 (Insightful Corporation, Seattle, Wash). Infarct Detection The authors had full access to and take full responsibility for the Both DE-CMR techniques were performed in all 238 patients. integrity of the data. All authors have read and agree to the The overall diagnostic performance of the 2 techniques is manuscript as written. displayed in Figure 1A. Standard imaging performed well in this highly selected cohort with a sensitivity of 98% and Results specificity of 100% for the diagnosis of MI. Patient Characteristics The sensitivity of the subsecond technique was modestly Baseline characteristics for both patient groups are shown in lower at 87% (PϽ0.0001). Of the 17 MI patients who were Table 2. Slightly less than half (45%) of the infarct group missed, 2 were also missed by the standard technique. Of the

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Sievers et al MI Detection by Subsecond, Free-Breathing CMR 239

Figure 1. Sensitivity, specificity, and diagnostic accuracy of standard and subsecond DE-CMR for the detection of MI (A). The sensitiv- ity of subsecond imaging was reduced primarily in small or subendocardial infarcts (B).

remaining 15 patients, the majority (13/15; 87%) had either Global Infarct Size small (Յ10% of LV myocardium) or subendocardial (Յ50% Infarcts were generally smaller on the subsecond technique transmural extent) infarcts (Figure 1B). Two patients were compared with the standard technique (12.1Ϯ9.3% versus exceptions in that the missed infarcts were both large and 17.0Ϯ11.2% of LV; PϽ0.0001). By Bland-Altman analysis transmural. One had an acute infarct with an extensive (Figure 4A), the average underestimation of infarct size by no-reflow zone, and the other had a very patchy infarct. The the subsecond technique was 4.9% of LV, and the 95% limits sensitivity of the subsecond technique was lower for both of agreement were Ϫ7.1% and 17.0% of LV. The plot, acute (85% versus 98%; Pϭ0.005) and chronic infarcts (89% however, was partially skewed in that the bias was less for versus 99%; Pϭ0.008). Additionally, in patients with acute smaller infarct sizes. By linear regression analysis (Figure infarcts (nϭ61), the subsecond technique detected fewer 4B), the subsecond technique underestimated infarct size by no-reflow zones than the standard technique (14/61 [23%] 31Ϯ4% (slopeϭ0.69Ϯ0.04). When the infarcts were as- versus 25/61 [41]%; Pϭ0.0009). Representative patient im- signed to coronary territories according to the 17-segment ages, demonstrating concordance and discordance of hyper- model, infarct size was moderately underestimated by the enhancement patterns between the 2 techniques, are shown in subsecond technique in all 3 (left anterior descending coro- Figures 2 and 3. nary artery, 13.0Ϯ17.8% versus 18.4Ϯ23.9%; left circumflex The specificity of the subsecond technique was excellent at artery, 9.1Ϯ13.2% versus 12.8Ϯ14.3%; right coronary artery, 96%. In the few patients (nϭ4) in whom false-positive 13.8Ϯ16.3% versus 19.4Ϯ18.4% of LV; PϽ0.0001 for all). infarcts were detected, the presumed infarct region was near In the subcohort of 68 patients in whom interobserver vari- high-SI epicardial fat in all cases (Figure 3c). ability was assessed for the presence of hyperenhancement, there

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 240 Circulation January 16, 2007

Figure 2. Representative images in 3 patients showing concordant hyperenhancement patterns between standard and subsecond imaging.

was excellent agreement (␬ϭ0.97). The difference in mean arrhythmia. All 5 had atrial fibrillation and were from the infarct size between the 2 readers was not significant (0.4Ϯ3.3% infarct group. In these 5, infarcts were detected by both of LV; Pϭ0.3123). By Bland-Altman analysis, the 95% limits of techniques. agreement were Ϫ6.1% and 6.9% of LV. Quantitative Analysis Transmural Extent of Infarction Table 4 displays the quantitative analysis performed in the Table 3 displays the comparison of the transmural extent of randomly selected cohort of 21 patients with infarction infarction (TEI) between the 2 techniques on a regional basis. detected by both techniques. The SNR values of both infarct This analysis was based on 2294 regions (135 patientsϫ17 and remote myocardium were moderately lower for the regionsϭ2295 regions; in 1 patient the true apex, region 17, subsecond technique. Likewise, the CNR values of infarct- was not imaged). Overall, the concordance was high at 91% to-remote and infarct-to-cavity were also lower. The SI ratios (2083/2294). However, there was a systematic, albeit modest, demonstrate that the subsecond technique led to reduced underestimation of TEI by the subsecond technique (Figure differences between infarct and remote myocardium and 3B). Whereas 536 regions (23%) by the subsecond technique between infarct and LV cavity. had TEI scores less than that of the standard technique (360 of which were 1 grade less), only 150 (6.5%) had higher Discussion scores. In regard to the 337 regions that were deemed The main finding of this study is that MI can be accurately predominantly infarcted (Ͼ50% TEI) by the standard tech- detected (91% accuracy) with the use of an ultrafast DE- nique, 51 (15%) were considered predominantly viable CMR technique that can acquire snapshot images during free (Յ25% TEI) by the subsecond technique. breathing. Importantly, the technique was robust and pro- vided full coverage of the heart in Ͻ30 seconds, and Influence of Time After Contrast Injection, Heart specificity was excellent at 96%. However, there were some Rate, and Arrhythmia limitations compared with the standard approach. Sensitivity In the subgroup of 11 patients in whom the subsecond was 11% lower (87% versus 98%), and there was a tendency technique was performed immediately before and after the to underestimate both global infarct size and the transmural standard technique, infarct size was virtually identical extent of infarction. (22.1Ϯ9.8% versus 22.2Ϯ10.7% of LV). Bland-Altman anal- Despite these limitations, the development of subsecond ysis demonstrated narrow limits of agreement (Ϫ2.5% and imaging has important clinical implications. For instance, 2.3% of LV). The heart rate in patients in whom infarction although CMR is a highly attractive modality for the assess- was detected by the subsecond technique was not signifi- ment of MI and viability,22 some practical drawbacks poten- cantly different from those in whom infarction was missed tially limit the impact of this technology for general clinical (71.4Ϯ13.8 versus 70.7Ϯ15.4; Pϭ0.84). Only 5 patients had practice. In comparison to other imaging modalities, standard

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Sievers et al MI Detection by Subsecond, Free-Breathing CMR 241

Figure 3. Representative images in patients showing discordant hyperenhancement patterns between standard and subsecond imag- ing. A shows typical examples of infarcts that were missed by subsecond imaging (false-negative) either because of small size (patient 4), only subendocardial involvement that was difficult to distinguish from the LV cavity (patient 5), or the presence of a large central no-reflow zone (patient 6). B shows an infarct that was detected by both DE-CMR techniques; however, the transmural extent of infarc- tion was significantly less on subsecond compared with standard imaging (patient 7; also see cartoon). C shows 2 patients without MI who were incorrectly diagnosed with MI by subsecond imaging. In both cases, the presumed infarct region was near high-SI epicardial fat, either apically (patient 8) or near the atrioventricular groove (patient 9).

CMR is more complex: Patient and protocol setup times are we note that current methods for the detection of MI have longer, and multiple breath-holds and longer scanning times significant limitations. For example, according to the Euro- are necessary.23 These conditions limit clinical throughput pean Society of Cardiology/American College of Cardiology and the types of patients that can be scanned and increase the consensus document, the 12-lead ECG is one of the funda- complexity and length of training required for CMR opera- mental tools for the diagnosis of chronic MI.2 However, Q tors. On the other hand, single-shot CMR techniques such as waves, on which the diagnosis is based, are often absent in subsecond imaging could potentially overcome a number of small or lateral wall infarcts.7 In fact, in the present study, these drawbacks. As we have demonstrated in the present only 49% of patients with chronic MI (36/74) and 16% of study, breath-holding is not necessary, and although we patients with lateral wall MI (4/25) had Q waves. In contrast, performed subsecond imaging with ECG gating to allow a subsecond imaging had markedly higher sensitivity at 89% precise comparison of the 2 techniques, we note that gating is (66/74) and 80% (20/25), respectively. not required for single-shot techniques. Indeed, one could One potential reason for the disparity in diagnostic sensi- imagine a scenario in which a patient is given intravenous tivity between the 2 DE-CMR techniques may be the differ- gadolinium contrast while outside the scanner, then 5 to 10 minutes later is placed on the scanner table with minimal ence in spatial resolution. Typical voxel sizes for the subsec- preparation (no ECG leads, no breath-holding instructions). ond technique were 2.2-fold larger than for the standard Then a relatively fixed subsecond protocol is run without technique. This reduction in spatial resolution and subsequent adjustment of parameters for breath-hold time or ECG gating, increase in partial volume effects may in part be responsible with rapid assessment of MI performed in under a minute. In for both inconsistent identification of small infarcts and less this scenario, DE-CMR becomes a quick “push-button” accurate characterization of the transmural extent of infarc- technique with the ability to scan a wide range of patients, tion.4 Additionally, it is perhaps not surprising that no-reflow including those more acutely ill or with dyspnea. zones were particularly difficult to identify on the subsecond Although subsecond imaging had reduced sensitivity com- technique. Because partial volume effects are pronounced at pared with standard imaging, to place this result in context, the borders between different tissues, the multiple transitions

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 242 Circulation January 16, 2007

Figure 4. Bland-Altman (A) and regression analy- ses (B) comparing infarct size (percentage of LV) between standard and subsecond imaging. See text for details.

that characterize no-reflow zone–type MI—dark no-reflow infarct and LV cavity was of particular importance. Although, zone to bright infarct to dark noninfarcted myocardium— on average, SI differences between infarct and LV cavity within a limited space likely render this pathophysiology were modest for both techniques, for the standard technique more susceptible to partial volume effects (Figure 3A). Par- the infarct was brighter than cavity in the majority of patients tial volume effects also likely explain the 4 patients without (90%), whereas for the subsecond technique the relationship MI who were incorrectly scored positive by the subsecond was reversed, and cavity was brighter than infarct in 90%. In technique. In all 4 cases, the presumed infarct region was near the absence of infarction, the normal progression of SIs for epicardial fat, which is high in SI. both techniques is as follows: dark myocardium, gray transi- Another potential reason for the reduced sensitivity of the tion zone between myocardium and cavity, then bright cavity. subsecond technique may relate to differences in tissue In the presence of subendocardial infarction, the progression contrast between the 2 techniques. Quantitative analysis of SIs for the subsecond technique—dark epicardium, gray demonstrated reduced CNRs between infarct and remote subendocardial infarction, then bright cavity—was occasion- myocardium and between infarct and LV cavity. Interest- ally difficult to distinguish from normal. In hindsight, in ingly, in retrospect, we found that tissue contrast between many cases of “missed” subendocardial infarction by subsec-

TABLE 3. Regional Analysis

Standard (TEI)

Subsecond (TEI) 0% 1%–25% 26%–50% 51%–75% 76%–100% Total 0% 1344* 164* 92 12 10 1622 1%–25% 50* 90* 71* 22 7 240 26%–50% 13 29* 62* 55* 33 192 51%–75% 6 6 20* 51* 70* 153 76%–100% 1 3 6 16* 61* 87 Total 1414 292 251 156 181 2294 *TEI scores that are concordant between the 2 techniques (defined as TEI scores that are within 1 category of each other).

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Sievers et al MI Detection by Subsecond, Free-Breathing CMR 243

TABLE 4. Quantitative Analysis same) and the potentially large benefit in patients who have their scans prematurely terminated before standard imaging is Standard Subsecond P completed (because of fatigue or illness) or in patients for n* 21 21 ⅐⅐⅐ whom findings can be confirmed when motion artifacts are SNR troublesome on the standard images. Infarct 24.8Ϯ7.2 15.3Ϯ7.7 Ͻ0.0001 Remote 4.7Ϯ1.2 3.3Ϯ0.8 0.0003 Sources of Funding CNR This study was supported by grants R01-HL64726 (Dr Kim), R01- HL63268 (Dr Judd), and K02-HL04394 (Dr Judd). Infarct-to-remote 20.1Ϯ6.6 12.0Ϯ7.2 Ͻ0.0001 Infarct-to-cavity 3.6Ϯ3.7 Ϫ2.5Ϯ2.7 Ͻ0.0001 Disclosures Ratio Drs Kim and Judd are inventors of a US patent on delayed Infarct/remote 5.1Ϯ1.4 4.3Ϯ1.6 0.0037 enhancement MRI, which is owned by Northwestern University. Drs Kim and Judd have previously received grant support from Siemens Ϯ Ϯ Ͻ Infarct/cavity 1.1 0.2 0.9 0.1 0.0001 Medical Solutions and have been consultants for Mallinckrodt/Tyco, *Twenty-one patients from the infarct group who had hyperenhancement on Inc. Dr Rehwald is an employee of Siemens Medical Systems, Inc. both techniques were randomly selected for quantitative analysis. References ond imaging, the infarct was clearly distinct and brighter than 1. National Heart, Lung, and Blood Institute. Morbidity and Mortality: 2004 Chartbook on Cardiovascular, Lung, and Blood Disease. Bethesda, Md: noninfarcted myocardium. Instead, the error was in consid- National Heart, Lung, and Blood Institute; 2004. ering the subendocardial infarct as part of the LV cavity 2. Myocardial infarction redefined: a consensus document of the Joint (Figure 3A, patient 5). The use of a multicontrast version of European Society of Cardiology/American College of Cardiology Com- DE-CMR that acquires both T1 and T2 contrasts within a mittee for the Redefinition of Myocardial Infarction. Eur Heart J. 2000; 21:1502–1513. single acquisition may improve the discrimination between 3. Cox CJ. Return to normal of the electrocardiogram after myocardial MI and LV cavity.24 infarction. Lancet. 1967;1:1194–1197. In part, the differences in tissue contrast may be due to 4. Kim RJ, Fieno DS, Parrish TB, Harris K, Chen EL, Simonetti O, Bundy J, Finn JP, Klocke FJ, Judd RM. Relationship of MRI delayed contrast inherent properties of the standard and subsecond pulse enhancement to irreversible injury, infarct age, and contractile function. sequences. For both sequences, data readout is preceded by Circulation. 1999;100:1992–2002. an inversion-recovery pulse for strong T1 weighting. How- 5. Wagner A, Mahrholdt H, Thomson L, Hager S, Meinhardt G, Rehwald ever, unlike the spoiled gradient-echo readout of the standard W, Parker M, Shah D, Sechtem U, Kim RJ, Judd RM. Effects of time, dose, and inversion time for acute myocardial infarct size measurements sequence, the steady state free-precession readout of the based on magnetic resonance imaging-delayed contrast enhancement. subsecond sequence can also lead to significant T2 weighting, J Am Coll Cardiol. 2006;47:2027–2033. especially at higher flip angles.11,25 Steady state free preces- 6. Wu E, Judd RM, Vargas JD, Klocke FJ, Bonow RO, Kim RJ. Visual- isation of presence, location, and transmural extent of healed Q-wave and sion was used because it provides intrinsically high signal- non-Q-wave myocardial infarction. Lancet. 2001;357:21–28. to-noise ratios.11,25 This advantage was utilized by incorpo- 7. Moon JC, De Arenaza DP, Elkington AG, Taneja AK, John AS, Wang D, rating high bandwidth (9-fold higher than standard) and Janardhanan R, Senior R, Lahiri A, Poole-Wilson PA, Pennell DJ. The parallel image acquisition,14 both of which reduce SNR, to pathologic basis of Q-wave and non-Q-wave myocardial infarction: a cardiovascular magnetic resonance study. J Am Coll Cardiol. 2004;44: provide sufficient temporal and spatial resolution. New ap- 554–560. proaches in which SNR is improved further by averaging 8. Wagner A, Mahrholdt H, Holly TA, Elliott MD, Regenfus M, Parker M, multiple single-shot images after motion correction may hold Klocke FJ, Bonow RO, Kim RJ, Judd RM. Contrast-enhanced MRI and 26 routine single photon emission computed tomography (SPECT) perfusion promise. imaging for detection of subendocardial myocardial infarcts: an imaging In summary, subsecond imaging has both advantages and study. Lancet. 2003;361:374–379. disadvantages compared with standard imaging. Therefore, 9. Li W, Li BS, Polzin JA, Mai VM, Prasad PV, Edelman RR. Myocardial the results suggest that the choice of technique should be delayed enhancement imaging using inversion recovery single-shot steady-state free precession: initial experience. J Magn Reson Imaging. based on multiple factors, including logistical issues at the 2004;20:327–330. CMR center, the technical knowledge level of the CMR 10. Wilson PW, D’Agostino RB, Levy D, Belanger AM, Silbershatz H, operators, and patient-specific concerns such as the presence Kannel WB. Prediction of coronary heart disease using risk factor cate- gories. Circulation. 1998;97:1837–1847. of arrhythmia, ability to breath-hold, and acuity of illness. 11. Simonetti OP, Kim RJ, Fieno DS, Hillenbrand HB, Wu E, Bundy JM, Despite some limitations, the development of subsecond Finn JP, Judd RM. An improved MR imaging technique for the visual- imaging is a significant advance for the clinical practice of ization of myocardial infarction. Radiology. 2001;218:215–223. CMR. Now there is a potential alternative to images with 12. Kim RJ, Shah DJ, Judd RM. How we perform delayed enhancement imaging. J Cardiovasc Magn Reson. 2003;5:505–514. motion artifacts or no images at all in patients who were not 13. Mahrholdt H, Wagner A, Holly TA, Elliott MD, Bonow RO, Kim RJ, referred to CMR because of dyspnea or inability to undergo Judd RM. Reproducibility of chronic infarct size measurement by a prolonged examination. Indeed, the speed and ease of use of contrast-enhanced magnetic resonance imaging. Circulation. 2002;106: 2322–2327. subsecond imaging are such that even in patients for whom 14. Sodickson DK, McKenzie CA. A generalized approach to parallel the standard technique should be preferred, eg, for the magnetic resonance imaging. Med Phys. 2001;28:1629–1643. evaluation of myocardial viability for revascularization deci- 15. Prineas R, Crow RS. The Minnesota Code Manual of Electrocardio- sions, the common practice in our laboratory is to perform graphic Findings. Littleton, Mass: John Wright & Sons Ltd; 1982. 16. Cerqueira MD, Weissman NJ, Dilsizian V, Jacobs AK, Kaul S, Laskey subsecond imaging first. This preference is based on the WK, Pennell DJ, Rumberger JA, Ryan T, Verani MS. Standardized minimal cost (because the total scan time is virtually the myocardial segmentation and nomenclature for tomographic imaging of

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 244 Circulation January 16, 2007

the heart: a statement for healthcare professionals from the Cardiac 22. Fuster V, Kim RJ. Frontiers in cardiovascular magnetic resonance. Cir- Imaging Committee of the Council on Clinical Cardiology of the culation. 2005;112:135–144. American Heart Association. Circulation. 2002;105:539–542. 23. Lardo AC, Cordeiro MA, Silva C, Amado LC, George RT, Saliaris AP, 17. Klem I, Heitner JF, Shah DJ, Sketch MH Jr, Behar V, Weinsaft J, Cawley Schuleri KH, Fernandes VR, Zviman M, Nazarian S, Halperin HR, Wu P, Parker M, Elliott M, Judd RM, Kim RJ. Improved detection of KC, Hare JM, Lima JA. Contrast-enhanced multidetector computed coronary artery disease by stress perfusion cardiovascular magnetic res- tomography viability imaging after myocardial infarction: character- onance with the use of delayed enhancement infarction imaging. JAm ization of myocyte death, microvascular obstruction, and chronic scar. Coll Cardiol. 2006;47:1630–1638. Circulation. 2006;113:394–404. 24. Kellman P, Chung YC, Simonetti OP, McVeigh ER, Arai AE. Multi- 18. Henkelman RM. Measurement of signal intensities in the presence of contrast delayed enhancement provides improved contrast between myo- noise in MR images. Med Phys. 1985;12:232–233. cardial infarction and blood pool. J Magn Reson Imaging. 2005;22: 19. Constantinides CD, Atalar E, McVeigh ER. Signal-to-noise mea- 605–613. surements in magnitude images from NMR phased arrays. Magn Reson 25. Oppelt A, Graumann R, Barfuss H, Fischer H, Hartl W, Schajor W. Eine Med. 1997;38:852–857. neue schnelle Pulssequenz fuer die Kernspintomographie. Electromedica. 20. Pruessmann KP, Weiger M, Scheidegger MB, Boesiger P. SENSE: sen- 1986;54:15–18. sitivity encoding for fast MRI. Magn Reson Med. 1999;42:952–962. 26. Kellman P, Larson AC, Hsu LY, Chung YC, Simonetti OP, McVeigh ER, 21. Bland JM, Altman DG. Statistical methods for assessing agreement Arai AE. Motion-corrected free-breathing delayed enhancement imaging between two methods of clinical measurement. Lancet. 1986;1:307–310. of myocardial infarction. Magn Reson Med. 2004;53:194–200.

CLINICAL PERSPECTIVE Cardiovascular magnetic resonance (CMR) is a highly attractive modality for the assessment of myocardial infarction and viability because of high spatial resolution and accuracy. However, some practical drawbacks potentially limit the impact of this technology for general clinical use. In comparison to some imaging modalities, standard CMR is more complex: Patient and protocol setup times are longer, and multiple breath-holds and longer scanning times are necessary. These conditions limit clinical throughput and the types of patients that can be scanned and increase the complexity and length of CMR training. A new delayed contrast-enhancement CMR technique that can acquire subsecond, “snapshot” images during free breathing (subsecond technique) could potentially overcome a number of these drawbacks and provide myocardial infarction imaging with complete left ventricular coverage in Ͻ30 seconds. In the present study, we compared the accuracy of subsecond delayed contrast-enhancement CMR with standard delayed contrast-enhancement CMR in 238 patients. We found that with the subsecond technique, the sensitivity for detecting myocardial infarction was mildly reduced (87%), but specificity and accuracy were excellent (96% and 91%, respectively). There was a tendency to underestimate the transmural extent of infarction, but overall, the techniques were highly concordant. Together, these results demonstrate that myocardial infarction can be detected rapidly by subsecond delayed contrast-enhancement CMR during free breathing with high accuracy. The clinical implication is that delayed contrast-enhancement CMR can become a quick “push-button” technique with the ability to scan a wide range of patients, including those who are more acutely ill, those with dyspnea, or those unable to undergo a prolonged examination. Moreover, clinical throughput could be increased multifold.

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Molecular Cardiology

Tumor Necrosis Factor-␣ Induces Endothelial Dysfunction in Leprdb Mice

Xue Gao, MD, PhD*; Souad Belmadani, PhD*; Andrea Picchi, MD*; Xiangbin Xu, PhD; Barry J. Potter, PhD; Neera Tewari-Singh, PhD; Stefano Capobianco, MD; William M. Chilian, PhD; Cuihua Zhang, MD, PhD

Background—We hypothesized that the inflammatory cytokine tumor necrosis factor-␣ (TNF) produces endothelial dysfunction in type 2 diabetes. Methods and Results—In m Leprdb control mice, sodium nitroprusside and acetylcholine induced dose-dependent vasodilation, and dilation to acetylcholine was blocked by the NO synthase inhibitor NG-monomethyl-L-arginine. In type 2 diabetic (Leprdb) mice, acetylcholine- or flow-induced dilation was blunted compared with m Leprdb, but sodium nitroprusside produced comparable dilation. In Leprdb mice null for TNF (dbTNFϪ/dbTNFϪ), dilation to acetylcholine or flow was greater than in diabetic Leprdb mice and comparable to that in controls. Plasma concentration of TNF was significantly increased in Leprdb versus m Leprdb mice. Real-time polymerase chain reaction and Western blotting showed that mRNA and protein expression of TNF and nuclear factor-␬B were higher in Leprdb mice than in controls. Administration of anti-TNF or soluble receptor of advanced glycation end products attenuated nuclear factor-␬B and TNF expression in the Leprdb mice. Immunostaining results show that TNF in mouse heart is localized predominantly in vascular smooth muscle cells rather than in endothelial cells and macrophages. Superoxide generation was elevated in vessels from Leprdb mice versus controls. Administration of the superoxide scavenger TEMPOL, NAD(P)H oxidase inhibitor (apocynin), or anti-TNF restored endothelium-dependent dilation in Leprdb mice. NAD(P)H oxidase activity, protein expression of nitrotyrosine, and hydrogen peroxide production were increased in Leprdb mice (compared with controls), but these variables were restored to control levels by anti-TNF. Conclusion—Advanced glycation end products/receptor of advanced glycation end products and nuclear factor-␬B signaling play pivotal roles in TNF expression through an increase in circulating and/or local vascular TNF production in the Leprdb mouse with type 2 diabetes. Increases in TNF expression induce activation of NAD(P)H oxidase and production of reactive oxidative species, leading to endothelial dysfunction in type 2 diabetes. (Circulation. 2007;115: 245-254.) Key Words: acetylcholine Ⅲ coronary disease Ⅲ inflammation Ⅲ microcirculation Ⅲ diabetes mellitus Ⅲ endothelium Ⅲ vasodilation

linical and experimental studies have demonstrated that Clinical Perspective p 254 Ccardiac function is compromised and cardiovascular diseases are increased in type 2 diabetes, suggesting that The ligand-activated transcription factor belonging to the alterations in cardiac tissue metabolism are responsible for nuclear receptor family, peroxisome proliferator-activated this impairment.1 In diabetic humans, vasodilation of coro- receptor-␥,3 is a regulator of lipid and glucose metabolism nary arteries was also altered after pharmacological (acetyl- and therefore is the target of insulin-sensitizing drugs, such as choline) or mechanical (cold test) stimuli, but these abnor- thiazolidinediones, which are frequently used to treat meta- malities of large vessels were not associated with bolic complications associated with type 2 diabetes mellitus.4 angiographic lesions and are independent of other cardiovas- Despite these connections, there has been no link established cular risk factors,2 suggesting impaired endothelial function between the vascular pathology of diabetes and the existent without any anatomic lesions. However, the exact mecha- inflammation. Tumor necrosis factor (TNF) is a proinflam- nisms underlying type 2 diabetes–induced impaired vasodi- matory cytokine that has been implicated in the pathogenesis lation remain unresolved, and there is no scientific consensus. of septic, traumatic, and hypovolemic shock–associated car-

Received July 9, 2006; accepted November 2, 2006. From the Department of Veterinary Physiology and Pharmacology, Texas A&M University, College Station, Tex (X.G., A.P., X.X., N.T.-S., S.C., C.Z.); and Department of Physiology, Louisiana State University Health Sciences Center, New Orleans (S.B., B.J.P., W.M.C.). *The first 3 authors contributed equally to this work. Correspondence to Cuihua Zhang, MD, PhD, Mechael E. DeBakey Institute, College of Veterinary Medicine and Biomedical Sciences, Texas A&M University, College Station, TX 77843-4466. E-mail [email protected] © 2007 American Heart Association, Inc. Circulation is available at http://www.circulationaha.org DOI: 10.1161/CIRCULATIONAHA.106.650671 245 246 Circulation January 16, 2007 diac dysfunction, as well as cardiovascular diseases such as Leprdb, Leprdb, and dbTNFϪ/dbTNFϪ mice with food ad libitum at the acute myocardial infarction, chronic heart failure, atheroscle- same time (7 to 8 AM) every time. (Absence of a functional leptin receptor in Leprdb mouse makes food deprivation stressful; this rosis, viral myocarditis, and cardiac allograft rejection.5 A precluded a fasting protocol.) decrease in endothelium-dependent dilation occurs shortly Ϫ after the generation of superoxide (O2 ) radicals during Lipid Level 6–9 Ϫ Serum lipid level was measured with the Cholesterol/Cholesteryl reperfusion, suggesting that endothelial generation of O2 radicals acts as a triggering mechanism for endothelial Ester Quantitation Kit (Biovision) with the use of spectrophotometry. 8,9 Ϫ dysfunction. Moreover, O2 can lead to formation of other reactive oxidative species (ROS) such as hydrogen peroxide Insulin Resistance Ϫ Blood (Ϸ1 mL) was obtained by cardiac puncture with a syringe (H2O2) and peroxynitrite (ONOO ). The interaction of ad- vanced glycation end products (AGEs) with its receptor containing 24 mmol/L EDTA. Insulin resistance was determined by using the homeostasis model assessment, which estimates steady (RAGE) induces production of ROS, which can stimulate the state insulin resistance. cascade leading to nuclear factor-␬B (NF-␬B)–induced tran- scriptional events.10,11 NF-␬B induces expression of TNF.12 Blood Pressure Despite these connections, there has been no link established Blood pressure was monitored with the use of a MacLab/8 data acquisition system (AD Instruments, Milford, Mass) equipped with between the vascular pathology of diabetes and the existent an ETH 400 transducer amplifier via the femoral artery catheterized inflammation. Therefore, we propose that type 2 diabetes– with PE-10 polyethylene tubing. induced coronary endothelial dysfunction is mediated by TNF and that further investigation will reveal the causal mRNA Expression of TNF by Real-Time mechanisms. To test this, we evaluated the endothelium- Polymerase Chain Reaction dependent and -independent vasodilation, the circulating Total RNA was extracted from isolated coronary arterioles with the levels of TNF, and its expression at the wall of coronary use of Trizol reagent (Life Technologies Inc) and was processed db directly to cDNA synthesis with SuperScript III Reverse Transcrip- arterioles in type 2 diabetes (Lepr ), type 2 diabetes null for tase (Life Technologies Inc). The primers of TNF were designed Ϫ Ϫ TNF (dbTNF /dbTNF ), and the lean control (m Leprdb) mice. (primer 3 software) and synthesized (Qiagen).7,13 cDNA was ampli- We also tested the mechanisms by which TNF induces fied with a quantitative reverse transcription polymerase chain ␬ reaction kit with SYBR Green (Life Technologies Inc). Data were endothelial dysfunction, the role of NF- B and AGE/RAGE Ϫ CT Ϫ calculated by the 2 ⌬⌬ method7 and presented as fold change of signaling in the expression of TNF, and the role of ROS (O2 , db ␤ Ϫ transcripts for TNF gene in Lepr mice normalized to -actin, H2O2, ONOO ) in coronary arterioles in type 2 diabetes. compared with m Leprdb mice (defined as 1.0-fold). Methods Treatment With TNF Neutralization or Animal Models Soluble RAGE The neutralizing antibody to TNF (anti-TNF)14 is 2E2 monoclonal The procedures followed were in accordance with approved guide- antibody (2E2 monoclonal antibody 94021402; NCI Biological lines set by the Laboratory Animal Care Committee at Texas A&M Resources Branch). At 12 to 14 weeks of age, all mice received the University. Heterozygote controls (m Leprdb), homozygote type 2 Ϫ Ϫ neutralizing anti-TNF (2E2 monoclonal antibody; 0.625 mg/mL per diabetes (Leprdb), and Leprdb null for TNF (dbTNF /dbTNF ) mice were kilogram per day IP for 3 days). Dosage was based on our estimates purchased from the Jackson Laboratory (Bar Harbor, Me) and of TNF expression (in the low nanogram or picogram range); this is maintained on a normal rodent chow diet. Our studies used 12- to able to neutralize this amount of TNF 10- to 100-fold. 14-week-old, 15- to 25-g m Leprdb and 25- to 50-g Leprdb and Ϫ Ϫ Soluble RAGE (sRAGE), the extracellular two thirds of the dbTNF /dbTNF mice of either sex. We used the same strain (C57BL/ db TNFϪ TNFϪ receptor, binds AGEs and interferes with their ability to bind and 6J) of m Lepr and db /db mice to match the backgrounds of 15 db TNFϪ TNFϪ db activate cellular RAGE. We administered sRAGE (a gift from Dr Lepr mice. The cross (db /db ) of Lepr with TNF knockout ␮ db db Ann Marie Schmidt; 80 g IP per mouse per day) to m Lepr and mice is heterozygous for Lepr and homozygous for TNF knockout db Ϫ Ϫ Ϫ Ϫ Lepr mice for 10 days to determine whether RAGE affects the mice (TNF / ). These dbTNF /dbTNF mice show the phenotypes of expression of TNF-␣. hyperglycemia and obesity, the diabetic phenotype that is consistent with the penetrance of the leptin receptor mutation. The obese mice from the second round of breeding of Leprdb and TNFϪ/Ϫ were used Functional Assessment of Isolated in experimentation. Coronary Arterioles The techniques for identification and isolation of coronary microves- Measurement of Blood Parameters sels were described in detail previously.16 Briefly, coronary arte- ␮ db db Blood was obtained from vena cava after anesthesia with sodium rioles (40 to 100 m in diameter) from m Lepr , Lepr , and TNFϪ TNFϪ pentobarbital (50 mg/kg IP) and exposure of the vein. Blood was db /db mice were carefully dissected for in vitro study.17 The collected, and the plasma was stored at Ϫ80 C° until analysis. contributions of the NO pathway in these vasodilations were exam- ined by treating the vessels with the NO synthase inhibitor NG-mono- Plasma Concentration of TNF-␣ methyl-L-arginine (L-NMMA) (10 ␮mol/L, 30-minute incubation). TNF was measured with the use of a commercial kit, BIO-Plex To determine whether TNF was playing a role in endothelial cytokine assay (BIO-Plex mouse 3-plex assay, Bio-Rad Laborato- injury in type 2 diabetes, endothelial-dependent dilation (acetylcho- ries, Hercules, Calif). TNF concentrations were automatically calcu- line, 0.1 nmol/L to 10 ␮mol/L), endothelial-independent dilation lated by BIO-Plex Manager software with the use of a standard curve (sodium nitroprusside, 0.1 nmol/L to 10 ␮mol/L), and flow-induced derived from a recombinant cytokine standard. Values were ex- dilation (NO-mediated, endothelial-dependent, but agonist- pressed as picograms per milliliter. independent; 4 to 60 cm H2O) were assessed in coronary arterioles in m Leprdb, Leprdb, and Leprdb mice treated with anti-TNF and Blood Glucose dbTNFϪ/dbTNFϪ mice. Flow is established by the production of a We used an Accu-check compact glucometer (Roche Diagnostic pressure drop across the vessel and is linearly related to the pressure Ϫ GmbH, Mannheim, Germany) for measuring blood glucose in m drop. To determine the role of TNF and O2 anion in altered Gao et al Role of TNF-␣ in Type 2 Diabetes 247 vasoactive responses in type 2 diabetes, the aforementioned vasodi- Leprdb, or Leprdb mice treated with anti-TNF. Images were obtained Ϫ latory functions were examined in the presence of the O2 scavenger with the use of a Nikon fluorescence microscope (605-nm long-pass Ϫ TEMPOL (a membrane-permeable O2 dismutase mimetic; filter). Control and experimental tissues were placed on the same 1 mmol/L, 60-minute incubation). The contributions of NAD(P)H slide and processed under the same conditions. oxidase, xanthine oxidase, and mitochondrial respiratory chain in Ϫ Electron Paramagnetic Resonance Spectroscopy generating O2 were assessed by treating the vessels with an NAD(P)H oxidase inhibitor apocynin (10 ␮mol/L), a xanthine Superoxide quantification from the electron paramagnetic resonance oxidase inhibitor allopurinol (10 ␮mol/L), or the mitochondrial spectra was determined in the homogenate (4 to 6 isolated coronary 17 respiratory chain inhibitor rotenone (1 ␮mol/L) for a 60-minute arterioles) as described previously. incubation, separately. Ebselen (10 ␮mol/L) and catalase (1000 U/mL) were also used (60-minute incubation) to determine whether Measurement of H2O2 Ϫ ONOO and H2O2 were involved in endothelial dysfunction. All Production of H2O2 was determined by using the Assay Kit (R&D drugs were administered extraluminally in these functional studies. Systems). Serum was obtained from m Leprdb, Leprdb, and Leprdb mice treated with catalase (1000 U/mL IP for 3 days) and Leprdb NAD(P)H Oxidase Activity mice treated with anti-TNF (0.625 mg/mL IP per kilogram per day NAD(P)H oxidase activity was assayed in protein isolated from for 3 days). The H2O2 concentration was then determined by coronary arteriole extracts as initiated by the addition of 50 ␮mol/L measuring the optical density of the solution in each well (microplate N,N-dimethyl-9,9-biacridinium dinitrate (Lucigenin) (Sigma, St reader set to 550 nm). Louis, Mo) to the incubation mixture. Samples were counted immediately with the use of a tabletop luminometer, and fluores- Data Analysis cence values were averaged from 2 minutes of stable readings for At the end of each experiment, the vessel was relaxed with 100 that sample. Samples were run in duplicate, and the NAD(P)H ␮mol/L sodium nitroprusside to obtain its maximal diameter at 60 17 oxidase activity was normalized to the m Leprdb control group. cm H2O intraluminal pressure. All diameter changes in response to agonists were normalized to the vasodilation in response to 100 ␮mol/L sodium nitroprusside and expressed as a percentage of Protein Expression of TNF, Nitrotyrosine, and Ϯ ␬ maximal dilation. All data are presented as mean SEM. Statistical NF- B by Western Blot Analyses comparisons of vasomotor responses under various treatments were Coronary arteries (4 to 6 vessels per sample) were separately performed with 2-way ANOVA, and intergroup differences were homogenized and sonicated in lysis buffer (Cellytic MT Mammalian tested with Bonferroni inequality. Significance was accepted at Tissue Lysis/Extraction Reagent, Sigma). Protein concentrations PϽ0.05. were assessed with the use of BCA Protein Assay Kit (Pierce The authors had full access to and take responsibility for the Biotechnology, Inc, Rockford, Ill), and equal amounts of protein (40 integrity of the data. All authors have read and agree to the ␮g) were separated by SDS-PAGE and transferred to nitrocellulose manuscript as written. membranes (Hybond, Amersham Pharmacia Biotech, Uppsala, Swe- den).18,19 TNF, nitrotyrosine (an indicator for peroxynitrite-mediated Results tissue injury), and NF-␬B protein expressions were detected with the use of TNF primary antibodies (Santa Cruz), nitrotyrosine primary Plasma Concentrations of Glucose, Blood Pressure, antibodies (Abcam), and NF-␬B primary antibodies (Abcam) in m Body Weights, Abdominal Girth, Lipid Level, and Leprdb, Leprdb, and Leprdb mice treated with anti-TNF (0.625 mg/mL Insulin Resistance per kilogram per day IP for 3 days), the NAD(P)H oxidase inhibitor Plasma parameters were measured at 12 to 14 weeks in apocynin (100 mg/kg per day IP for 3 days), or sRAGE (80 ␮g/d IP for 10 days; to antagonize RAGE signaling). Horseradish peroxi- different strains of mice (Table). dase–conjugated goat anti-mouse was used as the secondary anti- body (Abcam). Signals were visualized by enhanced chemilumines- Plasma Concentration, mRNA, and Protein cence (ECL, Amersham) and quantified by Quantity One (Bio-Rad Expression of TNF Versadoc imaging system). Figure 1 shows the plasma concentration and mRNA expres- sion of TNF in isolated coronary arterioles of m Leprdb and Immunohistochemical Analyses Leprdb mice. To identify and localize proteins in sections of arteries or myocardial tissue, we used immunohistochemistry.19 Slides prepared from formalin-fixed hearts were incubated with blocking solution (BSA Cellular Source of TNF Expression in 3% in Tris buffer), then incubated with polyclonal goat antibody Type 2 Diabetes against TNF (R&D Systems, Minneapolis, Minn) and endothelial Markers for endothelial cells (von Willebrand factor; Figure cell marker, von Willebrand factor (DakoCytomation), smooth 2), vascular smooth muscle cells (␣-actin; Figure 2), or ␣ muscle -actin (1A4; Calbiochem), or macrophages (mouse anti-rat macrophages (mouse anti-rat CD68; data not shown) along CD68; Serotec), then incubated with a secondary fluorescent anti- body (Alexa Fluor 488 and Alexa Fluor 568; Molecular Probes, with TNF to establish the cell type expressing the TNF show db Carlsbad, Calif). Sections were finally mounted in an antifading that TNF in Lepr mice hearts was localized in vascular agent (Slowfade gold with DAPI, Molecular Probes), and then the smooth muscle cells (Figure 2). Experiments were performed slides were observed and analyzed with the use of a fluorescence without the primary antibodies to test whether or not staining ϫ microscope (Leica microscope with a 63 objective). For every specificity was related to the nonspecific binding of the section, a negative control (without primary antibody) was performed. secondary antibodies, which showed no staining in heart sections, indicating that the signals were due to specific ؊ Measurement of O2 binding of the primary antibody. Dihydroethidium Ϫ Role of TNF in Type 2 Diabetes–Induced The production of O2 was evaluated in isolated coronary arterioles (40 to 100 ␮m) with the oxidative fluorescent dye dihydro- Vascular Dysfunction 17 Ϫ ethidium. Dihydroethidium fluorescence for O2 in both endothe- To show NO dependency of acetylcholine-dependent dilation lial and smooth muscle layers of vessels was measured in m Leprdb, in m Leprdb mice, we studied responses to the agonist before 248 Circulation January 16, 2007

Plasma Parameters in the Different Strains of Mice

Groups m Leprdb Leprdb dbTNFϪ/dbTNFϪ LeprdbϩAb Body weight, g 24Ϯ549Ϯ7* 46Ϯ5* 47Ϯ4* Blood pressure, mm Hg 98Ϯ6 102Ϯ4 107Ϯ5 106Ϯ3 Abdominal girth, cm 8.9Ϯ2 14.6Ϯ3 15.2Ϯ4 15.6Ϯ3 Glucose, mg/dL 139Ϯ12 283.5Ϯ19* 298.5Ϯ22* 295.7Ϯ26* Lipid level, ␮g/mL 0.18Ϯ0.17 0.44Ϯ0.2* 0.40Ϯ0.03* 0.38Ϯ0.02* Insulin resistance 0.0Ϯ0.0 2.3Ϯ0.28* 1.6Ϯ0.15* 1.8Ϯ0.19* Glucose concentration and lipid level are higher in Leprdb,dbTNFϪ/dbTNFϪ, and Leprdb mice treated with anti-TNF (LeprdbϩAb) than in m Leprdb control mice (PϽ0.05; nϭ14). The glucose concentration increased in Leprdb and dbTNFϪ/dbTNFϪ compared with m Leprdb mice, but there was no difference in lipid level among these 3 groups. In dbTNFϪ/dbTNFϪ mice, no differences were observed in body weight, blood glucose, or lipid level compared with Leprdb mice. Higher values of insulin resistance (nϭ11) were found in Leprdb,dbTNFϪ/dbTNFϪ, and Leprdb mice treated with anti-TNF. Blood pressure did not differ among the 4 groups on the day of surgery (nϭ12). *PϽ0.05 vs m Leprdb control mice. and after treatment with L-NMMA (Figure 3). Dilation to in TNF-induced endothelial dysfunction during diabetes (Fig- acetylcholine was significantly attenuated after administra- ure 5). tion of L-NMMA in m Leprdb mice (Figure 3), which indicates that vasodilation to acetylcholine was NO mediated. In dbTNFϪ/dbTNFϪ or anti-TNF–treated Leprdb mice, Type 2 Diabetes–Induced Superoxide Production acetylcholine-induced vasodilation was greater than that in in Coronary Arterioles in Type 2 Diabetes diabetic Leprdb mice and was comparable to that in m Leprdb Figure 6A shows dihydroethidium fluorescence imaging of Ϫ controls (Figures 3 and 4), which provides further support for O2 in coronary arterioles. Setting the scanning threshold to our hypothesis that TNF plays a key role in endothelial obtain a clear background image of the blood vessel allowed dysfunction in diabetes. identification of the smooth muscle and endothelial layers. In control conditions (nondiabetic, ie, m Leprdb mice), dihydro- Ϫ Roles of ROS in Type 2 Diabetes–Induced ethidium oxidative fluorescence revealed sparse levels of O2 Vascular Dysfunction throughout the vessel wall. Figure 6B shows the results from Ϫ To establish the pathway for O2 production, we administered electron paramagnetic resonance spectroscopy to quantify the Ϫ Ϫ O2 scavenger with TEMPOL (Figure 5A), the NAD(P)H production of O2 and reflects agreement with the results oxidase inhibitor apocynin (Figure 5B), the xanthine oxidase obtained by dihydroethidium and discussed above. inhibitor allopurinol (10 ␮mol/L; data not shown), or the mitochondrial respiratory chain inhibitor rotenone (1 ␮mol/L; data not shown) to determine whether vasodilation to acetyl- Type 2 Diabetes Increased NAD(P)H choline would be restored in Leprdb mice. Administration of Oxidase Activity TEMPOL, apocynin, or anti–TNF-␣ restored impaired vaso- We examined NAD(P)H oxidase activity from isolated cor- dilation to acetylcholine in Leprdb mice, but allopurinol or onary arterioles in m Leprdb and Leprdb mice (Figure 7A). The rotenone did not (Figure 5A and 5B). treatment of anti-TNF (0.625 mg/mL per kilogram per day IP

We also examined the roles of H2O2 with H2O2 inhibitor for 3 days) or apocynin (100 mg/kg per day IP for 3 days) did catalase (1000 U/mL) and ONOOϪ with the ONOOϪ scav- not affect NAD(P)H oxidase activity in m Leprdb mice (data enger ebselen (10 ␮mol/L; glutathione peroxidase mimetic) not shown).

