Reviews/Commentaries/ADA Statements CONSENSUS STATEMENT

Screening for in Patients With Diabetes

1 4 JEROEN J. BAX, MD, PHD ROBERT O. BONOW, MD controlled trials was lacking. However, 2 5 LAWRENCE H. YOUNG, MD HELMUT O. STEINBERG, MD the perceived high prevalence of adverse 3 6 ROBERT L. FRYE, MD EUGENE J. BARRETT, MD, PHD cardiac outcomes underscored the need for considering diagnostic testing. The panel also suggested that additional re- Coronary artery disease (CAD) accounts for a large fraction of the morbidity, mortality, and cost search should be conducted to evaluate of diabetes. Recognizing this, nearly 10 years ago the American Diabetes Association published the recommendations articulated. a consensus recommendation that clinicians consider a risk factor–guided screening approach to Over the past 10 years, there has been early diagnosis of CAD in both symptomatic and asymptomatic patients. Subsequent clinical trial a greatly increased recognition, both in results have not supported those recommendations. Since the prior consensus statement, newer the medical community and among pa- imaging methods, such as coronary artery calcium scoring and noninvasive angiography with tients, of the prevalence and impact of computed tomography (CT) techniques, have come into use. These technologies, which allow quantitation of atherosclerotic burden and can predict risk of cardiac events, might provide an CAD among patients with diabetes. The approach to more widespread coronary screening. However, over this same time complicity of both conventional and in- interval, there has been recognition of diabetes as a (CVD) equivalent, flammatory risk markers in this increased clear demonstration that medical interventions should provide primary and secondary CVD risk risk has been extensively explored (3). reduction in diabetic populations, and suggestive evidence that percutaneous coronary revas- Perhaps most importantly, the benefit of cularization may not provide additive survival benefit to intensive medical management in both primary and secondary CVD risk patients with stable CAD. This additional evidence raises the question of whether documenting factor modification on cardiac outcomes asymptomatic atherosclerosis or in people with diabetes is warranted. More data has been proven in prospective interven- addressing this issue will be forthcoming from the BARI 2-D (Bypass Angioplasty Revascular- tional studies (4–9), and these results ization Investigation 2 Diabetes) trial. Until then, for patients with type 2 diabetes who are asymptomatic for CAD, we recommend that testing for atherosclerosis or ischemia, perhaps with have driven new guidelines for care in di- cardiac CT as the initial test, be reserved for those in whom medical treatment goals cannot be abetes (10). This may now be favorably met and for selected individuals in whom there is strong clinical suspicion of very-high-risk affecting population risks for events, at CAD. Better approaches to identify such individuals based on readily obtained clinical variables least in those with adequate access to care. are sorely needed. In clinical practice, the improved out- comes of aggressive medical therapy (11) Diabetes Care 30:2729–2736, 2007 are modifying the approach to treatment of patients with symptomatically less se- s coronary artery disease (CAD) is ing, and most critically, which patients vere CAD. the major cause of morbidity, mor- are particularly appropriate to be tested, Several studies of asymptomatic type A tality, and medical cost of diabetes the consensus panel recommended spe- 2 diabetic patients have specifically exam- (1), early diagnosis of CAD to prevent cialized CAD screening for patients ined whether risk factor burden (i.e., progression and clinical events has intui- thought to be at high risk. For the CAD number of risk factors) is predictive of in- tive appeal. In February 1998, the Amer- asymptomatic diabetic patient, the panel ducible ischemia (as determined by myo- ican Diabetes Association (ADA) particularly focused on risk factor burden cardial perfusion imaging), and these convened an expert panel to develop con- (i.e., the number of risk factors, not the have not supported the recommendation sensus positions addressing a series of severity of any one), baseline electrocar- of the 1998 consensus panel for screening questions related to coronary heart dis- diogram (ECG), and whether there was asymptomatic patients with two or more ease in both symptomatic and asymptom- clinical evidence of vascular disease at risk factors (12–14). At the same time, atic patients with diabetes (2). After other sites. The panel recognized that the evidence has accumulated regarding discussing the potential value of early di- positions articulated represented expert newer CAD diagnostic modalities, e.g., agnosis, the types and frequency of test- opinion, since evidence from well- CT angiography (15), coronary artery cal- ●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● cium scoring (16,17), and cardiac mag- netic resonance imaging (18), that are From the 1Department of