Figure 1. A, Fold change of TNF mRNA expres- sion (real-time polymerase chain reaction) is significantly higher in Leprdb mice than in m Leprdb mice controls. B, Plasma concentration of inflammatory cytokine TNF is significantly increased in Leprdb mice. Bars represent the increased expression of TNF in Leprdb mice as a percentage of control. Data represent meanϮSD; nϭ4. *PϽ0.05 vs lean m Leprdb mice. Gao et al Role of TNF-␣ in Type 2 Diabetes 249

Figure 2. Dual fluorescence combining TNF with markers for vascular smooth muscle and endothelial cells (␣-actin and von Willebrand factor) with the use of specific antibodies against ␣-actin and von Willebrand factor followed by fluo- rescent-labeled secondary antibodies. A, B, and C, Dual labeling of TNF (green) and ␣-actin (red) in m Leprdb (Ctl) mouse heart tissue. D, E, and F, Dual labeling of TNF (green) and ␣-actin (red) in Leprdb mouse heart tissue. The pink arrow shows the staining of TNF, and the gray arrow shows the staining of ␣-actin in vascular smooth muscle cells. F shows the colocalization of TNF (green) and ␣-actin (red). G, H, and J, Dual labeling of TNF (red) and von Willebrand factor (green) in Leprdb mouse heart tissue. The white arrow in I and K inset shows a specific endothelial staining by von Wille- brand factor antibody. J and K inset shows no colocalization of TNF and von Willebrand factor. The pink arrow in inset K shows that vascular smooth muscle cells are the predominant source of TNF in coronary arteries in Leprdb. L and M, Negative control; the blue arrow shows an absence of staining in vessels with the secondary antibodies. N shows nuclear staining with DAPI (blue) in Leprdb mouse heart tissue. Magnification ϫ63. Data shown are representative of 3 separate experiments.

NF-␬B Expression Increased and sRAGE Discussion Decreased TNF Expression in Type 2 Diabetes Our results suggest that the inflammatory cytokine TNF ␬ Western blotting shows the protein expression NF- B p65 and AGE/RAGE signaling lead to oxidative stress via db db (Figure 7B) in isolated coronary arterioles of m Lepr , Lepr , NAD(P)H activation and perhaps via activating NF-␬B, db and Lepr mice treated with apocynin or anti-TNF. Apocynin which in turn may lead to TNF expression inducing ␬ db or anti-TNF decreased the expression of NF- B in Lepr endothelial dysfunction in the Leprdb mouse, which is a db mice to levels similar to those observed in m Lepr . Figure model for obesity and type 2 diabetes. Importantly, our db db 7C shows TNF protein expression in m Lepr , Lepr , findings support the concept that TNF plays a pivotal role db db sRAGE-treated m Lepr , and sRAGE-treated Lepr mice. in endothelial dysfunction in type 2 diabetes based on the db sRAGE significantly reduced TNF expression in the Lepr following observations: antibody neutralization of TNF db mice from the untreated Lepr mice. prevented coronary endothelial dysfunction and reduced Ϫ Ϫ ONOO and O2 generation and formation of nitrotyrosine ؊ db Type 2 Diabetes–Induced ONOO and H2O2 in Lepr mice. Blockade of NAD(P)H oxidase mimicked Production in Coronary Arterioles Isolated From the actions of anti-TNF on endothelial function in Leprdb Leprdb Mice mice. Molecular evidence indicated that the expressions of Western blot analysis (Figure 8A) for nitrotyrosine in homog- TNF, NF-␬B, and NAD(P)H oxidase activity were signif- enates from m Leprdb, Leprdb, and anti-TNF–treated Leprdb icantly increased in Leprdb mice; however, sRAGE de- mice revealed significantly higher levels of nitrotyrosine in creased TNF protein expression, anti-TNF attenuated Leprdb mice, which was reduced to control (m Leprdb) levels NF-␬B expression, and administration of anti-TNF to db by anti-TNF. Figure 8B shows the elevations in H2O2 pro- Lepr mice resulted in NAD(P)H oxidase levels compara- duction in Leprdb mice compared with the control m Leprdb ble to those in the control. Our findings are consistent with

strain. Treatment with anti-TNF or catalase reduced H2O2 TNF involvement by activating NAD(P)H oxidase in production in the Leprdb mice. endothelial dysfunction in type 2 diabetes. 250 Circulation January 16, 2007

Figure 3. Isolated coronary arterioles from m Leprdb control mice dilated in a concentration-dependent manner to ace- tylcholine. The NO synthase inhibitor L-NMMA (10 ␮mol/L, 30 minutes; nϭ6) inhibited vasodilation to acetylcholine vs control (Figure 3A). In a manner similar to that used with L-NMMA, acetylcholine- induced vasodilation was blunted in coro- nary arterioles from Leprdb vs m Leprdb mice (nϭ5; B). In dbTNFϪ/dbTNFϪ mice, acetylcholine-induced vasodilation was greater than that in diabetic Leprdb mice and comparable to that in m Leprdb con- trols (nϭ5; C). Sodium nitroprusside–in- duced vasodilation was normal in m Leprdb, Leprdb, and dbTNFϪ/dbTNFϪ mice (D; nϭ9). *PϽ0.05 vs m Leprdb;#PϽ0.05 vs Leprdb mice.

Impaired Coronary Control in Type 2 Diabetes endothelial dysfunction induced by TNF in isolated vessels Recent evidence supports the idea that the effect of TNF is from lean control rats, confirming that the monoclonal NO dependent by producing a rapid inhibitory action on NO anti-TNF14 was specific. Impaired endothelium-dependent synthase in the endothelium via activation of a sphingomy- vasodilation was restored in Leprdb mice after the treatment elinase/ceramide signaling pathway; this mechanism purport- with anti-TNF, demonstrating the pivotal role of this inflam- edly mediates the action of TNF, thereby contributing to matory cytokine in the vascular pathology of type 2 diabetes. vascular endothelial dysfunction in coronary circulation un- This result is in agreement with those of previous studies der different pathological conditions with increased cyto- showing that TNF can decrease the release of endothelial NO kines.9,17–19 TNF expression was significantly increased in and induce impairment of endothelium-dependent relaxation Leprdb mice; anti-TNF restored NO-mediated coronary arte- in a variety of vascular beds.14,19 rial dilation in Leprdb mice but did not affect the endotheli- In the present study, endothelium-dependent vasodilation um-dependent vasodilation in lean controls. Our previous was attenuated in coronary arterioles from Leprdb versus m study20 shows that administration of anti-TNF protected Leprdb control mice. The present results provide direct evi-

Figure 4. Anti-TNF restored NO-mediated coronary arteriolar dilation in Leprdb mice (nϭ4) but did not affect the vasodilation to acetylcholine in m Leprdb mice (A; nϭ6). B shows that flow-induced dilation was abro- gated in Leprdb mice (nϭ5) but, similar to the response to acetylcholine, anti-TNF restored flow-induced dilation to near control levels (nϭ4). Anti-TNF did not affect dilation of con- trol vessels. Nonimmune antibody (control antibodies) did not affect the responses to acetylcholine (A) or flow (B) in m Leprdb or Leprdb mice. *PϽ0.05 vs m Leprdb;#PϽ0.05 vs Leprdb mice. ⌬P indicates change in pressure. Gao et al Role of TNF-␣ in Type 2 Diabetes 251

Figure 5. TEMPOL (1 mmol/L; A) and apocynin (10 ␮mol/L; B) restored vasodi- lation in Leprdb mice. Administration of ebselen (10 ␮mol/L; C) or catalase (1000 U/mL; D) also restored impaired vasodi- lation to acetylcholine in Leprdb mice. *PϽ0.05 vs m Leprdb;#PϽ0.05 vs Leprdb mice.

dence that type 2 diabetes is associated with impaired production as measured by dihydroethidium fluorescence and agonist-induced NO production and NO-mediated dilation in electron paramagnetic resonance. Moreover, the antagonism of the coronary microcirculation. We found increased TNF NAD(P)H oxidase virtually normalized endothelium-dependent mRNA expression (6-fold), plasma concentration of TNF vasodilatation. Our results strongly suggest that the pathway for (6-fold), and protein expression of TNF (Ͼ2.5-fold in small TNF-induced endothelial dysfunction is mediated by activation coronary arteries), but sRAGE decreased TNF expression in of NF-␬B, NAD(P)H oxidase, and the subsequent production of db Ϫ Lepr mice. Immunostaining results showed that TNF in O2 . On the contrary, we did not find any improvement in Leprdb mice heart is localized in vascular smooth muscle cells. endothelial function after incubation with rotenone and allopuri- We believe that ROS production in vascular endothelial cells nol, suggesting that mitochondria and the xanthine oxidase Ϫ and smooth muscle cells by TNF would limit NO bioavail- system are not likely to be the source of O2 production. Guzik 21 Ϫ ability and reduce NO-dependent dilation. It is provocative to et al measured O2 production in diabetic and nondiabetic note that despite the similarities in glucose, body weights, vessels in response to a range of potential oxidase inhibitors.13 lipid level, insulin resistance, and blood pressure in diabetic Their results show that oxypurinol and rotenone had minimal or TNFϪ TNFϪ Ϫ animals, endothelial function was better in db /db and modest effect on O2 production, whereas diphenylene iodo- in Leprdb mice treated with anti-TNF. This suggests that nium, an inhibitor of flavin-containing oxidases such as Ϫ AGE/RAGE signaling plays a pivotal role in attenuating TNF NAD(P)H oxidases, abolished O2 production. Our results protein expression, and TNF is the key cytokine that induced demonstrate that the production of TNF is basal to the process of endothelial dysfunction in type 2 diabetes. eliciting this oxidative stress. Guzik et al21 reported that the protein expression of the NAD(P)H oxidase subunits p22, p47, ؊ The Major Source of O2 Production in and p67-phox are significantly increased in diabetic human Type 2 Diabetes tissue versus normal control. We measured NAD(P)H oxidase Although there are multiple intracellular sources for formation of activity by determining reductions in superoxide production oxygen free radicals (eg, mitochondria, xanthine oxidase, after inhibition of NAD(P)H oxidase. This supports our func- NAD(P)H oxidase), our results support the idea that the major tional results, in which the experiments with apocynin suggest enzyme activated by TNF in type 2 diabetes is NAD(P)H that NAD(P)H oxidase might be a major enzymatic vascular oxidase. We can state with conviction that the major source of source of ROS in Leprdb mice. Ϫ Ϫ O2 was NAD(P)H oxidase in type 2 diabetes because the O2 production is postulated to be linked to TNF. Our Ϫ NAD(P)H oxidase inhibitor apocynin significantly reduced O2 results also provide insight into the basis for the endothelial 252 Circulation January 16, 2007

Figure 6. Dihydroethidium fluorescence (A) was markedly ele- vated in both endothelial (arrowhead) and vascular smooth mus- cle (arrow) cells of arterial sections in Leprdb mice (nϭ4) com- pared with m Leprdb mice (control). Anti-TNF significantly reduced the fluorescent signals in Leprdb mice. However, anti- TNF did not affect the fluorescent signals in m Leprdb mice (data not shown). Data shown are representative of 4 separate experi- ments. B shows the results from electron paramagnetic reso- Ϫ Ϫ nance spectroscopy to quantify the production of O2 .O2 pro- duction was higher in isolated coronary arteries from Leprdb mice vs m Leprdb mice (PϽ0.05). Administration of anti-TNF Ϫ reduced the production of O2 to the level observed in the m Leprdb controls; nϭ6 (B). *PϽ0.05 vs m Leprdb mice; #PϽ0.05 vs Leprdb mice. dysfunction induced by type 2 diabetes, namely, oxidative stress. The present study indicates that the model of type 2 diabetes increases TNF, which stimulates endothelial gener- Ϫ ation of O2 through activation of NAD(P)H oxidase in the endothelium and contributes to the endothelial dysfunction. To our knowledge, this is the first functional study to link the mechanism(s) of the model of type 2 diabetes—in terms of endothelial dysfunction—to TNF, the subsequent activation Ϫ of NAD(P)H oxidase, and thus the production of O2 in coronary artery endothelium. The link between TNF overex- Figure 7. NAD(P)H oxidase activity was higher in Leprdb mice Ϫ db ϭ pression and O2 production has been investigated by Zhang than in m Lepr control mice (n 8). Treatment by apocynin et al,13 who found that tiron, a cell-permeable O Ϫ scavenger, (100 mg/kg per day IP for 3 days) or by anti-TNF (0.625 mg/mL 2 per kilogram per day IP for 3 days) significantly decreased and polyethylene glycol–superoxide dismutase prevented NAD(P)H oxidase activity in Leprdb mice, separately. Bars repre- TNF-induced impairment of endothelium-dependent vasore- sent the increased NAD(P)H oxidase activity in Leprdb mice as a laxation in coronary arterioles. Others18,19 have shown that fold change of control, eg, 2 represents a doubling of NAD(P)H oxidase activity. The protein expression of NF-␬B p65 (B; nϭ5) TNF activates sphingosine kinase and, in vascular smooth was higher in Leprdb mice than in m Leprdb mice, but NF-␬B p65 muscle, sphingosine kinase activation leads to NADPH acti- expression in apocynin-treated or anti-TNF–treated (3 days; IP) vation. Our results are consistent with these studies and was similar in Leprdb mice compared with m Leprdb mice. C db support our hypotheses that TNF and O Ϫ are connected in the shows the elevations of TNF expression: [mt]2-fold in Lepr 2 mice. sRAGE decreased TNF expression in Leprdb mice but did production of oxidative stress and endothelial dysfunction in not affect TNF expression in m Leprdb mice. *PϽ0.05 vs m type 2 diabetes. Our results also indicate that there is a link Leprdb mice; #PϽ0.05 vs Leprdb mice. among NF-␬B, NAD(P)H oxidase, and TNF in type 2 diabetes. Specifically, the results suggest that AGE/RAGE this scheme is that oxidative stress begets further oxidative and TNF signaling in the diabetic mouse lead to oxidative stress in the obese diabetic animal, and this may explain the stress via NAD(P)H activation, activating NF-␬B, which in development and evolution of vascular pathology in this turn may activate more TNF expression. The implication of condition. Gao et al Role of TNF-␣ in Type 2 Diabetes 253

produces the formation of ONOOϪ, which is a very potent oxidant. We cannot measure this species because of its very short half-life but can measure an index of its formation by performing immunohistochemical analyses and Western blot- ting for nitrotyrosine. Our results show that the production of Ϫ db ONOO (nitrotyrosine) and H2O2 is higher in Lepr mice than in controls and that anti-TNF reduces the production of Ϫ ONOO and H2O2 in diabetic mice, which indicates that TNF Ϫ may also contribute to increasing ONOO and H2O2 in Ϫ Ϫ Ϫ addition to O2 , and like O2 , ONOO , and H2O2 may also participate in endothelial dysfunction in type 2 diabetes. Additionally, apocynin and catalase greatly attenuated the

signals for NAD(P)H oxidase activity and H2O2, thus show- ing specificity for the particular ROS in these experiments. Our study shows the existence of a causal link between TNF Ϫ expression and O2 production in the endothelial dysfunction Ϫ occurring in type 2 diabetes; increased O2 also leads to the Ϫ Ϫ Ϫ formation of ONOO , so that O2 , ONOO , and H2O2 all limit the bioavailability of NO in type 2 diabetes. In conclusion, we found that TNF overexpression impairs endothelium-dependent vasodilation in coronary arterioles of type 2 diabetic mice, and the impaired endothelial function can be restored toward normal by administration of TNF antibodies. The mechanism by which TNF affects endothelial function is through increased superoxide production by NAD(P)H oxidases, which in turn leads to a reduced NO bioactivity by direct scavenging. These results confirmed that TNF plays a pivotal role in the vascular pathology of type 2 diabetes and provides new findings in the understanding of interactions between TNF and inflammation, diabetes, and Figure 8. We analyzed bands of nitrotyrosine (N-Tyr) migrating atherosclerosis. These findings may provide further insight at 95 and 50 kDa (nϭ4). In graphs a and b, the results are into a novel therapeutic target for cardiovascular diseases shown for nitrotyrosine bands migrating at 95 and 50 kDa, respectively. A (a and b) shows that both 95- and 50-kDa bands associated with elevated levels of TNF. of nitrotyrosine are higher in Leprdb mice than in m Leprdb mice; after administration of anti-TNF (0.625 mg/mL per kilogram per Acknowledgments day IP for 3 days), both bands of nitrotyrosine are decreased in We thank Dr Ann Marie Schmidt for her generous proprietary gift of Leprdb mice. c shows a typical nitrotyrosine blot; the arrows point to the analyzed bands. *PϽ0.05 vs m Leprdb;#PϽ0.05 vs sRAGE. db Lepr mice. B shows that H2O2 production (nϭ6) is higher in Leprdb than in controls (m Leprdb). Anti-TNF reduced the produc- Sources of Funding db tion of H2O2 in Lepr mice. Catalase (1000 U/mL) greatly atten- This study was supported by grants from Pfizer Atorvastatin Re- db uated H2O2 production in Lepr mice without affecting H2O2 db search Award (2004-37), an American Heart Association Scientist production in m Lepr mice, thus showing specificity of the Development grant (110350047A), and a National Institutes of measurement for the particular H2O2 production. *PϽ0.05 vs m Leprdb mice; #PϽ0.05 vs Leprdb mice. Health grant (RO1-HL077566) to Dr Cuihua Zhang.

؊ Disclosures Role of ONOO Anion and H2O2 in None. Type 2 Diabetes Oxygen-derived free radicals impair endothelium-dependent References relaxation, and NO formed in the endothelium is inactivated 1. Mahgoub MA, Abd-Elfattah AS. Diabetes mellitus and cardiac function. Ϫ Mol Cell Biochem. 1998;180:59–64. by superoxide anion radical O2 to form a stable peroxynitrite Ϫ Ϫ 2. Nitenberg A, Paycha F, Ledoux S, Sachs R, Attali JR, Valensi P. anion, ONOO . The ONOO anion can produce further Coronary artery responses to physiological stimuli are improved by def- oxidative damage to cells because of its inherent potency and eroxamine but not by L-arginine in non-insulin-dependent diabetic 22 Ϫ patients with angiographically normal coronary arteries and no other risk stability. Previous studies have shown that O2 may be dismutated (spontaneously or enzymatically) to H O , which factors. Circulation. 1998;97:736–743. 2 2 3. Verges B. Clinical interest of PPARs ligands. Diabetes Metab. 2004; is another oxidant and therefore another possible mediator of 30:7–12. oxidant injury. Our data show that H2O2 is also involved in 4. Moller DE. New drug targets for type 2 diabetes and the metabolic this mechanism because the H O inhibitor catalase partially syndrome. Nature. 2001;414:821–827. 2 2 5. Meldrum DM, Cleveland JC, Cain BS. Increased myocardial tumor protected the impaired vasodilation induced by acetylcholine necrosis factor-alpha in a crystalloid-perfused model of cardiac ischemia- Ϫ in type 2 diabetes. In addition, the reaction of NO and O2 reperfusion injury. Ann Thorac Surg. 1998;65:439–443. 254 Circulation January 16, 2007

6. Downey JM, Omar B, Ooiwa H. Superoxide dismutase therapy for myo- 15. Feng L, Matsumoto C, Schwartz A, Schmidt AM, Stern DM, Pile- cardial ischemia. Free Radic Res Commun. 1991;12–13:703–720. Spellman J. Chronic vascular inflammation in patients with type 2 dia- 7. Livak KJ, Schmittgen TD. Analysis of relative gene expression data using betes: endothelial biopsy and RT-PCR analysis. Diabetes Care. 2005;28: Ϫ⌬⌬CT real-time quantitative PCR and the 2 method. Methods. 2001;25: 379–384. 402–408. 16. Kuo L, Davis MJ, Chilian WM. Myogenic activity in isolated subepi- 8. Salvemini D, Cuzzocrea S. Superoxide, superoxide dismutase and ische- cardial and subendocardial coronary arterioles. Am J Physiol. 1988;255: mic injury. Curr Opin Investig Drugs. 2002;3:886–895. H1558–H1562. 9. Lefer AM, Ma XL. Cytokines and growth factors in endothelial dys- function. Crit Care Med. 1993;21(suppl):S9–S14. 17. Zhang C, Xu X, Potter BJ, Wang W, Kuo L, Michael L, Bagby GJ, 10. Zou W, Amcheslavsky A, Takeshita S, Drissi H, Bar-Shavit Z. Chilian MW. TNF-alpha contributes to endothelial dysfunction in ische- TNF-alpha expression is transcriptionally regulated by RANK ligand. mia/reperfusion injury. Arterioscler Thromb Vasc Biol. 2006;26: J Cell Physiol. 2005;202:371–378. 475–480. 11. Tanaka N, Yonekura H, Yamagishi S, Fujimori H, Yamamoto Y, 18. Keller M, Lidington D, Vogel L. Sphingosine kinase functionally links Yamamoto H. The receptor for advanced glycation end products is elevated transmural pressure and increased reactive oxygen species for- induced by the glycation products themselves and tumor necrosis mation in resistance arteries. FASEB J. 2006;20:702–704. factor-alpha through nuclear factor-kappa B, and by 17beta-estradiol 19. Zhang C, Hein TW, Wang W, Ren Y, Shipley RD, Kuo L. Activation of through Sp-1 in human vascular endothelial cells. J Biol Chem. 2000; JNK and xanthine oxidase by TNF-␣ impairs nitric oxide-mediated 275:25781–25790. dilation of coronary arterioles. J Mol Cell Cardiol. 2006;40:247–257. 12. De Martin R, Hoeth M, Hofer-Warbinek R. The transcription factor 20. Picchi A, Gao X, Potter BJ. TNF alpha induces endothelial dysfunction in NF-kappa B and the regulation of vascular cell function. Arterioscler the prediabetic metabolic syndrome. Circ Res. 2006;99:69–77. Thromb Vasc Biol. 2000;20:E83–E88. 13. Zhang DX, Yi FX, Zou AP, Li PL. Role of ceramide in TNF-alpha 21. Guzik TJ, Mussa S, Gastaldi D, Sadowsky J, Ratnatunga C, Pillai R, induced impairment of endothelium-dependent vasorelaxation in Channon KM. Mechanisms of increased vascular superoxide production coronary arteries. Am J Physiol. 2002;283:H1785–H1794. in humans diabetes mellitus: role of NADPH oxidase and endothelial 14. Lattime EC, Stutman O. Thymic lymphomas mediate non-MHC- nitric oxide synthase. Circulation. 2002;105:1656–1662. restricted, TNF-dependent lysis of the murine sarcoma WEHI-164. Cell 22. Cai H. Hydrogen peroxide regulation of endothelial function: origins, Immunol. 1991;136:69–79. mechanisms, and consequences. Cardiovasc Res. 2005;68:26–36.

CLINICAL PERSPECTIVE Diabetes is one of the leading risk factors for the development of coronary artery and peripheral vascular diseases. Before vascular disease develops in diabetics, endothelial dysfunction occurs. In fact, endothelial dysfunction appears to be a hallmark underlying vascular disease of many etiologies. We believe that understanding endothelial dysfunction is critical because the progression of vascular disease may be halted if endothelial dysfunction is rectified. Our goal was to delineate a potential cause of endothelial dysfunction by testing the hypothesis that tumor necrosis factor-␣ (TNF-␣) induces inflammation responsible for endothelial dysfunction in type 2 diabetes. Our data revealed that endothelial function was normal in diabetic mice that were lacking TNF (TNF knockout in the Leprdb diabetic mouse). Moreover, we observed that diabetic mice have elevated expression of TNF, suggesting that this inflammatory cytokine produces, or at least contributes to, endothelial dysfunction in diabetes. We also found that the endothelial dysfunction by TNF in diabetes was related to excess production of the free radical superoxide. Finally, we observed that advanced glycosylation end products and the receptor for these products seem to amplify TNF expression in diabetes; thus, TNF and advanced glycation end products/receptor of advanced glycation end products signaling play pivotal roles in endothelial dysfunction in type 2 diabetes. Furthermore, our study may provide new approaches for the treatment of vasculopathy in type 2 diabetes, such as the possible use of antibodies against TNF or its receptors or decoy receptors for this inflammatory agent. Hyperhomocysteinemia Alters Cardiac Substrate Metabolism by Impairing Nitric Oxide Bioavailability Through Oxidative Stress

Nobuhiro Suematsu, MD, PhD; Caroline Ojaimi, PhD; Shintaro Kinugawa, MD, PhD; Zipping Wang, MS; Xiaobin Xu, MD; Akos Koller MD, PhD; Fabio A. Recchia, MD, PhD; Thomas H. Hintze, PhD

Background—Hyperhomocysteinemia (HHcy) has been considered a vascular disease associated with increased levels of oxidative stress that results in scavenging of NO. However, little is known of the impact of HHcy on cardiac function and especially myocardial metabolism. Methods and Results—L-Homocysteine was intravenously infused into conscious dogs, and the dogs were fed methionine to increase plasma homocysteine to 10 ␮mol/L for acute and 24 ␮mol/L for chronic HHcy. There was no significant change in hemodynamics with HHcy. Veratrine-induced, NO-dependent, coronary vasodilation (Bezold-Jarisch reflex) was reduced by 32% but was restored by simultaneous intravenous infusion of ascorbic acid or apocynin. Acute and chronic HHcy significantly increased uptake of glucose and lactate and decreased uptake of free fatty acid by the heart. HHcy significantly decreased bradykinin- or carbachol-induced reduction of myocardial oxygen consumption in vitro, and this effect was completely restored by coincubation with ascorbic acid, Tempol, or apocynin. Western blot analysis indicated an increase in Nox2 (82%) and a reduction in endothelial nitric oxide synthase (39%), phospho-endothelial nitric oxide synthase (39%), and superoxide dismutase-1 (45%). Microarray analysis of gene expression in heart tissue from chronic HHcy indicated a switch in cardiac phenotype to enzymes that metabolize glucose. Conclusions—HHcy directly modulates substrate use by the heart independent of changes in hemodynamics or ventricular function by reducing NO bioavailability through the generation of superoxide. The progression of cardiac or coronary heart disease associated with HHcy should be evaluated in light of the impact of alterations in the regulation of cardiac metabolism and substrate use. (Circulation. 2007;115:255-262.) Key Words: blood flow Ⅲ heart failure Ⅲ metabolism Ⅲ nitric oxide Ⅲ physiology

omocysteine (Hcy) is a sulfur-containing amino acid the bioavailability of NO in cultured bovine aortic endothelial Hthat is derived from dietary methionine. An elevation in cells. Ungvari et al9 have also shown that HHcy increases plasma Hcy, hyperhomocysteinemia (HHcy), has been shown tumor necrosis factor ␣ expression, which induces a proin- to be an independent risk factor for atherosclerosis, which flammatory vascular phenotype through oxidative stress in includes peripheral vascular disease,1 venous thrombosis,2 coronary arteries of Wistar rats. Although the precise mech- coronary artery disease,3 and cerebrovascular disease.4 Hcy anisms are still not well understood, a number of in vitro and concentrations are elevated in up to 30% of patients with in vivo studies in humans and in animals have indicated that atherosclerosis,3 and levels only 12% above the upper limit of HHcy causes vascular endothelial dysfunction that leads to normal are associated with a 3-fold increase in the risk of atherosclerosis mainly by increasing oxidative stress and acute myocardial infarction.5 In healthy adult humans, Bel- attenuating NO bioavailability. lamy et al6 have shown that oral methionine loading raises plasma Hcy and impairs flow-mediated endothelium- Clinical Perspective p 262 dependent vasodilation. Chambers et al7 also have shown that On the other hand, little is known about the effects of this effect is mediated through increased oxidative stress. HHcy on cardiac function and myocardial metabolism. In Upchurch et al8 have demonstrated with animal models that recent years, several investigators, including those participat- HHcy reduces glutathione peroxidase activity and decreases ing in the present study, have proposed that NO may play an

Received July 18, 2006; accepted October 13, 2006. From the Department of Physiology, New York Medical College, Valhalla, NY. The online-only Data Supplement, consisting of a table, is available with this article at http://circ.ahajournals.org/cgi/content/ full/CIRCULATIONAHA.106.652693/DC1. Address correspondence to Prof. Thomas H. Hintze, Department of Physiology, New York Medical College, Valhalla, NY 10595. E-mail [email protected] © 2007 American Heart Association, Inc. Circulation is available at http://www.circulationaha.org DOI: 10.1161/CIRCULATIONAHA.106.652693

Downloaded from circ.ahajournals.org 255at Mohammed Mahboob on June 10, 2007 256 Circulation January 16, 2007

important role in the regulation of cardiac metabolism. NO, trophotometer (Bio-Tek Instruments, Winooski, Vt) at 450 nm after which is a potent vasodilator, modulates mitochondrial res- addition of the substrate tetramethylbenzidine. piration in vivo10 and in vitro.11 NO attenuates mitochondrial Protocols respiration by inhibition of complexes I and II of the electron transport chain12–14 and by interactions with cytochrome Acute Hyperhomocysteinemia oxidase.13 Through a cGMP-mediated mechanism, NO also To examine the effects of acute HHcy on the regulation of cardiac 15 function and substrate use by NO, L-Homocysteine (Sigma-Aldrich) regulates glucose uptake by the heart. was intravenously infused into conscious dogs at 0.8 mg/min for 120 Recent studies have also detected elevated levels of super- minutes. oxide in HHcy, which results in impaired NO-dependent vasodilation.9,16–18 Such dysfunction of coronary endotheli- Effects of Hcy on Activation of Bezold-Jarisch Reflex by Veratrine um-mediated relaxation is a strong predictor of atheroscle- We measured the NO-dependent response of CBF to veratrine at 19 rotic disease and future cardiovascular events. We have also baseline, 40 minutes, 80 minutes, and 120 minutes during intrave- shown previously that HHcy dose-dependently inhibits the nous infusion of L-Hcy as described previously.25 After 120 minutes

NO-dependent regulation of cardiac O2 consumption in vitro of L-Hcy infusion, ascorbic acid was administered intravenously at through the generation of superoxide by NADPH oxidase in an initial dose of 2000 mg and followed by a constant infusion of 25 20 mg/min for 120 minutes (simultaneous with L-Hcy infusion) to rats and mice that were fed methionine for 5 to 9 weeks. scavenge superoxide as previously reported.25 However, the mechanism by which superoxide levels are increased in HHcy has not yet been fully elucidated. Because Chronic Hyperhomocysteinemia the hypertension of HHcy is modest at best and certainly less L-methionine (Sigma-Aldrich) was added to the drinking water (10 g/L) of dogs for 2 weeks. We performed the veratrine-induced than that after complete NO synthesis inhibition, it is plausi- activation of the Bezold-Jarish reflex, and we also collected blood ble that only certain aspects of the function of NO are altered; samples for the measurement of glucose, lactate, FFA, and Hcy ie, the threshold for effects on cardiac metabolism are less concentration. Apocynin, which was administered intravenously at a than those for vascular dysfunction. Therefore, the causes, dose of 10 mg/kg for 120 minutes, was used as an inhibitor of extent, and consequences of NO inactivation in HHcy on NADPH oxidase activation. Ascorbic acid was given on another day to scavenge superoxide. After these experiments, we euthanized the cardiac function, oxygen consumption, and substrate use all dogs and took hearts for the measurement of oxygen consumption, require further investigation. Thus the purpose of our present Western blotting, and gene expression. Control hearts were harvested study was to evaluate the potential impact of HHcy on cardiac from 7 similarly instrumented dogs. metabolism with special reference to the production of Calculation of Myocardial Oxygen Consumption, ATP superoxide and scavenging of NO. Equivalent, and Triple Product

Myocardial oxygen consumption (MVO2) was calculated by multi- Methods plying the arterial-coronary sinus difference in oxygen content by CBF. Assumed total ATP synthesis levels in both control and HHcy Surgical Preparation were calculated by the theoretical ATP yield from each substrate as Twenty male mongrel dogs (25 to 27 kg) were sedated with follows: glucose 36, lactate 18, and palmitate 129 (mol ATP/mol acepromazine maleate (1 mg/kg IM), anesthetized with sodium substrate). Triple product, an index of mechanically related oxygen pentobarbital (25 mg/kg IV), and ventilated with room air. A consumption, was calculated as LV systolic pressure multiplied by

thoracotomy was performed and catheters (Tygon) were placed in dP/dtmax and heart rate and expressed in units. the aorta, left atrium, and coronary sinus. A left ventricular (LV) pressure gauge (P 6.5, Konigsberg Instruments, Inc, Pasadena, Measurement of O2 Consumption in Cardiac Calif), a Doppler flow transducer (Craig Hartley), and a pair of Muscle Segments pacing electrodes were implanted as described previously.21 All Myocardial tissue was isolated from the LV free wall of hearts.

protocols were approved by the Institutional Animal Care and Use MVO2 was measured polarographically in vitro with a Clark-type Committee of New York Medical College. oxygen electrode (YSI-5331, Yellow Springs Instruments, Yellow Springs, Ohio). Tissue respiration was calculated as the rate of Hemodynamic Recordings decrease in oxygen concentration after the addition of muscle slices Arterial pressure, LV pressure, and left circumflex coronary blood with the assumption of an initial oxygen concentration of 224 flow (CBF) were measured, heart rate was monitored, and mean nmol/mL; tissue respiration was expressed as nanomoles of oxygen Ϫ arterial pressure and mean CBF were derived, as we described consumed per minute per gram of tissue. Cumulative doses (10 7 to Ϫ4 previously.22 10 mol/L) of 1 of 3 drugs: the B2 kinin receptor agonist BK, the muscarinic-2 receptor agonist carbachol, or the exogenous NO donor Cardiac Metabolites S-nitroso-N-acetyl-penicillamine were used. The O2 scavenger ascor- bic acid (10Ϫ3 mol/L), the superoxide dismutase (SOD) mimetic Blood samples from aorta and coronary sinus were collected in Ϫ Tempol (10 3 mol/L), or the NADPH oxidase inhibitor apocynin plastic syringes treated with heparin or EDTA. Blood gases, lactate, Ϫ (10 4 mol/L) were used as antioxidants.20 and glucose were measured with a blood gas analyzer (IL-682 CO-Oximeter). Free fatty acid (FFA) analysis was performed on Western Blot Analysis plasma from EDTA-treated samples with a colorimetric assay The preparation of protein samples from myocardial tissues was (NEFA C kit from Wako Diagnostics, Richmond Va) as described performed as described previously.20 Standard techniques for elec- previously.23 Triglyceride content in the myocardium was measured trophoresis (150 V) and Western blotting were used to separate as previously reported.24 proteins by molecular weight and transfer them to a polyvinylidene difluoride membrane (Amersham Pharmacia Biotech, Piscataway, Measurement of Plasma Hcy NJ). Antibodies to endothelial NO synthase (eNOS; Affinity Biore- Plasma Hcy was quantified with a Hcy microplate enzyme immu- agents, Golden, Colo; 1:1000 dilution), phospho-eNOS (serine-1177; noassay (Bio-Rad Laboratories, Hercules, Calif) as we have done Cell Signaling Technology, Danvers, Mass; 1:1000 dilution), Cu/Zn previously.20 Activity was assessed with a Power Wave 200 spec- SOD (SOD-1; Calbiochem, EMD Biosciences, Darmstadt, Germany;

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Suematsu et al Hyperhomocysteinemia, a Cardiac Metabolic Disease 257

TABLE 1. Hemodynamic Data

HHcy Dogs (nϭ10)

Control Dogs (nϭ10) 120 Minutes (nϭ10) 1 Week (nϭ10) 2 Weeks (nϭ10) Body weight, kg 25Ϯ0.7 25Ϯ0.7 24Ϯ0.8 24Ϯ0.8 Hemodynamic measurement LVSP, mm Hg 136Ϯ4.8 128Ϯ4.1 136Ϯ8.2 132Ϯ6.7 SBP, mm Hg 136Ϯ4.8 127Ϯ4.7 134Ϯ5.4 137Ϯ3.8 DBP, mm Hg 88Ϯ4.0 80Ϯ4.0 87Ϯ4.2 88Ϯ6.5 mAoP, mm Hg 110Ϯ3.8 103Ϯ3.9 109Ϯ6.3 111Ϯ5.4 CBF, mL/min 29.6Ϯ2.8 28.1Ϯ3.4 27.9Ϯ3.8 32.7Ϯ5.2

dP/dtmax 3407Ϯ245 3349Ϯ319 3064Ϯ220 3295Ϯ222 Heart rate, bpm 90Ϯ3.9 93Ϯ6.1 96Ϯ6.3 102Ϯ7.0 Summary data for the body weight, plasma Hcy concentration, and hemodynamic parameters of both control dogs and HHcy dogs (120 minutes, 1 week, and 2 weeks). LVSP indicates LV systolic pressure; SBP, systolic blood pressure; DBP, diastolic blood pressure; and mAoP, mean arterial pressure.