Cardiology, Leiden University Medical Center, Leiden, the Netherlands; the 2Departments of Medicine and Physiology, Section of Cardiovascular Medicine, Yale University School of being implemented in strategies to diag- Medicine, New Haven, Connecticut; the 3Department of Medicine, Cardiovascular Division, Mayo Clinic nose and stage CAD. These modalities College of Medicine, Rochester, Minnesota; the 4Department of Medicine, Division of Cardiology, North- likely have implications for diabetic as western University Feinberg School of Medicine, Chicago, Illinois; the 5Department of Medicine, Division of well as nondiabetic patients (2). How- 6 Endocrinology and Metabolism, Indiana University, Indianapolis, Indiana; and the Department of Medi- ever, data that could provide a robust “ev- cine, Division of Endocrinology and Metabolism, University of Virginia, Charlottesville, Virginia. Address correspondence and reprint requests to Eugene Barrett, MD, PhD, University of Virginia, Box idence-based” recommendation for CAD 801410, Charlottesville, VA 22903. E-mail: [email protected]. testing in diabetic patients are not avail- Abbreviations: ADA, American Diabetes Association; CAD, coronary artery disease; CT, computed able. Recognizing these issues, the ADA tomography; CVD, cardiovascular disease; ECG, electrocardiogram; MI, . recently convened an expert panel that A table elsewhere in this issue shows conventional and Syste`me International (SI) units and conversion factors for many substances. revisited the issue of screening for CAD in DOI: 10.2337/dc07-9927. diabetic patients. The panel approached © 2007 by the American Diabetes Association. this by addressing four specific questions:

DIABETES CARE, VOLUME 30, NUMBER 10, OCTOBER 2007 2729 Screening for CAD in diabetic patients

1. Which patients with diabetes are at in- cal surveillance, or revascularization of warrant careful consideration for identi- creased risk for adverse cardiovascular their CAD. However, clinical factors that fying patients at risk for events. outcomes? confer risk for adverse cardiac outcomes Other atherosclerotic vascular disease. 2. What are the implications of an early do not always predict which patients will Clinically evident atherosclerotic disease diagnosis of coronary ischemia or ath- have abnormal screening tests (13,14), involving lower-extremity, cerebral, re- erosclerosis? and negative screening tests in patients nal, or mesenteric arteries identifies a pa- 3. What tests, or sequence of tests, with diabetes do not uniformly confer a tient with diabetes who is at increased risk should be considered? With what fre- benign prognosis (20,21). Tests that de- for adverse cardiovascular outcomes and quencies should testing be done? tect inducible ischemia or assess athero- might have advanced coronary athero- 4. What further research is needed to sclerotic burden do not always identify sclerosis (24). In patients with claudica- evaluate the effectiveness of these rec- those patients at risk for plaque rupture tion or asymptomatic peripheral arterial ommendations? and thrombosis, which typically leads to disease, 90% of deaths are attributable to acute coronary events. Further research CAD (25). Clinical history is important to The goal was to provide practitioners with focusing on the biological properties of the determine the presence of vascular symp- information on the role of specialized car- vessel wall is needed to identify individuals toms (transient ischemic attack, mesen- diac testing in their CAD asymptomatic at high risk for acute coronary events, and teric ischemia, or claudication), as is the diabetic patients, recognizing that dia- such evidence might in the future guide physical exam for bruits and peripheral betic patients with CAD symptoms are treatment in asymptomatic individuals. pulses. A diminished ankle-brachial in- typically referred for cardiac testing. The Patients with clinical CAD are gener- dex is a sensitive indicator of increased panel recognized that some patients, who ally classified into three categories: low risk for future CVD events (26). may be at high risk for CAD, do not have risk, with a cardiac mortality risk Ͻ1% Microalbuminuria and chronic kidney access to care that would allow them to per year; high risk, with a cardiac mortal- disease. Microalbuminuria predicts in- benefit from this guidance or from the ity risk Ͼ3% per year; or intermediate creased risk for vascular disease compli- proven benefits of risk modification. risk, with a cardiac mortality risk between cations (15,17,27,28), as well as for the Clearly, priority resources are needed to 1 and 3% per year. The overall population progression to overt nephropathy, in address these access disparities. The panel of asymptomatic patients with type 2 di- patients with type 1 or type 2 diabetes. also recognized the need for educational abetes falls into the intermediate cardiac Patients with type 1 diabetes and dimin- interventions, directed toward diabetic risk category. In UK Prospective Diabetes ished renal function often develop exten- patients and their