1:5000 dilution), Mn SOD (SOD-2; BD Transduction; 1:10 000 Results dilution), or one of the following subunits of NADPH oxidase: p67phox (BD Transduction, BD Biosciences, San Jose, Calif; 1:2000 Plasma Homocysteine dilution), p22phox (Santa Cruz Biotechnology, Inc, Santa Cruz, Calif; Hcy levels of control were 4.4Ϯ0.2 ␮mol/L; infusion of 1:2000 dilution), Nox2 (gp91phox; BD Transduction; 1:1000 dilution), L-Hcy for 120 minutes significantly increased Hcy levels to p47phox (Santa Cruz Biotechnology, Inc; 1:1000 dilution) or phospho- 8.9Ϯ0.5 ␮mol/L, and Hcy levels finally reached 10.1Ϯ1.8 p47 (phospho-tyrosine antibody; Upstate, Millipore Corp, Billerica, ␮mol/L after 90 more minutes during the simultaneous Mass; 1:5000 dilution) were used. infusion of ascorbic acid (PϽ0.01, nϭ10). Addition of RNA Isolation and Microarray Analysis methionine to diet for 2 weeks increased Hcy concentration to Total cardiac RNA was extracted from the LV of the control and 24.3Ϯ9.6 ␮mol/L (PϽ0.01, nϭ10). methionine-fed groups (nϭ4) as described previously.26 RNA qual- ity was assessed by electrophoresis with the Agilent Bioanalyzer Hemodynamic Measurements in Dogs 2100 (Agilent Technologies, Santa Clara, Calif). RNA from each All hemodynamic data and body weight are shown in Table sample was used to generate a high-fidelity cDNA with the labeling 1. LV systolic pressure, systolic arterial blood pressure, protocols for sample preparation recommended by Affymetrix. The microarray labeling, hybridization, and analysis procedures were diastolic arterial blood pressure, mean arterial pressure, mean described previously in more detail.26 Each sample was individually CBF, dP/dtmax, and heart rate in resting conscious dogs did not subjected to gene expression analysis via the Affymetrix Canine differ significantly between control and HHcy. Genome Array. The GeneChip Canine Genome Array was used and contains Ͼ23 813 Canis familiaris probe sets to monitor gene Effects of Veratrine on the Coronary Circulation expression for Ͼ16 000 full-length C. familiaris transcripts. Se- Figure 1 shows the representative recordings of veratrine- quence information for the GeneChip Canine Genome Array in- induced, NO-dependent, coronary vasodilation (Bezold- cludes public content from GenBank (release 137.0, August 2003), Jarisch reflex) during acute HHcy. Veratrine administration at dbEST (October 2003), and proprietary beagle sequence content ␮ licensed from LION Bioscience AG. All the hybridization data have a dose of 5 g/kg caused significant increases in CBF been submitted to the National Center for Biotechnology Information (76Ϯ19%, nϭ7). Infusion of L-Hcy significantly decreased (NCBI) Gene Expression Omnibus database (http://www.ncbi.nlm. this response by 42%, and this effect was completely reversed nih.gov/geo) with GEO Accession Numbers for series GSE5260. by the simultaneous infusion of ascorbic acid (PϽ0.05, nϭ7). Similar to acute HHcy, 2 weeks of methionine-enhanced diet Data Analysis significantly attenuated this response by 32% (PϽ0.05, nϭ8) Ϯ All data are presented as mean SEM. In the in vivo studies, the (Figure 2), which was reversed by the simultaneous infusion responses are the peak after administration of veratrine. Statistical of ascorbic acid (nϭ5) or apocynin (nϭ6). significance of differences was determined with Student t test for each peak response, and differences between groups were deter- mined with repeated-measures ANOVA. Changes are considered Effects of Hcy on Energy Metabolism significant at a value of PϽ0.05. The short-term infusion of Hcy significantly increased glu- Determination of statistical significance for changes in gene cose uptake (1.47Ϯ0.72 to 7.48Ϯ1.58 ␮mol/min, PϽ0.01, expression was performed in GeneTraffic with a t test and with nϭ7) and also decreased FFA uptake (7.38Ϯ1.09 to variance stabilization. Differences were considered statistically sig- 3.88Ϯ1.08 ␮mol/min, PϽ0.05, nϭ6). Lactate uptake tended Յ Ϯ nificant at a nominal significance of P 0.05 and at least a 2-fold to increase (3.76Ϯ2.94 to 12.15Ϯ6.21 ␮mol/min, Pϭ0.25, change in expression between control and dogs fed methionine for 2 nϭ6), but this did not reach the statistical significance. All of weeks. The authors had full access to the data and take responsibility for this substrate switching occurred in the absence of hyperten- the integrity of the data. All authors have read and agree to the sion or any change in CBF, which indicates a lack of effect of manuscript as written. HHcy on resistance vessels. Figure 3 shows the relationship

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 258 Circulation January 16, 2007

Figure 1. A, Representative recordings of veratrine-induced (5 ␮g/kg), NO-dependent, coronary vasodilation (Bezold-Jarisch Figure 3. Relationship between Hcy concentration and the sub- reflex) in acute HHcy. The response to veratrine was reduced strate switching from FFA (nϭ6) to glucose (nϭ7) or lactate after 120 minutes of infusion of Hcy. This response was (nϭ7) in chronic HHcy. *PϽ0.05, **PϽ0.01 for difference from restored by simultaneous infusion of ascorbic acid. ‘ver indi- control. cates the point of veratrine injection; bar ϭ10 seconds. B, Sum- mary data for the veratrine-induced, NO-dependent, coronary vasodilation in the acute HHcy. Values are meanϮSEM, versus cardiac triglyceride with 5 paired samples *PϽ0.05 for difference from control (nϭ7). (yϭ1.19ϫϪ1.77, rϭ0.72, Pϭ0.16). Thus there is only a weak and variable correlation between the reduction in between Hcy concentration and substrate uptake with chronic plasma FFA and cardiac triglyceride, perhaps in part because HHcy. In chronic HHcy, both glucose uptake (2.91Ϯ0.48 to of the small sample size and the brief duration of methionine- 9.49Ϯ0.96 ␮mol/min, PϽ0.01, nϭ7) and lactate uptake enhanced diet. (3.91Ϯ2.34 to 16.02Ϯ4.10 ␮mol/min, PϽ0.05, nϭ7) were increased (PϽ0.05, nϭ7 each). On the other hand, FFA Calculation of MVO2, ATP Equivalent, and uptake was decreased (7.67Ϯ0.77 to 1.88Ϯ0.96 ␮mol/min, Triple Product PϽ0.01, nϭ5). Arterial concentrations of each substrate at During chronic HHcy, there was no change in the triple Ϯ Ϯ control were glucose; 4.7Ϯ0.1 ␮mol/L, lactate; 1.1Ϯ0.2 product (416 45 to 469 62 U), no change in MVO2 Ϯ Ϯ ␮mol/L, FFA; 0.71Ϯ0.04 ␮mol/L, and none of these changed (3.32 0.28 to 3.30 0.38 mL/min), and no change in esti- Ϯ Ϯ ␮ in HHcy. To determine the significance of reduced FFA mated ATP concentration (1354 124 to 834 201 mol), as uptake, we measured cardiac triglyceride in a control group well as no change in the arteriovenous coronary NOx differ- Ϯ Ϫ Ϯ ␮ and a group fed methionine for 2 weeks. Triglyceride levels ence (0.75 0.64 to 0.11 0.24 mol/min). Thus the switch were 5.0Ϯ1.1 ␮mol/g wet wt in normal dog heart and in substrate did not limit cardiac function. 4.0Ϯ0.7 ␮mol/g wet wt in dogs fed methionine (PϽ0.2). We also performed linear regression analysis of plasma FFA Effects of Hcy on MVO2 in Tissue Cumulative doses of BK caused concentration-dependent

decreases in MVO2 in control heart (Ϫ39.7Ϯ1.4%, nϭ6).

BK-induced reduction in MVO2 was significantly attenuated by HHcy (Ϫ19.6Ϯ2.5%, PϽ0.01, nϭ6). The inhibitory

effects of HHcy on MVO2 in response to BK were completely restored with co-incubation with ascorbic acid (Ϫ29.5Ϯ2.3%), tempol (Ϫ29.3Ϯ1.7%), or apocynin (Ϫ31.6Ϯ3.0%), respectively (PϽ0.05, nϭ6). In the same

manner, carbachol-induced reduction in MVO2 was also significantly decreased by HHcy (Ϫ19.7Ϯ2.1%, PϽ0.01, nϭ6) and was reversed by ascorbic acid (Ϫ33.0Ϯ1.3%), tempol (Ϫ32.9Ϯ2.1%) or apocynin (Ϫ29.0Ϯ3.6%), respec- tively (PϽ0.05, nϭ6). These data are summarized in Figure 4. In contrast to reduction induced by BK and carbachol,

S-nitroso-N-acetyl-penicillamine–induced reduction in MVO2 (Ϫ37.1Ϯ1.6%) was not affected by HHcy (Ϫ33.2Ϯ3.7%) Figure 2. Summary data for the veratrine-induced, (nϭ6). NO-dependent, coronary vasodilation during chronic HHcy. There was significant reduction in the coronary vasodilation Western Blot Analysis caused by veratrine, which was restored by infusion of either phox ascorbic acid or apocynin. Values are meanϮSEM, *PϽ0.05 for There was an upregulation of Nox2 (gp91 ) during chronic difference from control (nϭ8). HHcy accompanied by a reduction in eNOS and phospho-

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Suematsu et al Hyperhomocysteinemia, a Cardiac Metabolic Disease 259

Figure 6. Fold-change of representative genes that are differen- tially regulated (PϽ0.05) in chronic HHcy. A, Junctional sarco- plasmic reticulum; B, myosin heavy chain class II, DLA alpha chain; C, skeletal muscle myosin light chain 2 mRNA; D, phos- Figure 4. Changes in oxygen consumption in cardiac muscle phatidic acid phosphatase type 2B; E, pyruvate kinase M; F, segments. Bradykinin (A) and carbachol (B) were used to stimu- NADH dehydrogenase:ubiquinone Fe-S; G, GAPDH; H, cardiac ryanodine receptor2;I,slow myosin heavy chain (MYH7); J, late an endogeneous NO release. Co-incubation with ascorbic 2ϩ acid (10Ϫ3 mol/L), tempol (10Ϫ3 mol/L), or apocynin (10Ϫ4 mol/L) sarcoplasmic reticulum Ca -transport ATPase; K, myoglobin. restored the inhibitory effects of HHcy on oxygen consumption. Values are fold-change compared with their expression in the *PϽ0.05, **PϽ0.01 for difference from control (nϭ6 each). ⅙ control heart. indicates control; ⅷ, HHcy; , HHcyϩascorbic acid; }, HHcyϩapocynin; Ⅲ, HHcyϩtempol. these are shown in Figure 6 and Table 2. All genes that are significantly increased or decreased can be found in the eNOS (39%), SOD-1 and phospho-p47 (73%, Figure 5) online-only Data Supplement, Table I. protein expression. There was no change in p22phox (9%), p67phox (19%), SOD-2 (4%), or p47phox (6%). Thus these may Discussion also serve as loading controls. The major findings of the present study are as follows. (1) Both acute and chronic HHcy caused a shift of cardiac Comparison of Gene Expression During HHcy substrate uptake from FFA to carbohydrate. These alterations The microarray data revealed that 459 genes were differen- were not accompanied by any hemodynamic changes (work) Յ Ϯ tially expressed (P 0.05; fold change of at least 2). One or by alterations in calculated ATP availability or oxygen hundred twenty-seven genes were expressed at significantly consumption. (2) Both acute and chronic HHcy reduced greater levels in HHcy, and 332 were significantly reduced in veratrine-induced, NO-dependent, coronary vasodilation HHcy. Examples of these genes are shown in Figure 6 and (Bezold-Jarisch reflex). This effect was restored by infusion Table 2. As expected, a number of cytokines and immune of ascorbic acid or apocynin. (3) The ability of BK or response genes were regulated. These included serum amy- carbachol to reduce oxygen consumption in vitro was abro- loid A protein (12.0-fold), interleukin-8 (9.4-fold), selectin gated in chronic HHcy. This effect was restored by antioxi- (5.0-fold), tissue inhibitor of metalloproteinases (3.6-fold), dants. (4) Protein expression of Nox2 (gp91phox) was upregu- CD48 (2.4-fold), and macrophage inflammatory protein 3 lated, whereas it was downregulated for eNOS, phospho- (2.3-fold). Unexpectedly, a substantial number of cardiac and eNOS, and SOD-1, which indicates potential sources of metabolic genes were differentially regulated. Examples of superoxide anion. (5) There was a switch in cardiac gene expression toward a new cardiac phenotype. In the present study, the serum concentration of Hcy was elevated to 24.3Ϯ9.6 ␮mol/L, which is considered mild clinical HHcy. Under these experimental conditions, we did not see any changes in hemodynamics or LV function. Therefore, all significant changes we observed were caused by the direct effects of HHcy rather than by the altered cardiac work associated with hypertension or heart failure. This may be the result of the limited exposure to HHcy over 2 weeks and may be evidence of the beginning of a process that leads to altered cardiac dynamics. It should be noted that there was an alteration in control of the coronary circulation at this time because the response to veratrine was signifi- cantly reduced. Our data with the gene chips also indicated Figure 5. A, Representative Western blots of gp91phox, SOD-1, what is perhaps an early step that contributes to cardiac and eNOS from normal (lanes 1, 3, 5, and 7) and HHcy (lanes 2, dysfunction. Be that as it may, the present is the first study to 4, 6, and 8) hearts. B, Summary data for the densitometric anal- ysis of gp91phox, SOD-1, and eNOS from HHcy. Data are indicate altered cardiac substrate uptake in HHcy, until this expressed as arbitrary units. time indicative of cardiac disease.

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 260 Circulation January 16, 2007

TABLE 2. Gene Expressions in Dogs With HHcy after short-term exercise training22 in proportion to the altered release of NO and upregulation or downregulation of eNOS. Affy Probe Set IDs Fold Change Gene Title The present results indicate that veratrine administration at a 1583409_s_at 10.1 complement component C3 dose of 5 ␮g/kg significantly increased CBF (76%). Infusion 1586890_at 2.68 mannosidase, alpha, class 1A, member 1 of L-Hcy decreased this response by 42%, and the response 1600896_at 2.48 cytochrome P450, family 4, subfamily A, was completely reversed by infusion of ascorbic acid. Simi- polypeptide 11 larly, after 2 weeks of a diet high in methionine, there was a 1601853_at 2.13 5-lipoxygenase activating protein 32% reduction of this response, which was completely 1603487_s_at Ϫ2.19 chloride channel 3 restored by the infusion of either ascorbic acid or apocynin. 1585912_at Ϫ2.27 NADH-ubiquinone oxidoreductase 23-kDa Thus, there appear to be both short-term and longer-term subunit effects of HHcy mediated in part by changes in protein and 1583828_at Ϫ2.27 phosphoglycerate mutase 1 gene expression. Recent studies in animals and humans have 1583300_at Ϫ2.3 UDP-glucose pyrophosphorylase 2 isoform a demonstrated that mild HHcy induces endothelial dysfunc- 7,16,17,27,28 6 1585886_at Ϫ2.31 adenylate cyclase tion. Flow-mediated vasodilation as well as vaso- dilation induced by acetylcholine is impaired in patients with 1586746_at Ϫ2.35 thioredoxin HHcy,29 which indicates alterations in NO-mediated re- Ϫ 1601385_at 2.57 kynurenine–oxoglutarate transaminase I sponses. Hcy may decrease NO bioavailability through the 1604432_at Ϫ2.71 prenylcysteine oxidase 1 formation of disulfides and the generation of hydrogen 1597496_s_at Ϫ2.95 isocitrate dehydrogenase 2 (NADPϩ) peroxide30 and superoxide anion.31 Therefore, the decrease in 1583576_at Ϫ3.07 prostaglandin D2 synthase 21kDa the coronary vasodilation induced by the activation of the 1600018_at Ϫ3.16 plasma glutathione peroxidase precursor Bezold-Jarisch reflex in HHcy is thought to be caused by (GSHPx-P) impaired NO bioavailability through oxidative stress. Indeed, 1586610_at Ϫ3.18 2-oxoisovalerate dehydrogenase beta this reduction of the reflex recovered completely with simul- subunit taneous infusion of ascorbic acid or apocynin in both acute 1583490_x_at Ϫ3.41 NADH-ubiquinone oxidoreductase 23-kDa and chronic HHcy. subunit We have previously reported the potential role of NO in the 1603260_at Ϫ3.48 titin control of myocardial metabolism with both in vivo and in 32 1606045_at Ϫ3.48 transaldolase 1 vitro models. Recchia et al found a significant increase in 1588525_at Ϫ3.56 hydroxysteroid dehydrogenase like 2 myocardial glucose uptake and decrease in FFA uptake associated with a fall in cardiac NO production in conscious 1592430_at Ϫ3.66 methionine adenosyltransferase II, alpha dogs with pacing-induced heart failure. Moreover, this shift Ϫ 1603583_s_at 4 glycogen phosphorylase in substrate use was also observed in normal dogs with 1594663_at Ϫ4.53 Rab12 protein short-term pharmacological inhibition of NO synthase with 1596903_at Ϫ4.53 cardiomyopathy-associated 5 nitro-L-arginine, which was completely reversed by a NO 1583122_at Ϫ4.56 C-myc binding protein Mm-1 donor.23 Tada et al15 have shown that glucose uptake is 1590675_at Ϫ4.63 NAD(P) transhydrogenase elevated in Langendorff-perfused hearts from eNOS knock- 1586101_at Ϫ4.66 ATP synthase, Hϩ transporting, out mice, a condition that was mimicked by NOS inhibition in mitochondrial F0 complex hearts from wild-type mice. Furthermore, this increase in 1596406_at Ϫ6.15 mitogen-activated protein kinase kinase glucose uptake was normalized by 8-Br-cGMP or S-nitroso- kinase 15 N-acetyl-penicillamine, which indicates that cardiac NO reg- 1590259_at Ϫ6.59 thioredoxin interacting protein ulates myocardial glucose uptake via a cGMP-dependent 1605998_at Ϫ6.96 myosin, heavy polypeptide 6, cardiac mechanism. These previous reports suggest that NO plays an muscle, alpha important role in the regulation of myocardial substrate use. 1587642_s_at Ϫ7.21 ATPase, Caϩϩ transporting, cardiac Our current study shows that HHcy causes a change in muscle, slow twitch 2 cardiac energy substrate utilization in the absence of any 1586083_x_at Ϫ7.36 NADH dehydrogenase (ubiquinone) 1 hemodynamic changes. Taking these findings into account, we conclude that HHcy directly alters cardiac metabolism by 1604159_at Ϫ7.78 ubiquinol-cytochrome-c reductase complex core protein 2 impairing NO bioavailability at least partly through oxidative stress. List of genes that are upregulated or downregulated in chronic HHcy. Affy probe set IDs refers to the Affymetrix probe set identifier. In a normal metabolic state, FFA is the major substrate for the heart and provides the highest yield of ATP compared In 1867, von Bezold and Hirt observed that an intravenous with glucose or lactate. On the other hand, downregulation of injection of veratrum alkaloids caused a decrease in blood myocardial FFA oxidization and accelerated glucose oxidiza- pressure and heart rate, which was then named the Bezold- tion has been reported in a number of models of heart failure, Jarisch reflex.25 Activation of this reflex also causes coronary such as pacing-induced heart failure in dogs,24,32 myocardial vasodilation. Our recent studies have indicated that the infarction in rats,33 and also in idiopathic dilated cardiomy- coronary vasodilation induced by activation of the Bezold- opathy in humans.34,35 This substrate switching may be Jarisch reflex is entirely NO-dependent and is selectively considered either to be a marker of heart disease or to be a impaired after pacing-induced heart failure21 and enhanced contributor to the development and progression of heart

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Suematsu et al Hyperhomocysteinemia, a Cardiac Metabolic Disease 261

disease. In terms of evidence of heart disease, we did not find atherosclerosis or hypertension have not been investigated. any mechanical stress on the heart because of HHcy. Triple Our results indicate, however, that the progression of cardiac

product (work), MVO2, or calculated ATP did not change. or coronary disease associated with HHcy should be viewed This means that we may have observed the very early stage of in light of alterations in the regulation of cardiac metabolism, heart disease with HHcy. Indeed, we only examined HHcy substrate use, and gene expression. Indeed, a recent clinical for 2 weeks, even though HHcy in humans is a chronic, study37 has also shown that the treatment of HHcy by the

long-lasting, and gradually progressive condition. Microarray combination of folic acid and vitamins B6 and B12 did not data revealed that some of the genes that regulate cardiac reduce the risk of major cardiovascular events in patients with structure or contractile proteins were unexpectedly down- vascular disease, even though it significantly lowered Hcy regulated after methionine-enhanced diet for 2 weeks despite levels. Even though the plasma level of Hcy is low or the the fact that there was no change in hemodynamics. This duration of HHcy is short, the potential that HHcy has direct implies that HHcy exerts direct effects on the heart and causes effects on the heart and in that way may increase the risk of morphological changes that might contribute to the onset and future cardiovascular events should be taken into account the progression of heart disease. during the assessment of patients with HHcy. We found an increase in Nox2 (gp91phox) and reduction in eNOS, phospho-eNOS, SOD-1, and phospho-p47. Because phox phox Acknowledgments there was no change in other proteins, ie, p47 , p22 , We are thankful to Dr William Stanley, Case Western Reserve phox p67 , or SOD-2, it is unlikely that alterations in loading University, for the analysis of cardiac triglycerides. determined the results. This is especially true for p47phox and phospho-p47. Funding Sources Even though the precise mechanisms are still not fully This work was supported by NIH grants P01-HL-43023, R01-HL- elucidated, previous studies have proposed that HHcy causes 50142 and HL-63129 to Dr Hintze, P01-HL-74237 to Dr Recchia, endothelial dysfunction through oxidative stress and attenu- and by American Heart Association grant 0555897T to Dr Koller. ates NO bioavailability. Upchurch et al8 have shown that HHcy reduces glutathione peroxidase activity and decreases Disclosures NO bioavailability in cultured bovine aortic endothelial cells. None. Weiss et al36 have also demonstrated with glutathione perox- idase transgenic mice that overexpression of glutathione References 1. Cheng SW, Ting AC, Wong J. Fasting total plasma homocysteine and peroxidase can compensate for the adverse effects of HHcy atherosclerotic peripheral vascular disease. Ann Vasc Surg. 1997;11: on endothelial function, which indicates that these effects are 217–223. at least partly mediated by oxidative inactivation of NO. 2. den Heijer M, Koster T, Blom HJ, Bos GM, Briet E, Reitsma PH, Interestingly, 2 superoxide scavenging mechanisms, thiore- Vandenbroucke JP, Rosendaal FR. Hyperhomocysteinemia as a risk factor for deep-vein thrombosis. N Engl J Med. 1996;334:759–762. doxin and glutathione, are downregulated in chronic HHcy 3. Clarke R, Daly L, Robinson K, Naughten E, Cahalane S, Fowler B, (Table 2 and online Data Supplement, Table I). In this Graham I. Hyperhomocysteinemia: an independent risk factor for context, the present study indicates (1) that restoration of the vascular disease. N Engl J Med. 1991;324:1149–1155. coronary vasodilation to veratrine by apocynin and (2) 4. Parnetti L, Caso V, Santucci A, Corea F, Lanari A, Floridi A, Conte C, phox Bottiglieri T. Mild hyperhomocysteinemia is a risk-factor in all etio- upregulation of Nox2 (gp91 ) and downregulation of logical subtypes of stroke. Neurol Sci. 2004;25:13–17. SOD-1, which further implicates increased superoxide pro- 5. Nygard O, Nordrehaug JE, Refsum H, Ueland PM, Farstad M, Vollset duction by the NADPH oxidase. SE. Plasma homocysteine levels and mortality in patients with coronary One limitation of the present study is that we did not use artery disease. N Engl J Med. 1997;337:230–236. 6. Bellamy MF, McDowell IF, Ramsey MW, Brownlee M, Bones C, isotopic tracers, so we could not determine the metabolic fate Newcombe RG, Lewis MJ. Hyperhomocysteinemia after an oral of FFA, lactate, and glucose. The large changes in energy methionine load acutely impairs endothelial function in healthy adults. substrate uptake that we found, however, are strong indicators Circulation. 1998;98:1848–1852. of a marked alteration in cardiac metabolism. Moreover, it is 7. Chambers JC, McGregor A, Jean-Marie J, Obeid OA, Kooner JS. Dem- conceivable that, at the steady state, the rate of substrate onstration of rapid onset vascular endothelial dysfunction after hyperho- mocysteinemia: an effect reversible with vitamin C therapy. Circulation. uptake equals the rate of utilization for energy production. 1999;99:1156–1160. Rates of ATP production were calculated based on this 8. Upchurch GR Jr, Welch GN, Fabian AJ, Freedman JE, Johnson JL, assumption and should be considered as approximate values. Keaney JF Jr, Loscalzo J. Homocyst(e)ine decreases bioavailable nitric In conclusion, the present study is the first evidence of the oxide by a mechanism involving glutathione peroxidase. J Biol Chem. 1997;272:17012–17017. direct effect of HHcy on the heart (ie, the substrate switching 9. Ungvari Z, Csiszar A, Edwards JG, Kaminski PM, Wolin MS, Kaley G, from FFA to glucose) by impairing NO bioavailability Koller A. Increased superoxide production in coronary arteries in hyper- through oxidative stress, independent of changes in cardiac homocysteinemia: role of tumor necrosis factor-alpha, NAD(P)H oxidase, function. These alterations could be seen after only 2 weeks and inducible nitric oxide synthase. Arterioscler Thromb Vasc Biol. 2003;23:418–424. of methionine-enhanced diet or even during Hcy infusion at a 10. Shen W, Xu X, Ochoa M, Zhao G, Wolin MS, Hintze TH. Role of nitric rate that results in a plasma concentration of 10.1Ϯ1.8 oxide in the regulation of oxygen consumption in conscious dogs. Circ ␮mol/L, which is lower than the defined clinical threshold for Res. 1994;75:1086–1095. HHcy. HHcy has until this time been considered a vascular 11. Xie YW, Shen W, Zhao G, Xu X, Wolin MS, Hintze TH. Role of endothelium-derived nitric oxide in the modulation of canine myocardial disease rather than a cardiac disease, and the direct effects of mitochondrial respiration in vitro. Implications for the development of HHcy on hearts independent of the secondary effects such as heart failure. Circ Res. 1996;79:381–387.

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 262 Circulation January 16, 2007

12. Stadler J, Billiar TR, Curran RD, Stuehr DJ, Ochoa JB, Simmons RL. endothelial stunning after acute pressure overload in the conscious dog is Effect of exogenous and endogenous nitric oxide on mitochondrial res- caused by oxidant processes: the role of angiotensin II type 1 receptor and piration of rat hepatocytes. Am J Physiol. 1991;260:C910–C916. NAD(P)H oxidase. Circulation. 2003;108:2934–2940. 13. Borutaite V, Brown GC. Rapid reduction of nitric oxide by mitochondria, 26. Ojaimi C, Li W, Kinugawa S, Post H, Csiszar A, Pacher P, Kaley G, and reversible inhibition of mitochondrial respiration by nitric oxide. Hintze TH. Transcriptional basis for exercise limitation in male eNOS- Biochem J. 1996;315:295–299. knockout mice with age: heart failure and the fetal phenotype. 14. Granger DL, Lehninger AL. Sites of inhibition of mitochondrial electron Am J Physiol Heart Circ Physiol. 2005;289:H1399–H1407. transport in macrophage-injured neoplastic cells. J Cell Biol. 1982;95: 27. Kanani PM, Sinkey CA, Browning RL, Allaman M, Knapp HR, Haynes 527–535. WG. Role of oxidant stress in endothelial dysfunction produced by 15. Tada H, Thompson CI, Recchia FA, Loke KE, Ochoa M, Smith CJ, experimental hyperhomocyst(e)inemia in humans. Circulation 1999;100: Shesely EG, Kaley G, Hintze TH. Myocardial glucose uptake is regulated 1161–1168. by nitric oxide via endothelial nitric oxide synthase in Langendorff mouse 28. Virdis A, Ghiadoni L, Cardinal H, Favilla S, Duranti P, Birindelli R, heart. Circ Res. 2000;86:270–274. Magagna A, Bernini G, Salvetti G, Taddei S, Salvetti A. Mechanisms 16. Ungvari Z, Pacher P, Rischak K, Szollar L, Koller A. Dysfunction of responsible for endothelial dysfunction induced by fasting hyperhomo- nitric oxide mediation in isolated rat arterioles with methionine diet- cystinemia in normotensive subjects and patients with essential hyper- induced hyperhomocysteinemia. Arterioscler Thromb Vasc Biol. 1999; tension. J Am Coll Cardiol. 2001;38:1106–1115. 19:1899–1904. 29. Lentz SR, Sobey CG, Piegors DJ, Bhopatkar MY, Faraci FM, Malinow 17. Eberhardt RT, Forgione MA, Cap A, Leopold JA, Rudd MA, Trolliet M, MR, Heistad DD. Vascular dysfunction in monkeys with diet-induced Heydrick S, Stark R, Klings ES, Moldovan NI, Yaghoubi M, hyperhomocyst(e)inemia. J Clin Invest 1996;98:24–29. Goldschmidt-Clermont PJ, Farber HW, Cohen R, Loscalzo J. Endothelial 30. Starkebaum G, Harlan JM. Endothelial cell injury due to copper- dysfunction in a murine model of mild hyperhomocyst(e)inemia. J Clin catalyzed hydrogen peroxide generation from homocysteine. J Clin Invest 2000;106:483–491. Invest. 1986;77:1370–1376. 18. Dayal S, Arning E, Bottiglieri T, Boger RH, Sigmund CD, Faraci FM, 31. Lang D, Kredan MB, Moat SJ, Hussain SA, Powell CA, Bellamy MF, Lentz SR. Cerebral vascular dysfunction mediated by superoxide in Powers HJ, Lewis MJ. Homocysteine-induced inhibition of endothelium- hyperhomocysteinemic mice. Stroke. 2004;35:1957–1962. dependent relaxation in rabbit aorta: role for superoxide anions. Arte- 19. Schachinger V, Britten MB, Zeiher AM. Prognostic impact of coronary rioscler Thromb Vasc Biol. 2000;20:422–427. vasodilator dysfunction on adverse long-term outcome of coronary heart 32. Recchia FA, McConnell PI, Bernstein RD, Vogel TR, Xu X, Hintze TH. disease. Circulation. 2000;101:1899–1906. Reduced nitric oxide production and altered myocardial metabolism 20. Becker JS, Adler A, Schneeberger A, Huang H, Wang Z, Walsh E, Koller during the decompensation of pacing-induced heart failure in the con- A, Hintze TH. Hyperhomocysteinemia, a cardiac metabolic disease, role scious dog. Circ Res. 1998;83:969–979. of nitric oxide and the p22 phox subunit of NADPH oxidase. Circulation. 33. Remondino A, Rosenblatt-Velin N, Montessuit C, Tardy I, Papageorgiou 2005;111:2112–2118. I, Dorsaz PA, Jorge-Costa M, Lerch R. Altered expression of proteins of 21. Zhao G, Shen W, Xu X, Ochoa M, Bernstein R, Hintze TH. Selective metabolic regulation during remodeling of the left ventricle after myo- impairment of vagally mediated, nitric oxide-dependent coronary vaso- cardial infarction. J Mol Cell Cardiol. 2000;32:2025–2034. dilation in conscious dogs after pacing-induced heart failure. Circulation. 34. Sack MN, Rader TA, Park S, Bastin J, McCune SA, Kelly DP. Fatty acid 1995;91:2655–2663. oxidation enzyme gene expression is downregulated in the failing heart. 22. Zhao G, Zhang X, Xu X, Ochoa M, Hintze TH. Short-term exercise Circulation. 1996;94:2837–2842. training enhances reflex cholinergic nitric oxide-dependent coronary 35. Davila-Roman VG, Vedala G, Herrero P, de las Fuentes L, Rogers JG, vasodilation in conscious dogs. Circ Res. 1997;80:868–876. Kelly DP, Gropler RJ. Altered myocardial fatty acid and glucose metab- 23. Recchia FA, McConnell PI, Loke KE, Xu X, Ochoa M, Hintze TH. Nitric olism in idiopathic dilated cardiomyopathy. J Am Coll Cardiol. 2002;40: oxide controls cardiac substrate utilization in the conscious dog. Car- 271–277. diovasc Res. 1999;44:325–332. 36. Weiss N, Zhang YY, Heydrick S, Bierl C, Loscalzo J. Overexpression of 24. Osorio JC, Stanley WC, Linke A, Castellari M, Diep QN, Panchal AR, cellular glutathione peroxidase rescues homocyst(e)ine-induced endothe- Hintze TH, Lopaschuk GD, Recchia FA. Impaired myocardial fatty acid lial dysfunction. Proc Natl Acad Sci U S A. 2001;98:12503–12508. oxidation and reduced protein expression of retinoid X receptor-alpha in 37. Lonn E, Yusuf S, Arnold MJ, Sheridan P, Pogue J, Micks M, McQueen pacing-induced heart failure. Circulation. 2002;106:606–612. MJ, Probstfield J, Fodor G, Held C, Genest J Jr. Homocysteine lowering 25. Kinugawa S, Post H, Kaminski PM, Zhang X, Xu X, Huang H, Recchia with folic acid and B vitamins in vascular disease. N Engl J Med. FA, Ochoa M, Wolin MS, Kaley G, Hintze TH. Coronary microvascular 2006;354:1567–1577.

CLINICAL PERSPECTIVE Our studies indicate that hyperhomocysteinemia, which was once considered a vascular disease, should now also be considered a cardiac disease. As such, it is important to recognize the potential involvement of alterations in the control of cardiac metabolism to this disease process. The present study supports the conclusion that the scavenging of NO by superoxide anion in hyperhomocysteinemia results in a switch of cardiac substrate use from fatty acids to lactate and glucose and dysregulation of myocardial oxygen consumption. Furthermore, after just 2 weeks of methionine-enhanced diet, gene chip analysis indicates changes in cardiac gene expression, some of which are quite surprising. Thus, our present study draws attention to the fact that clinicians should consider hyperhomocysteinemia a cardiac metabolic disease as well as a vascular disease.

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 CONTROVERSIES IN CARDIOVASCULAR MEDICINE Is renal artery stenting the correct treatment of renal artery stenosis?

The Case for Renal Artery Stenting for Treatment of Renal Artery Stenosis Christopher J. Cooper, MD; Timothy P. Murphy, MD

he case favoring renal artery stenting for individuals epinephrine, and is implicated in end-organ damage in the with renal artery hypertension is largely circumstan- heart3 and kidney.4 RAS is suggested to cause 2 types of Ttial. At best, the clinical evidence presented in this hypertension. With unilateral RAS and a normally perfused discussion is derived primarily from nonrandomized cohort and normally functioning contralateral kidney, blood pressure studies. It would certainly be easier to argue that medical elevation is referred to as “renin dependent” and is charac- therapy is preferred for such individuals because there are 3 terized by increased peripheral resistance.5,6 In this circum- published randomized clinical trials that concluded just that stance, renin and angiotensin levels remain elevated, but and none that support renal artery intervention. Nonetheless, volume expansion is limited by natriuresis of the contralateral there is considerable evidence to support the role for revas- normally functioning kidney.6 Importantly, although renin cularization in general, and stenting specifically, as an im- levels are elevated, the value of peripheral or even renal vein portant adjunctive therapy to medical therapy in the care of renin values is limited by substantial overlap with patients patients with renal artery stenosis (RAS). The argument has 3 having essential hypertension.7,8 principal components: observations about the impact on When stenoses are bilateral or when the contralateral cardiovascular physiology, end-organ effects, and natural kidney is absent or is dysfunctional, intravascular volume history. increases and renin secretion decreases over a period of 5 to Response by Dworkin and Jamerson p 270 10 days.6,9–13 Without the natriuretic effect of a normally perfused contralateral kidney, hypertension is maintained by RAS and Hypertension volume expansion. This scenario is described as 1-kidney RAS is associated with and is an important cause of secon- 1-clip Goldblatt hypertension. dary hypertension. RAS causes endocrine activation with In addition, the sympathetic and central nervous systems release of renin from renal juxtaglomerular cells. Renin contribute to hypertension in RAS.14–16 RAS also causes catalyzes the breakdown of angiotensinogen to angiotensin I. increased production of vasoactive reactive oxygen species, Angiotensin I is transformed by angiotensin-converting en- which increases vasomotor tone.17 Ischemic injury of the zyme into angiotensin II, and angiotensin II promotes the affected kidney(s), hypertensive nephrosclerosis of the con- release of aldosterone from the adrenal cortex.1 Angiotensin tralateral kidney, or hypertrophy of the peripheral vasculature II is a potent vasoconstrictor,2 substantially more potent than also may be important contributors to sustained hypertension

The opinions expressed in this article are not necessarily those of the editors or of the American Heart Association. From the Department of Medicine, University of Toledo, Toledo (C.J.C.), and Department of Diagnostic Imaging, Brown University, Vascular Disease Research Center, Rhode Island Hospital, Providence (T.P.M.). Correspondence to Christopher J. Cooper, MD, Department of Medicine, University of Toledo, 3000 Arlington Ave, Hospital Room No. 1192, Toledo, OH 43614–2598. E-mail [email protected] (Circulation. 2007;115:263-270.) © 2007 American Heart Association, Inc. Circulation is available at http://www.circulationaha.org DOI: 10.1161/CIRCULATIONAHA.106.619015

Downloaded from circ.ahajournals.org 263at Mohammed Mahboob on June 10, 2007 264 Circulation January 16, 2007

with RAS.6 Abnormal endothelium-dependent vasodilation required for kidney tissue metabolism.35 Thus, chronic ische- also appears to be found and may be associated with mia does not damage renal tissue simply by lack of oxygen increased oxidative stress.17 Recent work suggests that pri- delivery.36 Chronic renal failure can be the result of severe mary hyperaldosteronism may contribute to the persistence of global ischemia, but nephrosclerosis also occurs in the hypertension in a significant minority of patients who appear nonstenotic kidney, perhaps mediated by hypertension, a to be unresponsive to revascularization.18 vasculotoxic effect of renin,37,38 or by angiotensin II through Thus, it is clear that hypertension is an important conse- its interaction with endothelin-1, platelet-derived growth quence of RAS. Antihypertensive medical therapy may ad- factor-␤, and transforming growth factor-␤.4 Gobe et al39 dress one or several of these mechanisms, but revasculariza- studied the cellular events related to unilateral RAS when the tion with endovascular stenting may be an efficient kidney underwent progressive atrophy. During the initial mechanism to downregulate many of these changes.15,17 As phase (2 to 5 days), tubular cell death resulted from both an example, successful renal artery revascularization may necrosis and apoptosis. During the later phase (10 to 20 days), decrease muscle sympathetic nerve activity,15 as does dener- renal atrophy progressed, and cell death resulted from apo- vation of the affected kidney,19 and may improve endothelial ptosis alone. After reversal of RAS, evidence of regeneration, function.17,20 consisting of hypertrophy and hyperplasia, was found. These findings imply that chronic ischemia is a dynamic process RAS and Kidney Function comprising not only an adaptation to reduced blood flow but Although end-stage renal disease is uncommon in the short also a potential for tubular cell regeneration. term,21 historical data suggest that up to 27% of those with RAS develop chronic renal failure within 6 years.22 Impor- RAS and Cardiovascular Events tantly, several investigators have reported that renal artery In patients with RAS, the risk of cardiovascular morbidity revascularization is associated with stabilization or improve- and mortality appears to be substantial. Wollenweber et al22 ment in renal function.23–26 described a 6-year cardiovascular event–free survival of 53%, The adverse effect of ischemia on the kidney is well with risk related to the severity of the renal stenosis. Several established experimentally and clinically. The role of athero- others have suggested that the risk of adverse cardiovascular sclerotic RAS in the genesis of renal dysfunction is contro- events is high and occurs in excess of the severity of versial, however. Over 1 to 4 years, atherosclerotic stenoses hypertension.40–42 More recently, a significant decrease in often progress, some to occlusion.22,27–29 RAS is associated 4-year survival was seen in patients with incidental RAS with loss of renal size, a reasonable but crude measure of undergoing coronary angiography.43 Thus, the risk of cardio- renal function.29 In patients with significant (Ͼ60%) RAS, 1 vascular events appears to be high in RAS, and blood of 4 ipsilateral kidneys demonstrates atrophy of Ͼ1cmin pressure control may be a poor surrogate for clinical length,30,31 whereas significant loss of renal size is uncom- outcomes. mon without RAS.30 Occlusion is associated with substantial It is not known whether the high cardiovascular event rate loss of renal size and function.32 seen in patients with atherosclerotic RAS is attributable to the Several investigators have been unable to demonstrate a effects of renal ischemia and subsequent neuroendocrine relationship between stenosis severity and renal function.33,34 activation or is simply a marker for advanced atherosclerosis Although one might conclude from this finding that stenoses and cardiovascular risk. A biologically plausible link is do not cause renal dysfunction, such a conclusion would be present, however, between renal ischemia and subsequent unfounded. Fundamentally, the kidney requires blood flow to events that may be independent of blood pressure per se. function. Clearly, there are factors beyond percent stenosis Angiotensin II is implicated in smooth muscle proliferation, that influence function. Some are intrinsic to RAS, including plaque rupture, endothelial dysfunction, and inhibition of the duration of the insult, atheroemboli, hypertensive neph- fibrinolysis.3 Angiotensin II also promotes medial and cardiac rosclerosis of the contralateral kidney, activation of the myocyte hypertrophy.6,44–51 It is important to note that renin-angiotensin system, and effect of the stenosis (including myocardial hypertrophy occurs when angiotensin II is present lesion length, minimal lumen diameter, etc) on renal blood even when blood pressure is controlled.52 Angiotensin II flow and intrarenal pressure. Additionally, other factors such interacts with other peptides like endothelin, transforming as essential hypertension, diabetes, concomitant medications, growth factor-␤, and platelet-derived growth factor-␤, each generalized atherosclerosis progression, and aging play a role of which is implicated in end-organ damage, ventricular in determining overall renal function. hypertrophy, and vascular hypertrophy.4,51–53 Excess aldoste- Most research has ascribed a direct role of the renin- rone has been related to extracellular matrix and collagen angiotensin system in ischemic nephropathy. Angiotensin II deposition and therefore to myocardial fibrosis.54 induces efferent arteriolar constriction, which aids in the Although the mechanism(s) responsible for the relationship maintenance of glomerular filtration rate. Kidneys remain between RAS and congestive heart failure (CHF) are not well viable with blood flows and pressures below that required for characterized, there is little doubt that ventricular hypertro- glomerular filtration because Ͻ10% of oxygen delivery is phy, sustained hypertension, activation of the renin-angioten-

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Cooper and Murphy Medical Therapy of RAS 265

sin and sympathetic nervous systems, and volume retention process with statins, antiplatelet therapy, antihypertensive associated with renal ischemia are likely important contribu- therapies, and hypoglycemic agents in the presence of diabe- tors. Recent work suggests a relationship between brain tes. Inhibition of sympathetic activation and the renin-angio- natriuretic peptide and RAS, with high brain natriuretic tensin-aldosterone system with ␤-blockers and angiotensin- peptide levels (Ͼ80 pg/mL) associated with better outcomes converting enzyme inhibitors is a very successful strategy for after renal artery intervention.55 Angiotensin II, endothelin, or patients with cardiovascular diseases in several settings, an insulin-like growth factor, and sympathetic activation all including acute myocardial infarction, CHF, and hyperten- may be involved in the process of ventricular thickening and sion. In each of these settings, treatment has demonstrated stiffening.46–49,53 Angiotensin has been implicated in the significant benefits as measured by reductions in clinically mechanism of cardiac hypertrophy and is known to induce important events. Renin-angiotensin-aldosterone system– protein synthesis in myocardial cells,50 even when the mean inhibiting drugs also are indicated for retarding the progres- arterial pressure is lowered.52 Endothelin also has been 75,76 implicated in cardiac hypertrophy in RAS.51 Interestingly, the sion to end-stage renal disease to preserve renal function. use of endothelin receptor blockers reverses left ventricular Although there are no outcome data with antihypertensive hypertrophy in rats with RAS.48,51 medical therapy for treatment of RAS specifically, clinical Remodeling of the left ventricular wall and peripheral studies suggest that renin-angiotensin-aldosterone system in- arteries in response to renovascular hypertension has been hibition decreases cardiovascular events in other high-risk noted. In patients with RAS with hypertension, cardiac patient groups.77,78 remodeling permits maintenance of normal cardiac function Antihypertensive medical therapies can be associated with despite increased left ventricular wall stress resulting from adverse health effects, however, such as lack of energy or systemic hypertension.56 Ventricular performance, deter- chronic cough, which may affect quality of life and reduce mined by afterload, chamber size, mass index, and functional compliance.79–81 Fifty percent to 70% of such medications shortening or contractility, often is abnormal in people with are discontinued or changed within 6 months.82 Adherence to renovascular hypertension.5,57 published guidelines for treatment of hypertension is poor, Congestive heart failure is common in patients undergoing often with significant adverse impact on patients and the 58–60 renal artery revascularization, just as RAS is common in healthcare system.83,84 Other factors that may limit the effec- patients presenting with CHF.61 Left ventricular hypertrophy tiveness of medical therapy include temporal variations in and decreased contractility, risk factors for development of drug levels and variable absorption.85,86 Finally, as described overt CHF,62 are significantly more common in those with above, renal ischemia increases activity in a number of other renovascular hypertension than essential hypertension,44,63 even when matched for age and gender.5 RAS is prevalent in pathogenic systems that may have deleterious cardiovascular those with CHF and may be implicated in its cause or severity or renal effects. Whether medical therapy alone is sufficient in a large proportion of patients. In a series of patients Ͼ70 to minimize risk in patients with atherosclerotic RAS is not years of age presenting with New York Heart Association known. class II to IV heart failure, 34% were found to have stenosis of at least 50% involving at least 1 renal artery.61 Intervention Cohort Studies As discussed, left ventricular hypertrophy is an important An important feature of revascularization is the ability to risk factor for cardiovascular events and cardiovascular death obviate the root cause of the entire process: the stenotic renal in both the presence and absence of hypertension.64–71 Re- artery. Importantly, others have demonstrated that relief of gression of left ventricular hypertrophy is favorable prognos- stenosis with revascularization results in a reduction in tically,69 and left ventricular mass index, a measure of renin-angiotensin-aldosterone system and sympathetic activ- ventricular hypertrophy, has been shown to decrease after ity.15 What is entirely unclear is whether pharmacological renal artery revascularization.72 When CHF accompanies therapy directed at one or several of these pathways is equally RAS, renal artery revascularization often is associated with effective or whether a reduction in activation attributable to 60,73 symptomatic improvement. Currently, this may be under- revascularization leads to improvements in long-term clinical recognized in clinical practice. There certainly appears to be outcome. some confusion about the management of patients with RAS In 1974, Hunt et al87 described lower rates of mortality, and CHF; the recent American Heart Association/American stroke, myocardial infarction, and azotemia and better blood College of Cardiology consensus document on the manage- pressure control in surgically revascularized patients than in a ment of peripheral vascular disease74 suggests limited support for screening patients with CHF but substantial support for comparison medical group despite more severe hypertension treatment of RAS in patients with CHF, despite a paucity of in the revascularized subjects. Interestingly, in a more con- evidence to support either position. temporary series comparing nonrandomized groups of pa- tients with RAS and hypertension treated with stents or Limitations of Medical Therapy for RAS medications, Pizzolo et al88 reported a survival treatment Undoubtedly, all patients that have RAS require lifelong effect in the intervention group similar to the experience of medical therapy to manage the underlying atherosclerotic Hunt et al 30 years previously, with an 87% probability of

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 266 Circulation January 16, 2007

5-year survival in the intervention group versus 67% in the benefits, then screening programs for high-, moderate-, or medical management group. potentially even low-risk individuals become advisable. In A number of single-center registries of stenting89 and a contrast, if the benefit is small or marginal, then screening recent multicenter registry of stenting for failed balloon and treatment may be reserved for those at high risk. angioplasty90 suggest substantial benefits related primarily to Furthermore, identifying subsets that derive greater or lesser improved blood pressure control. In contrast, the 3 multi- benefit or even harm becomes possible when the relative center trials of percutaneous balloon renal angioplasty benefits and risks can be identified. (PTRA) without stenting failed to demonstrate a blood Undoubtedly, some patients with atherosclerotic RAS ap- pressure benefit.91–93 Importantly, the randomized clinical pear to have a very favorable response to revascularization as trials done thus far have substantial weaknesses that have determined by improvement in blood pressure or kidney been the subject of considerable debate and discussion. The function. Previous investigators have suggested that the largest randomized trial comparing drug treatment and PTRA response to revascularization may be predicted by the pres- was the Dutch Renal Artery Stenosis Intervention Coopera- ence of microvascular dysfunction measured with the renal tive (DRASTIC) study. In this study, 106 patients were resistive index100 or by the presence of neurohumoral activa- randomly assigned to treatment by PTRA or medical therapy. tion.55 Fundamentally, though, each of these observations is The study design was such that refractory patients in the drug limited by the absence of a contrasting group of medically therapy group were allowed to receive balloon angioplasty if managed patients who are concurrent and carefully con- their blood pressure control was inadequate. With 44% of the trolled. Thus, all such observations suggest but are unable to drug therapy group receiving PTRA, the study contrasted establish the relative benefit of revascularization or the true provisional angioplasty against angioplasty. Despite the au- value of the baseline test strategy to predict outcome with thors’ assertion that PTRA in addition to drug therapy differing treatments. This limitation, the absence of a robust provided “little benefit,” patients in the PTRA group were and appropriate control group, has led to false conclusions, less likely to have deterioration of their blood pressure later disproved with appropriate randomized trials, about the control or renal artery occlusion during 12 months of follow- utility of a wide variety of clinical therapies, including up. Two other studies of balloon angioplasty versus medical surgery for Ménière’s disease,101 warfarin for acute myocar- therapy were confounded by similar issues and failed to dial infarction,102 portosystemic shunts,103 and antiarrhythmic demonstrate differences in blood pressure control.79,82 Impor- therapy after myocardial infarction.104 tantly, none of the studies comparing PTRA and drug therapy This is likely to be important for patients with renal artery thus far have included substantial numbers of patients who hypertension. In fact, in the Scottish and Newcastle series,92 received stents.93 Renal artery stent placement is standard 55 patients with renovascular hypertension were randomized today and substantially improves technical and long-term to angioplasty or medical therapy after a run-in period of 4 clinical outcomes compared with angioplasty alone,89,94,95 weeks. It was noted that significant decreases in blood especially for ostial stenoses,96 which make up 80% of pressure occurred before the initiation of treatment. Thus, the atherosclerotic stenoses.97 In a meta-analysis of 1322 pa- effect of revascularization on blood pressure, observed in tients, stent placement had a higher technical success rate and cohort studies, may be confounded by regression to the mean. a lower restenosis rate than did PTRA (98% versus 77% and 17% versus 26%, respectively; PϽ0.001) and a higher cure How Should Patients Be Managed? rate for hypertension.94 These issues have pragmatic implications for the various physicians who care for patients with ischemic renal disease. Limitations of Nonrandomized Evidence Many in the interventional community feel compelled to There is little controversy that stenting is an effective means revascularize renal stenoses, even when found incidentally. In of achieving durable revascularization of the renal artery. 2000, an estimated 19 800 renal angioplasties were per- Similarly, there is a wealth of uncontrolled study data formed for Medicare beneficiaries alone at an estimated cost demonstrating that some patients receive substantial benefits of $200 million.105 From 1996 to 2000, the annual number of in blood pressure control and renal function.89,94,98,99 How- renal stent procedures has increased 364%.105 ever, renal function worsens in some individuals after revas- In contrast, such enthusiasm is largely absent among cularization, and blood pressure may not improve. The nephrologists and others who routinely care for patients with current controversy is driven by a singular key factor: the lack chronic kidney diseases. Importantly, without widespread of an appropriate control group to allow the magnitude of the screening by the medical community, it is apparent that most benefit derived from stenting to be measured. Does stent patients with renovascular disease never come to attention. revascularization markedly reduce the risk of cardiovascular Recent work suggests that Ϸ7% of individuals Ͼ65 years of and renal events, does it modestly reduce events, or are the age have significant RAS.106 Clearly, the majority are never risks of stenting balanced or even overwhelmed by compli- identified within our communities, and far fewer are treated cations? This determination has fundamental implications in with either appropriate medical therapy or revascularization the decisions that physicians make. If stenting has substantial and medical therapy. This nihilistic approach of “no screen,

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Cooper and Murphy Medical Therapy of RAS 267

no treat” may be proximate to the occurrence of a number of 6. Korner PI. Cardiovascular hypertrophy and hypertension: causes and preventable cardiovascular and renal events. consequences. Blood Press. 1995;2:6–16. 7. Martin LG, Cork RD, Wells JO. Renal vein renin analysis: limitations of Is there an exit to this conundrum? Yes. The National Heart, its use in predicting benefit from percutaneous angioplasty. Cardiovasc Lung, and Blood Institute has funded a pivotal trial, Cardiovas- Intervent Radiol. 1993;16:76–80. cular Outcomes in Renal Atherosclerotic Lesions (CORAL), 8. Schreij G, van Es PN, Schiffers PM, Lavrijssen AT, de Leeuw PW. “Captopril test,” with blood pressure and peripheral renin as response that is currently randomizing 1080 patients to optimal medical variables in hypertensive patients with suspected renal artery stenosis. J therapy versus optimal medical therapy with stenting, with a Hum Hypertens. 1995;9:741–746. primary outcome consisting of a composite of attributable 9. Liard JF, Cowley AW Jr, McCaa RE, McCaa CS, Guyton AC. Renin, cardiovascular and renal events (www.coralclinicaltrial.org). At aldosterone, body fluid volumes, and the baroreceptor reflex in the development and reversal of Goldblatt hypertension in conscious dogs. the present time, however, examining the published data pro- Circ Res. 1974;34:549–560. vides little confidence that either intervention or medical man- 10. Ayers CR. Renin-angiotensin-aldosterone system in hypertension: the agement is superior for any given individual. clinical usefulness of measuring components of this system. Va Med Mon (1918). 1969;96:312–314. 11. Bianchi G, Campolo L, Vegeto A, Pietra V, Piazza U. The value of Conclusions plasma renin concentration per se, and in relation to plasma and extra- The case in favor of renal artery stenting is highly circum- cellular fluid volume in diagnosis and prognosis of human renovascular stantial but directionally consistent with important benefits of hypertension. Clin Sci. 1970;39:559–576. 12. Anderson WP, Korner PI, Angus JA, Johnston CI. Contribution of revascularization. Many of the physiological changes linked stenosis resistance to the rise in total peripheral resistance during exper- to adverse cardiovascular events may be improved with the imental renal hypertension in conscious dogs. Clin Sci (Lond). 1981;61: durable relief of renal ischemia provided by stent therapy, 663–670. 13. Schoenberg SO, Bock M, Kallinowski F, Just A. Correlation of hemo- including hypertension, sympathetic nervous system activa- dynamic impact and morphologic degree of renal artery stenosis in a tion, oxidative stress, and direct hormone-mediated end- canine model. J Am Soc Nephrol. 2000;11:2190–2198. organ damage. In contrast, medication regimens may not be 14. Johansson M, Elam M, Rundqvist B, Eisenhofer G, Herlitz H, Lambert G, Friberg P. Increased sympathetic nerve activity in renovascular as effective, and their use may be limited by side effects that hypertension. Circulation. 1999;99:2537–2542. influence compliance. Nonrandomized comparisons suggest 15. Miyajima E, Yamada Y, Yoshida Y, Matsukawa T, Shionoiri H, superior survival with revascularization compared with med- Tochikubo O, Ishii M. Muscle sympathetic nerve activity in reno- ical therapy.18,87 The published randomized trials that have vascular hypertension and primary aldosteronism. Hypertension. 1991; 17:1057–1062. compared blood-pressure lowering between patients who are 16. Mathias CJ, Kooner JS, Peart S. Neurogenic components of hyper- treated medically and those treated with renal artery angio- tension in human renal artery stenosis. Clin Exp Hypertens A. 1987;9: plasty have shown no benefit of balloon angioplasty, a 293–306. 17. Higashi Y, Sasaki S, Nakagawa K, Matsuura H, Oshima T, Chayama K. therapy of largely historical interest for the treatment of Endothelial function and oxidative stress in renovascular hypertension. atherosclerotic RAS. The current debate about the role of N Engl J Med. 2002;346:1954–1962. stent revascularization will undoubtedly continue until pow- 18. Pizzolo F, Pavan C, Guarini P, Trabetti E, Girelli D, Corrocher R, Olivieri O. Primary hyperaldosteronism: a frequent cause of residual erful and adequately controlled clinical trials are available hypertension after successful endovascular treatment of renal artery that place the absolute and relative benefits of revasculariza- disease. J Hypertens. 2005;23:2041–2047. tion within the context of meaningful clinical events. 19. Ye S, Ozgur B, Campese VM. Renal afferent impulses, the posterior hypothalamus, and hypertension in rats with chronic renal failure. Kidney Int. 1997;51:722–727. Disclosures 20. Converse RL Jr, Jacobsen TN, Jost CM, Toto RD, Grayburn PA, Drs Cooper and Murphy are supported by NIH U01HL071556 and Obregon TM, Fouad-Tarazi F, Victor RG. Paradoxical withdrawal of have grant funding from AstraZeneca, Pfizer, and Cordis corpora- reflex vasoconstriction as a cause of hemodialysis-induced hypotension. tions. Dr Cooper also has grant funding from Centocor Corp. J Clin Invest. 1992;90:1657–1665. 21. Chabova V, Schirger A, Stanson AW, McKusick MA, Textor SC. Outcomes of atherosclerotic renal artery stenosis managed without re- References vascularization. Mayo Clin Proc. 2000;75:437–444. 1. Goldblatt H, Lynch J, Hanzal RF, Summerville WW. Studies on exper- 22. Wollenweber J, Sheps SG, Davis GD. Clinical course of atherosclerotic imental hypertension: production of persistent elevation of systolic renovascular disease. Am J Cardiol. 1968;21:60–71. blood pressure by means of renal ischemia. J Exp Med. 1934;59: 23. La Batide-Alanore A, Azizi M, Froissart M, Raynaud A, Plouin PF. Split 347–379. Abstract. renal function outcome after renal angioplasty in patients with unilateral 2. Van Velzer DA, Burge CH, Morris GC Jr. Arteriosclerotic narrowing of renal artery stenosis. J Am Soc Nephrol. 2001;12:1235–1241. renal arteries associated with hypertension. Am J Roentgenol Radium 24. Watson PS, Hadjipetrou P, Cox SV, Piemonte TC, Eisenhauer AC. Ther Nucl Med. 1961;86:807–818. Effect of renal artery stenting on renal function and size in patients with 3. Lonn EM, Yusuf S, Jha P, Montague TJ, Teo KK, Benedict CR, Pitt B. atherosclerotic renovascular disease. Circulation. 2000;102:1671–1677. Emerging role of angiotensin-converting enzyme inhibitors in cardiac 25. Harden PN, MacLeod MJ, Rodger RS, Baxter GM, Connell JM, and vascular protection. Circulation. 1994;90:2056–2069. Dominiczak AF, Junor BJ, Briggs JD, Moss JG. Effect of renal-artery 4. Meyrier A. Vascular mechanisms of renal fibrosis: vasculone- stenting on progression of renovascular renal failure. Lancet. 1997;349: phropathies and arterial hypertension. Bull Acad Natl Med. 1999;183: 1133–1136. 33–45. 26. Dean RH, Kieffer RW, Smith BM, Oates JA, Nadeau JH, Hollifield JW, 5. Vensel LA, Devereux RB, Pickering TG, Herrold EM, Borer JS, Laragh DuPont WD. Renovascular hypertension: anatomic and renal function JH. Cardiac structure and function in renovascular hypertension changes during drug therapy. Arch Surg. 1981;116:1408–1415. produced by unilateral and bilateral renal artery stenosis. Am J Cardiol. 27. Meaney TF, Dustan HP, McCormack LJ. Natural history of renal arterial 1986;58:575–582. disease. Radiology. 1968;91:881–887.

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 268 Circulation January 16, 2007

28. Tollefson DF, Ernst CB. Natural history of atherosclerotic renal artery 51. Ehmke H, Faulhaber J, Munter K, Kirchengast M, Wiesner RJ. Chronic stenosis associated with aortic disease. J Vasc Surg. 1991;14:327–331. ETA receptor blockade attenuates cardiac hypertrophy independently of 29. Schreiber MJ, Pohl MA, Novick AC. The natural history of athero- blood pressure effects in renovascular hypertensive rats. Hypertension. sclerotic and fibrous renal artery disease. Urol Clin North Am. 1984;11: 1999;33:954–960. 383–392. 52. Phillips PA. Interaction between endothelin and angiotensin II. Clin Exp 30. Guzman RP, Zierler RE, Isaacson JA, Bergelin RO, Strandness DE Jr. Pharmacol Physiol. 1999;26:517–518. Renal atrophy and arterial stenosis: a prospective study with duplex 53. Iwanaga Y, Kihara Y, Inagaki K, Onozawa Y, Yoneda T, Kataoka K, ultrasound. Hypertension. 1994;23:346–350. Sasayama S. Differential effects of angiotensin II versus endothelin-1 31. Caps MT, Zierler RE, Polissar NL, Bergelin RO, Beach KW, inhibitions in hypertrophic left ventricular myocardium during transition Cantwell-Gab K, Casadei A, Davidson RC, Strandness DE Jr. Risk of to heart failure. Circulation. 2001;104:606–612. atrophy in kidneys with atherosclerotic renal artery stenosis. Kidney Int 54. Pessina AC, Sacchetto A, Rossi GP. Left ventricular anatomy and 1998;53:735–742. function in primary aldosteronism and renovascular hypertension. Adv 32. Zierler RE, Bergelin RO, Isaacson JA, Strandness DE Jr. Natural history Exp Med Biol. 1997;432:63–69. of atherosclerotic renal artery stenosis: a prospective study with duplex 55. Silva JA, Chan AW, White CJ, Collins TJ, Jenkins JS, Reilly JP, Ramee ultrasonography. J Vasc Surg. 1994;19:250–257. SR. Elevated brain natriuretic peptide predicts blood pressure response 33. Wright JR, Shurrab AE, Cheung C, Waldek S, O’Donoghue DJ, Foley after stent revascularization in patients with renal artery stenosis. Cir- RN, Mamtora H, Kalra PA. A prospective study of the determinants of culation. 2005;111:328–333. renal functional outcome and mortality in atherosclerotic renovascular 56. Broughton A, Korner PI. Left ventricular pump function in renal hyper- disease. Am J Kidney Dis. 2002;39:1153–1161. tensive dogs with cardiac hypertrophy. Am J Physiol. 1986;251: 34. Suresh M, Laboi P, Mamtora H, Kalra PA. Relationship of renal dys- H1260–H1266. function to proximal arterial disease severity in atherosclerotic reno- 57. Kannel WB. Role of blood pressure in cardiovascular disease: the vascular disease. Nephrol Dial Transplant. 2000;15:631–636. Framingham study. Angiology. 1975;26:1–14. 35. Morgan T, Wilson M, Johnston W, Clunie GJ, Gordon R. Restoration of 58. Meissner MD, Wilson AR, Jessup M. Renal artery stenosis in heart renal function by arterial surgery. Lancet. 1974;1:653–656. failure. Am J Cardiol. 1988;62:1307–1308. 36. Textor SC. Pathophysiology of renal failure in renovascular disease. 59. Olin JW. Atherosclerotic renovascular disease: diagnosis and man- Am J Kidney Dis. 1994;24:642–651. agement. Curr Opin Cardiol. 1990;5:659–665. 37. Volpe M, Camargo MJ, Mueller FB, Campbell WG Jr, Sealey JE, 60. Khosla S, White CJ, Collins TJ, Jenkins JS, Shaw D, Ramee SR. Effects Pecker MS, Sosa RE, Laragh JH. Relation of plasma renin to end organ of renal artery stent implantation in patients with renovascular hyper- damage and to protection of Kϩ feeding in stroke-prone hypertensive tension presenting with unstable angina or congestive heart failure. rats. Hypertension. 1990;15:318–326. Am J Cardiol. 1997;80:363–366. 38. Teunissen KE, Postma CT, van Jaarsveld BC, Derkx FH, Thien T. 61. MacDowall P, Kalra PA, O’Donoghue DJ, Waldek S, Mamtora H, Endothelin and active renin levels in essential hypertension and hyper- Brown K. Risk of morbidity from renovascular disease in elderly tension with renal artery stenosis before and after percutaneous trans- patients with congestive cardiac failure. Lancet. 1998;352:13–16. luminal renal angioplasty. J Hypertens. 1997;15:1791–1796. 62. MacMahon S, Peto R, Cutler J, Collins R, Sorlie P, Neaton J, Abbott R, 39. Gobe GC, Axelsen RA, Searle JW. Cellular events in experimental Godwin J, Dyer A, Stamler J. Blood pressure, stroke, and coronary heart unilateral ischemic renal atrophy and in regeneration after contralateral disease, part 1: prolonged differences in blood pressure: prospective nephrectomy. Lab Invest. 1990;63:770–779. observational studies corrected for the regression dilution bias. Lancet. 40. Sheps SG, Osmundson PJ, Hunt JC, Schirger A, Fairbairn JF 2nd. 1990;335:765–774. Hypertension and renal artery stenosis: serial observations on 54 patients 63. Sorensen MB, Bille Brahe NE. Preoperative haemodynamic evaluation treated medically. Clin Pharmacol Ther. 1965;6:700–709. of patients submitted for renal artery stenosis. Acta Anaesthesiol Scand. 41. Isles C, Main J, O’Connell J, Brown I, Findlay J, Stewart R, Wilkinson 1984;28:334–338. R. Survival associated with renovascular disease in Glasgow and New- 64. MacMahon S, Collins G, Rautaharju P, Cutler J, Neaton J, Prineas R, castle: a collaborative study. Scott Med J. 1990;35:70–73. Crow R, Stamler J. Electrocardiographic left ventricular hypertrophy 42. Valentine RJ, Myers SI, Miller GL, Lopez MA, Clagett GP. Detection and effects of antihypertensive drug therapy in hypertensive participants of unsuspected renal artery stenoses in patients with abdominal aortic in the Multiple Risk Factor Intervention Trial. Am J Cardiol. 1989;63: aneurysms: refined indications for preoperative aortography. Ann Vasc 202–210. Surg. 1993;7:220–224. 65. Culleton BF, Larson MG, Wilson PW, Evans JC, Parfrey PS, Levy D. 43. Conlon PJ, Little MA, Pieper K, Mark DB. Severity of renal vascular Cardiovascular disease and mortality in a community-based cohort with disease predicts mortality in patients undergoing coronary angiography. mild renal insufficiency. Kidney Int. 1999;56:2214–2219. Kidney Int. 2001;60:1490–1497. 66. Levy D, Garrison RJ, Savage DD, Kannel WB, Castelli WP. Prognostic 44. Losito A, Fagugli RM, Zampi I, Parente B, de Rango P, Giordano G, implications of echocardiographically determined left ventricular mass Cao P. Comparison of target organ damage in renovascular and essential in the Framingham Heart Study. N Engl J Med. 1990;322:1561–1566. hypertension. Am J Hypertens. 1996;9:1062–1067. 67. Kannel WB. Left ventricular hypertrophy as a risk factor: the Fra- 45. de Simone G, Devereux RB, Volpe M, Camargo MJ, Wallerson DC, mingham experience. J Hypertens Suppl. 1991;9:S3–S8. Laragh JH. Relation of left ventricular hypertrophy, afterload, and con- 68. Kannel WB, Cobb J. Left ventricular hypertrophy and mortality: results tractility to left ventricular performance in Goldblatt hypertension. from the Framingham study. Cardiology. 1992;81:291–298. Am J Hypertens. 1992;5:292–301. 69. Devereux RB, Okin PM, Roman MJ. Left ventricular hypertrophy as a 46. Gavras H, Lever AF, Brown JJ, Macadam RF, Robertson JI. Acute renal surrogate end-point in hypertension. Clin Exp Hypertens. 1999;21: failure, tubular necrosis, and myocardial infarction induced in the rabbit 583–593. by intravenous angiotensin II. Lancet. 1971;2:19–22. 70. Sullivan JM, Vander Zwaag RV, el-Zeky F, Ramanathan KB, Mirvis 47. Yamazaki T, Shiojima I, Komuro I, Nagai R, Yazaki Y. Involvement of DM. Left ventricular hypertrophy: effect on survival. J Am Coll Cardiol. the renin-angiotensin system in the development of left ventricular 1993;22:508–513. hypertrophy and dysfunction. J Hypertens. 1994;12:S153–S157. 71. Brown MJ, Castaigne A, de Leeuw PW, Mancia G, Palmer CR, 48. Hocher B, George I, Rebstock J, Bauch A, Schwarz A, Neumayer HH, Rosenthal T, Ruilope LM. Influence of diabetes and type of hyper- Bauer C. Endothelin system-dependent cardiac remodeling in reno- tension on response to antihypertensive treatment. Hypertension. 2000; vascular hypertension. Hypertension. 1999;33:816–822. 35:1038–1042. 49. Wahlander H, Isgaard J, Jennische E, Friberg P. Left ventricular 72. Symonides B, Chodakowska J, Januszewicz A, Lapinski M, insulin-like growth factor I increases in early renal hypertension. Hyper- Januszewicz M, Rowinski O, Szmidt J, Kuch-Wocial A, Kurzyna M, tension. 1992;19:25–32. Malek G, Berent H, Szmigielski C, Januszewicz W. Effects of the 50. Robertson AL Jr, Khairallah PA. Angiotensin II: rapid localization in correction of renal artery stenosis on blood pressure, renal function and nuclei of smooth and cardiac muscle. Science. 1971;172:1138–1139. left ventricular morphology. Blood Press. 1999;8:141–150.

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Cooper and Murphy Medical Therapy of RAS 269

73. Missouris CG, Belli AM, MacGregor GA. “Apparent” heart failure: a deletion polymorphism and endovascular revascularization. J Vasc Surg. syndrome caused by renal artery stenoses. Heart. 2000;83:152–155. 2004;39:140–147. 74. Hirsch AT, Haskal ZJ, Hertzer NR, Bakal CW, Creager MA, Halperin 90. Rees CR, Palmaz JC, Becker GJ, Ehrman KO, Richter GM, Noeldge G, JL, Hiratzka LF, Murphy WR, Olin JW, Puschett JB, Rosenfield KA, Katzen BT, Dake MD, Schwarten DE. Palmaz stent in atherosclerotic Sacks D, Stanley JC, Taylor LM Jr, White CJ, White J, White RA, stenoses involving the ostia of the renal arteries: preliminary report of a Antman EM, Smith SC Jr, Adams CD, Anderson JL, Faxon DP, Fuster multicenter study. Radiology. 1991;181:507–514. V, Gibbons RJ, Hunt SA, Jacobs AK, Nishimura R, Ornato JP, Page RL, 91. Plouin PF, Chatellier G, Darne B, Raynaud A. Blood pressure outcome Riegel B. ACC/AHA 2005 practice guidelines for the management of of angioplasty in atherosclerotic renal artery stenosis: a randomized trial: patients with peripheral arterial disease. Circulation. 2006;113: Essai Multicentrique Medicaments vs Angioplastie (EMMA) Study e463–e654. Group. Hypertension. 1998;31:823–829. 75. Brenner BM, Cooper ME, de Zeeuw D, Keane WF, Mitch WE, Parving 92. Webster J, Marshall F, Abdalla M, Dominiczak A, Edwards R, Isles CG, HH, Remuzzi G, Snapinn SM, Zhang Z, Shahinfar S. Effects of losartan Loose H, Main J, Padfield P, Russell IT, Walker B, Watson M, on renal and cardiovascular outcomes in patients with type 2 diabetes Wilkinson R. Randomised comparison of percutaneous angioplasty vs continued medical therapy for hypertensive patients with atheromatous and nephropathy. N Engl J Med. 2001;345:861–869. renal artery stenosis: Scottish and Newcastle Renal Artery Stenosis 76. Lewis EJ, Hunsicker LG, Clarke WR, Berl T, Pohl MA, Lewis JB, Ritz Collaborative Group. J Hum Hypertens. 1998;12:329–335. E, Atkins RC, Rohde R, Raz I. Renoprotective effect of the angioten- 93. van Jaarsveld BC, Krijnen P, Pieterman H, Derkx FH, Deinum J, Postma sin-receptor antagonist irbesartan in patients with nephropathy due to CT, Dees A, Woittiez AJ, Bartelink AK, Man in ’t Veld AJ, Schalekamp type 2 diabetes. N Engl J Med. 2001;345:851–860. MA. The effect of balloon angioplasty on hypertension in atherosclerotic 77. Dagenais GR, Yusuf S, Bourassa MG, Yi Q, Bosch J, Lonn EM, Kouz renal-artery stenosis: Dutch Renal Artery Stenosis Intervention Coop- S, Grover J. Effects of ramipril on coronary events in high-risk persons: erative Study Group. N Engl J Med. 2000;342:1007–1014. results of the Heart Outcomes Prevention Evaluation Study. Circulation. 94. Leertouwer TC, Gussenhoven EJ, Bosch JL, van Jaarsveld BC, van Dijk 2001;104:522–526. LC, Deinum J, Man In ’t Veld AJ. Stent placement for renal arterial 78. Dahlof B, Zanchetti A, Diez J, Nicholls MG, Yu CM, Barrios V, Aurup stenosis: where do we stand? A meta-analysis. Radiology. 2000;216: P, Smith RD, Johansson M. Effects of losartan and atenolol on left 78–85. ventricular mass and neurohormonal profile in patients with essential 95. Hennequin LM, Joffre FG, Rousseau HP, Aziza R, Tregant P, Bernadet hypertension and left ventricular hypertrophy. J Hypertens. 2002;20: P, Salvador M, Chamontin B. Renal artery stent placement: long-term 1855–1864. results with the Wallstent endoprosthesis. Radiology. 1994;191: 79. Curb JD, Borhani NO, Blaszkowski TP, Zimbaldi N, Fotiu S, Williams 713–719. W. Long-term surveillance for adverse effects of antihypertensive drugs. 96. Baumgartner I, von Aesch K, Do DD, Triller J, Birrer M, Mahler F. JAMA. 1985;253:3263–3268. Stent placement in ostial and nonostial atherosclerotic renal arterial 80. MRC trial of treatment of mild hypertension: principal results: Medical stenoses: a prospective follow-up study. Radiology. 2000;216:498–505. Research Council Working Party. BMJ (Clin Res Ed). 1985;291:97–104. 97. Kaatee R, Beek FJ, Verschuyl EJ, vd Ven PJ, Beutler JJ, van Schaik JP, 81. Stewart AL, Greenfield S, Hays RD, Wells K, Rogers WH, Berry SD, Mali WP. Atherosclerotic renal artery stenosis: ostial or truncal? McGlynn EA, Ware JE Jr. Functional status and well-being of patients Radiology. 1996;199:637–640. with chronic conditions: results from the Medical Outcomes Study. 98. Burket MW, Cooper CJ, Kennedy DJ, Brewster PS, Ansel GM, Moore JAMA. 1989;262:907–913. JA, Venkatesan J, Henrich WL. Renal artery angioplasty and stent 82. Oparil S, Calhoun DA. The calcium antagonists in the 1990s: an placement: predictors of a favorable outcome. Am Heart J. 2000;139: overview. Am J Hypertens. 1991;4:396S–405S. 64–71. 83. Siegel D, Lopez J, Meier J, Cunningham F. Changes in the pharma- 99. Harding MB, Smith LR, Himmelstein SI, Harrison K, Phillips HR, cologic treatment of hypertension in the Department of Veterans Affairs Schwab SJ, Hermiller JB, Davidson CJ, Bashore TM. Renal artery 1997–1999: decreased use of calcium antagonists and increased use of stenosis: prevalence and associated risk factors in patients undergoing beta-blockers and thiazide diuretics. Am J Hypertens. 2001;14:957–962. routine cardiac catheterization. J Am Soc Nephrol. 1992;2:1608–1616. 84. Okano GJ, Rascati KL, Wilson JP, Remund DD. A comparison of 100. Radermacher J, Chavan A, Bleck J, Vitzthum A, Stoess B, Gebel MJ, Galanski M, Koch KM, Haller H. Use of Doppler ultrasonography to antihypertensive medication utilization before and after guideline predict the outcome of therapy for renal-artery stenosis. N Engl J Med. changes using the Department of Defense Prescription Database. Ann 2001;344:410–417. Pharmacother. 1999;33:548–553. 101. Snow JB Jr, Kimmelman CP. Assessment of surgical procedures for 85. Campbell DJ. Endogenous angiotensin II levels and the mechanism of Meniere’s disease. Laryngoscope. 1979;89:737–747. action of angiotensin-converting enzyme inhibitors and angiotensin 102. Chalmers TC, Matta RJ, Smith H Jr, Kunzler AM. Evidence favoring the receptor type 1 antagonists. Clin Exp Pharmacol Physiol. 1996;3: use of anticoagulants in the hospital phase of acute myocardial S125–S131. infarction. N Engl J Med. 1977;297:1091–1096. 86. Wolny A, Clozel JP, Rein J, Mory P, Vogt P, Turino M, Kiowski W, 103. Grace ND, Muench H, Chalmers TC. The present status of shunts for Fischli W. Functional and biochemical analysis of angiotensin portal hypertension in cirrhosis. Gastroenterology. 1966;50:684–691. II-forming pathways in the human heart. Circ Res. 1997;80:219–227. 104. Preliminary report: effect of encainide and flecainide on mortality in a 87. Hunt JC, Sheps SG, Harrison EG Jr, Strong CG, Bernatz PE. Renal and randomized trial of arrhythmia suppression after myocardial infarction: renovascular hypertension: a reasoned approach to diagnosis and man- the Cardiac Arrhythmia Suppression Trial (CAST) Investigators. N Engl agement. Arch Intern Med. 1974;133:988–999. J Med. 1989;321:406–412. 88. Blum U, Krumme B, Flugel P, Gabelmann A, Lehnert T, Buitrago- 105. Murphy TP, Soares G, Kim M. Increase in utilization of percutaneous Tellez C, Schollmeyer P, Langer M. Treatment of ostial renal-artery renal artery interventions by Medicare beneficiaries, 1996–2000. AJR stenoses with vascular endoprostheses after unsuccessful balloon angio- Am J Roentgenol. 2004;183:561–568. plasty. N Engl J Med. 1997;336:459–465. 106. Hansen KJ, Edwards MS, Craven TE, Cherr GS, Jackson SA, Appel RG, 89. Pizzolo F, Mansueto G, Minniti S, Mazzi M, Trabetti E, Girelli D, Burke GL, Dean RH. Prevalence of renovascular disease in the elderly: Corrocher R, Olivieri O. Renovascular disease: effect of ACE gene a population-based study. J Vasc Surg. 2002;36:443–451.

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 270 Circulation January 16, 2007

Response to Cooper and Murphy Lance D. Dworkin, MD; Kenneth A. Jamerson, MD

Drs Cooper and Murphy admit that observational data are often misleading, citing several famous examples of therapies that initially appeared to be markedly beneficial but were found to be completely without value or even harmful when examined in randomized trials. Nevertheless, they go on to present several uncontrolled case series of patients with renal artery stenosis that appear to show a benefit of revascularization. These data should be viewed with skepticism and are at best hypothesis generating. The other main arguments in favor of revascularization are theoretical. Renal ischemia leads to a pattern of neurohumoral activation involving the renin-angiotensin-aldosterone system, the systemic nervous system, and other circulating mediators and cytokines. What is lacking, however, is evidence that revascularization reverses these changes and is more effective than medical therapy, which also blocks the same systems in this regard. In fact, there are many reasons why revascularization may not be effective. Most patients are elderly with advanced atherosclerotic disease in other vascular beds and end-organ damage, including cardiac hypertrophy, ischemic cardiomyopathy, myocardial fibrosis, hypertensive glomerular sclerosis, tubulointerstitial fibrosis, atheroemboli, and microcerebral and macrocerebral vascular disease. These changes are likely to drive adverse outcomes even if renal perfusion improves. Finally, the authors acknowledge that the 3 randomized controlled trials, albeit flawed, fail to demonstrate any advantage of angioplasty and stenting over medical therapy. If medical decision making is to be evidence based, then angioplasty and stenting should be rarely, if ever, performed.