families, dealing with Study (UKPDS), relatively healthy pa- sive atherosclerosis. Patients with type 2 early symptom recognition and response, tients recently diagnosed with type 2 dia- diabetes with chronic kidney disease due as well as lifestyle modifications that di- betes had a 10-year mortality rate of to diabetic nephropathy have a very high minish CVD risk. 18.9%, with an incidence of myocardial risk for MI and cardiac death, with 40% Finally, in these discussions, it was infarction (MI) of 17.4%, including fatal experiencing a cardiac complication over recognized that the preponderance of MI of 8.9% (22). In middle-aged subjects a 5-year period (25). Microalbuminuria data regarding both testing and risk factor with newly detected diet-controlled diabe- has been a predictor of inducible ischemia modification derive from studies of type 2 tes in the Diabetes Intervention Study, 15% in some (29), but not all (14), studies of diabetes. However, there is evidence from had evidence of myocardial infarction by 11 asymptomatic patients with diabetes. epidemiological studies that people with years of follow-up (23). However, within Abnormal resting ECG. Asymptomatic type 1 diabetes have significantly higher such groups, there is a range of risk that patients with type 2 diabetes occasionally CAD event rates than age-matched con- would be helpful to further understand. have evidence of previously unrecognized trol subjects; these rates exceed 1% per The 1998 Consensus Development MI on resting ECG, including abnormal year after age 45 years and 3% per year Conference proposed a group of patients Q-waves or deep T-wave inversions. after age 55 years (19). These estimates with diabetes who might have the highest These findings or the presence of a left- likely do not reflect the impact of newer yield from cardiac testing, those with bundle branch block usually trigger eval- risk-reduction interventions, but there is multiple CVD risk factors. Unfortunately, uation for CAD and inducible ischemia, no reason to suspect that this population recent studies have shown that the bur- as would be pursued in a patient with an- would respond less well to these than ei- den of traditional categorical risk factors gina. Testing in these patients should ther the type 2 diabetic or nondiabetic does not predict inducible ischemia on probably not be considered “screening,” populations. When appropriate, we point nuclear or echocardiographic myocardial but rather evaluation of an objective ab- out how positions articulated herein may perfusion imaging (13,14). normality for clinical reasons. However, apply to people with type 1 diabetes. In the absence of symptomatic CAD, nonspecific ST-T wave changes also are a clinical features that help to identify the strong predictor of inducible ischemia in 1. Which patients with diabetes are patient with increased risk for myocardial asymptomatic diabetic patients (12). at increased risk for adverse infarction or cardiac death include evi- Autonomic neuropathy. Cardiovascu- cardiovascular outcomes and should dence of other atherosclerotic vascular lar autonomic neuropathy is associated be screened? disease, renal disease, abnormal resting with a poor overall prognosis in patients The strategy of screening patients with di- electrocardiogram, diabetes complica- with type 2 diabetes (30). The mecha- abetes for advanced asymptomatic CAD is tions including autonomic neuropathy, nisms that confer the high risk are poorly motivated by the goal of identifying pa- age, sex, and both traditional and novel understood but may include impairment tients with high cardiac risk whose out- cardiac risk factors. Although these fac- in ischemia awareness, delaying the diag- comes might be improved through more tors might not specifically predict the nosis of CAD, or hemodynamic lability aggressive risk factor modification, medi- presence of inducible ischemia, they still due to blunted parasympathetic activa-

2730 DIABETES CARE, VOLUME 30, NUMBER 10, OCTOBER 2007 Bax and Associates tion. Autonomic neuropathy might also in these patients and when present por- erosclerotic involvement of the coronary be a parallel consequence of cardiac risk tends a very poor prognosis (42). arteries may now be made with measure- factors, including hyperglycemia, dyslip- Multiple cardiac risk factors. Patients ment of the amount of coronary calcifica- idemia, and renal disease. Autonomic with type 2 diabetes often have multiple tion. CT coronary angiography is also neuropathy was a major predictor of in- cardiac risk factors, including hyperten- being used to define the atherosclerotic ducible ischemia in the DIAD (Detection sion in ϳ50–60% of individuals, dyslip- burden, including noncalcified athero- of Ischemia in Asymptomatic Diabetics) idemia, inactivity, smoking, and sclerotic disease, and to estimate the de- study (14) and has been associated with abdominal obesity. Multiple risk factors gree of narrowing of individual lesions. abnormal cardiac test findings in other in the same patient substantially increase Documentation of coronary atherosclero- (31), but