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 CONTROVERSIES IN CARDIOVASCULAR MEDICINE

Case Against Angioplasty and Stenting of Atherosclerotic Renal Artery Stenosis Lance D. Dworkin, MD; Kenneth A. Jamerson, MD

therosclerotic renal artery stenosis (RAS) is a rela- progression to end-stage renal failure, or (3) in patients with tively common problem, estimated to affect between congestive heart failure to prevent recurrent exacerbations A1% and 5% of patients with hypertension.1,2 Given requiring hospitalization. The present article discusses these the high prevalence of hypertension, it follows that there are “indications” and reviews the data suggesting that revascu- from 2 million to 4 million individuals with RAS in the larization is no better than medical therapy in these settings. United States alone. A challenging feature of the disease is Furthermore, just as interventional techniques progress, ad- the high incidence of adverse cardiovascular events in af- vances in medical therapy have significantly improved out- fected individuals compared with those with normal renal comes for patients treated conservatively, making it less arteries, which also presents a potential opportunity. On the likely that revascularization will provide significant benefits one hand, there may be such a large burden of atherosclerotic to future patients. disease by the time patients present that it is too late for an intervention in a single vascular bed to significantly alter Does Revascularization Reduce outcomes. On the other hand, it is possible that an effective Blood Pressure? intervention might reduce adverse events, which has led some There are 3 published randomized prospective clinical trials physicians to take an aggressive approach to diagnosing and comparing angioplasty with or without stenting to medical correcting RAS. therapy in patients with atherosclerotic RAS3–5 (Table 1). Response by Cooper and Murphy p 276 Most would agree that all 3 studies are seriously flawed. First, the definition of clinically important renal vascular disease In fact, although individual patients sometimes appear to was overly inclusive. Although most practitioners believe that benefit, whether renal artery revascularization reduces ad- lesions Ͻ70% often are clinically insignificant, patients with verse cardiovascular and renal events is currently unknown. only a Ն50% stenosis were enrolled, and the proportion of Nevertheless, stenting of RAS is performed commonly, with subjects with mild stenoses in the 50% to 70% range was not Ϸ40 000 procedures reported each year in the United States specified. The studies also were marred by a high crossover alone. Most often, revascularization is undertaken in 1 of 3 rate. In the largest trial, the Dutch Renal Artery Stenosis clinical settings and with a specific goal: (1) in patients with Intervention Cooperative (DRASTIC) study, Ϸ40% of sub- resistant hypertension to achieve better blood pressure con- jects crossed over from the medical to the angioplasty arm trol, (2) in patients with renal insufficiency to prevent within the first 3 months. Nevertheless, under intention to

The opinions expressed in this article are not necessarily those of the editors or of the American Heart Association. From Brown Medical School, Providence, RI (L.D.D.), and University of Michigan, Ann Arbor (K.A.J.). Clinical trial registration information—URL: http://www.clinicaltrials.gov. Unique identifier: NCT00081731. Correspondence to Lance D. Dworkin, MD, Rhode Island Hospital, 593 Eddy St, Providence, RI 02903. E-mail [email protected] (Circulation. 2007;115:271-276.) © 2007 American Heart Association, Inc. Circulation is available at http://www.circulationaha.org DOI: 10.1161/CIRCULATIONAHA.106.619031

Downloaded from circ.ahajournals.org 271at Mohammed Mahboob on June 10, 2007 272 Circulation January 16, 2007

TABLE 1. Clinical Trials Comparing Renal Artery Angioplasty With Medical Therapy

Follow-Up, ⌬BP, Number of Antihypertensive Study Subjects Enrolled, n mo mm Hg Drugs Needed to Control BP Webster et al Uni RAS3 Med,13;PTA,14 6 Ϫ8/Ϫ6 2 Ϫ9/Ϫ5 Webster et al Bi RAS3 Med,12;PTA,16 6 Ϫ2/Ϫ2 2 Ϫ19/Ϫ4 Plouin et al4 Med,26;PTA,23 6 Ϫ8/Ϫ5 2 Ϫ12/Ϫ10 Van Jaarsveld et al5 Med,50;PTA,56 12 Same 2 BP indicates blood pressure; Uni, unilateral; Bi, bilateral; Med, medical therapy; and PTA, percutaneous transluminal angioplasty.

treat, patients were still analyzed as part of the group to which normal to nearly zero when, at most, a 50% reduction might they were initially randomized. Perhaps most important, be expected.13 Similarly, nuclear studies in patients with comparatively less attention was paid to the medical regimen unilateral RAS reveal that glomerular filtration rate in the that patients received; however, this regimen needs to be nonstenotic kidney often is the same as or even lower than robust so that it does not bias the data in favor of the that in the kidney distal to a stenosis.14 The lack of correlation intervention. In particular, evidence suggests that blockade of between the severity of renal arterial disease and kidney the renin-angiotensin-aldosterone system is highly effective function probably relates to the presence of ischemic damage in controlling blood pressure and may be critical to improving in the poststenotic kidney and hypertensive injury in the renal and cardiovascular outcomes as well.6 These drugs were nonstenotic kidney. Thus, in many patients with RAS, intrin- not available, their use was not mandated, or their use was sic kidney damage rather than the vascular stenosis per se restricted as a result of the fear of acute renal failure in these accounts for the renal functional impairment. This phenom- trials, however. enon also explains why filtration rate often fails to improve Given the shortcomings of the data, it is fair to say that these significantly after revascularization. Regarding the impact of studies are at best not interpretable, neither supporting nor revascularization on kidney function, observational studies refuting the potential benefits of revascularization. Nevertheless, document that kidney function may stabilize or improve in none of the studies showed a benefit of revascularization over some subjects. On the other hand, angioplasty and stenting medical therapy in reducing blood pressure. At best, the number may cause immediate and sometimes irreversible declines in of antihypertensive medications needed to control blood pres- kidney function, either from contrast nephropathy or athero- sure tended to decline, and almost all patients continued to embolic disease. In the largest prospective randomized trial, require medication. In part, this may relate to the fact that with creatinine clearance tended to improve in patients treated long-standing hypertension, secondary processes such as vascu- medically or with angioplasty, and there was no difference in lar remodeling, atherosclerosis, ischemic damage to the postste- kidney function between the 2 treatment groups after 1 year. notic kidney, and hypertensive injury to the nonstenotic kidney In another retrospective series of Ͼ300 patients with RAS ensue and sustain hypertension.7–9 Revascularization may fail to and impaired kidney function who underwent surgical revas- cure hypertension when stenosis is longstanding and these cularization, sustained reductions in serum creatinine were secondary processes dominate. Also of note is the significant observed in only Ϸ25% of patients. Serum creatinine was complication rate with angioplasty and stenting, reported to be essentially unchanged in more than half the subjects and was from 7% to 15% and including outcomes as serious as death or significantly increased in Ϸ20%, so there was no net im- rapid progression to end-stage renal disease. In the final risk- provement in kidney function for the group as a whole.15 benefit analysis, the advantage of taking 3 as opposed to 4 blood Thus, a few patients benefit while most are either not helped pressure medications must be weighed against these risks.10,11 or even harmed. That revascularization reduces the incidence of end-stage renal disease in patients with RAS has never Does Renal Artery Stenting Preserve been shown. Kidney Function? Although atherosclerotic renal artery lesions tend to progress Does Renal Artery Stenting Prevent with time, relatively few arteries go to complete occlusion Exacerbations of Congestive Heart Failure? within a 5-year period.12 In the absence of complete occlu- Episodes of severe congestive heart failure, so-called flash sion, there is little or no correlation between the degree of pulmonary edema, are associated with RAS and sometimes anatomic stenosis and glomerular filtration rate. In 1 study in are cited as an indication for revascularization. Gray and patients with unilateral RAS and an apparently normal coworkers16 examined the number of admissions for conges- contralateral kidney, glomerular filtration rate ranged from tive heart failure and the New York Heart Association

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Dworkin and Jamerson Medical Therapy of RAS 273

TABLE 2. Optimal Medical Therapy for Atherosclerotic RAS

Risk Factor Recommendation Reference Hypertension Target BP Յ140/90 mm Hg, JNC VII20 Յ130/80 mm Hg with diabetes or proteinuria; ARB as first-line agent; monitor potassium and creatinine Dyslipidemia Target LDL Ͻ100 mg/dL, consider Ͻ70 mg/dL 28 Diabetes Target HgbA1c Ͻ7 mg/dL; foot and eye care Antiplatelet agent Aspirin, clopidogrel, or ticlopidine 29, 30 Smoking Cessation 31, 32 Chronic kidney disease Tight control of blood pressure, dyslipidemia, diabetes; NKF DOQI manage anemia, hyperparathyroidism BP indicates blood pressure; ARB, angiotensin receptor blockade; LDL, low-density lipoprotein; JNC VII, Joint National Committee on the Prevention, Detection, Evaluation, and Treatment of High Blood Pressure VII; and NKF DOQI, National Kidney Foundation Disease Quality Initiatives.

functional class in 39 patients who underwent renal artery Hypertension Treatment stenting for recurrent episodes of congestive heart failure. Although it is well known that lowering blood pressure can Hospitalizations for congestive heart failure declined signif- prevent adverse cardiovascular events, 2 fundamental ques- icantly in the year after stenting, and New York Heart tions remain: Is there an ideal blood pressure target that Association functional class also improved. The number of confers maximal cardiovascular protection, and is there a patients receiving angiotensin-converting enzyme inhibitors specific antihypertensive regimen that provides cardiovascu- also increased from 15% to 50%, however, which may have lar benefits beyond just lowering blood pressure? contributed to the improved outcome. Because the study was uncontrolled, the real impact of revascularization is impossi- Is There an Ideal Blood Pressure Target? ble to determine. At present, there are no prospective ran- Although any method of reducing blood pressure is lifesaving domized data demonstrating that angioplasty and stenting in malignant and accelerated hypertension, clinical trials were reduce admissions for severe congestive heart failure, or any needed to demonstrate the value of therapy of less severe other cardiovascular event, compared with medical therapy hypertension. In 1967, the Veterans Administration Cooper- alone. ative Study was the first to demonstrate the value of treating elevated blood pressure (diastolic blood pressure Reducing the Risk of Cardiovascular Disease Ͼ110 mm Hg) in preventing cardiovascular morbidity (74% in RAS: The Medical Approach reduction in events per year).17 Subsequent trials have pro- gressively lowered the bar until currently a target pressure Optimal Medical Therapy: The Cardiovascular Ͻ140/90 mm Hg is recommended, particularly for hyperten- Outcomes in Renal Atherosclerotic Lesions Trial sive diabetic patients18 or those with proteinuric renal dis- Cardiovascular Outcomes in Renal Atherosclerotic Lesions ease.19 In CORAL, the target blood pressure is Ͻ140/ (CORAL) is a large, multicenter, randomized, prospective 90 mm Hg overall but Ͻ130/80 mm Hg for diabetics or those trial that is comparing the effects of angioplasty with stenting with proteinuric renal disease.20 and optimal medical therapy to medical therapy alone on a composite of adverse cardiovascular and renal events. In Use of Angiotensin II Receptor Blockade as designing the CORAL study, the investigators realized that Initial Therapy previous trials in RAS often lacked a rigorous medical In CORAL, all patients will receive an angiotensin II type 1 approach that effectively managed the multiple risk factors receptor antagonist as the first-line antihypertensive agent. for cardiovascular and renal disease progression that affect Because the renin-angiotensin-aldosterone system is acti- this population. Accordingly, the CORAL trial uses a medical vated in many patients with renal vascular disease, drugs that treatment intervention that includes tight control of blood block the system are often highly effective in controlling pressure, treatment of dyslipidemia and diabetes, smoking blood pressure.21 The hypothesis that newer drugs could cessation, administration of an antiplatelet agent, and atten- confer additional cardiovascular benefits beyond lowering tion to the complications of renal insufficiency (Table 2). blood pressure compared with older drugs like diuretics and Although the effect of these interventions on outcomes in ␤-blockers has been tested in recent trials. With few excep- patients with RAS has not been specifically tested, the tions, it appears that lowering blood pressure with any agent CORAL treatment algorithm is based on current, evidence- confers benefit, although none of these trials examined based practice guidelines. Data from other populations sug- patients with known RAS.22,23 gest that these medical interventions can produce profound The newer generation of hypertension clinical trials tests reductions in cardiovascular event rates. the impact of combination therapy. In the Anglo-

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 274 Circulation January 16, 2007

TABLE 3. Hypertension Drug Algorithm in the CORAL Trial

First-line agent Angiotensin receptor antagonist Begin with low dose and titrate up Monitor serum creatinine and potassium First-line agent if ARB ACE inhibitor Begin with low dose and titrate up Monitor serum creatinine and potassium not tolerated Second-line agent Thiazide diuretic Begin with low dose and titrate up Monitor serum creatinine and potassium Combinations with ARB/ACE may be available Use loop diuretics for patients with serum creatinine Ն2 mg/dL Third-line agents Calcium channel blocker Selection may be based on comorbidities (Table 2) ␤-Blocker Keep titrating up and adding third-line agents until patient reaches goal ␣-Blocker Vasodilator ARB indicates angiotensin receptor blockade; ACE, angiotensin-converting enzyme.

Scandinavian Cardiac Outcomes Trial (ASCOT), an angio- Third Report of the Expert Panel on Detection, Evaluation, tensin-converting enzyme inhibitor/calcium channel blocker and Treatment of High Blood Cholesterol in Adults (Adult regimen improved cardiovascular mortality compared with a Treatment Panel III), RAS is considered a coronary artery ␤-blocker/diuretic.24 Anticipating that most patients with disease equivalent in terms of cardiovascular risk. Thus, RAS also will require combination therapy, the blood pres- lowering low-density lipoprotein cholesterol to at least Ͻ100 sure treatment algorithm in CORAL is a stepped-care ap- mg/dL is the goal of therapy, with some suggesting a target proach, with multiple classes of agents added sequentially low-density lipoprotein of Ͻ70 mg/dL.28 until the target blood pressure is reached (Table 3). In addition to preventing cardiovascular events, a number Diabetes Mellitus of trials suggest that blocking the renin-angiotensin system The CORAL approach to diabetes will acknowledge the clear reduces proteinuria and delays the onset of end-stage renal evidence that tight glucose control to an HbA1c of Ͻ7 mg/dL disease. In part, the observed risk reductions in progressive with oral agents and/or insulin is associated with reductions in nephropathy form the basis for the use of an angiotensin II microvascular and macrovascular complications.29,30 Addition- type 1 receptor antagonist as first-line therapy in CORAL. ally, medical nutrition therapy, multidisciplinary foot care (par- Despite the potential benefits, use of renin-angiotensin ticularly for patients with peripheral vascular disease, a common system–blocking drugs is controversial in patients with RAS. comorbidity in RAS), eye care to prevent and treat diabetic Of particular concern is the risk of acute renal failure. retinopathy, and physical activity are recommended. Blocking the system can cause precipitous declines in glo- merular filtration rate in the poststenotic kidney, which may Chronic Renal Insufficiency cause significant acute renal failure if the RAS affects both Chronic kidney disease is common in RAS. The guidelines kidneys or a solitary functioning kidney.25,26 That these established by the National Kidney Foundation Kidney Disease concerns may be somewhat overemphasized is suggested by Quality Initiatives (www.kidney.org/professionals/KDOQI/ the fact that the actual incidence of acute renal failure with guidelines.cfm) form the underpinnings to the approach to the renin-angiotensin system–blocking drugs is quite low, affect- management of chronic kidney disease in CORAL. ing Ͻ5% of these patients.27 In addition, acute renal failure in this setting is usually immediately reversible on cessation of Antiplatelet Agents the medication and therefore without long-term adverse Although there are no direct data in patients with RAS, effects. Also of concern is the risk that renin-angiotensin administration of an antiplatelet agent is required in CORAL system blockade will promote progression of ischemic ne- and recommended for all patients with RAS.31,32 phropathy in the poststenotic kidney, which has been ob- served in animal studies. On the other hand, these drugs may Cumulative Impact of the Medical actually preserve glomerular structure and function in the Therapy Intervention contralateral kidney in patients with unilateral stenosis.24 Aggressive treatment of comorbidities and atherosclerotic risk Taken together, these data suggest that renin-angiotensin factors has tremendous potential to reduce cardiovascular risk in system inhibition may have important therapeutic benefits in patients with RAS. Yusuf et al33 summarize the potential patients with renovascular disease that are independent of the cumulative risk reduction resulting from 4 interventions (aspirin, effect on blood pressure. ␤-blocker, ACE inhibitor, statin) for prevention of myocardial Dyslipidemia Treatment infarction. Although the effects of interventions may not be Although no specific evidence exists for patients with reno- additive in this manner, a multifaceted medical approach is a vascular disease, according to the guidelines, including the powerful tool for preventing adverse cardiovascular events. The

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Cooper and Murphy Medical Therapy of RAS 265

sin and sympathetic nervous systems, and volume retention process with statins, antiplatelet therapy, antihypertensive associated with renal ischemia are likely important contribu- therapies, and hypoglycemic agents in the presence of diabe- tors. Recent work suggests a relationship between brain tes. Inhibition of sympathetic activation and the renin-angio- natriuretic peptide and RAS, with high brain natriuretic tensin-aldosterone system with ␤-blockers and angiotensin- peptide levels (Ͼ80 pg/mL) associated with better outcomes converting enzyme inhibitors is a very successful strategy for after renal artery intervention.55 Angiotensin II, endothelin, or patients with cardiovascular diseases in several settings, an insulin-like growth factor, and sympathetic activation all including acute myocardial infarction, CHF, and hyperten- may be involved in the process of ventricular thickening and sion. In each of these settings, treatment has demonstrated stiffening.46–49,53 Angiotensin has been implicated in the significant benefits as measured by reductions in clinically mechanism of cardiac hypertrophy and is known to induce important events. Renin-angiotensin-aldosterone system– protein synthesis in myocardial cells,50 even when the mean inhibiting drugs also are indicated for retarding the progres- arterial pressure is lowered.52 Endothelin also has been 75,76 implicated in cardiac hypertrophy in RAS.51 Interestingly, the sion to end-stage renal disease to preserve renal function. use of endothelin receptor blockers reverses left ventricular Although there are no outcome data with antihypertensive hypertrophy in rats with RAS.48,51 medical therapy for treatment of RAS specifically, clinical Remodeling of the left ventricular wall and peripheral studies suggest that renin-angiotensin-aldosterone system in- arteries in response to renovascular hypertension has been hibition decreases cardiovascular events in other high-risk noted. In patients with RAS with hypertension, cardiac patient groups.77,78 remodeling permits maintenance of normal cardiac function Antihypertensive medical therapies can be associated with despite increased left ventricular wall stress resulting from adverse health effects, however, such as lack of energy or systemic hypertension.56 Ventricular performance, deter- chronic cough, which may affect quality of life and reduce mined by afterload, chamber size, mass index, and functional compliance.79–81 Fifty percent to 70% of such medications shortening or contractility, often is abnormal in people with are discontinued or changed within 6 months.82 Adherence to renovascular hypertension.5,57 published guidelines for treatment of hypertension is poor, Congestive heart failure is common in patients undergoing often with significant adverse impact on patients and the 58–60 renal artery revascularization, just as RAS is common in healthcare system.83,84 Other factors that may limit the effec- patients presenting with CHF.61 Left ventricular hypertrophy tiveness of medical therapy include temporal variations in and decreased contractility, risk factors for development of drug levels and variable absorption.85,86 Finally, as described overt CHF,62 are significantly more common in those with above, renal ischemia increases activity in a number of other renovascular hypertension than essential hypertension,44,63 even when matched for age and gender.5 RAS is prevalent in pathogenic systems that may have deleterious cardiovascular those with CHF and may be implicated in its cause or severity or renal effects. Whether medical therapy alone is sufficient in a large proportion of patients. In a series of patients Ͼ70 to minimize risk in patients with atherosclerotic RAS is not years of age presenting with New York Heart Association known. class II to IV heart failure, 34% were found to have stenosis of at least 50% involving at least 1 renal artery.61 Intervention Cohort Studies As discussed, left ventricular hypertrophy is an important An important feature of revascularization is the ability to risk factor for cardiovascular events and cardiovascular death obviate the root cause of the entire process: the stenotic renal in both the presence and absence of hypertension.64–71 Re- artery. Importantly, others have demonstrated that relief of gression of left ventricular hypertrophy is favorable prognos- stenosis with revascularization results in a reduction in tically,69 and left ventricular mass index, a measure of renin-angiotensin-aldosterone system and sympathetic activ- ventricular hypertrophy, has been shown to decrease after ity.15 What is entirely unclear is whether pharmacological renal artery revascularization.72 When CHF accompanies therapy directed at one or several of these pathways is equally RAS, renal artery revascularization often is associated with effective or whether a reduction in activation attributable to 60,73 symptomatic improvement. Currently, this may be under- revascularization leads to improvements in long-term clinical recognized in clinical practice. There certainly appears to be outcome. some confusion about the management of patients with RAS In 1974, Hunt et al87 described lower rates of mortality, and CHF; the recent American Heart Association/American stroke, myocardial infarction, and azotemia and better blood College of Cardiology consensus document on the manage- pressure control in surgically revascularized patients than in a ment of peripheral vascular disease74 suggests limited support for screening patients with CHF but substantial support for comparison medical group despite more severe hypertension treatment of RAS in patients with CHF, despite a paucity of in the revascularized subjects. Interestingly, in a more con- evidence to support either position. temporary series comparing nonrandomized groups of pa- tients with RAS and hypertension treated with stents or Limitations of Medical Therapy for RAS medications, Pizzolo et al88 reported a survival treatment Undoubtedly, all patients that have RAS require lifelong effect in the intervention group similar to the experience of medical therapy to manage the underlying atherosclerotic Hunt et al 30 years previously, with an 87% probability of

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 266 Circulation January 16, 2007

5-year survival in the intervention group versus 67% in the benefits, then screening programs for high-, moderate-, or medical management group. potentially even low-risk individuals become advisable. In A number of single-center registries of stenting89 and a contrast, if the benefit is small or marginal, then screening recent multicenter registry of stenting for failed balloon and treatment may be reserved for those at high risk. angioplasty90 suggest substantial benefits related primarily to Furthermore, identifying subsets that derive greater or lesser improved blood pressure control. In contrast, the 3 multi- benefit or even harm becomes possible when the relative center trials of percutaneous balloon renal angioplasty benefits and risks can be identified. (PTRA) without stenting failed to demonstrate a blood Undoubtedly, some patients with atherosclerotic RAS ap- pressure benefit.91–93 Importantly, the randomized clinical pear to have a very favorable response to revascularization as trials done thus far have substantial weaknesses that have determined by improvement in blood pressure or kidney been the subject of considerable debate and discussion. The function. Previous investigators have suggested that the largest randomized trial comparing drug treatment and PTRA response to revascularization may be predicted by the pres- was the Dutch Renal Artery Stenosis Intervention Coopera- ence of microvascular dysfunction measured with the renal tive (DRASTIC) study. In this study, 106 patients were resistive index100 or by the presence of neurohumoral activa- randomly assigned to treatment by PTRA or medical therapy. tion.55 Fundamentally, though, each of these observations is The study design was such that refractory patients in the drug limited by the absence of a contrasting group of medically therapy group were allowed to receive balloon angioplasty if managed patients who are concurrent and carefully con- their blood pressure control was inadequate. With 44% of the trolled. Thus, all such observations suggest but are unable to drug therapy group receiving PTRA, the study contrasted establish the relative benefit of revascularization or the true provisional angioplasty against angioplasty. Despite the au- value of the baseline test strategy to predict outcome with thors’ assertion that PTRA in addition to drug therapy differing treatments. This limitation, the absence of a robust provided “little benefit,” patients in the PTRA group were and appropriate control group, has led to false conclusions, less likely to have deterioration of their blood pressure later disproved with appropriate randomized trials, about the control or renal artery occlusion during 12 months of follow- utility of a wide variety of clinical therapies, including up. Two other studies of balloon angioplasty versus medical surgery for Ménière’s disease,101 warfarin for acute myocar- therapy were confounded by similar issues and failed to dial infarction,102 portosystemic shunts,103 and antiarrhythmic demonstrate differences in blood pressure control.79,82 Impor- therapy after myocardial infarction.104 tantly, none of the studies comparing PTRA and drug therapy This is likely to be important for patients with renal artery thus far have included substantial numbers of patients who hypertension. In fact, in the Scottish and Newcastle series,92 received stents.93 Renal artery stent placement is standard 55 patients with renovascular hypertension were randomized today and substantially improves technical and long-term to angioplasty or medical therapy after a run-in period of 4 clinical outcomes compared with angioplasty alone,89,94,95 weeks. It was noted that significant decreases in blood especially for ostial stenoses,96 which make up 80% of pressure occurred before the initiation of treatment. Thus, the atherosclerotic stenoses.97 In a meta-analysis of 1322 pa- effect of revascularization on blood pressure, observed in tients, stent placement had a higher technical success rate and cohort studies, may be confounded by regression to the mean. a lower restenosis rate than did PTRA (98% versus 77% and 17% versus 26%, respectively; PϽ0.001) and a higher cure How Should Patients Be Managed? rate for hypertension.94 These issues have pragmatic implications for the various physicians who care for patients with ischemic renal disease. Limitations of Nonrandomized Evidence Many in the interventional community feel compelled to There is little controversy that stenting is an effective means revascularize renal stenoses, even when found incidentally. In of achieving durable revascularization of the renal artery. 2000, an estimated 19 800 renal angioplasties were per- Similarly, there is a wealth of uncontrolled study data formed for Medicare beneficiaries alone at an estimated cost demonstrating that some patients receive substantial benefits of $200 million.105 From 1996 to 2000, the annual number of in blood pressure control and renal function.89,94,98,99 How- renal stent procedures has increased 364%.105 ever, renal function worsens in some individuals after revas- In contrast, such enthusiasm is largely absent among cularization, and blood pressure may not improve. The nephrologists and others who routinely care for patients with current controversy is driven by a singular key factor: the lack chronic kidney diseases. Importantly, without widespread of an appropriate control group to allow the magnitude of the screening by the medical community, it is apparent that most benefit derived from stenting to be measured. Does stent patients with renovascular disease never come to attention. revascularization markedly reduce the risk of cardiovascular Recent work suggests that Ϸ7% of individuals Ͼ65 years of and renal events, does it modestly reduce events, or are the age have significant RAS.106 Clearly, the majority are never risks of stenting balanced or even overwhelmed by compli- identified within our communities, and far fewer are treated cations? This determination has fundamental implications in with either appropriate medical therapy or revascularization the decisions that physicians make. If stenting has substantial and medical therapy. This nihilistic approach of “no screen,

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Congenital Heart Disease for the Adult Cardiologist

Ebstein’s Anomaly

Christine H. Attenhofer Jost, MD; Heidi M. Connolly, MD; Joseph A. Dearani, MD; William D. Edwards, MD; Gordon K. Danielson, MD

bstein’s anomaly is a rare congenital heart disorder lar shelf between the inlet and trabecular zones of the right Eoccurring in Ϸ1 per 200 000 live births and account- ventricle. ing for Ͻ1% of all cases of congenital heart disease.2–6 The anterior leaflet is generally redundant and may contain This anomaly was described by Wilhelm Ebstein in 1866 several fenestrations.6 Its chordae tendineae are generally in a report titled, “Concerning a very rare case of insuffi- short and poorly formed. Moreover, the anterior leaflet of the ciency of the tricuspid valve caused by a congenital tricuspid valve may be severely deformed, so that the only malformation.”7,8 The patient was a 19-year-old cyanotic mobile leaflet tissue is displaced into the right ventricular man with dyspnea, palpitations, jugular venous distension, outflow tract, where it may cause obstruction or form a large and cardiomegaly.7,8 At autopsy, Ebstein described an sail-like intracavitary curtain. Typical autopsy examples of enlarged and fenestrated anterior leaflet of the tricuspid Ebstein’s anomaly are shown in Figures 3 and 4. valve. The posterior and septal leaflets were hypoplastic, In Ebstein’s anomaly, the right ventricle is divided into 2 thickened, and adherent to the right ventricle. There was regions: the part directly involved with the malformation (ie, the also a thinned and dilated atrialized portion of the right inlet portion), which is functionally integrated with the right ventricle, an enlarged right atrium, and a patent foramen atrium, and the part that is not involved by the anomaly, which ovale9 (Figure 1). By 1950, only 3 cases of this anomaly consists of the other 2 components of the right ventricle, namely had been published.8,10,11 the trabecular and outlet portions, that constitute the functional right ventricle. The “atrialized” portion of the right ventricle (ie, Pathological Anatomy the inlet component) can become disproportionately dilated and In the normal heart, the tricuspid valve has 3 leaflets: may account for more than half of the right ventricular volume anterior, posterior, and septal.12,13 Ebstein’s anomaly is a in extreme cases instead of the usual one third of the total right malformation of the tricuspid valve and right ventricle ventricular volume. There is often marked dilatation of the true characterized by (1) adherence of the septal and posterior tricuspid valve annulus, which is not displaced, and a large leaflets to the underlying myocardium (failure of delami- chamber separating this true annulus from the functional right nation, namely splitting of the tissue by detachment of the ventricle (atrialized portion of the right ventricle)6,12 (Figure 2). inner layer during embryologic development); (2) down- The right coronary artery demarcates the level of the true ward (apical) displacement of the functional annulus annulus and may become kinked during plication annuloplasty (septalϾposteriorϾanterior); (3) dilation of the “atrial- procedures. ized” portion of the right ventricle, with various degrees of Two thirds of hearts with Ebstein’s anomaly show hypertrophy and thinning of the wall; (4) redundancy, dilated right ventricles. Dilatation often involves not only fenestrations, and tethering of the anterior leaflet; and (5) the atrialized inlet portion of the right ventricle but also the dilation of the right atrioventricular junction (true tricuspid functional right ventricular apex and outflow tract. In some annulus).6,14 cases, right ventricular dilatation is so marked that the The apical displacement of the hinge point of the valve in ventricular septum bulges leftward, compressing the left Ebstein’s anomaly from the atrioventricular ring is shown in ventricular chamber.6 In such cases, the short-axis view Figure 2.12,15–17 The point of maximal displacement is at the demonstrates a circular right ventricle and a crescentic left commissure between the posterior and septal leaflets of the ventricle. In extreme cases, episodic left ventricular out- tricuspid valve.16 In normal human hearts, the downward flow tract obstruction can occur. displacement of the septal and posterior leaflets in relation to the anterior mitral valve leaflet is Ͻ8 mm/m2 body surface Nomenclature and Classification area.6 The spectrum of the malformation in Ebstein’s anom- We use 2 approaches in describing the anatomic severity of aly may range from only minimal displacement of the septal Ebstein’s anomaly. The first approach is based on the and posterior leaflets to an imperforate membrane or muscu- echocardiographic appearance in which the abnormality is

From the Divisions of Cardiovascular Diseases (C.H.A.J., H.M.C.), Cardiovascular Surgery (J.A.D.), and Anatomic Pathology (W.D.E.), Mayo Clinic, Rochester, Minn. Dr Danielson is an emeritus member, Division of Cardiovascular Surgery, Mayo Clinic, Rochester, Minn. This article is based in part on a previously published manuscript.1 Used with permission. Reprint requests to Heidi M. Connolly, MD, Division of Cardiovascular Diseases, Mayo Clinic, 200 First St SW, Rochester MN, 55905. (Circulation. 2007;115:277-285.) © 2007 American Heart Association, Inc. Circulation is available at http://www.circulationaha.org DOI: 10.1161/CIRCULATIONAHA.106.619338 277

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 278 Circulation January 16, 2007

Figure 2. Top, Normal tricuspid valve with anterior, posterior, and septal leaflets in 1 plane. Middle, Tricuspid valve in right- sided Ebstein’s anomaly showing displacement of posterior and septal leaflets; maximal displacement is at the crux of the pos- terior and septal leaflets. Bottom, Tricuspid valve in left-sided Ebstein’s anomaly; the displacement of leaflets is similar to that in the right-sided anomaly. From Anderson et al.16 Used with permission of the Mayo Foundation for Medical Education and Research.

gow Outcome Scale, with grades 1 to 4.19 The ratio of the combined area of the right atrium and atrialized right ventri- Figure 1. Figure from Ebstein’s original case report. The right cle is compared with that of the functional right ventricle and atrium and right ventricle are shown opened along the right bor- left heart (ratio Ͻ0.5, grade 1; ratio of 0.5 to 0.99, grade 2; der beginning at the superior vena cava. A, Right atrium; B, ratio of 1.0 to 1.49, grade 3; ratio Ն1.5, grade 4). right ventricle; b, valve; I, rudimentary septal leaflet of tricuspid valve with its chordae tendineae, which insert on the endocar- dium of the ventricular septum; r, opening through which one Prevalence and Genetic Factors can get into the right conus arteriosus, and in the opposite The leaflets of the tricuspid valve develop equally from the direction, one can get into the sac that is formed by membrane h, h’, and posterior part of endocardium of ventricular septum o. endocardial cushion tissues and the myocardium.13 The leaf- From Mann and Lie.7 Used with permission of the Mayo Foun- lets and tensile apparatus of the atrioventricular valves are dation for Medical Education and Research. formed by a process of delamination of the inner layers of the inlet zone of the ventricles. In Ebstein’s anomaly, delamina- tion of the tricuspid valve leaflets fails to occur, but the described as anatomically mild, moderate, or severe. The mechanism for this is not entirely understood.16 amount of displacement and tethering of the leaflets and the degree of right ventricular dilatation are assessed. This classification is imprecise but simple. Our second approach is to describe the exact anatomy of each of the involved structures of the heart as visualized at operation. This nomen- clature system emphasizes characteristics that surgeons find important when considering repair versus replacement of the tricuspid valve.12 In 1988, Carpentier et al18 proposed the following classi- fication of Ebstein’s anomaly: type A, the volume of the true right ventricle is adequate; type B, a large atrialized compo- nent of the right ventricle exists, but the anterior leaflet of the tricuspid valve moves freely; type C, the anterior leaflet is severely restricted in its movement and may cause significant obstruction of the right ventricular outflow tract; and type D, almost complete atrialization of the ventricle except for a small infundibular component. Figure 3. Marked cardiomegaly caused by right-sided chamber 19 Celermajer et al described an echocardiographic grading dilatation in a 67-year-old man with severe Ebstein’s anomaly, score for neonates with Ebstein’s anomaly, extended Glas- with normal heart at right for comparison (anterior view).

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Attenhofer Jost et al Ebstein’s Anomaly 279

Ebstein’s anomaly.14,15,27–30 We recently reported features resembling noncompaction in 3 patients with the anomaly.31 Since then, we have analyzed 106 consecutive patients who had Ebstein’s anomaly and found left-sided heart abnormal- ities in 39%; 18% of these patients had left ventricular dysplasia resembling noncompaction.32 Most patients with congenitally corrected transposition of the great arteries have an abnormal systemic tricuspid valve, which fulfills the criteria for Ebstein’s anomaly in 15% to 50% of cases.16,33–35 It is unclear whether the fundamental nature of the anomaly is identical in concordant and discor- dant atrioventricular connections.36,37 The morphological right ventricle is rarely dilated in congenitally corrected transposition.14,36

Physiology The functional impairment of the right ventricle and regurgi- tation of the tricuspid valve retard forward flow of blood through the right side of the heart. In addition, during contraction of the atrium, the atrialized portion of the right ventricle balloons out and acts as a passive reservoir, decreas- ing the volume of ejected blood. The overall effect on the Figure 4. Severe Ebstein’s malformation of tricuspid valve right atrium is dilatation, increasing the size of an interatrial (4-chamber view) showing marked downward displacement of communication. Tricuspid regurgitation increases by annular shelf-like posterior leaflet with attachment to underlying free wall dilatation.14 Associated heart disease in Ebstein’s anomaly by numerous muscular stumps (arrows), markedly dilated atrial- ized portion of right ventricle (ARV), small functional portion of has a further effect on physiology. right ventricle (RV), leftward bowing of ventricular septum, and marked dilatation of right atrium (RA). LA indicates left atrium; LV, left ventricle. Clinical Features The cardinal symptoms in Ebstein’s anomaly are cyanosis, There are heterogeneous genetic factors in Ebstein’s anom- right-sided heart failure, arrhythmias, and sudden cardiac aly. Case-control studies suggest genetic, reproductive, and death. The hemodynamic variations and clinical presentation environmental risk factors (eg, the anomaly is more common depend on age at presentation, anatomic severity, hemody- in twins, in those with a family history of congenital heart namics, and degree of right-to-left interatrial shunting.38 disease, and in those with maternal exposure to benzodiaz- On examination, the jugular venous pulse rarely shows a epines).4 Maternal lithium therapy can rarely lead to Ebstein’s large V wave despite severe regurgitation of the tricuspid anomaly in the offspring.20 Most cases are sporadic; familial valve because the large right atrium engulfs the increased Ebstein’s anomaly is rare. volume. A widely and persistently split second heart sound In a genetic study of 26 families with Ebstein’s anomaly, and several added sounds are typical.38 A systolic murmur 93 of 120 first-degree relatives were evaluated.21 No case of may be audible. Digital clubbing depends on the degree of the anomaly was found, but 2 first-degree relatives had cyanosis.38 ventricular septal defects, and another, who died at 7 months, Ebstein’s anomaly is a common lesion referred for fetal was said to have had congenital heart disease. Rare cases of echocardiography because severe forms may lead to cardio- cardiac transcription factor NKX2.5 mutations, 10p13-p14 megaly, hydrops, and tachyarrhythmias.39,40 deletion, and 1p34.3-p36.11 deletion have been described in Neonates with Ebstein’s anomaly may present with cyano- the anomaly.22–24 sis, congestive heart failure caused by regurgitation of the Associated Cardiac Malformations in tricuspid valve, and marked cardiomegaly.39 Symptomatic Ebstein’s Anomaly children with Ebstein’s anomaly may have progressive right- An interatrial communication is present in 80% to 94% of sided heart failure, but most will reach adolescence and patients with Ebstein’s anomaly.25,26 Additional associated adulthood. Ͼ anomalies include bicuspid or atretic aortic valves, pulmo- Children 10 years of age and adults often present with nary atresia or hypoplastic pulmonary artery, subaortic ste- arrhythmias.19 Adults also present with progressive cyanosis, nosis, coarctation, mitral valve prolapse, accessory mitral decreasing exercise tolerance, fatigue, or right-sided heart valve tissue or muscle bands of the left ventricle, ventricular failure. In the presence of an interatrial communication, the septal defects, and pulmonary stenosis.1 risk of paradoxical embolization, brain abscess, and sudden Abnormalities of left ventricular morphology and function, death increases.19 Exercise tolerance is dependent on heart as well as other left-sided heart lesions, also occur in size and oxygen saturation.41,42

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 280 Circulation January 16, 2007

atrial enlargement, as well as complete or incomplete right 38 bundle-branch block. The R waves in leads V1 and V2 are small. Bizarre morphologies of the terminal QRS pattern result from infra-Hisian conduction disturbance and abnormal activation of the atrialized right ventricle.46 A typical ECG is shown in Figure 6. Complete heart block is rare in Ebstein’s anomaly, but first-degree atrioventricular block occurs in 42% of patients because of right atrial enlargement and structural abnormal- ities of the atrioventricular conduction system.38,47 The atrio- ventricular node may be compressed and the central fibrous body abnormally formed. The right bundle branch may be abnormal or show marked fibrosis (or both).15,16,48 The downward displacement of the septal leaflet of the tricuspid valve is associated with discontinuity of the central Figure 5. Example of an echocardiogram (4-chamber view, fibrous body and septal atrioventricular ring with direct apex down) of a patient with severe Ebstein’s anomaly showing muscular connections, thus creating a potential substrate for a grossly displaced septal leaflet (arrow). The anterior leaflet is accessory atrioventricular connections and pre-excitation.5,6 severely tethered and nearly immobile. The functional right ven- From 6% to 36% of patients with Ebstein’s anomaly have Ն1 tricle (RV) is small. ARV indicates atrialized right ventricle; LA, left atrium; LV, left ventricle; and RA, right atrium. accessory pathways,26,38,46,49,50 and most accessory pathways are located around the orifice of the malformed tricuspid valve.25,27,47 Correct identification and treatment of accessory Diagnostic Evaluation pathways are essential and may help to prevent sudden Echocardiography cardiac death. Paroxysmal tachyarrhythmias in Ebstein’s Echocardiography, the diagnostic test of choice for Ebstein’s anomaly are based on typical, fast-conducting atrioventricular anomaly, has largely obviated cardiac catheterization38,43 accessory pathways with both antegrade and retrograde con- (Figure 5). Echocardiography allows accurate evaluation of duction properties in most patients.46 In addition, wide QRS the tricuspid valve leaflets and the size and function of the tachycardia over a septal accessory atrioventricular pathway, cardiac chambers. ventricular tachycardia, or flutter, as well as ectopic atrial The principal feature of Ebstein’s anomaly is apical dis- tachycardia, atrial flutter, and atrial fibrillation, can occur.46,50 placement of the septal leaflet of the tricuspid valve from the Atrial fibrillation and atrial flutter are most likely caused by insertion of the anterior leaflet of the mitral valve by at least secondary alterations of the right atrial myocardium from 8 mm/m2 body surface area.6 Tethering of the tricuspid valve previous cardiac surgery or are postoperative as a result of is present if there are at least 3 accessory attachments of the incisional atrial tachycardia.46 leaflet to the ventricular wall, causing restricted motion of the leaflet.39 Marked enlargement of the right atrium and atrial- Chest Radiography ized right ventricle is present when the combined area of the The cardiac silhouette may vary from almost normal to the right atrium and atrialized right ventricle is larger than the typical Ebstein’s anomaly configuration consisting of a combined area of the functional right ventricle, left atrium, globe-shaped heart with a narrow waist similar to that seen and left ventricle measured in the apical 4-chamber view at with pericardial effusion (Figure 7). Vascularity of the pul- end diastole.19 The site and degree of regurgitation of the monary fields is either normal or decreased. A cardiothoracic tricuspid valve and the feasibility of valve repair also are ratio Ͼ0.65 carries a poor prognosis. assessed with echocardiography.14 Cine magnetic resonance imaging may be used to assess Cardiac Catheterization ventricular size and function when echocardiographic image Diagnostic cardiac catheterization is rarely necessary in quality is inadequate.44,45 patients with Ebstein’s anomaly, other than for preoperative coronary angiography. Right ventricular and pulmonary ar- Electrocardiography tery pressures are usually normal in patients with the anom- The ECG is abnormal in most patients with Ebstein’s anom- aly, although the right ventricular end-diastolic pressure may aly. It may show tall and broad P waves as a result of right be increased. Right atrial pressure may be normal despite

Figure 6. ECG of a patient with severe Ebstein’s anomaly showing the typical changes, with prolongation of the PR interval (226 ms), right bundle-branch block, and somewhat bizarre configura- tion of the QRS complex.

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Attenhofer Jost et al Ebstein’s Anomaly 281

in patients with the anomaly than in those with structurally normal hearts, and the risk of recurrence is increased.46,52–54 Supraventricular tachyarrhythmia associated with Ebstein’s anomaly also can be ablated at the time of operative repair.55,56

Surgical Options In 1959, repair of the tricuspid valve was reported in 2 patients who had Ebstein’s anomaly; both died.57 Successful operative intervention for regurgitation of the tricuspid valve in patients with the anomaly was first described in 1962; the valve was replaced.58 The initial publication on patients with Ebstein’s anomaly undergoing tricuspid valve replacement reported a surgical mortality of 54%.49 Similar high early mortality and unsatisfactory late results for tricuspid valve repair by the methods available at that time were described.57,59 Between April 1972 and January 2005, 540 consecutive patients with Ebstein’s anomaly were operated on at Mayo Clinic Rochester. The age at operation ranged from 2 months to 79.1 years (median, 20 years). Of those having a tricuspid valve procedure, valve reconstruction was possible in 34.4%, Figure 7. Chest radiograph of a patient who had Ebstein’s anomaly with severe tricuspid regurgitation and a small atrial and valve replacement (usually bioprosthesis) was performed septal defect before tricuspid valve surgery. This typical image in 65.6%. There were 29 early deaths (5.4%). Late results of shows cardiomegaly, a narrow waist, and a cardiothoracic ratio the first 323 operations have been reviewed.60 There were 23 of 0.56. late deaths (7.6%) during a follow-up extending to 25 years (mean, 7.1 years). severe regurgitation of the tricuspid valve, especially if the Our initial repair technique reported in 1979 consisted of right atrium is markedly dilated. Oximetry may show sys- plication of the free wall of the atrialized portion of the right temic arterial desaturation in the presence of an interatrial ventricle, posterior tricuspid annuloplasty, and right reduction communication and right-to-left shunting. atrioplasty.61 The repair is based on the construction of a monocuspid valve using the anterior leaflet. Since the initial Management report, we have incorporated various modifications of tricus- Medical pid valve repair, depending on the numerous variants encoun- Any patient with Ebstein’s anomaly needs to be evaluated tered with the anatomy of Ebstein’s anomaly.14,62 Our current regularly by a cardiologist who has expertise in congenital repair usually involves bringing the anterior papillary mus- heart disease. Prophylaxis for endocarditis is recommended cle(s) toward the ventricular septum, thus facilitating coapta- despite its low risk in the anomaly. tion of the leading edge of the anterior leaflet with the Physical activity recommendations are summarized by ventricular septum (Figure 8). Generally, an anteroposterior Task Force 1 on Congenital Heart Disease.51 Athletes with tricuspid purse-string annuloplasty is used, and atrialized mild Ebstein’s anomaly, nearly normal heart size, and no right ventricular plication or resection is performed selec- arrhythmias can participate in all sports. Athletes with severe tively. This results in a tricuspid valve repair at the level of Ebstein’s anomaly are precluded from sports unless the the functional annulus, in contrast to our original repair, anomaly has been optimally repaired, the heart size is nearly which brought the functional annulus up to the true annulus. normal, and no history of arrhythmias exists. We believe the 2 most important features that enable a Patients with Ebstein’s anomaly and cardiac failure who successful, durable repair are a free leading edge of the are not candidates for surgery are treated with standard heart anterior leaflet and at least 50% delamination of the anterior failure therapy, including diuretics and digoxin. The efficacy leaflet. of angiotensin-converting enzyme inhibitors in patients with In 1988, Carpentier et al18 proposed a repair that used Ebstein’s anomaly who have right-sided heart failure is mobilization of the anterior leaflet of the tricuspid valve. For unproved. Medical management of arrhythmias should be their types B and C, temporary detachment of the anterior individualized and combined with operative or catheter-based leaflet and adjacent part of the posterior leaflet was followed intervention. by longitudinal plication of the atrialized ventricle and adjacent right atrium, repositioning of the anterior and pos- Catheter Ablation terior leaflets to cover the orifice area at the normal level, and Electrophysiological evaluation and radiofrequency ablation remodeling and reinforcement of the tricuspid annulus with a of symptomatic accessory pathway(s) should be performed prosthetic ring. This repair was reported in 191 patients when feasible in patients with Ebstein’s anomaly who have (meanϮSD age, 24Ϯ15 years).64 The early mortality rate was tachyarrhythmias. Catheter ablation has a lower success rate 9%, and the mean late survival rate at 20 years was 82%Ϯ5%.

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 282 Circulation January 16, 2007

leaflet serves as an opposing structure for coaptation of the reconstructed atrioventricular valve.65 Posterior annular plication without plication of the atrial- ized right ventricle and prophylactic cavopulmonary connec- tion are additional surgical options for nonneonatal Ebstein’s anomaly; however, no long-term follow-up data are avail- able.66 The benefit of adding a bidirectional cavopulmonary shunt after tricuspid valve repair or replacement to reduce right ventricular volume load in selected patients with Eb- stein’s anomaly is unclear. We use the bidirectional cavopul- monary shunt when the right ventricle is markedly dilated and functioning poorly. This allows a reduced volume load on the right ventricle and improves preload to the left ventricle. Left atrial and pulmonary artery pressures must be low for the shunt to provide hemodynamic benefit. In our experience, a modified Fontan procedure is very rarely required for patients with Ebstein’s anomaly who present after infancy. It has not been shown whether tricuspid valve repair or replacement has the better long-term outcome, nor is it known whether bioprostheses are preferable to mechanical prosthe- ses for tricuspid valve replacement in Ebstein’s anomaly. One study has suggested that valve repair is less durable in adults than in children.67 It is known that tricuspid bioprostheses in Ebstein’s anomaly have greater durability than in other cardiac positions, especially for pediatric patients; the mean rate of freedom from bioprosthesis replacement was 80.6Ϯ7.6% in a study with up to 17.8 years of follow-up (mean, 4.5 years).68 In a series of 294 patients with Ebstein’s anomaly, the difference in freedom from reoperation at 12 years for tricuspid valve repair versus replacement and for bioprosthetic versus mechanical valve prostheses was not significant.68 Currently, we prefer valve repair, when feasible, over valve replacement because repair has the potential of Figure 8. Diagram of the tricuspid valve repair technique cur- being more durable and avoids the potential complications of rently used for Ebstein’s anomaly. A, Two papillary muscles valve prostheses. Some of our early patients are doing well, arise from the free wall of the right ventricle, with short chordal Ͼ attachments to the leading edge of the anterior leaflet. The sep- free of reoperation 20 years after valve repair. We generally tal leaflet is diminutive and only a ridge of tissue. The posterior prefer bioprostheses over mechanical prostheses for Ebstein’s leaflet is not well formed and is adherent to the underlying anomaly, reserving the latter for those who are already taking endocardium. A small patent foramen ovale is present. B, C, anticoagulants for another indication.14 The method we cur- The base of each papillary muscle is moved toward the ventric- ular septum at the appropriate level with horizontal mattress rently use for tricuspid valve replacement in Ebstein’s anom- sutures backed with felt pledgets. The patent foramen ovale is aly is shown in Figure 9. closed by direct suture. D, The posterior angle of the tricuspid Although a decrease in atrial tachyarrhythmias after stan- orifice is closed by bringing the right side of the anterior leaflet 69 down to the septum and plicating the nonfunctional posterior dard repair of Ebstein’s anomaly generally has been noted, leaflet in the process. E, A posterior annuloplasty is performed we prefer to combine repair with directed antiarrhythmia to narrow the diameter of the tricuspid annulus. The coronary procedures.14,53,56 This can be accomplished without an in- sinus marks the posterior and leftward extent of the annulo- crease in operative mortality and with freedom from arrhyth- plasty. F, An anterior purse-string annuloplasty is performed to further narrow the tricuspid annulus. This annuloplasty stitch is mia recurrence in 75% of patients with atrial flutter or tied down over a 25-mm valve sizer in an adult to prevent tri- fibrillation and in up to 100% of patients with accessory cuspid stenosis. G, Completed repair that allows the anterior pathway–mediated tachycardia or atrioventricular nodal re- 63 leaflet to function as a monocuspid valve. From Dearani et al. 55,70 Used with permission of Mayo Foundation for Medical Educa- entry tachycardia. tion and Research. Symptomatic neonates with Ebstein’s anomaly have a poor prognosis. Marked cardiac enlargement, advanced echocar- It is unclear whether late problems will develop because of diographic severity score, cyanosis, and severe regurgitation devitalized tricuspid valve tissue related to reattachment. of the tricuspid valve all predict neonatal death without Another repair technique is characterized by reintegration surgery.19,71 Biventricular repairs in combination with correc- of the atrialized chamber into the right ventricular cavity tion of all associated cardiac defects are feasible, and midterm (called ventricularization). Ventricularization can be obtained results are good.71 Conversion to a single-ventricle approach by orthotopic transposition of the detached septal and poste- for symptomatic neonates also has been advocated.72 Results rior leaflets of the tricuspid valve. The reimplanted septal of tricuspid valve repair in young children have been reported

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Attenhofer Jost et al Ebstein’s Anomaly 283

Pacing Permanent pacing is required for 3.7% of patients with Ebstein’s anomaly, most commonly for atrioventricular block and rarely for sinus node dysfunction.75 In the presence of a tricuspid valve prosthesis, the ventricular lead for permanent DDD pacing usually is placed epicardially or through the coronary sinus or a cardiac vein. Alternatively, a previously placed transvenous ventricular lead may be sutured outside the prosthesis sewing ring at the time of valve replacement. Placement of a transvenous ventricular lead through a bio- prosthesis is effective but less desirable because of the possibility of propping open one of the valve cusps, thus creating regurgitation of the tricuspid valve. This complica- tion can be minimized by use of transesophageal echocardio- graphic monitoring to ensure that the lead lies safely in a commissure between the valve cusps.

Natural History and Long-Term Sequelae Figure 9. Diagram of technique for tricuspid valve replacement Several studies have reported on the natural history of in Ebstein’s anomaly. A, The valve suture line is placed on the Ebstein’s anomaly.49,76–78 The largest of these studies, report- atrial side of the membranous septum and atrioventricular (AV) node to avoid injury to the conduction system. The suture line is ing on 505 patients with Ebstein’s anomaly, was published also deviated cephalad to the tricuspid annulus posterolaterally Ͼ30 years ago.49 The study consisted primarily of patients when the tissues are thin to avoid injury to the right coronary between 1 and 25 years of age, with 67 patients Ͼ25 years artery. When sufficient distance between the coronary sinus and Ͻ Ͻ the AV node exists, the coronary sinus may be left on the atrial and only 35 patients 1 year of age. Of the infants 1 year side of the suture line. B, The sutures are tied with the heart of age, 72% were in heart failure, but for 81% of the others, perfused and beating to ensure that a conducted rhythm is pre- growth and development during infancy were average or 63 served. From Dearani et al. Used with permission of Mayo good. In addition, 71% of children and adolescents and 60% Foundation for Medical Education and Research. of adults were classified as New York Heart Association functional class I or II.49 A high mortality rate from conges- and demonstrate low early mortality and good durability at tive heart failure was noted during the first few months of life; late follow-up.73 subsequently, mortality plateaued at an average of 12% scattered uniformly throughout childhood and adolescence. Indications for Operation Of those who had surgical treatment, 54% did not survive the Observation alone is advised for asymptomatic patients with operation. This study was published before the echocardio- no right-to-left shunting and only mild cardiomegaly. Chil- graphic era and thus does not reflect the current Ebstein’s dren who have survived infancy generally do well for several anomaly population. years, and surgery can be postponed until symptoms appear, Of all neonates with the diagnosis of Ebstein’s anomaly, cyanosis becomes evident, or paradoxical emboli occur. 20% to 40% do not survive 1 month, and Ͻ50% survive to 5 Deliberations about an operation should begin if evidence of years.79–82 deterioration exists, such as progressive increase in right heart Celermajer et al19 reviewed 220 cases of Ebstein’s anomaly size, reduction in systolic function, or appearance of ventric- with 1 to 34 years of follow-up. Actuarial survival for all ular or atrial tachyarrhythmias. However, once symptoms live-born patients was 67% at 1 year and 59% at 10 years. progress to New York Heart Association functional class III Predictors of death were echocardiographic grade of severity at presentation (relative risk increased by 2.7 for each or IV, medical management has little to offer, surgical risks increase in grade), fetal presentation, and right ventricular increase, and operation is clearly indicated. A biventricular outflow tract obstruction. reconstruction is feasible for most patients. A 1.5 ventricle In rare cases, patients with Ebstein’s anomaly live Ͼ70 repair can be applied to the failing right ventricle. Heart years, but 1 reported patient died at 85 years of age.38 A transplantation is reserved for patients with severe biventric- reassessment of the prognosis of Ebstein’s anomaly is appro- ular dysfunction. priate in the current era of refined cardiovascular Some patients with cyanosis on exercise who have a shunt intervention. at the atrial level but only mild or moderate regurgitation of the tricuspid valve may benefit from device closure to Conclusions alleviate cyanosis and to prevent paradoxical emboli. Some Ebstein’s anomaly is a complex congenital anomaly with a centers commonly perform such procedures either as a staged broad anatomic and clinical spectrum. Management is com- approach or for long-term palliation.74 The degree of tricuspid plex and must be individualized. Precise knowledge about the valve regurgitation must be assessed carefully, however, different anatomic and hemodynamic variables, associated because closure of an atrial septal defect alone may worsen malformations, and management options is essential. Thus, it right ventricular dysfunction. is important that patients with Ebstein’s anomaly be evalu-

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 284 Circulation January 16, 2007

ated regularly by a cardiologist who has expertise in congen- 22. Benson DW, Silberbach GM, Kavanaugh-McHugh A, Cottrill C, Zhang ital heart disease.83–85 With better management strategies, it is Y, Riggs S, Smalls O, Johnson MC, Watson MS, Seidman JG, Seidman CE, Plowden J, Kugler JD. Mutations in the cardiac transcription factor hoped that survival of patients with this anomaly of all ages NKX2.5 affect diverse cardiac developmental pathways. J Clin Invest. will continue to improve. 1999;104:1567–1573. 23. Yatsenko SA, Yatsenko AN, Szigeti K, Craigen WJ, Stankiewicz P, Acknowledgment Cheung SW, Lupski JR. Interstitial deletion of 10p and atrial septal defect Editing, proofreading, and reference verification were provided by in DiGeorge 2 syndrome. Clin Genet. 2004;66:128–136. 24. Yang H, Lee CL, Young DC, Shortliffe M, Yu W, Wright JR. A rare case the Section of Scientific Publications, Mayo Clinic. of interstitial del(1)(p34.3p36.11) diagnosed prenatally. Fetal Pediatr Pathol. 2004;23:251–255. Disclosures 25. Danielson GK, Driscoll DJ, Mair DD, Warnes CA, Oliver WC Jr. None. Operative treatment of Ebstein’s anomaly. J Thorac Cardiovasc Surg. 1992;104:1195–1202. References 26. Brickner ME, Hillis LD, Lange RA. Congenital heart disease in adults: 1. Attenhofer Jost CH, Connolly HM, Edwards WD, Hayes D, Warnes CA, second of two parts. N Engl J Med. 2000;342:334–342. Danielson GK. Ebstein’s anomaly: review of a multifaceted congenital 27. Daliento L, Angelini A, Ho SY, Frescura C, Turrini P, Baratella MC, cardiac condition. Swiss Med Wkly. 2005;135:269–281. Thiene G, Anderson RH. Angiographic and morphologic features of the 2. Keith JD, Rowe RD, Vlad P. Heart Disease in Infancy and Childhood. left ventricle in Ebstein’s malformation. Am J Cardiol. 1997;80: New York, NY: Macmillan Co; 1958. 1051–1059. 3. Perloff JK. The Clinical Recognition of Congenital Heart Disease. 5th ed. 28. Benson LN, Child JS, Schwaiger M, Perloff JK, Schelbert HR. Left Philadelphia, Pa: WB Saunders; 2003. ventricular geometry and function in adults with Ebstein’s anomaly of the 4. Correa-Villasenor A, Ferencz C, Neill CA, Wilson PD, Boughman JA, for tricuspid valve. Circulation. 1987;75:353–359. the Baltimore-Washington Infant Study Group. Ebstein’s malformation of 29. Saxena A, Fong LV, Tristam M, Ackery DM, Keeton BR. Late nonin- the tricuspid valve: genetic and environmental factors. Teratology. 1994; vasive evaluation of cardiac performance in mildly symptomatic older 50:137–147. patients with Ebstein’s anomaly of tricuspid valve: role of radionuclide 5. Frescura C, Angelini A, Daliento L, Thiene G. Morphological aspects of imaging. J Am Coll Cardiol. 1991;17:182–186. Ebstein’s anomaly in adults. Thorac Cardiovasc Surg. 2000;48:203–208. 30. Castaneda-Zuniga W, Nath HP, Moller JH, Edwards JE. Left-sided 6. Edwards WD. Embryology and pathologic features of Ebstein’s anomaly. anomalies in Ebstein’s malformation of the tricuspid valve. Pediatr Prog Pediatr Cardiol. 1993;2:5–15. Cardiol. 1982;3:181–185. 7. Mann RJ, Lie JT. The life story of Wilhelm Ebstein (1836–1912) and his 31. Attenhofer Jost CH, Connolly HM, Warnes CA, O’Leary P, Tajik AJ, almost overlooked description of a congenital heart disease. Mayo Clin Pellikka PA, Seward JB. Noncompacted myocardium in Ebstein’s Proc. 1979;54:197–204. anomaly: initial description in three patients. J Am Soc Echocardiogr. 8. van Son JA, Konstantinov IE, Zimmermann V. Wilhelm Ebstein and 2004;17:677–80. Ebstein’s malformation. Eur J Cardiothorac Surg. 2001;20:1082–1085. 32. Attenhofer Jost CH, Connolly HM, O’Leary PW, Warnes CA, Tajik AJ, 9. Ebstein W. Ueber einen sehr seltenen Fall von Insufficienz der Valvula Seward JB. Left heart lesions in patients with Ebstein anomaly. Mayo tricuspidalis, bedingt durch eine angeborene hochgradige Missbildung Clin Proc. 2005;80:361–368. derselben. Arch Anat Physiol. 1866;238–255. 33. van Son JA, Danielson GK, Huhta JC, Warnes CA, Edwards WD, Schaff 10. Engle MA, Payne TP, Bruins C, Taussig HB. Ebstein’s anomaly of the HV, Puga FJ, Ilstrup DM. Late results of systemic atrioventricular valve tricuspid valve: report of three cases and analysis of clinical syndrome. replacement in corrected transposition. J Thorac Cardiovasc Surg. 1995; Circulation. 1950;1:1246–1260. 109:642–652. 11. Soloff LA, Stauffer HM, Zatuchni J. Ebstein’s disease: report of the first 34. Celermajer DS, Cullen S, Deanfield JE, Sullivan ID. Congenitally cor- case diagnosed during life. Am J Med Sci. 1951;222:554–561. rected transposition and Ebstein’s anomaly of the systemic atrioventric- 12. Dearani JA, Danielson GK. Congenital Heart Surgery Nomenclature and ular valve: association with aortic arch obstruction. J Am Coll Cardiol. Database Project: Ebstein’s anomaly and tricuspid valve disease. Ann 1991;18:1056–1058. Thorac Surg. 2000;69(suppl):S106–S117. 35. Dekker A, Mehrizi A, Vengsarkar AS. Corrected transposition of the 13. Lamers WH, Viragh S, Wessels A, Moorman AF, Anderson RH. For- great vessels with Ebstein malformation of the left atrioventricular valve: mation of the tricuspid valve in the human heart. Circulation. 1995;91: an embryologic analysis and two case reports. Circulation. 1965;31: 111–121. 119–126. 14. Dearani JA, Danielson GK. Ebstein’s anomaly of the tricuspid valve. In: 36. Silverman NH, Gerlis LM, Horowitz ES, Ho SY, Neches WH, Anderson Mavroudis C, Backer CL, eds. Pediatric Cardiac Surgery. 3rd ed. Phil- RH. Pathologic elucidation of the echocardiographic features of Ebstein’s adelphia, Pa: Mosby; 2003:524–536. malformation of the morphologically tricuspid valve in discordant atrio- 15. Lev M, Liberthson RR, Joseph RH, Seten CE, Eckner FA, Kunske RD, ventricular connections. Am J Cardiol. 1995;76:1277–1283. Miller RA. The pathologic anatomy of Ebstein’s disease. Arch Pathol. 37. Anderson KR, Danielson GK, McGoon DC, Lie JT. Ebstein’s anomaly of 1970;90:334–343. the left-sided tricuspid valve: pathological anatomy of the valvular mal- 16. Anderson KR, Zuberbuhler JR, Anderson RH, Becker AE, Lie JT. Mor- formation. Circulation. 1978;58(suppl):I-87–I-91. phologic spectrum of Ebstein’s anomaly of the heart: a review. Mayo Clin Proc. 1979;54:174–180. 38. Giuliani ER, Fuster V, Brandenburg RO, Mair DD. Ebstein’s anomaly: 17. Zuberbuhler JR, Allwork SP, Anderson RH. The spectrum of Ebstein’s the clinical features and natural history of Ebstein’s anomaly of the anomaly of the tricuspid valve. J Thorac Cardiovasc Surg. 1979;77: tricuspid valve. Mayo Clin Proc. 1979;54:163–173. 202–211. 39. Roberson DA, Silverman NH. Ebstein’s anomaly: echocardiographic and 18. Carpentier A, Chauvaud S, Mace L, Relland J, Mihaileanu S, Marino JP, clinical features in the fetus and neonate. J Am Coll Cardiol. 1989;14: Abry B, Guibourt P. A new reconstructive operation for Ebstein’s 1300–1307. anomaly of the tricuspid valve. J Thorac Cardiovasc Surg. 1988;96: 40. Oberhoffer R, Cook AC, Lang D, Sharland G, Allan LD, Fagg NL, 92–101. Anderson RH. Correlation between echocardiographic and morphological 19. Celermajer DS, Bull C, Till JA, Cullen S, Vassillikos VP, Sullivan ID, investigations of lesions of the tricuspid valve diagnosed during fetal life. Allan L, Nihoyannopoulos P, Somerville J, Deanfield JE. Ebstein’s Br Heart J. 1992;68:580–585. anomaly: presentation and outcome from fetus to adult. J Am Coll 41. MacLellan-Tobert SG, Driscoll DJ, Mottram CD, Mahoney DW, Wollan Cardiol. 1994;23:170–176. PC, Danielson GK. Exercise tolerance in patients with Ebstein’s anomaly. 20. Cohen LS, Friedman JM, Jefferson JW, Johnson EM, Weiner ML. A J Am Coll Cardiol. 1997;29:1615–1622. reevaluation of risk of in utero exposure to lithium. JAMA. 1994;271: 42. Driscoll DJ, Mottram CD, Danielson GK. Spectrum of exercise intol- 146–150. erance in 45 patients with Ebstein’s anomaly and observations on exercise 21. Emanuel R, O’Brien K, Ng R. Ebstein’s anomaly: genetic study of 26 tolerance in 11 patients after surgical repair. J Am Coll Cardiol. 1988; families. Br Heart J. 1976;38:5–7. 11:831–836.

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Attenhofer Jost et al Ebstein’s Anomaly 285

43. Shiina A, Seward JB, Edwards WD, Hagler DJ, Tajik AJ. Two- 66. Hancock Friesen CL, Chen R, Howlett JG, Ross DB. Posterior annular dimensional echocardiographic spectrum of Ebstein’s anomaly: detailed plication: tricuspid valve repair in Ebstein’s anomaly. Ann Thorac Surg. anatomic assessment. J Am Coll Cardiol. 1984;3:356–370. 2004;77:2167–2171. 44. Eustace S, Kruskal JB, Hartnell GG. Ebstein’s anomaly presenting in 67. Chen JM, Mosca RS, Altmann K, Printz BF, Targoff K, Mazzeo PA, adulthood: the role of cine magnetic resonance imaging in diagnosis. Clin Quaegebeur JM. Early and medium-term results for repair of Ebstein Radiol. 1994;49:690–692. anomaly. J Thorac Cardiovasc Surg. 2004;127:990–998. 45. Fisher MR, Lipton MJ, Higgins CB. Magnetic resonance imaging and 68. Kiziltan HT, Theodoro DA, Warnes CA, O’Leary PW, Anderson BJ, computed tomography in congenital heart disease. Semin Roentgenol. Danielson GK. Late results of bioprosthetic tricuspid valve replacement 1985;20:272–282. in Ebstein’s anomaly. Ann Thorac Surg. 1998;66:1539–1545. 46. Hebe J. Ebstein’s anomaly in adults: arrhythmias: diagnosis and thera- 69. Ahel V, Kilvain S, Rozmanic V, Taylor JF, Vukas D. Right atrial peutic approach. Thorac Cardiovasc Surg. 2000;48:214–219. reduction for tachyarrhythmias in Ebstein’s anomaly in infancy. Tex 47. Ho SY, Goltz D, McCarthy K, Cook AC, Connell MG, Smith A, Heart Inst J. 2001;28:297–300. Anderson RH. The atrioventricular junctions in Ebstein malformation. 70. Bockeria L, Golukhova E, Dadasheva M, Revishvili A, Levant A, Bazaev V, Rzaev F, Kakuchaya T. Advantages and disadvantages of one-stage Heart. 2000;83:444–449. and two-stage surgery for arrhythmias and Ebstein’s anomaly. Eur J Car- 48. Anderson KR, Lie JT. The right ventricular myocardium in Ebstein’s diothorac Surg. 2005;28:536–540. anomaly: a morphometric histopathologic study. Mayo Clin Proc. 1979; 71. Knott-Craig CJ, Overholt ED, Ward KE, Ringewald JM, Baker SS, 54:181–184. Razook JD. Repair of Ebstein’s anomaly in the symptomatic neonate: an 49. Watson H. Natural history of Ebstein’s anomaly of tricuspid valve in evolution of technique with 7-year follow-up. Ann Thorac Surg. 2002; childhood and adolescence: an international co-operative study of 505 73:1786–1792. cases. Br Heart J. 1974;36:417–427. 72. Starnes VA, Pitlick PT, Bernstein D, Griffin ML, Choy M, Shumway NE. 50. Smith WM, Gallagher JJ, Kerr CR, Sealy WC, Kasell JH, Benson DW Jr, Ebstein’s anomaly appearing in the neonate: a new surgical approach. Reiter MJ, Sterba R, Grant AO. The electrophysiologic basis and man- J Thorac Cardiovasc Surg. 1991;101:1082–1087. agement of symptomatic recurrent tachycardia in patients with Ebstein’s 73. Boston US, Dearani JA, O’Leary PW, Driscoll DJ, Danielson GK. Tri- anomaly of the tricuspid valve. Am J Cardiol. 1982;49:1223–1234. cuspid valve repair for Ebstein’s anomaly in young children: a 30-year 51. Graham TP Jr, Bricker JT, James FW, Strong WB. Twenty-Sixth experience. Ann Thorac Surg. 2006;81:690–696. Bethesda Conference: recommendations for determining eligibility for 74. Atiq M, Lai L, Lee KJ, Benson LN. Transcatheter closure of atrial septal competition in athletes with cardiovascular abnormalities: Task Force 1: defects in children with a hypoplastic right ventricle. Catheter Cardiovasc congenital heart disease. Med Sci Sports Exerc. 1994;26(suppl): Interv. 2005;64:112–116. S246–S253. 75. Allen MR, Hayes DL, Warnes CA, Danielson GK. Permanent pacing in 52. Reich JD, Auld D, Hulse E, Sullivan K, Campbell R, for the Pediatric Ebstein’s anomaly. Pacing Clin Electrophysiol. 1997;20:1243–1246. Electrophysiology Society. The Pediatric Radiofrequency Ablation Reg- 76. Kumar AE, Fyler DC, Miettinen OS, Nadas AS. Ebstein’s anomaly: istry’s experience with Ebstein’s anomaly. J Cardiovasc Electrophysiol. clinical profile and natural history. Am J Cardiol. 1971;28:84–95. 1998;9:1370–1377. 77. Bialostozky D, Horwitz S, Espino-Vela J. Ebstein’s malformation of the 53. Cappato R, Schluter M, Weiss C, Antz M, Koschyk DH, Hofmann T, tricuspid valve: a review of 65 cases. Am J Cardiol. 1972;29:826–836. Kuck KH. Radiofrequency current catheter ablation of accessory atrio- 78. Attie F, Rosas M, Rijlaarsdam M, Buendia A, Zabal C, Kuri J, Granados ventricular pathways in Ebstein’s anomaly. Circulation. 1996;94: N. The adult patient with Ebstein anomaly: outcome in 72 unoperated 376–383. patients. Medicine (Baltimore). 2000;79:27–36. 54. Chetaille P, Walsh EP, Triedman JK. Outcomes of radiofrequency 79. McElhinney DB, Salvin JW, Colan SD, Thiagarajan R, Crawford EC, catheter ablation of atrioventricular reciprocating tachycardia in patients Marcus EN, del Nido PJ, Tworetzky W. Improving outcomes in fetuses with congenital heart disease. Heart Rhythm. 2004;1:168–173. and neonates with congenital displacement (Ebstein’s malformation) or 55. Khositseth A, Danielson GK, Dearani JA, Munger TM, Porter CJ. dysplasia of the tricuspid valve. Am J Cardiol. 2005;96:582–586. Supraventricular tachyarrhythmias in Ebstein anomaly: management and 80. Celermajer DS, Dodd SM, Greenwald SE, Wyse RK, Deanfield JE. outcome. J Thorac Cardiovasc Surg. 2004;128:826–833. Morbid anatomy in neonates with Ebstein’s anomaly of the tricuspid 56. Greason KL, Dearani JA, Theodoro DA, Porter CB, Warnes CA, valve: pathophysiologic and clinical implications. J Am Coll Cardiol. 1992;19:1049–1053. Danielson GK. Surgical management of atrial tachyarrhythmias asso- 81. Yetman AT, Freedom RM, McCrindle BW. Outcome in cyanotic ciated with congenital cardiac anomalies: Mayo Clinic experience. Semin neonates with Ebstein’s anomaly. Am J Cardiol. 1998;81:749–754. Thorac Cardiovasc Surg Pediatr Card Surg Annu. 2003;6:59–71. 82. Celermajer DS, Cullen S, Sullivan ID, Spiegelhalter DJ, Wyse RK, 57. Schiebler GL, Adams P Jr, Anderson RC, Amplatz K, Lester RG. Clinical Deanfield JE. Outcome in neonates with Ebstein’s anomaly. J Am Coll study of twenty-three cases of Ebstein’s anomaly of the tricuspid valve. Cardiol. 1992;19:1041–1046. Circulation . 1959;19:165–187. 83. Deanfield J, Thaulow E, Warnes C, Webb G, Kolbel F, Hoffman A, 58. Barnard CN, Schrire V. Surgical correction of Ebstein’s malformation Sorenson K, Kaemmerer H, Thilen U, Bink-Boelkens M, Iserin L, Daliento with prosthetic tricuspid valve. Surgery. 1963;54:302–308. L, Silove E, Redington A, Vouhe P, Priori S, Alonso MA, Blanc J-J, Budaj 59. Hardy KL, May IA, Webster CA, Kimball KG. Ebstein’s anomaly: a A, Cowie M, Deckers J, Fernandez Burgos E, Lekakis J, Lindahl B, Mazzotta functional concept and successful definitive repair. J Thorac Cardiovasc G, Morais J, Oto A, Smiseth O, Trappe HJ, Klein W, Blömstrom-Lundqvist Surg. 1964;48:927–940. C, de Backer G, Hradec J, Parkhomenko A, Presbitero P, Torbicki A, for the 60. Theodoro DA, Danielson GK, Kiziltan HT, Driscoll DJ, Mair DD, Task Force on the Management of Grown Up Congenital Heart Disease of Warnes CA, Anderson BJ. Surgical management of Ebstein’s anomaly: a the European Society of Cardiology, ESC Committee for Practice 25-year experience. Circulation. 1997;96(suppl I):I–507. Abstract. Guidelines. Management of grown up congenital heart disease. Eur Heart J. 61. Danielson GK, Maloney JD, Devloo RA. Surgical repair of Ebstein’s 2003;24:1035–1084. anomaly. Mayo Clin Proc. 1979;54:185–192. 84. Therrien J, Dore A, Gersony W, Iserin L, Liberthson R, Meijboom F, 62. Dearani JA, Danielson GK. Tricuspid valve repair for Ebstein’s anomaly. Colman JM, Oechslin E, Taylor D, Perloff J, Somerville J, Webb GD, for Oper Tech Thorac Cardiovasc Surg. 2003;8:188–192. the Canadian Cardiovascular Society. CCS Consensus Conference 2001 63. Dearani JA, Danielson GK. Ebstein’s anomaly. In: Sellke FW, del Nido update: recommendations for the management of adults with congenital PJ, Swanson SJ, eds. Sabiston & Spencer Surgery of the Chest. 7th ed. heart disease: part I. Can J Cardiol. 2001;17:940–959. Philadelphia, Pa: Elsevier Saunders; 2005;2:2223–2235. 85. Landzberg MJ, Murphy DJ Jr., Davidson WR Jr, Jarcho JA, Krumholz 64. Chauvaud S, Berrebi A, d’Attellis N, Mousseaux E, Hernigou A, Carpentier HM, Mayer JE Jr, Mee RB, Sahn DJ, Van Hare GF, Webb GD, Williams A. Ebstein’s anomaly: repair based on functional analysis. Eur J Cardio- RG. Task Force 4: organization of delivery systems for adults with thorac Surg. 2003;23:525–531. congenital heart disease. J Am Coll Cardiol. 2001;37:1187–1193. 65. Ullmann MV, Born S, Sebening C, Gorenflo M, Ulmer HE, Hagl S. Ventricularization of the atrialized chamber: a concept of Ebstein’s KEY WORDS: atrium Ⅲ heart defects, congenital Ⅲ hemodynamics anomaly repair. Ann Thorac Surg. 2004;78:918–924. Ⅲ pathology Ⅲ pediatrics

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Images in Cardiovascular Medicine

Electrocardiogram, Echocardiography, and Magnetic Resonance Imaging Characteristics in Uhl’s Disease

Akli Otmani, MD; Laurent Leborgne, MD; Cédric Renard, MD; Houssam Bakkour, MD; Serge Quenum, MD; Christophe Tribouilloy, MD; Jean-Luc Rey, MD

hl’s disease is a very rare affection that was originally Transverse T1-weighted conventional spin-echo sequences Udescribed in 19521; few cases have been reported in the (Figure 3) showed an extremely dilated thin-wall right ventricle literature. It is caused by isolated right ventricular enlarge- and absence of trabeculation. The right atrium was dilated and ment and failure with partial or total absence of right hypertrophied. There was no fat signal in the right ventricular ventricular myocardium. It was first considered to be a wall as seen in arrhythmogenic ventricular dysplasia. The septal congenital development failure in the human embryo in its and left ventricular myocardium were preserved. Cine magnetic early stages. However, more recent publications incriminate resonance imaging (Figure 4, Movie III) demonstrated an apoptotic anomalies for right ventricular muscle agenesy.2–3 absence of right ventricular wall motion. Surgical correction has been attempted, without success.4 The patient was treated with ␤-blocker and diuretics and is Cardiac transplantation may be proposed when patients be- free of symptoms at 2 years. In this case, the ECG showed come refractory to medical treatment. right ventricular electrical activity, indicating persistent myo- We report the case of a 22-year-old woman without any cardial tissue. medical history who was referred to us for dyspnea and syncope. She did not complain of palpitation and had no manifestation of right heart failure. Twelve-lead ECG (Figure Disclosures 1) showed right atrial hypertrophy and an exaggerated, and None. not anteriorly described, fragmented ventricular depolariza- tion, which corresponded to very slow conduction in the right ventricle. Chest x-ray revealed enlargement of the cardiac silhouette with a prominent right-ventricle contour (Figure 2). References Neither supraventricular nor ventricular arrhythmia was ob- 1. Uhl HSM. A previously undescribed congenital formation of the heart: served during long-term ambulatory ECG monitoring. Trans- almost total absence of the myocardium of the right ventricle. Bull Johns Hopkins Hosp. 1952;91:197–205. thoracic (Movie I) and transesophageal echocardiography 2. Uhl HSM. Uhl’s anomaly revisited. Circulation. 1996;93:1483–1484. (Movie II) depicted a huge hypokinetic right ventricle with 3. James TN, Nichols MM, Sapire DW, DiPatre PL, Lopez SM. Complete severe tricuspid regurgitation. There was no right obstacle heart block and fatal right ventricular failure in an infant. Circulation. and no argument for Ebstein disease. The right atrium was 1996;93:1588–1600. 4. Abe T, Kuribayashi R, Sato M, Nioch S, Abe S. Congenital hypoplasia of also dilated as a consequence of the right cardiomyopathy and the right ventricular myocardium (Uhl’s anomaly): a case report and severe tricuspid regurgitation. review of the literature. J Cardiovasc Surg. 1973;14:431–437.

From the Cardio Vascular Center, University of Picardie, Amiens, France. The online-only Data Supplement, which contains 3 movies, can be found at http://circ.ahajournals.org/cgi/content/full/115/2/e11/DC1. Correspondence to Dr Akli Otmani, CHU Amiens SUD, Service de Cardiologie A, 80054 Amiens Cedex 01, France. E-mail [email protected] (Circulation. 2007;115:e11-e12.) © 2007 American Heart Association, Inc. Circulation is available at http://www.circulationaha.org DOI: 10.1161/CIRCULATIONAHA.106.630061

e11

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 e12 Circulation January 16, 2007

Figure 1. Twelve-lead ECG (A) and V2 lead (B) showing right atrial hypertrophy and fragmented QRS. Although the left ventricle is acti- vated normally, right ventricular depolarization is delayed and prolonged, with the right ventricle being the last to be activated. Total QRS duration is 180 ms.

Figure 2. Chest x-ray showing enlargement of the cardiac sil- houette with a prominent right-ventricle contour (arrow). Figure 3. Transverse T1-weighted spin-echo sequences show- ing very dilated thin-wall right ventricle and no trabeculation. The right atrium is dilated and hypertrophied. The left ventricle is normal.

Figure 4. Cine magnetic resonance imaging showing no right ventricular wall motion in short-axis sequences.

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Images in Cardiovascular Medicine

Magnetic Resonance Imaging of a Posttraumatic Myocardial Infarction and Ventricular Septal Defect With a Closure Device in Place

Stephanie L. Jun, MD; Nikhil K. Chanani, MD; Phillip Moore, MD; Charles B. Higgins, MD

16-year-old boy sustained blunt trauma as a passenger of the scar is substantially larger than the defect. Cardiac injuries Ain a high-speed motor vehicle accident 2 years earlier. from blunt chest trauma include acute myocardial infarction, Before the accident, he was healthy with no known cardiac valvular disruption, pericardial trauma, myocardial contusion, or disease or murmur. On admission after the accident, he was rupture. Ventricular septal defects attributable to blunt trauma noted to have a loud IV/VI holosystolic murmur along the left are rare and may be caused by extremely elevated intrathoracic sternal border. One month after admission, an echocardio- pressures or by compression of the heart between the sternum gram revealed a 1.5-cm muscular ventricular septal defect and the spine, resulting in septal contusion or rupture. An with a left-to-right shunt, enlargement of the right atrium and additional hypothesis implicates tearing of a coronary artery with ventricle, elevated pulmonary arterial pressure, and moderate thrombosis or spasm leading to an acute myocardial infarction tricuspid regurgitation (Figure 1 and Movie I). and postinfarction rupture.1,2 Traumatic ventricular septal de- Two years after the incident, the patient underwent percu- fects tend to occur in the muscular septum. Large and hemody- taneous closure of the ventricular septal defect with a 14-mm namically significant or symptomatic traumatic ventricular sep- Amplatzer septal occluder. The Qp:Qs ratio before device tal defects can be treated with surgical repair or percutaneously deployment was 2.8:1, and after closure, the Qp:Qs ratio was with a closure device.3,4 1.1:1. Cardiac magnetic resonance imaging was performed to evaluate the extent of the ventricular septal defect and any residual postprocedural left-to-right shunting. In the region of the muscular ventricular septal defect, Disclosures cardiac magnetic resonance imaging demonstrates delayed None. enhancement of the septum, which is consistent with readily identifiable myocardial infarction despite the adjacent metal- lic closure device. Additional cine images demonstrate a References small residual left-to-right shunt across the ventricular septal 1. Tiao GM, Gritffith PM, Szmuszkovicz JR, Mahour GH. Cardiac and great defect and akinetic-to-dyskinetic motion of the septal seg- vessel injuries in children after blunt trauma: an institutional review. ment of the left ventricle (Figures 2 and 3 and Movies II). J Pediatr Surg. 2000;35:1656–1660. Right ventricular enlargement and tricuspid regurgitation are 2. Genoni M, Jenni R, Turina M. Traumatic ventricular septal defect. Heart. 1997;78:316–318. also identified (Figure 4 and Movie II). 3. Rollins MD, Koehler RP, Stevens MH, Walsh KJ, Doty DB, Price RS, The current case shows that delayed-contrast cardiac mag- Allen TL. Traumatic ventricular septal defect: case report and review of netic resonance imaging could demonstrate the chronic in- the English literature since 1970. J Trauma. 2005;58:175–180. farct (fibrotic scar) despite the adjacent metallic closure 4. Schwalm S, Hijazi Z, Sugeng L, Lang R. Interventional pediatric cardi- ology: percutaneous closure of a post-traumatic muscular ventricular device; indeed, the enhanced segment is located between the septal defect using an Amplatzer duct occluder. J Invasive Cardiol. 2 limbs of the device. This case also illustrates that the extent 2005;17:100–103.

From the Department of Radiology (S.L.J., C.B.H.) and the Department of Pediatrics (N.K.C., P.M.), Division of Cardiology, University of California, San Francisco. The online-only Data Supplement, which contains 4 movies, can be found at http://circ.ahajournals.org/cgi/content/full/115/2/e13/DC1. Correspondence to Stephanie L. Jun, MD, Department of Radiology, 505 Parnassus Ave, Suite M-391, San Francisco, CA 94143–0628. E-mail [email protected] (Circulation. 2007;115:e13-e15.) © 2007 American Heart Association, Inc. Circulation is available at http://www.circulationaha.org DOI: 10.1161/CIRCULATIONAHA.106.631275

e13

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 e14 Circulation January 16, 2007

Figure 3. Delayed contrast enhancement in the horizontal long- axis plane demonstrates a larger extent of the septal infarct.

Figure 1. Delayed contrast enhancement in the short-axis plane demonstrates both sides of the closure device, spanning the ventricular septal defect, and delayed enhancement between the limbs of the device, corresponding to chronic infarct.

Figure 2. Delayed contrast enhancement in the short-axis plane demonstrates that the extent of the infarct is substantially larger than the ventricular septal defect.

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Jun et al MRI of a Posttraumatic Ventricular Septal Defect e15

Figure 4. Still image from cine magnetic resonance imaging in the axial plane demonstrates dyskinetic septal motion and a larger sig- nal void from the closure device than is seen on the delayed-enhancement images.

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Images in Cardiovascular Medicine

Sinus of Valsalva Aneurysm With Right Ventricular Outflow Tract Obstruction Evaluation With Doppler, Real-Time 3-Dimensional and Contrast Echocardiography

Eli V. Gelfand, MD; Dorota Bzymek, RDCS; Michael T. Johnstone, MD

78-year-old man presented with new exertional dys- ischemia attributable to coronary compression.1 Clinically sig- Apnea. Physical examination demonstrated a parasternal nificant right ventricular outflow tract obstruction from a sinus thrill that was associated with a systolic murmur, and mild of Valsalva aneurysm is uncommon, with only 2 previously edema of the lower extremities. Transthoracic 2-dimensional published cases found in the literature.2,3 Two-dimensional and real-time 3-dimensional echocardiography demonstrated Doppler echocardiography easily makes this diagnosis, but a 5.3ϫ4.5-cm aneurysm of the right sinus of Valsalva real-time 3-dimensional echocardiography allows for complete (Figures 1 and 2 and Movie I). There was moderate to severe visualization of the aneurysm neck and its relation to the aortic regurgitation. The aneurysm protruded into the right surrounding structures, including the aortic valve. To our knowl- ventricular outflow tract, and color Doppler showed turbulent edge, preoperative evaluation of the unruptured aneurysm of the flow around the aneurysm with a peak systolic pressure sinus of Valsalva with transthoracic real-time 3-dimensional gradient of 49 mm Hg (Figure 3 and Movie II). Imaging after echocardiography has not previously been described. The use of intravenous injection of perflutren ultrasound contrast agent intravenous echo contrast in this case allowed us to prove that demonstrated the partition of the right ventricle by the the structure in the right ventricular outflow tract was vascular. aneurysm (Figure 4 and Movie III). When injected into a Echo contrast consists of lipid microspheres filled with an inert peripheral vein, the contrast was seen to opacify the right fluorocarbon gas. Unlike relatively large bubbles of agitated heart, the left atrium, and then the left ventricle. Finally, the saline, the smaller bubble size of dedicated ultrasound contrast contrast was seen to fill the ascending aorta and the body of allows for transpulmonary passage of the agent and imaging of the aneurysm (Figure 4 and Movie III). The timing of contrast the left-sided structures. appearance proved that the aneurysm communicated with the aorta but not with the pulmonary artery. The patient underwent elective repair of the aortic aneu- Disclosures rysm with a prosthetic patch. The patient’s aortic valve was Dr Gelfand serves on the Speaker’s Bureau for Pfizer. Dr Johnstone replaced with a pericardial bioprosthesis. Intraoperatively, the served on the Speaker’s Bureau for Wyeth and AstraZeneca. Ms. aneurysm neck was shown to be distinct from the right Bzymek has no conflicts of interest to disclose. coronary artery ostium, and therefore, button excision and reimplantation of the right coronary artery were not required. He was discharged home on the sixth postoperative day, and on a subsequent outpatient visit reported that his shortness of References breath had significantly improved. 1. Feldman DN, Roman MJ. Aneurysms of the sinuses of Valsalva. Car- diology. 2006;106:73–81. Aneurysms of the sinus of Valsalva are rare anomalies that 2. Kiefaber RW, Tabakin BS, Coffin LH, Gibson TC. Unruptured sinus of are usually diagnosed after an acute rupture into an adjacent Valsalva aneurysm with right ventricular outflow obstruction diagnosed cardiac structure. Prior to rupture, aneurysms of the sinus of by two-dimensional and Doppler echocardiography. J Am Coll Cardiol. Valsalva may present with conduction-system abnormalities 1986;7:438–442. 3. Thankachen R, Gnanamuthu R, Doshi H, Shukla V, Korula RJ. attributable to erosion into the interventricular septum, throm- Unruptured aneurysm of the sinus of Valsalva presenting with right boembolism originating in the aneurysm sac, and myocardial ventricular outflow obstruction. Tex Heart Inst J. 2003;30:152–154.

From the Cardiovascular Division, Beth Israel Deaconess Medical Center and Harvard Medical School, Boston, Mass. The online-only Data Supplement, which contains movies, can be found at http://circ.ahajournals.org/cgi/content/full/115/2/e16/DC1. Correspondence to Dr. Eli V. Gelfand, Cardiovascular Division, 330 Brookline Ave, RW-453, Boston, MA 02215. E-mail [email protected] (Circulation. 2007;115:e16-e17.) © 2007 American Heart Association, Inc. Circulation is available at http://www.circulationaha.org DOI: 10.1161/CIRCULATIONAHA.106.641043

e16

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Gelfand et al Sinus of Valsalva Aneurysm With RVOT Obstruction e17

Figure 1. Parasternal short-axis view of right sinus of Valsalva aneurysm. AoV indicates aortic valve; An, body of the aneu- rysm; PA, main pulmonary artery; RA, right atrium; and RV, right Figure 3. Color and continuous-wave Doppler demonstration of ventricle. the right ventricular outflow tract obstruction. LV indicates left ventricle; RV, right ventricle; and An, body of the aneurysm.

Figure 4. Parasternal long-axis images after administration of intravenous echo contrast, demonstrating the border between the aneurysm and the body of the right ventricle. AscAo indi- Figure 2. Real-time 3-dimensional view of the aneurysm neck, cates ascending aorta; An, body of the aneurysm; LA, left as viewed from the left. AscAo indicates ascending aorta; RCC, atrium; LV, left ventricle; and RV, right ventricle. right coronary aortic valve cusp.

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Correspondence

Letter by Ben-Dov and Bursztyn Regarding Because clinic BP was used to screen and guide the treatment Article, “Role of Diuretics in the Prevention of of ALLHAT subjects, it is possible that in patients achieving Heart Failure: The Antihypertensive and similar clinic BP, ambulatory BP may have differed according to Lipid-Lowering Treatment to Prevent Heart the treatment assigned. This, in turn, could lead to dissimilar outcomes, including differences in heart failure incidence. Per- Attack Trial” haps a future large-scale hypertension trial should also involve To the Editor: ambulatory BP monitoring, at least in a subset of patients. In their article, Davis et al1 for the Antihypertensive and Lipid- Lowering Treatment to Prevent Heart Attack Trial (ALLHAT) Collaborative Research Group characterize the differences in heart failure incidence according to treatment assignment in patients enrolled to ALLHAT. On the basis of time-dependent analyses and Disclosures comparison with previous trials, they conclude that blood pressure None. (BP) differences alone cannot account for the superior prevention of heart failure afforded by chlorthalidone compared with lisinopril Iddo Z. Ben-Dov, MD, MSc and amlodipine (and also doxazosin2). We suggest that ambulatory Nephrology and Hypertension Services BP differences may bridge the gap between the achieved clinic BP Hadassah–Hebrew University Medical Center and the ensuing outcome. In both treated and untreated hypertensive Ein-Karem Campus, subjects, ambulatory BP has been shown to predict mortality Jerusalem, Israel beyond clinic BP. Furthermore, nonclassic BP parameters, such as Michael Bursztyn, MD the sleep-related dip and the white-coat phenomenon, have emerged Department of Internal Medicine as prognosticators. Hadassah–Hebrew University Medical Center For example, in our cohort of subjects referred for 24-hour Mount-Scopus Campus, ambulatory monitoring, adjustment for baseline covariates dis- Jerusalem, Israel closed higher ambulatory pressures during sleep in subjects treated with ␣-blockers, whereas their manual BP tended to be lower.3 Assuming the same is true in ALLHAT subjects, this 1. Davis BR, Piller LB, Cutler JA, Furberg C, Dunn K, Franklin S, Goff D, Leenen F, Mohiuddin S, Papademetriou V, Proschan M, Ellsworth A, Golden J, Colon may partly explain the superiority of chlorthalidone despite the P, Crow R; Antihypertensive and Lipid-Lowering Treatment to Prevent Heart metabolic advantages of doxazosin. Attack Trial Collaborative Research Group. Role of diuretics in the prevention of In addition, it has been shown that volume-overload states heart failure: the Antihypertensive and Lipid-Lowering Treatment to Prevent such as salt-sensitive hypertension, diabetic nephropathy, and Heart Attack Trial. Circulation. 2006;113:2201–2210. congestive heart failure are characterized by a nondipping BP 2. Major cardiovascular events in hypertensive patients randomized to dox- pattern.4 Diuretics have been convincingly shown to reverse azosin vs chlorthalidone: the antihypertensive and lipid-lowering nondippers to dippers in such conditions.5 Conversely, we and treatment to prevent heart attack trial (ALLHAT). ALLHAT Collabo- others have shown that ␣-blockers, which cause volume over- rative Research Group. JAMA. 2000;283:1967–1975. load, are associated with decreased dipping.3 In our cohort, the 3. Ben-Dov IZ, Ben-Arie L, Mekler J, Bursztyn M. How should patients covariates’ adjusted odds ratio for nondipping among ␣-blocker treated with alpha-blockers be followed? Insights from an ambulatory blood pressure monitoring database. J Hypertens. 2006;24:861–865. treated subjects was 1.7 (95% confidence interval, 1.1 to 2.8), ␤ 4. Ben-Dov IZ, Bursztyn M. Ambulatory blood pressure monitoring may be whereas -blockers, calcium blockers, and renin–angiotensin a prognostic marker in people with heart failure: commentary 1. Evid antagonists were not associated with dipping. Diuretics conveyed Based Cardiovasc Med. 2005;9:197–199. a reduced odds ratio for nondipping, at 0.65 (95% confidence 5. Uzu T, Kimura G. Diuretics shift circadian rhythm of blood pressure from interval, 0.45 to 0.94).3 nondipper to dipper in essential hypertension. Circulation. 1999;100:1635–1638.

(Circulation. 2007;115:e18.) © 2007 American Heart Association, Inc. Circulation is available at http://www.circulationaha.org DOI: 10.1161/CIRCULATIONAHA.106.639617

e18

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Correspondence

Response to Letter Regarding Article, “Role of Jeffrey A. Cutler, MD, MPH Diuretics in the Prevention of Heart Failure: The Michael Proschan, PhD Antihypertensive and Lipid-Lowering Treatment to National, Heart, Lung, and Blood Institute Prevent Heart Attack Trial (ALLHAT)” Bethesda, Md We thank Drs Ben-Dov and Bursztyn for their thoughtful letter Curt Furberg, MD, PhD regarding our article1 and the role of clinic versus ambulatory blood pressure in assessing the relative impact of antihyperten- David Goff, MD, PhD sive treatments on clinical outcomes, especially heart failure. The Wake Forest University School of Medicine Antihypertensive and Lipid-Lowering Treatment to Prevent Winston-Salem, NC Heart Attack Trial (ALLHAT) was designed to allow for Stanley Franklin, MD recruitment and retention of a large number of patients, with University of California at Irvine minimal burden to the providers and study participants and a cost Irvine, Calif structure that could be achieved and sustained over a long period of follow-up. As such, there was much information not collected Frans Leenen, MD, PhD within ALLHAT that could have provided insight into the University of Ottawa Heart Institute study’s results.2 Recording of ambulatory blood pressure on a Ottawa, Canada subset of the ALLHAT patients, in addition to clinic blood pressure, might have revealed greater blood pressure differences Syed Mohiuddin, MD than were observed. One such ambulatory blood pressure sub- Creighton Cardiac Center study was conducted within the framework of the Health Out- Omaha, Neb comes Prevention Evaluation Trial.3 This study showed greater Vasilios Papademetriou, MD falls in ambulatory blood pressure, especially at night, than office blood pressure for those on ramipril versus those on placebo. Veterans Affairs Medical Center Therefore, we agree that future large-scale hypertension trials Washington, DC comparing treatments should include a substudy of ambulatory Allan Ellsworth, PharmD blood pressure monitoring. University of Washington Seattle, Wash Disclosures Dr Davis has worked as a consultant for Takeda, Merck, and John Golden, MD GlaxoSmithKline. Dr Franklin has served on speakers’ bu- Penderbrook Medical Center reaus for Boehringer Ingelheim, Merck, and Bristol-Myers Fairfax, Va Squibb; as an expert witness for La Follette, Johnson, De Pedro Colon, MD Haas, Fesler & Ames, Los Angeles, Calif, and Gordon, Centro Cardiovascular de Caguas Thomas, Honeywell, Malanca, Peterson, & Daheim LLP, Caguas, Puerto Rico Seattle, Wash; and as a consultant for AtCor Medical, Inc. Dr. Leenen has received honoraria from Pfizer; has ownership Richard Crow, MD interest in Bristol-Myers Squibb, Merck, Johnson and John- University of Minnesota son, and Schering Plough; and has worked as a consultant for Minneapolis, Minn Pfizer. Dr Mohiuddin has worked on research grants funded by Health Future Foundation, Nebraska Tobacco Settlement Biomedical Research Development Grants, and AstraZeneca; The authors listed above represent the ALLHAT Collaborative has served on the speakers’ bureau for AstraZeneca; has Research Group. received honoraria from the American College of Cardiology and AstraZeneca; has ownership interest in Pfizer, Johnson 1. Davis BR, Piller LB, Cutler JA, Furberg C, Dunn K, Franklin S, Goff D, and Johnson, and Abbott Laboratories; and has served as a Leenen F, Mohiuddin S, Papademetriou V, Proschan M, Ellsworth A, consultant for Pfizer and AstraZeneca. Dr Papademetriou has Golden J, Colon P, Crow R; Antihypertensive and Lipid-Lowering received research grants from AstraZeneca; has worked on the Treatment to Prevent Heart Attack Trial Collaborative Research Group. speakers’ bureau for AstraZeneca; and has received honoraria Role of diuretics in the prevention of heart failure: the Antihypertensive from AstraZeneca. The other authors report no conflicts of and Lipid-Lowering Treatment to Prevent Heart Attack Trial. Circu- interest. lation. 2006;113:2201–2210. 2. Davis BR, Furberg CD, Wright JT, Cutler JA, Whelton P; ALLHAT Barry R. Davis, MD, PhD Collaborative Research Group. ALLHAT: setting the record straight. Ann Linda B. Piller, MD, MPH Intern Med. 2004;141:39–46. Kay Dunn, PhD 3. Svensson P, de Faire U, Sleight P, Yusuf S, Östergren J. Comparative University of Texas School of Public Health effects of ramipril on ambulatory and office blood pressures: a HOPE Houston, Tex substudy. Hypertension. 2001;38:e28–e32.

(Circulation. 2007;115:e19.) © 2007 American Heart Association, Inc. Circulation is available at http://www.circulationaha.org DOI: 10.1161/CIRCULATIONAHA.106.651539

e19

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Acknowledgment of Reviewers

The editors express appreciation to the following referees who served from January 1, 2006 - June 30, 2006.

Einari Aavik Amjad AlMahameed Birgit Assmus Jean-Pierre Bassand Antonio Abbate Martin A. Alpert Carmela Asteria Craig T. Basson Jinnette Dawn Abbott Pierre Amarenco Brad Astor Wendy W. Batenburg Hazem Abdulhussien John A. Ambrose Thanos Athanasiou Eric R. Bates Jun-ichi Abe Giuseppe Ambrosio Dianne L. Atkins Philip M.W. Bath Yoshiteru Abe Khawaja Afzal Ammar Håvard Attramadal Deante Naquae Battle E. Dale Abel Inder S. Anand Andrew M. Atz Kenneth A. Bauer George S. Abela Frank Andersohn Heinrich J. Audebert Johann Bauersachs Benjamin S. Abella Garnet L. Anderson Angelo Auricchio Kenneth Lee Baughman Alexandre Abizaid H. Vernon Anderson Gerard P. Aurigemma Helmut Baumgartner M. Roselle Abraham Jeffrey L. Anderson Pablo Avanzas Ralf W. Baumgartner Pierre Abraham Mark E. Anderson Boaz Avitall William Baumgartner Theodore P. Abraham Neil Anderson Abraham Aviv Jeroen J. Bax Charles S. Abrams Page A.W. Anderson Alberto Avolio B. Timothy Baxter Hugues Abriel Peter G. Anderson Hakan Ay Gary F. Baxter Elias Abrutyn Robert H. Anderson Clerio Francisco Azevedo Antoni Bayés de Luna Vinod Achan Todd J. Anderson Ricardo Azziz W. Scott Beattie Stephan Achenbach Arne Kristian Andreassen Heinrich F. Becker Tayfun Acil Maria Grazia Andreassi Fritz H. Bach Karsten Becker Michael Acker Felicita Andreotti Emile Bacha Richard C. Becker Michael J. Ackerman Vicente Andres Robert J. Bache Frank Beckers Volker Adams Annalisa Angelini Carl Backer Joshua A. Beckman Philip A. Ades Djillali Annane Peter Backx Solomon Behar Marina Afanasyeva Brian H. Annex Larry Baddour Alexa Beiser Vahid Afshar-Kharghan Jack Ansell Michael Bader Bernard Belhassen Ali R. Afzal Frederic Anselme Juan Jose Badimon Jonathan N. Bella Sachin Agarwal Jovan P. Antovic Stephen F. Badylak George A. Beller Salvatore Agati Charles Antzelevitch Emilia Bagiella Marek Belohlavek Pierfrancesco Agostoni Piero Anversa Hossein Baharvand David G. Benditt Eustachio Agricola Toshihisa Anzai Donald S. Baim Frank M. Bengel Shaheeda Ahmed Fred S. Apple Alison E. Baird Kristina Bengtsson Enrico Aidala Annelies E. Aquarius O. Bakker Ivor J. Benjamin Elena Aikawa Andrew E. Arai George L. Bakris Goetz Benndorf Masanori Aikawa Aloir Queiroz Araujo Stephan Baldus Ralf Benndorf William C. Aird Eloisa Arbustini J. Timothy Baldwin Joel S. Bennett Nadine Ajzenberg Stephen L. Archer Christie M. Ballantyne Neal Benowitz Fadi G. Akar Moshe Arditi Jean-Luc Balligand D. Woodrow Benson Shahab A. Akhter Ross Arena Aigul T. Baltabaeva Pascal A. Berdat Olakunle O. Akinboboye Edgar Argulian Ko Bando Robert A. Berg Joseantonio Josean Robert A. Ariens Rupak K. Banerjee Alan K. Berger Alarconduque Gary C. Armitage Eddy Barasch Peter B. Berger Petar Alaupovic Ehrin Johnson Armstrong Giuseppe Barbaro Rudolf Berger John J. Albers Paul W. Armstrong Silvia Stella Barbieri Zekarias Berhane Christine M. Albert Donna K. Arnett John C. Barefoot Bradford C. Berk Jeffrey Albert Alice Arnold Philip M. Barger Judith A. Berliner Michelle Albert Margaret A. Arstall Susan M. Barman Stephen A. Bernard Gabriel Aldea Helen M. Arthur Adrian G. Barnett Michael C. Berndt Michael H. Alderman Thiruma Arumugam Jose A. Barrabes Daniel Bernstein Alexey N. Aleshin Michael Arzt William H. Barry Donald M. Bers Barbara Alexander Salman Ashfaq Robyn J. Barst Giuseppe S. Berton M. Yvonne Alexander Hiroshi Ashikaga Philip Barter Michel E. Bertrand Mark E. Alexander Samuel J. Asirvatham Malgorzata Zofia Bartnik Patricia J.M. Best Lindsey D. Allan Christopher D. Askew Matthias Barton Connie R. Bezzina Hugh D. Allen Peter Aspelin Paul J.R. Barton Vikas Bhalla Maurits A. Allessie Abid R. Assali Peter Bärtsch Dwaipayan Bharadwaj Matthew A. Allison Folkert W. Asselbergs Riyaz Bashir Saroja Bharati Kevin C. Allman Gerd Assmann Theodore A. Bass Aruni Bhatnagar

e20

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Acknowledgment of Reviewers e21

Deepak Bhatnagar Paul F. Bray Umberto Campia Kai Chen Deepak L. Bhatt Roger E. Breitbart Hannia Campos Peng-Sheng Chen Italo Biaggioni Ole A. Breithardt Jaume J.C.R. Candell Shih-Ann Chen Giuseppe Bianchi Robert T. Brennan Ann E. Canfield Yiu-Fai Chen Giorgio M. Biasi Kate M. Brett Christopher P. Cannon Bin Cheng Luigi Marzio Biasucci Thomas Breymann Richard O. Cannon Glenn M. Chertow Kirsten Bibbins-Domingo Mark E. Brezinski Charles E. Canter Shu Chien Lawrence F. Bielak John Bridge John M. Canty John S. Child J. Thomas Bigger David Brieger Yihai Cao William M. Chilian Diane E. Bild Michael R. Bristow Zemin Cao Michael T. Chin George E. Billman Gernot Brockmann Louis Caplan Benedetta Diamante Chiodini Ofer Binah Bruce R. Brodie Thomas P. Cappola Julio A. Chirinos Bernd R. Binder Sergey V. Brodsky Alessandro M. Capponi Randolph W. Chitwood, Jr. Joyce Bischoff Jana Brodszki Joseph A. Caprini Ray C-J Chiu Nanette H. Bishopric Ulrich Broeckel Blase A. Carabello Eunyoung Cho John A. Bittl Robert D. Brook Stephane G. Carlier Aram V. Chobanian Vera Bittner Maria Mori Brooks Christopher S. Carlson Anand Chockalingam Kristina Bjorklund B. Greg Brown Mark David Carlson Michael A. Choma Henry W. Blackburn Clive M. Brown Peter Carmeliet Zhao-Zhong Chong Eugene H. Blackstone Peter Brubaker Mercedes R. Carnethon Roger Chou Stefan Blankenberg Martina Brueckmann Robert M. Carney Chi-Ming Chow W. Matthijs Blankesteijn Josep Brugada Oscar A. Carretero Phil Chowienczyk Peter C. Block Pedro Brugada John D. Carroll Maryanne R. Chrisant James A. Blumenthal Bruce H. Brundage M.D. Carroll Karkos D. Christos Roger S. Blumenthal Eric J. Brunner Joseph P. Carrozza Sarah Chua Christoph Bode John D. Brunzell Barbara Casadei Sumeet S. Chugh William E. Boden Robert M. Bryan Juan Pablo Casas Mina Chung Rosemary S. Bubien Manfred Boehm Franco Casazza Moo K. Chung John Buckwalter Matthijs Boekholdt Lisa A. Cassis Francesco Cipollone Matthew J. Budoff Cornelis Boersma Lawrence Castellani Leslie Cirome T. Jared Bunch Rainer H. Boger Martha Cathcart Kieran Clarke Gregory L. Burke William Boisvert Mark J. Caulfield Robert Clarke John C. Burnett Ann Bolger Alison C. Cave John G.F. Cleland Jane C. Burns Roberto Bolli Frank Cecchin Ton J. Cleophas Ivo Buschmann Marvin O. Boluyt David S. Celermajer Alexander Clowes Rudi Busse Raoul Bonan Bojan Cercek William T. Clusin Eckhart Bu¨ssemaker Nikolaos Bonaros Matteo Cesari Stuart M. Cobbe Peter M. Buttrick Robert O. Bonow Hugues Chabriat William A. Coetzee John Byrne George W. Booz Claudia U. Chae Christopher S. Coffey Graham B. Byrnes Jeffrey S. Borer Bernard R. Chaitman Jay D. Coffman Marcel Borgers Howard Cabral Subrata Chakrabarti David J. Cohen Martin Borggrefe Brian S. Caffo John Chalmers Marc Cohen John Boscardin Hua (Linda) Cai Hunter Clay Champion Michael V. Cohen Rene´ M. Botnar Jianming Cai Chang-Chuan Chan Richard A. Cohen Tomaso Bottio Weijing Cai Krishnaswamy Chandrasekaran Jay N. Cohn Elias H. Botvinick Michael E. Cain Anthony Christopher Chang Lawrence H. Cohn Mark M. Boucek Paolo Calabro Chih-Jen Chang Steven D. Colan Chantal M. Boulanger Antonio Maria Calafiore Ruey-Kang R. Chang Désiré Collen Anne Bouloumie Angelino Calderone Keith M. Channon Alan R. Collins Henri M. Bounameaux James H. Caldwell Richard Chappell Sathiakar Paul Collison Martial G. Bourassa David A. Calhoun Panithaya Chareonthaitawee Antonio Colombo Daniel F. Bowen-Pope Robert M. Califf Chris J. Charles David Colquhoun Neil E. Bowles Giuseppina Caligiuri Israel F. Charo Gianluigi Condorelli Penelope A. Boyden Hugh Calkins John C. Chatham John E. Connett Mark R. Boyett David J. Callans Kanu Chatterjee Michael S. Conte T. Douglas Bradley Francois A. Cambien Sarwat Chaudhry C. Richard Conti Ralf P. Brandes Duke Cameron Faisal H. Cheema Thomas Cook Ruediger C. Braun-Dullaeus Vicky A. Cameron Melvin D. Cheitlin John P. Cooke Joel Braunstein Paolo G. Camici Alex F. Chen Joshua M. Cooper Eugene Braunwald A. John Camm Chunguang Chen Leslie T. Cooper George A. Bray Malcolm Campbell Frederick Y. Chen Mark E. Cooper

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 e22 Acknowledgment of Reviewers

Richard S. Cooper Tirone E. David Augusto Filippo Di Jay M. Edelberg Giuseppe Coppolino Amy J. Davidoff Castelnuovo Igor R. Efimov Domenico Corrado Karina W. Davidson Carlo Di Mario Oliver Eickelberg James P. Corsetti Michael H. Davidson Vincenzo Di Marzo Benjamin W. Eidem Roberto Corti Victor G. Davila-Roman Timm Dickfeld John W. Eikelboom Joseph S. Coselli Roger B. Davis Sean P. Didion Mark J. Eisenberg Francesco Cosentino Buddhadeb Dawn Gerhard-Paul Diller Mickey S. Eisenberg Andry F. Costa Jeffrey D. Dawson Rodney J. Dilley Eric L. Eisenstein Marco A. Costa Ian N.M. Day Vasken Dilsizian John A. Eisman Salvatore Costa Sharlene M. Day John P. DiMarco Daniel T. Eitzman John M. Costello Dirk De Bacquer Stefanie Dimmeler Amir Elami Lisa C. Costello-Boerrigter Jacques de Bakker Frank A. Dinenno Khalid Elased Patrick Anthony Coughlin Mark de Caestecker Michael Diringer John A. Elefteriades David Couper Raffaele De Caterina Luc Djousse Abdou Elhendy James W. Covell Sarah D. de Ferranti Dobromir Dobrev Jonathan L. Eliason Adrian Covic Gaetano M. De Ferrari Douglas W. Dockery Michael ‘Tex’ Eliasziw Douglas B. Cowan Pim J. de Feyter Kimberly Dodge-Kafka Uri Elkayam Allen W. Cowley Judy B. de Haan Christoph Dodt Mitchell S. Elkind Dermot Cox Jan W. de Jong Clare M. Dollery Kenneth A. Ellenbogen Ciprian M. Crainiceanu Paulus T.V.M. de Jong Michael J. Domanski Myrvin Ellestad Jill P. Crandall Dominique P. de Kleijn Anna F. Dominiczak C. Gregory Elliott James D. Crapo Chris L. de Korte J. Kevin Donahue Perry M. Elliott Michael H. Crawford James A. de Lemos Rosario Donato Randall P. Ellis Sybil Crawford Michel de Lorgeril Jing-Fei Dong Stephen G. Ellis Filippo Crea Moniek P.M. de Maat Annie Dore Masao Endoh Mark A. Creager Jo G. de Mey Alessandro Doria Stefan Engeli Alain G. Cribier Albert de Roos Andrea Doria Jan E. Engvall Mark L. Entman Michael H. Criqui Luc de Saint Martin Pedro D’Orléans-Juste Janet B. Croft Gerald W. Dorn II Laurence M. Epstein Leon J. De Windt Kevin D. Croft Robert Neil Doughty Stephen E. Epstein Barbara J. Deal John Robert Crouse Pamela S. Douglas Raimund R. Erbel Deborah Dean Richard S. Crow James M. Downey Benedek Erdos John Eric Deanfield J. Kennedy Cruickshank Douglas E. Drachman Margaretha Eriksson Arjun Deb Jian Cui Luciano Ferreira Drager Paul Ernsberger G. William Dec Bruce F. Culleton Helmut Drexler Sabine Ernst Jeanne M. De Cara Anne B. Curtis Daniel L. Dries Georg Ertl John R. Dedman Jeptha P. Curtis Junbao Du Denis Escande Joep C. Defesche Linda K. Curtiss Dayue Duan Thomas Eschenhagen Christoph Dehnert Mary Cushman Anne M. Dubin Ahmad Esmaillzadeh Ranjan Deka Daniele M. Cusi Anique Ducharme Mark A. Espeland Pedro J. Del Nido Donald E. Cutlip Samuel C. Dudley Mohammed Rafique Essop Etienne Delacre´taz Antoine Dunac N.A. Mark Estes III Patrice Delafontaine Antoine DaCosta Brian W. Duncan Susan A. Everson-Rose Claude Delcayre Michael W. Dae Mark E. Dunlap Gordon A. Ewy Mat J.A.P. Daemen Brian P. Delisle Frank Dunn Derek V. Exner Darshan Dalal Christian Delles Daniel Duprez Michael D. Ezekowitz James E. Dalen Mario Delmar Jocelyn Dupuis Ronald L. Dalman Linda L. Demer William Durante Robert H. Fagard Jan Kristian Damas Serkalem Demissie Greg Dusting Erling Falk Nicolas Danchin Martin den Heijer Susan Duval Rodney H. Falk Haim D. Danenberg Susan W. Denfield Peter Dyck James A. Fallavollita Jean-Marie Daniel Lamaziere Christophe Depre Vladimir Dzavik James C. Fang Stephen R. Daniels Claudia Derian Tomasz Dziedzic Margaret C. Fang A.H. Jan Danser Akshay S. Desai Francesco Fantin Lara Danziger-Isakov Christopher A. De Souza Kim A. Eagle Harrison W. Farber John H. Dark Gernot Desoye Elaine D. Eaker Michel Farnier Victor Darley-Usmar Jean-Pierre Despres John W. Eaton Rossella Fattori Dipak K. Das Alexander Deten Robert T. Eberhardt William P. Fay Sandeep Das Patricia Detmers Franz R. Eberli Zahi A. Fayad Stella S. Daskalopoulou Gabrielle de Veber Andrea D. Eckhart Jocelyne Fayn Harold L. Dauerman Vitantonio Di Bello Joris Ector Sergio Fazio Anthony P. Davenport Marcelo F. Di Carli Saadia Eddahibi Daniel I. Feig

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Acknowledgment of Reviewers e23

Jeffrey A. Feinstein Vance G. Fowler David C. Gaze Martin E. Goldman Gregory Feld Caroline S. Fox J. Michael Gaziano Pascal J. Goldschmidt Arthur M. Feldman Keith A.A. Fox Raul J. Gazmuri James A. Goldstein Ross D. Feldman Gabriele Fragasso Carolyn L. Geczy Larry B. Goldstein Ted Feldman Alain Fraisse Bruce D. Gelb Sidney Goldstein G. Michael Felker Lawrence H. Frame Caroline Genco Jonathan Golledge David Felson Charles W. Francis Thomas L. Gentles Michael H. Gollob Qingping Feng Gary S. Francis Sarah Jane George Gershon Golomb Robert R. Fenichel Nikolaos G. Frangogiannis Lior Gepstein Celso E. Gomez-Sanchez Riccardo R. Fenici Sohrab Fratz Bernhard L. Gerber Philimon Gona Maros Ferencik Gunilla Nordin Fredrikson A. Martin Gerdes Luís F. Gonçalves James J. Ferguson Michael Freed Bernard J. Gersh Mark O. Goodarzi T. Bruce Ferguson David S. Freedman Deborah R. Gersony Aasha S. Gopal Francisco Fernandez-Aviles Gregory L. Freeman Welton M. Gersony John Gorcsan Olivier Feron Mason Wright Freeman Edward P. Gerstenfeld Ian L. Gordon Christiane Ferran Balz Frei Robert E. Gerszten Joel M. Gore Ele Ferrannini Michael P. Frenneaux Tal Geva Agnes Gorlach Alberto U. Ferrari Ulrich H. Frey Henry Gewirtz Joseph H. Gorman Carlos M. Ferrario Linda F. Fried Michael Gewitz Philippe Gosse Robert E. Ferrell Paul A. Friedman Mihai Gheorghiade Avrum Gotlieb Albert Ferro M. Kent Froberg Marzia Giaccardi Shinya Goto Andreas Festa James B. Froehlich Carlo Giansante Roberta A. Gottlieb Nathalie M. Fiaschi-Taesch Victor Froelicher Gary H. Gibbons Stephen S. Gottlieb Loren J. Field Edward D. Frohlich Raymond J. Gibbons K. Lance Gould David S. Fieno Jiri J. Frohlich C. Michael Gibson Andrew A. Grace Janos G. Filep Lars Frost Frank C. Gibson Michael Grace Juan F. Granada Jeffrey R. Fineman Andrea Frustaci Samuel S. Gidding Robert L. Frye Christopher B. Granger Toren Finkel Stephan Gielen Bianca Fuhrman D. Neil Granger John Paul Finn Raymond D. Gilbert Yasushi Fujio Peter J. Grant Peter Fischbach Wayne R. Giles Hisayoshi Fujiwara Guido Grassi Tim A. Fischell Linda D. Gillam John W. Funder Claus Højbjerg Gravholt Michael C. Fishbein A. Marc Gillinov Curt D. Furberg Richard A. Gray Edward A. Fisher Matthew W. Gillman Anthony P. Furnary Paul A. Grayburn Patrick W. Fisher Richard F. Gillum Valentin Fuster David R. Greaves Glenn I. Fishman J. Rod Gimbel Darren C. Greenwood Garret A. FitzGerald Jeffrey S. Ginsberg William H. Gaasch Edward W. Gregg Frank A. Flachskampf Myron D. Ginsberg Alain-Pierre Gadeau Andreas Greinacher Greg C. Flaker Frank J. Giordano James V. Gainer Kathy K. Griendling Scott D. Flamm Dario Giugliano Fiorenzo Gaita Brian P. Griffin Marcus D. Flather Michael M. Givertz Mark E. Galantowicz Peter W. Groeneveld Jerome L. Fleg David Gjertson Catharine R. Gale William J. Groh Mark T. Gladwin Ingrid Fleming Augusto Gallino Gil J. Gross John S. Floras Jonas B. Galper Stanton A. Glantz Paul D. Grossfeld James S. Floyd Apoor S. Gami Christopher Glass Eugene A. Grossi Alan M. Fogelman Tji-Joong C. Gan Stephen J. Glatt Eberhard Grube Robert N. Foley Ronald Gangnon Kathryn A. Glatter Scott M. Grundy Aaron R. Folsom Peter Ganz Mark Glickman Gary L. Grunkemeier John D. Folts Feng Gao Charles J. Glueck Maurizio D. Guazzi Gregg C. Fonarow Lawrence A. Garcia Robert J. Glynn Vilmundur G. Gudnason Vivian A. Fonseca Mario J. Garcia Alan S. Go Martha Gulati Luigi Fontana Julius M. Gardin Ian F. Godsland Mahesh P. Gupta JoAnne Micale Foody Helena M. Gardiner Els Goetghebur Rishi Gupta Thomas Force Sheila M. Gardiner Andreas Goette Paul A. Gurbel Ian Ford Vidu Garg David C. Goff Enrique P. Gurfinkel Robert D. Foreman Daniel Gaudet Noyan Gokce Tomasz J. Guzik Daniel Forman Glenn R. Gaudette Diane R. Gold Helena K. Gylling Tom Forsen Mario Gaudino Robert J. Goldberg Ulrich Forstermann Kimberlee Gauvreau Ronald B. Goldberg Donald C. Haas Elizabeth Bowen Fortescue Haralambos P. Gavras Jeffrey J. Goldberger Hans U. Haering Elyse Foster Meinrad Gawaz Uri Goldbourt Steven M. Haffner F. Gerald Fowkes J. William Gaynor Samuel Z. Goldhaber Dominik Georg Haider

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 e24 Acknowledgment of Reviewers

David E. Haines Keith Henry Daphne T. Hsu Jorge E. Jalil Michel Haissaguerre Moonseong Heo Willa A. Hsueh Ik-Kyung Jang Roger J. Hajjar Dirk Hermann Chengcheng Hu Joseph S. Janicki Julian P.J. Halcox Ramon C. Hermida Frank B. Hu Jan Janousek Andrew Halestrap Adrian F. Hernandez Gang Hu Warren R. Janowitz Alistair S. Hall Amy Herring Howard Hu Maurits A. Jansen Jesper Hallas David M. Herrington Kurt Huber Joseph A.M.J.L. Janssen Hermann Haller Howard C. Herrmann Sally Ann Huber Stefan P. Janssens Michael Hallman Ray E. Hershberger Murray Walker Huff Craig T. January Alfred P. Hallstrom Charles A. Herzog James C. Huhta James Louis Januzzi Henry R. Halperin Otto M. Hess Heikki V. Huikuri David L. Jardine Perry V. Halushka Gerd Heusch Philippe P. Hujoel Patrick Y. Jay Rainer Hambrecht William R. Hiatt Richard A. Humes David Jegger Mohamed H. Hamdan Catherine L. Higgins Karin H. Humphries David Philip Jenkins Pavel Hamet Charles B. Higgins William Gregory Hundley Allen Jeremias Robert L. Hamlin Per Hildebrandt Joseph Hung Jamie Yancey Jeremy Christian W. Hamm John S. Hill Judy Hung Mariell Jessup H. Kirk Hammond Jonathan M. Hill Donald B. Hunninghake Xavier Jeunemaitre Masayuki Hanaoka Joseph A. Hill Sharon Ann Hunt Ishwarlal Jialal Graeme J. Hankey Nicholas S. Hill Steven C. Hunt Xian-Cheng Jiang Anthony J. Hanley L. David Hillis Mansoor Husain Bernd Jilma Edward L. Hannan Gerhard Hindricks Sabah N.A. Hussain Zhezhen Jin Mark A. Hanson Aroon D. Hingorani Adolph M. Hutter B. Delia Johnson Goran K. Hansson Thomas H. Hintze Paul M. Hwang Bruce D. Johnson Violet I. Haraszthy Yoshitaka Hirooka Elaine M. Hylek Jason L. Johnson David T. Hardman Karen K. Hirschi Julie A. Johnson Guido Iaccarino Karen Johnson David L. Hare Valeria Hirschler Raymond E. Ideker S. Claiborne Johnston Joshua M. Hare John W. Hirshfeld Salim F. Idriss Daniel W. Jones Stephen B. Harrap James E. Hixson Richard G. IJzerman Gregory T. Jones Joanne S. Harrell Mark A. Hlatky Peter B. Imrey Peter Lloyd Jones Robert A. Harrington Kalon Ho Akihiro Imazi Peter J.H. Jones William S. Harris Michael Ho Ciro Indolfi Robert H. Jones David G. Harrison Nyssa Hoch Mel Ingber Steven P. Jones Rachel Harrison Judith S. Hochman Julie R. Ingelfinger W. Keith Jones Pamela Hartigan Julien I. Hoffman David A. Ingram, Jr. Habo J. Jongsma Hali A. Hartmann Udo Hoffmann Joanne S. Ingwall Ulrich P. Jorde David Hasdai Polly A. Hofmann Taku Inoue Michihisa Jougasaki Gerd Hasenfuss Peter Ho¨glund Teruo Inoue Michael J. Joyner Vic Hasselblad Thomas Hohlfeld John P. Ioannidis Robert M. Judd Paul J. Hauptman Brian D. Hoit Carlos Iribarren Amy C. Justice Derek J. Hausenloy Judd E. Hollander Shun Ishibashi Richard J. Havel David R. Holmes Mitsuaki Isobe Alan H. Kadish Edward P. Havranek Uffe Holmskov Eric M. Isselbacher Anthony Kafatos Nat Hawkins Paul Holvoet Peter Ivanovich Richard Kahn William G. Haynes Gerhard A. Holzapfel Susan Lee Ivey Jan Kajstura Daniel Hayoz Yuling Hong D. Dunbar Ivy Jonathan M. Kalman Julian Haywood Paul N. Hopkins Joseph L. Izzo Marjolein H. Kamphuis Stanley L. Hazen Richard Hopkins David E. Kandzari John P. Headrick Susan Hopkins Wael A. Jaber Ronald J. Kanter David Heber William E. Hopkins Christopher L. Jackson Jørgen K. Kanters Robert A. Hegele Uta C. Hoppe Alice K. Jacobs Emmanouil Ioannis Paul A. Heidenreich Lisa K. Hornberger Bradley S. Jacobs Kapetanakis Jay W. Heinecke John D. Horowitz David R. Jacobs Edward L. Kaplan Gerardo Heiss James Hosking Marshall L. Jacobs Norman M. Kaplan E. Kevin Heist Naohisa Hosomi Paul Jacques Richard H. Karas Donald D. Heistad Jan Laws Houghton Edgar T. Jaeggi John M. Karemaker Gary V. Heller Steven R. Houser Tazeen H. Jafar Joel S. Karliner Joan Heller Brown George Howard Allan S. Jaffe Aly Karsan Harry Hemingway Reuben Howden Farouc A. Jaffer Carlos S. Kase Brenda Hemmelgarn Vicky Y. Hoymans Mukesh Kumar Jain David Alan Kass Jeroen Hendrikse Mihail Hristov Jose Jalife Ghassan Kassab

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Acknowledgment of Reviewers e25

Adnan Kastrati Werner Koch Shenghan Lai Robert A. Levine Sekar Kathiresan Wolfgang Koenig Wyman W. Lai Bodo Levkau Masahiko Kato Sunao Kojima Buu-Khanh Lam Wayne C. Levy Hugo A. Katus Theodore J. Kolias John J. Lamberti Robert A. Lew Zvonimir S. Katusic Tatsuya Komaru Nicolas Lamblin E. Douglas Lewandowski Irene L. Katzan Issei Komuro Eric G. Lancelot Sarah Lewington Philipp A. Kaufmann Takahisa Kondo Michael J. Landzberg Martin M. LeWinter Sanjiv Kaul Marvin A. Konstam Peter Lang Eldrin F. Lewis Rae-Ellen Kavey Stavros V. Konstantinides Roberto M. Lang Chaoyang Li Ziya Kaya Michael C. Kontos Alexandra J. Lansky Chuanfu Li Friederike K. Keating Bruce A. Koplan Pierre Lantelme Debiao Li Bernard Keavney Maria S. Kopp Emma K. Larkin Jennifer S. Li Roya Kelishadi Ran Kornowski John C. LaRosa Liang Li Mary Kelley Rebecca Koscik Eric Larose Ronald A. Li Daniel P. Kelly Mikhail Kosiborod Terje S. Larsen Xiao-hui Li Ralph A. Kelly Sawa Kostin Torben B. Larsen James K. Liao Richard E. Kerber John B. Kostis Warren K. Laskey Ronglih Liao Dean J. Kereiakes Theodore A. Kotchen Robert D. Lasley Youlian Liao Morton J. Kern Thomas Erling Kottke Johan P.E. Lassus Peter Libby Steven J. Keteyian Petri T. Kovanen Roberto Latini Joseph R. Libonati Paul Khairy Charlotte Kragelund Chu-Pak Lau Andrew H. Lichtman Bobby V. Khan Dara L. Kraitchman Wei C. Lau Judith H. Lichtman Dirar S. Khoury Jonathan Krakoff Michael S. Lauer David Lichtstein Shukri F. Khuri Christopher M. Kramer Stephane Laurent Stephen B. Liggett Stefan Kiechl Ronald M. Krauss Kenneth R. Laurita CheeChew Lim Jan T. Kielstein Nancy R. Kressin Peter C. Laussen Joao A. Lima Hyo-Soo Kim Reinhold Kreutz Carl Lavie Jing Ping Lin Monty Krieger Nancy Kim Catherine M. Law Thomas M. Lincoln Eswar Krishnan Raymond J. Kim Debbie A. Lawlor Lars Lind Itzhak Kronzon Akinori Kimura Harold L Lazar Bertil Lindahl Henry Krum Spencer B. King Thu H. Le JoAnn Lindenfeld Isao Kubota Kevin E. Kip Thierry H. LeJemtel Jonathan R. Lindner Karl-Heinz Kuck Margaret L. Kirby Alexander Leaf Karl H. Lindner Peter J. Kudenchuk Paulus F. Kirchhof Scott Lear Bruce Lindsay Karen Kuehl Joel A. Kirsh Alexander Wolfgang Leber Ken A. Lindstedt Harald P. Ku¨hl Lorrie A. Kirshenbaum Nathan K. LeBrasseur Mark S. Link Hartmut Kuhn Ajay J. Kirtane Franck Lebrin MacRae F. Linton Petra Kuijpers Caroline Kistorp Amanda J. Lee Gregory Y.H. Lip Taco Kuijpers Masafumi Kitakaze E.T. Lee Michael J. Lipinski Marrick L. Kukin Dalane W. Kitzman Hon-Chi Lee H.H. Lipowsky Thomas J. Kulik Jorge R. Kizer I-Min Lee William C. Little Lewis H. Kuller John K. Kjekshus Richard T. Lee Sheldon E. Litwin Koichiro Kumagai David J. Lefer Kiang Liu Andre G. Kleber Raman Krishna Kumar Donald M. Lloyd-Jones Franz Xaver Kleber Karen M. Kuntz Michael H. Lehmann Neal S. Kleiman Richard E. Kuntz Stephan E. Lehnart Amanda Lochner Allan L. Klein Lih Kuo Norbert Leitinger James E. Lock George J. Klein Tobias Kurth Paul LeLorier Martin J. Lohse Irwin Klein Jeffrey T. Kuvin Daniel Lemogoum Federico Lombardi Liviu Klein David J. Kwiatkowski Pedro A. Lemos Anne-Marie Lompre Charles S. Kleinman On-Hing Kwok Karl B. Lemstrom Barry London Elizabeth S. Klings Raymond Y. Kwong Steven R. Lentz Gérard M. London Francis J. Klocke Martin B. Leon Gary D. Lopaschuk Robert A. Kloner Markku Laakso Antonio Maria Leone John J. Lopez John L. Knight David E. Laaksonen Hélène Lepetit Jose Lopez-Sendon Stephen B. Knisley Carlos Alberto Labarrere John J. Lepore Carlos Lorenzo Anne A. Knowlton Daniel Lackland Annarosa Leri Douglas W. Losordo Kirk U. Knowlton Karl J. Lackner Amir Lerman William R. Lovallo Sarah S. Knox Francisca Lago Bruce B. Lerman Gordon D. Lowe Dennis T. Ko Roger J. Laham Michelle Letarte Jacobus Lubsen Lars Kober Chao-Qiang Lai Adeera Levin Benedict R. Lucchesi Walter J. Koch Huichuan Lai Glenn N. Levine Russell V. Luepker

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 e26 Acknowledgment of Reviewers

Friedrich C. Luft Ulrich Martin Silke Meiners David A. Morrow Juha Lund James B. Martins Christa Meisinger Richard F. Mortensen Jon O. Lundberg Thomas H. Marwick Gerhard W. Meissner Richard Mortensen Kathryn Lunetta Nikolaus Marx Luis G. Melo Mauro Moscucci Keith G. Lurie Attilio Maseri Philippe Menasche Debra K. Moser Frederick A. Masoudi Lisa A. Mendes Jeffrey W. Moses Xin Liang Ma Joseph M. Massaro Armando J. Mendez Ivan P. Moskowitz Christoph Maack Barry M. Massie Maurizio Menichelli Arthur J. Moss Peter Simon Macdonald Idit Matot Vandana Menon Jean-Jacques Mourad Scott M. MacDonnell Hiroaki Matsubara Jean-Jacques Mercadier Lemuel A. Moye Rajiv D. Machado Reiko Matsui Olaf Mercier Thomas Muenzel Charles A. Mack Rumiko Matsuoka Yahye Merhi Andreas Mugge Christopher Mack Yuji Matsuzawa C. Noel Merz Andrew Mugglin Michael J. Mack Francesco Mattace-Raso Franz H. Messerli Joseph B. Muhlestein Wendy J. Mack Christian M. Matter Luisa Mestroni Debabrata Mukherjee Todd A. MacKenzie Heinrich Mattle David G. Meyers Rupak Mukherjee Rachel H. Mackey Laura Mauri J.A. Michaels James E. Muller Michael I. Mackness Constantine Mavroudis Evangelos D. Michelakis Jochen Muller-Ehmsen Kenneth N. MacLean Kevin H. Mayo David Milan Tomoatsu Mune William Robb MacLellan Bongani Mawethu Mayosi D. Douglas Miller Neal Indu Muni Briain D. MacNeill Todor N. Mazgalev D. Craig Miller Paul Muntner Colin H. Macphee Nathalie M. Mazure Francis J. Miller Daniel J. Murphy Paolo Madeddu Brendan F. McAdam Leslie W. Miller Elizabeth Murphy Joren C. Madsen Peter McCarron Todd D. Miller Charles E. Murry Aldo P. Maggioni Patrick M. McCarthy Virginia M. Miller Anthony J. Muslin William T. Mahle Leslie Ain McClure Erich Minar Robert J. Myerburg Lynn Mahony Daniel D. Myers Michael V. McConnell Gary S. Mintz Heiko Mahrholdt Jonathan Myers Timothy McConnell Seema Mital Willibald Maier Brian W. McCrindle Yoshihide Mitani William H. Maisel Elizabeth G. Nabel Peter A. McCullough Gary F. Mitchell Amy S. Major Gerald V. Naccarelli Doff B. McElhinney Richard N. Mitchell Sumit R. Majumdar Koonlawee Nademanee Andrew D. McGavigan Arnold Mitnitski Jonathan C. Makielski Vinay Nadkarni Daniel McGee Friedrich Mittermayer David J. Malenka James M. Naessens Ashwani Malhotra Paul G. McGlinchey Murray A. Mittleman Ryozo Nagai Ziad Mallat Michael McGoon Yoko Miyasaka Noritoshi Nagaya Giuseppe Mancia John L. McGregor Daria Mochly-Rosen Eike Nagel Donna M. Mancini Gordon T. McInnes Gordon Moe Sherif F. Nagueh G.B. John Mancini William J. McKenna Sharon M. Moe Matthias Nahrendorf Jayawant N. Mandrekar T.A. McKinsey Emile R. Mohler, III Hiroshi Nakagawa Roberto Manfredini Vallerie V. McLaughlin David J. Moliterno Shinsuke Nakayama Douglas L. Mann C. Alex McMahan Jeffery D. Molkentin Brahmajee K. Nallamothu Giovanni E. Mann John J.V. McMurray Tom E. Mollnes Bin Nan Teri A. Manolio Elizabeth M. McNally Gilles Montalescot Carlo Napolitano Moussa Mansour Coleen Ann McNamara Nicola Montano Tasneem Zehra Naqvi Eduardo Marba´n Robert L. McNamara Alan R. Moody Sanjiv M. Narayan Simona Marchesi Walter McNicholas James C. Moon Martyn P. Nash Francis E. Marchlinski David D. McPherson John William Moore Andrea Natale Metra Marco Patrick S. McQuillen Nicholas D. Moore Mark A. Nathanson Frank I. Marcus Tim C. McQuinn Samia Mora Stanley Nattel Andrew O. Maree Charles F. McTiernan Alfredo Morabia L. Gabriel Navar Kenneth B. Margulies Mandeep R. Mehra Martin Morad Frank Naya Daniel B. Mark Roxana Mehran Fred Morady James D. Neaton John R. Marler Jawahar L. Mehta Pierre Moreau Krassen Nedeltchev Barry J. Maron Nehal N. Mehta Thomas M. Morgan David P. Nelson David J. Maron Rajendra H. Mehta Marie-Claude Morice Donna S. Neuberg Martin Maron Shamir R. Mehta Carlos A. Morillo Ellis J. Neufeld Steven P. Marso Bernhard Meier Anthony P. Morise David E. Newby Steven B. Marston Christoph R. Meier Gregory E. Morley L. Kristin Newby Fabio Martelli Hans Meier-Ewert Maria A. Moro Anne B. Newman Edward T. Martin James B. Meigs Nicholas W. Morrell Mark F. Newman Randolph P. Martin Erik J. Meijboom Brian J. Morris Christopher H. Newton-Cheh

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Acknowledgment of Reviewers e27

Ludwig Neyses Clive Osmond Richard P. Phipps Harry Rakowski Wilmer W. Nichols Catherine M. Otto Mariann R. Piano Alessandra Rampazzo Andrew C. Nicholson Noriyuki Ouchi Michael H. Picard Gwendalyn J. Randolph Georg Nickenig Ronald J. Oudiz J. Geoffrey Pickering L. Vijay Rao Antonino Nicoletti Feifan Ouyang Massimo F. Piepoli Elliot Rapaport James T. Niemann Michel Ovize Luc A. Pierard Nora R. Ratcliffe Christoph A. Nienaber Al Ozonoff Gabriele Piffaretti Pentti M. Rautaharju Bernhard Nieswandt Nico H.J. Pijls Flavia Ravelli Petros Nihoyannopoulos Francis D. Pagani Louise Pilote Peter B. Raven Sigrid Nikol Massimo Pagani Yigal M. Pinto Ursula Ravens Rick A. Nishimura Richard L. Page Iraklis I. Pipinos Katya Ravid Steven E. Nissen Myunghee Cho Paik Tobias Pischon Reza S. Razavi Stefano Nistri Igor F. Palacios Bertram Pitt Gerald M. Reaven Koichiro Niwa Paolo Palatini Jonathan F. Plehn Peter D. Reaven Jacqueline A. Noonan Wulf Palinski Edward F. Plow Fabio A. Recchia Sharon-Lise T. Normand Colin N.A. Palmer Stuart J. Pocock Margaret M. Redfield C.M. Norris Vittorio Palmieri Karsten Pohle Andrew N. Redington William C. Nugent Demosthenes Panagiotakos Gerald M. Pohost Guy L. Reed Natesa G. Pandian Ju¨rg Nussberger Roberto Pola Kathryn L. Reed Vernon Shane Pankratz Ottar Kjell Nygard Joseph F. Polak Jalees Rehman Nazareno Paolocci Jaimie W. Polson Johan Reiber Steve Oberste Carlo Pappone Judit Pongracz Nathaniel Reichek Eoin O’Brien Stephen Paridon Philip A. Poole-Wilson Muredach Reilly Ira S. Ockene John T. Parissis Michael Poon Olaf Reinhartz Jacob O. Odeberg John D. Parker Jeffrey J. Popma Peter J. Reiser Christopher J. O’Donnell Thomas G. Parker Zoran B. Popovic Karen Remsberg Valerie B. O’Donnell Michael Parmacek Richard L. Popp Sybille Rex Erwin N. Oechslin Juan C. Parodi Thomas R. Porter Kathyrn M. Rexrode Peter Oettgen Vincenzo Pasceri Michael A. Portman Bernardo J. Reyes Patrick T. O’Gara Ayan Patel Ronald J. Portman Jonathan Rhodes Hisao Ogawa Manesh R. Patel Wendy S. Post Josep Ribalta Yoshihiro Ogawa Carlo Patrono Lucilla Poston Ken Rice Sang Woo Oh Richard D. Patten Neil Poulter Jean-Paul Richalet Ann M. O’Hare Cam Patterson Janet T. Powell Vincent Richard E. Magnus Ohman Eugene Patterson Sumanth D. Prabhu A. Mark Richards John Ohrvik Stephen G. Pauker Ashwin Prakash Yoram Richter Richard G. Ohye Walter J. Paulus Abhiram Prasad Paul M Ridker Angel Javier Ois Andrew J. Peacock Domenico Pratico Charanjit S. Rihal Kaie M. Ojamaa Jeffrey M. Pearl Susan J. Pressler Alain Rivard Peter M. Okin Justin D. Pearlman Matthew J. Price Damiano Rizzoni Jeffrey E. Olgin Shayn M. Peirce Frits W. Prinzen Jeffrey Robbins Brian Olshansky Michael Pencina Silvia G. Priori Robert Roberts Eric N. Olson Marc S. Penn Kirkwood A. Pritchard, Jr. William C. Roberts Lyle J. Olson Len A. Pennacchio Bruce M. Psaty Timothy M. Olson Dudley J. Pennell Patrick Pullicino Sander J. Robins Patrick G. O’Malley Carl J. Pepine Reed Pyeritz Richard B. Robinson William W. O’Neill Mark B. Pepys Kalevi Pyorala Gaël Y. Rochefort Andrew T.L. Ong Joseph K. Perloff Howard A. Rockman Henry Ooi Sharina D. Person Lu Qi Richard J. Rodeheffer Robert A.J. Oostendorp Inga Peter Paul H. Quax Alfredo E. Rodriguez Suzanne Oparil Karlheinz Peter Miguel A. Quiñones Carlos Jose Rodriguez Tobias Opthof Annette Peters Arshed A. Quyyumi Alicia Rodriguez-Pla E. John Orav Nicholas S. Peters Matthew T. Roe Marek Orban Eric D. Peterson LeRoy E. Rabbani Ellen Roecker Jose M. Ordovas Gail E. Peterson Frank E. Rademakers Veronique L. Roger John A. Ormiston J. Thomas Peterson Daniel J. Rader Campbell Rogers Joseph P. Ornato Eva Petkova Paolo Raggi Gregg Rokosh Brian O’Rourke George John Philippides Shahbudin H. Rahimtoola Mary J. Roman Michael F. O’Rourke Christopher O. Phillips Leopoldo Raij Mario Romano Robert A. O’Rourke David R. Phillips Elaine W. Raines Patrick Romano Kristina Orth-Gomer Gerald B. Phillips Olli T. Raitakari Mats Ro¨nnback Kevin O’Shaughnessy M. Ian Phillips Nalini M. Rajamannan Wayne D. Rosamond

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 e28 Acknowledgment of Reviewers

Jonathan Rosand Leslie A. Saxon Ralph Shabetai Stefan Soderberg Eric A. Rose Fakhredin Sayed-Tabatabaei Robert E. Shaddy Kyoko Soejima Noel R. Rose Fernando Scaglia Ajay M. Shah John Solaro Michael R. Rosen Rosari Scalia Maully J. Shah Scott D. Solomon David S. Rosenbaum Pierre-Yves Scarabin Pravin M. Shah Virend K. Somers Marjorie Rosenberg Volker Schachinger Prediman K. Shah Pierre Sonveaux Frits R. Rosendaal Saul Schaefer David M. Shahian Dan Sorescu Clive Rosendorff Hartzell V. Schaff Richard P. Shannon Vincent L. Sorrell Uri Rosenschein David V. Schaffer Oz M. Shapira Janet Soul David N. Rosenthal Wolfgang Schaper Arya M. Sharma James R. Sowers Eric Rosenthal Robert G. Schaub Leslee J. Shaw Jonas Spaak Gian Paolo D. Rossi Melvin M. Scheinman Fadia T. Shaya Madison S. Spach Michael Roth Sebastian M. Schellong Amanda M. Shearman Arthur A. Spector Florence Rothenberg Benjamin J. Scherlag Frank Shellock J. David Spence Peter M. Rothwell Ralph Theo Schermuly James Shepherd William H. Spencer Anne H. Rowley Urs Scherrer Lianne Sheppard Sarah Spiegel Sylvain Roy Thomas M. Schiele Guo-Ping Shi Francis G. Spinale Prabir Roy-Chaudhury Ernesto Schiffrin Weibin Shi David H. Spodick Speranza Rubattu Stefan Schildknecht Hiroaki Shimokawa Matthew L. Springer Frederick L. Ruberg Martin Schillinger Atsushi Shioi Martin G. St. John Sutton Lewis J. Rubin Martina Schinke Ichiro Shiojima Bart Staels Alicja Regina Rudnicka Markus Peter Schlaich Kalyanam Shivkumar Jan A. Staessen John S. Rumsfeld Thomas Schlosser Michael G. Shlipak Gregory L. Stahl Marschall S. Runge Alvin Schmaier Sergey V. Shmelkov Bruce S. Stambler Frank Ruschitzka Axel Schmermund Yehuda Shoenfeld Jonathan S. Stamler Raymond R. Russell Chris Schmid Gunjan Shukla Meir J. Stampfer Ann-Marie Schmidt Domenic A. Sica Kenneth Stanley Tobias Saam Sebastian Schneeweiss Robert J. Siegel William C. Stanley Samir Saba David J. Schneider Ulrich Sigwart Randall C. Starling Roger A. Sabbadini Michael D. Schneider Michael J. Silka Vaughn A. Starnes Joseph F. Sabik Frederick J. Schoen Jonathan Silva Petter A. Steen Ralph L. Sacco Uwe Schonbeck Marc A. Silver Paul Steendijk Michael N. Sack Kevin A. Schulman Jean-Sebastien Silvestre Norbert Stefan Frank M. Sacks Junichi Sadoshima Heinz-Peter Schultheiss Robert D. Simari Philippe Gabriel Steg Michel E. Safar Rainer Schulz Joel A. Simon Coen D. Stehouwer Jeffrey E. Saffitz Richard Schulz Michael Simons Evan A. Stein Robert D. Safian P. Christian Schulze Paul C. Simpson Julia Steinberger David J. Sahn Eric Schulze-Bahr Alan R. Sinaiko George Steiner Yoshihiko Saito Arnold Schwartz Charles F. Sing David A. Steinman Gerald Salen Gary L. Schwartz Donald Singer Kurt R. Stenmark Veikko Salomaa Kenneth A. Schwartz Bramah N. Singh Peter Stenvinkel Flora Sam Ketty Schwartz Krishna Singh Elizabeth A. Stephenson Nilesh J. Samani Robert Stockton Schwartz Steven N. Singh Kent Stephenson Allen M. Samarel Ronald G. Schwartz Tajinder P. Singh Andrew Steptoe Jonathan M. Samet Stephen Mark Schwartz Lawrence I. Sinoway Michael P. Stern Jane-Lise Samuel Ernst R. Schwarz Cesare R. Sirtori Naftali Stern Timothy Allen Sanborn Robert A. Schweikert Allan C. Skanes Lynne Warner Stevenson Stephen P. Sanders Udo Sechtem Susan A. Slaugenhaupt Duncan J. Stewart Giuseppe Sangiorgi Artyom Sedrakyan Lynn A. Sleeper Kerry J. Stewart L. Fernando Santana Patrick Segers Peter Sleight Roland Stocker John Lewis Sapp Christine E. Seidman Michael A. Sloan Henri E.J.H. Stoffers Dennis Sarabi Jonathan G. Seidman Eric J. Smart Karen Stokes Ian J. Sarembock Christian Seiler George Davey Smith Monika Stoll Mark J. Sarnak Frank W. Sellke Nicholas L. Smith Gregg W. Stone Masataka Sata Gregg L. Semenza Timothy William Smith Peter H. Stone Naveed Sattar Chris Sempos Jos F.M. Smits Michael Stoto J. Philip Saul Roxy Senior Steve Snapinn Bodo E. Strauer Peter Savage Susan M. Sereika Allan D. Sniderman Gary Striker Stephen G. Sawada Charles N. Serhan Harold Snieder Erik S.G. Stroes Tatsuya Sawamura Patrick W. Serruys Burton E. Sobel Allan D. Struthers Douglas B. Sawyer Howard D. Sesso Piotr Sobieszczyk Matthias Stuber

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Acknowledgment of Reviewers e29

Therese A. Stukel Norma Terrin Viola Vaccarino Dirk H. Walter James T. Stull Andre Terzic Patrick J.T. Vallance Thomas Walther Krishnankutty Sudhir Ravi Thadhani Glen S. Van Arsdell Oliver Wang Atsushi Sugiyama Pierre Theroux Ruud M.A. Van de Wal Paul J. Wang Samy Suissa Perumal Thiagarajan Frans J. Van de Werf Qing Wang Galina K. Sukhova Ravi R. Thiagarajan Johanna Gerarda Vanderbom Thomas J. Wang Jerome L. Sullivan Matthias Thielmann Yvonne T. Van der Schouw Wenyu Wang Gere Sunder-Plassmann David Thiemann George F. Van Hare Xiaohong Wang Mark A. Sussman Gaetano Thiene Ton G. van Leeuwen S. Goya Wannamethee George Ross Sutherland Dierk Thomas Joost P. van Melle Carole A. Warnes Jill Suttles James D. Thomas Dirk J. van Veldhuisen Manabu Watanabe Alex Sutton Karen E. Thomas David R. Van Wagoner Nozomi Watanabe Kim Sutton-Tyrrell Jamie I. Vandenberg Shane R. Thomas David D. Waters Maarten J. Suttorp Raymond Camille Vanholder William Thomas Hugh Watkins Atsuko Suzuki Paul M. Vanhoutte Mary Lou Thompson Karol E. Watson Yuichiro J. Suzuki Gerald Vanzetto Paul D. Thompson Stephanie W. Watts Lars G. Svensson Cristina Varas-Lorenzo Richard Thompson Sergio Waxman Elisabet Svenungsson Elizabeth A. Varga Eric Thorin W. Douglas Weaver Karl Swedberg Nerea Varo Kent Thornburg David J. Webb Michael O. Sweeney Stephen F. Vatner John G. Webb Rong Tian Charles D. Swerdlow Douglas E. Vaughan Steven A. Webber Uwe J.F. Tietge Bernard Swynghedauw John P. Veinot Christian Weber Laurence Tiret Lynda Anne Szczech Richard C. Venema Karl T. Weber Asa Tivesten Paolo Verdecchia Rainer Weber Koichi Tabayashi Martin D. Tobin Mervyn Vergouwen Keith A. Webster Stefano Taddei Geoffrey H. Tofler Freek W.A. Verheugt William Weglicki Heinrich Taegtmeyer Gordon F. Tomaselli Subodh Verma Xander H. Wehrens David P. Taggart Marcello Tonelli Richard L. Verrier Kevin Wei Masato (Mike) Takahashi Peter Tontonoz Francesco Versaci L. Wei Yasuchika Takeishi Eric J. Topol Sara Vesely Wei Wei Alan R. Tall David F. Torchiana George W. Vetrovec Max Harry Weil Nagara Tamaki G. Wesley Vick Christian Torp-Pedersen Janice Weinberg Walter A. Tan Flordeliza S. Villanueva Rhian M. Touyz Milton C. Weinstein Toshihiro Tanaka Francisco Villarreal Jeffrey A. Towbin Neal L. Weintraub Ron Tanel Karen A. Vincent Jonathan N. Townend William S. Weintraub W.H. Wilson Tang Hans Vink Russell P. Tracy E. Kenneth Weir Weihong Tang Jakob Vinten-Johansen Pierluigi Tricoci Richard D. Weisel Xian-Liang Tang Renu Virmani John K. Triedman Myron L. Weisfeldt Rajendra K. Tangirala Sami Viskin Chinmay M. Trivedi Neil J. Weissman Lloyd Y. Tani Gary A. Visner Gerardus C. Tromp Hein J. Wellens Laszlo B. Tanko Antonio Vitarelli Han-Mou Tsai Ernst Wellnhofer David Tanne Mara Z. Vitolins EtsukoTsuda David J. Welsh Felix C. Tanner Eric Vittinghoff Yukiomi Tsuji Frederick G. Welt Herbert B. Tanowitz Gus J. Vlahakes Jack V. Tu Paramjit S. Tappia Rozemarijn Vliegenthart Nanette Kass Wenger Katherine L. Tucker Jean-Claude Tardif Robert A. Vogel Rene R. Wenzel Igor I. Tulevski Robert B. Tate Arnold von Eckardstein Rachel Werner Katherine R. Tuttle Mark B. Taubman Roland von Ka¨nel Rainer Wessely Allen J. Taylor E. Murat Tuzcu Marc A. Vos Malcolm West Doris A. Taylor James S. Tweddell Robert Voswinckel Arthur E. Weyman Herman A. Taylor Sari Voutilainen J. Marcus Wharton J. Andrew Taylor Per Magne Ueland John Wharton W. Robert Taylor Thor Ueland Atsuyuki Wada Donald C. Whitaker Alain Tedgui Cuno S.P.M. Uiterwaal Galen S. Wagner Christopher J. White Usha Tedrow Michael R. Ujhelyi Ichiro Wakabayashi Halina White Tom Teerlink Thomas Unger Ron Waksman Richard H. White David F. Teitel Gilbert R. Upchurch Albert L. Waldo Roger D. White George Tellides Philip Urban Jeffery W. Walker Patrick L. Whitlow Michiel Ten Hove Zsolt Urban Lars Wallentin J. Lindsay Whitton Alexander Tenenbaum Konrad Urbanek Reidar Wallin Thomas Wichter Leena Tenkanen Elaine Urbina Edward P. Walsh Samuel A. Wickline Masaru Terai Masuko Ushio-Fukai Simon J. Walsh Julian Widder

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 e30 Acknowledgment of Reviewers

Robert E. Widdop Scott R. Willoughby D. George Wyse Chun Yuan Jean A. Wiedeman Andrew M. Wilson Ralf Zahn Rob F. Wiegerinck Peter W. Wilson Jay S. Yadav Jose Zamorano William Wijns Stephan Windecker Norikazu Yamada David J. Wilber Karl Winkler Alberto Zanchetti Yoshiji Yamada Wojciech Zareba Josiah N. Wlcox Jacqueline C.M. Witteman Hiromi Yanagisawa Robert Y. Zee Arthur A.M. Wilde Carin Wittnich Clyde W. Yancy Marco A. Zenati Ian B. Wilkinson Bernhard J. Witzenbichler Shu Ye Andrey G. Zenovich Bruce L. Wilkoff Cordula M. Wolf Edward T.H. Yeh Bjorn Zethelius Andrew R. Willan Matthew R. Wolff Edward Yellin Jianyi Zhang Howard J. Willens Roswitha M. Wolfram Derek M. Yellon Yingyi Zhang James T. Willerson Michael S. Wolin Midori Anne Yenari Zefeng Zhang Walter C. Willett Paul Wolkowicz Shaw-Fang Yet David Zhao Bryan Williams Nathan D. Wong Alan C. Yeung Xinsheng Zhu David O. Williams Tak Ming Wong Paul G. Yock Susan J. Zieman Gordon H. Williams Y. Joseph Woo Mitsuhiro Yokota Felix Zijlstra Ken Williams Mark A. Wood Mitsuhiro Yokoyama Michael R. Zile Kevin Jon Williams Elizabeth A. Woodcock Shinji Yokoyama Frank Zimmerman Paul T. Williams Samuel Donald Wright Young-sup Yoon Douglas P. Zipes Redford B. Williams Stephen A. Wright Kiyoshi Yoshida William A. Zoghbi Sarah J. Williams Kenneth K. Wu James Benard Young Irving H. Zucker Scott M. Williams Joanna Wykrzykowska Lawrence H. Young Jay L. Zweier

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Correction

In the editorial, “Prediction and Prevention of Chemotherapy-Induced Cardiomyopathy: Can It Be Done?” by Granger that appeared in the December 5, 2006, issue (Circulation. 2006;114:2432– 2433), a few corrections did not appear in the second page of the final article.

The second sentence of the first full paragraph should read, “Chemotherapy-induced cardio- toxicity is not well understood, but it is believed, at least with anthracyclines, to be caused partly by generation of free radicals, mitochondrial dysfunction, iron- and calcium-handling abnormal- ities, and resulting apoptosis.”

The following acknowledgment should have been added: I thank Marc Jolicoeur, MD, for his thoughtful review of this editorial.

The correct Reference 7 is van Dalen EC, van der Pal HJ, Caron HN, Kremer LC. Different dosage schedules for reducing cardiotoxicity in cancer patients receiving anthracycline chemo- therapy. Cochrane Database Syst Rev. 2006;(4):CD005008.

The online version of this article has been corrected. The publisher regrets this error. DOI: 10.1161/CIRCULATIONAHA.106.180989

(Circulation. 2007;115:e31.) © 2007 American Heart Association, Inc. Circulation is available at http://www.circulationaha.org DOI: 10.1161/CIRCULATIONAHA.106.180989

e31

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Circulation January 16, 2007 f5

Spotlight on Gino Gerosa, MD History of Cardiology: Bloodletting Dr Gino Gerosa, University of Padova From ancient times to the early 20th century, Medical School, Padova, Italy speaks about bloodletting has been seen as the answer to his career and special interests. many cardiovascular conditions. Page f6 Page f8

European Perspectives in Cardiology

is that the market here is less attractive than in the United Viewpoint: Pharmaceutical States. Dr Vasella explains, “We don’t have a single European market, either from the point of view of pricing or Innovation in Europe from the point of view of regulation. In the short term, this

probably doesn’t directly affect pharmaceutical innovation Circulation: European Perspectives Daniel Vasella, MD, chairman overall, because it is balanced out by opportunities in other and CEO of Novartis AG, warns countries such as the United States, where there is an attrac- that changes are needed if tive market situation and rapid access. But European patients pay the price through delayed access to better treatments.” Europe is to remain a key site The European Union is, in theory, a single market, with a for pharmaceutical innovation. free flow of products between countries, but in reality there He explains his concerns to are differences in pricing between EU countries because prices Emma Baines, MSc. are set by individual governments. This combination can mean that more money is made by intermediaries from a par- ince he took up his role as chief executive at the newly ticular drug than by the pharmaceutical firms themselves. Sformed Novartis in 1996, Dr Daniel Vasella has been According to Dr Vasella, this is a serious problem. “If you credited with building the company into a market leader. His have a situation where a third party is profiting from parallel tactic has been to invest heavily in innovation, spending far importing and is making more money than the company that more on research and development than other pharmaceutical invented, developed, and marketed the drug, then there is companies. As a result, Novartis has had more drug approvals something wrong with the system.” in the past 5 years than any other pharmaceutical company. Another area in which Dr Vasella feels Europe is losing In the past, much of the basic research needed for phar- out to the United States and other countries is in the amount maceutical innovation has been carried out in Europe, but of high-quality basic research that it supports. This is also Novartis, despite being based in Switzerland, moved its related to how easy it is for researchers working in academia research headquarters to the United States in 2002. And, says to bring the results of their work to the market. “Academic Dr Vasella, “Unless changes are introduced to the way institutions are a great source of innovation,” says Dr research is funded and regulated in Europe, investment in Vasella. “In the United States, the professors and scientists pharmaceutical innovation will continue to leave Europe, at are given more freedom than in Europe, and they have a serious cost to the health and wealth of European nations.” access to funds and venture capital that allow them to create According to Dr Vasella, the multiple layers of bureau- their own start-up companies. This means that top scientists cracy in Europe mean that new products take longer to come are more likely to be retained in academia.” to market than in other countries, such as the United States. “However,” Dr Vasella says, “this pattern is changing. Despite “This increases the costs of launching new drugs for the European researchers having access to fewer resources than pharmaceutical companies,” he explains. their US colleagues, there are now as many biotech start-up Dr Vasella says, “European regulators have started creat- companies in Europe as in the United States. There seems to ing ‘artificial delays’ in how quickly patients get access to be more courage, guts, and optimism in Europe than we new drugs, because of the implementation of cost-benefit saw in the past. If you are going into biotech, you have to be studies and through delaying the reimbursement of a new willing to fail, and in Europe this has traditionally been less drug by up to 3 years.” These delays, he believes, put obsta- well tolerated than in the United States.” cles in the way of evaluating efficacy and safety and slow But the most serious advantage that US researchers have down the final stages of drug development, making it less over their European equivalents, from the point of view cost-effective for pharmaceutical companies to develop drugs of pharmaceutical companies looking for innovation in in Europe than elsewhere. biomedical research, is the amount of money that the US Another obstacle to pharmaceutical innovation in Europe government puts into basic research through funding to the

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 f6 Circulation January 16, 2007

National Institutes of Health. “The National Institutes of emerges in the Far East. “The next wave of scientists and Health have a budget well in excess of $30 million a year,” Dr world-class research will come from China,” he predicts. Vasella points out. “This generates research activities I consider “The Chinese government is investing money in research extraordinary and with which Europe cannot compete.” and pursuing policies that will make it an attractive place to A similar Europe-wide funding mechanism could rein- do research. If you find a country where there is an attractive state Europe at the heart of pharmaceutical innovation, Dr market, where there are a sufficient number of scientists of Vasella believes. “Creating a centralised funding mechanism top quality, where there are clear and transparent rules being with an amount of funds comparable with the National implemented from the point of view of regulation and intel- Institutes of Health would be an important step in making lectual property, and where there is a good quality of life so Europe competitive,” he says. “But, instead of promoting that people want to go and stay there, then you probably egalitarianism, the decisions about European funding for have a winning proposition.” research would need to start promoting elitism based on He concludes, “This is why research has been attractive in achievement. This would mean handing out grants solely on Europe in the past, and why, since World War II, it is increas- the basis of the quality of research programmes, rather than ingly carried out in the United States, and it is the reason by nationality or region.” He goes on to say, “There are many that, in the future, we will be going to China.” great people working here in Europe. If we can ensure that the best are recognised and promoted and given opportu- Emma Baines is a freelance medical writer.

Circulation: European Perspectives nities, we will be more competitive.” Dr Vasella adds that although at the moment the United The opinions expressed in Circulation: European States is Europe’s biggest competitor for pharmaceutical Perspectives in Cardiology are not necessarily those of the investment, it will not be long before serious competition editors or of the American Heart Association.

“mentor for excellence.” Spotlight: Dr Gerosa says, “I always dreamed of becoming a heart surgeon, and I became fascinated with the heart operations Gino Gerosa, MD I watched while in the 4th year of medical school at the University of Verona, Italy. The year after, I went to London and spent some time as a visiting student with Dr Gino Gerosa is associate Donald Ross at the National Heart Hospital, and that was professor of Cardiac Surgery, the final click.” Dr Gerosa explains, “He was not only a chief of the Cardiac Surgery mentor but a man with an incredible stamina, superb Department, and director of surgical skill, an unexhausted scientific curiosity—and an the School of Specialisation of elegant taste to fully enjoy the spirit of life.” Cardiac Surgery at the Dr Gerosa mentions Sir Magdi Yacoub, FRS, FRCS, as University of Padova Medical a man who inspired the idea of the scientist–surgeon. Another influence was Dr Gerosa’s Italian mentor, Dino School, Padova, Italy. He Casarotto, MD, a professor at the Institute of Cardiac speaks to Mark Nicholls about Surgery at the University of Padova, “who taught me the his career and special interests. full dedication needed for this profession.” Dr Gerosa was in residency at Verona University, Italy, r Gino Gerosa is a member of the Italian Society for from 1983 to 1988, and he worked as a registrar at the DCardiac Surgery, the New Technology Committee of Department of Cardiac Surgery at Verona University Hospital the European Association for Cardio-Thoracic Surgery, and from 1989 to 1992. He then moved to the Department of the American Association for Thoracic Surgery. He now Cardiac Surgery, Padova University Hospital, where he operates at the leading edge of cardiac surgery, with specific worked until 2000 before taking up his current position. interests in robotic surgery, transplantation, artificial hearts, In recent years, he has developed his specialist interests minimally invasive surgery, tissue engineering, and regen- and was involved in performing the first Italian totally erative medicine. endoscopic robot-enhanced coronary bypass on a beating But, he also recalls the days when he was learning his heart in 2001. He was the first Italian—and one of the first skills. “I wondered whether I was physically and mentally in Europe—to perform totally endoscopic robot-enhanced equipped to sustain the stresses of specialising in heart pulmonary vein isolation for lone atrial fibrillation. surgery and able to perform those operations with accept- “Robotic surgery is solo surgery,” he explains (see Figure). able results.” “You have to reinvent your surgical technique and skill It was time spent with Donald Ross, DSc, FRCS, at the because you don’t have an assistant helping you through National Heart Hospital in London, United Kingdom in the procedure. On the other side of the coin, the fine degree the 1980s that helped Dr Gerosa focus on a career in car- of movement offered by the telemanipulated instruments diac surgery. He regards Dr Ross as an inspiration and a allows you to surgically perform in a way that is impossible

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 f7 Circulation January 16, 2007

for human hands.” He recalls one occasion in particular. “I still remember the exact moment when I completed the totally endoscopic coro- nary artery bypass anastomosis. Such joy and Dr Gino Gerosa satisfaction, and all the team members were clapping.” But Dr Gerosa admits there have been disap- pointments. “Unfortunately, the great expecta- tion for cardiac surgeons to be competitive with interventional cardiologists in terms of less invasiveness for myocardial revascularisation has been substantially frustrated by the intro- duction of drug-eluting stents.” He has clear views on where the future lies in response to what he calls the “less than satisfactory” results of drug-eluting stents, and he advocates a return to a hybrid procedure for treating patients with

coronary disease. This would be a left internal Circulation: European Perspectives mammary artery anastomosis to the left anterior descending coronary artery, combined with Robotic surgery is solo surgery, says Dr Gerosa, as you have to reinvent your surgical technique and skill because you don’t have an assistant to help you. stents on the remaining affected vessels, thus getting the best results from both techniques. one of the most rapidly evolving fields,” he says, “where car- Dr Gerosa also expresses great satisfaction as he recalls diac surgeons will be able to perform by combining new performing one of the first stem cell transplantations in a surgical techniques, cell therapies, and mechanical devices.” patient with ischaemic cardiomyopathy in 2002. “I have Yet, alongside all the innovation and technical advance- vivid memories of those days and of that patient who had ment, Dr Gerosa also believes that cardiac surgeons in so much trust in the ability of such therapy to solve his ill- Western countries should put more effort into providing ness,” he says. “In Padova, we are still working in that field, cardiac surgery care for patients in developing countries, and we are exploring the ability of stem cells to recreate along the lines of Sir Magdi Yacoub’s Chain of Hope human valves and to regenerate the failed myocardium. organisation.1 When I think of the success and failure in this field, I feel Talking about his personal life, Dr Gerosa says he was that we can compare this experience to the first heart trans- born in Rovereto, in the mountains in northern Italy. Away plantation performed in 1967.” from work, he enjoys skiing and golf. He is married to Janine He refers to the expectation, enthusiasm, and hype after Gerosa, whom he met while training in London, United the first human heart transplantation that was followed Kingdom, and he takes great pride in their 3 children by “rather unsatisfactory” results. This meant that in the Edoardo (age 15), Carlo Andrea (age 12), and Filippo Alberto 1970s, only centres in Cape Town, South Africa and (age 6). “Fatherhood is a lot more complicated and reward- Stanford, United States continued to perform heart trans- ing than cardiac surgery,” he says with a smile. plants for some years. “At the beginning of the 1980s, the He describes his children as his proudest achievement, discovery of cyclosporine totally modified the results, but he also lists his 1991 article2 that reviewed the experi- improving organ and patient survival, and the scenario ence of the Ross operation (a procedure where the patient’s changed completely,” he explains. “I believe that with stem normal pulmonary valve is removed and used to replace a cells we are mimicking that experience. We need to clarify diseased aortic valve, and the pulmonary valve is then replaced some things and find the missing pieces to complete the with a pulmonary homograft) in children. He said it was the puzzle. We can then properly use stem cells to solve the first article postulating the ability of the pulmonary autograft problem of the failing heart. We are still in a phase of full to grow after transplantation into the aortic position. research, but there is light at the end of the tunnel.” Mark Nicholls is a freelance medical journalist. Dr Gerosa feels that cardiac surgery is about to enter References a new phase. “I believe that we are at the end of cardiac 1. Nicholls M. Chain of hope. Circulation. 2006;114:f134–f136. surgery as we have interpreted it in the last two decades. 2. Gerosa G, McKay R, Davies J, Ross DN. Comparison of the aortic Cardiac surgeons are retraining to perform and optimise homograft and the pulmonary autograft for aortic valve or root minimally invasive cardiac and intravascular procedures. In replacement in children. J Thorac Cardiovasc Surg. 1991;102:51–60. addition, cardiac surgeons should acquire the knowledge to employ gene-enhanced cell therapy and tissue-engineering Editor: Thomas F. Lüscher, MD, FRCP, FACC techniques.” Managing Editor: Keith Barnard, MB, BS, MRCS, LRCP Dr Gerosa believes that a shortage of organ donors is We welcome your comments. E-mail the managing editor at [email protected] driving the need to find additional solutions. “Heart failure is

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007 Circulation January 16, 2007 f5

Spotlight on Gino Gerosa, MD History of Cardiology: Bloodletting Dr Gino Gerosa, University of Padova From ancient times to the early 20th century, Medical School, Padova, Italy speaks about bloodletting has been seen as the answer to his career and special interests. many cardiovascular conditions. Page f6 Page f8

European Perspectives in Cardiology

is that the market here is less attractive than in the United Viewpoint: Pharmaceutical States. Dr Vasella explains, “We don’t have a single European market, either from the point of view of pricing or Innovation in Europe from the point of view of regulation. In the short term, this

probably doesn’t directly affect pharmaceutical innovation Circulation: European Perspectives Daniel Vasella, MD, chairman overall, because it is balanced out by opportunities in other and CEO of Novartis AG, warns countries such as the United States, where there is an attrac- that changes are needed if tive market situation and rapid access. But European patients pay the price through delayed access to better treatments.” Europe is to remain a key site The European Union is, in theory, a single market, with a for pharmaceutical innovation. free flow of products between countries, but in reality there He explains his concerns to are differences in pricing between EU countries because prices Emma Baines, MSc. are set by individual governments. This combination can mean that more money is made by intermediaries from a par- ince he took up his role as chief executive at the newly ticular drug than by the pharmaceutical firms themselves. Sformed Novartis in 1996, Dr Daniel Vasella has been According to Dr Vasella, this is a serious problem. “If you credited with building the company into a market leader. His have a situation where a third party is profiting from parallel tactic has been to invest heavily in innovation, spending far importing and is making more money than the company that more on research and development than other pharmaceutical invented, developed, and marketed the drug, then there is companies. As a result, Novartis has had more drug approvals something wrong with the system.” in the past 5 years than any other pharmaceutical company. Another area in which Dr Vasella feels Europe is losing In the past, much of the basic research needed for phar- out to the United States and other countries is in the amount maceutical innovation has been carried out in Europe, but of high-quality basic research that it supports. This is also Novartis, despite being based in Switzerland, moved its related to how easy it is for researchers working in academia research headquarters to the United States in 2002. And, says to bring the results of their work to the market. “Academic Dr Vasella, “Unless changes are introduced to the way institutions are a great source of innovation,” says Dr research is funded and regulated in Europe, investment in Vasella. “In the United States, the professors and scientists pharmaceutical innovation will continue to leave Europe, at are given more freedom than in Europe, and they have a serious cost to the health and wealth of European nations.” access to funds and venture capital that allow them to create According to Dr Vasella, the multiple layers of bureau- their own start-up companies. This means that top scientists cracy in Europe mean that new products take longer to come are more likely to be retained in academia.” to market than in other countries, such as the United States. “However,” Dr Vasella says, “this pattern is changing. Despite “This increases the costs of launching new drugs for the European researchers having access to fewer resources than pharmaceutical companies,” he explains. their US colleagues, there are now as many biotech start-up Dr Vasella says, “European regulators have started creat- companies in Europe as in the United States. There seems to ing ‘artificial delays’ in how quickly patients get access to be more courage, guts, and optimism in Europe than we new drugs, because of the implementation of cost-benefit saw in the past. If you are going into biotech, you have to be studies and through delaying the reimbursement of a new willing to fail, and in Europe this has traditionally been less drug by up to 3 years.” These delays, he believes, put obsta- well tolerated than in the United States.” cles in the way of evaluating efficacy and safety and slow But the most serious advantage that US researchers have down the final stages of drug development, making it less over their European equivalents, from the point of view cost-effective for pharmaceutical companies to develop drugs of pharmaceutical companies looking for innovation in in Europe than elsewhere. biomedical research, is the amount of money that the US Another obstacle to pharmaceutical innovation in Europe government puts into basic research through funding to the

Downloaded from circ.ahajournals.org at Mohammed Mahboob on June 10, 2007