not all (32), studies. The ADA the overall cardiovascular risk (43). Fur- sis with noninvasive imaging has attracted has recently recommended screening for thermore, intervention directed at multi- the attention of patients, physicians, and cardiac autonomic neuropathy, at least ple risk factors significantly improves the general public. While the images are with a history and an examination for cardiovascular prognosis (44). Despite quite impressive and provide objective signs of autonomic dysfunction, begin- the rationale that these patients are logical evidence of coronary atherosclerosis, how ning at diagnosis of type 2 diabetes or 5 candidates for screening, recent CAD these tests should influence management years after the diagnosis of type 1 diabetes screening studies in type 2 diabetes have decisions is uncertain. For example, in the (33). The possibility of cardiac autonomic been unable to link the number of risk asymptomatic middle-aged or older dia- neuropathy should be considered in the factors to inducible ischemia on perfusion betic patient who already receives inten- presence of unexplained tachycardia, or- imaging (14). This may reflect the inabil- sive atherosclerotic risk factor therapy, thostatic hypertension and/or hypoten- ity of these studies to account for the se- “routine” imaging of the sion, and other autonomic or peripheral verity, duration, and effect of treatment of is not necessary to make decisions regard- neuropathies. both lipid abnormalities and hyperten- ing treatment for well-established risk Retinopathy. Although a manifestation sion in patients with long-standing type 2 factors. of microvascular disease, diabetic reti- diabetes. Coronary atherosclerosis may exist nopathy is also an indicator of risk for There remains a need to improve our with or without flow limitation in the epi- CAD in both type 1 and type 2 diabetes ability to identify, on the basis of readily cardial coronary arteries. Flow limitation (34,35). In clinical studies, retinopathy available clinical data, those patients at is a major consequence of epicardial cor- has been associated with inducible isch- highest risk for CVD events who would be onary atherosclerosis and is the basis of priority candidates for additional diag- “ischemic heart disease.” There is a large emia in some (36), but not all, screening nostic and/or therapeutic interventions. body of evidence documenting the rela- studies. While simple categorical risk factor bur- tionship between indexes of myocardial Hyperglycemia. Hyperglycemia is a den has not proven to effectively discrim- ischemia by noninvasive testing and the stronger predictor of microvascular dis- inate which asymptomatic diabetic presence of high-risk coronary anatomy. ease than atherosclerotic macrovascular patients will or will not have ischemia on Autopsy studies have documented a disease in people with diabetes (37,38). stress testing (14), it is still possible that greater prevalence of severe multivessel In clinical trials, interventions to improve risk factor burden might predict risk of CAD among patients with diabetes com- glycemic control reduced coronary events CVD events in individual patients. Efforts pared with those without diabetes, even in type 1 and type 2 populations (37,39) have been made using data from Framing- in the absence of prior symptoms or clin- and mortality in a type 1 population (39). ham (which included fewer than 400 di- ical evidence of disease (49). Ischemic ab- Hence, chronic undertreated hyperglyce- abetic subjects), the UKPDS (which normalities on noninvasive testing mia could be viewed as a risk factor for included only newly diagnosed diabetic predict outcomes better than presence or CVD. subjects and excluded patients with sig- absence of (21). Thus, the pre- Age and sex. Although diabetes in- nificant comorbidities), and other popu- sumed benefit to the individual asymp- creases relative cardiovascular risk more lations (45) to develop models that tomatic patient of assessing the presence in women than men, the absolute risk of identify individuals at highest risk for car- and extent of myocardial ischemia is to cardiovascular events is still higher in diovascular events. These efforts have identify those patients who would have a men than women (40). Age is an impor- been only modestly successful (45,46). survival benefit from coronary artery revas- tant determinant of cardiovascular risk, Another prediction model, Archimedes, cularization, specifically those with left and the prevalence of inducible ischemia available on the ADA Web site as Diabetes main or severe multivessel disease with a is significantly higher in patients with PHD, has proved useful in predicting large area of jeopardized myocardium. type 2 diabetes over the age of 65 years population clinical outcomes in response Patients with diabetes have significant (41). to specific interventions used in clinical risk for atherosclerotic vascular disease, Unexplained dyspnea. Frequently pro- trials (47,48). Such tools may provide and aggressive treatment of risk factors is viders are uncertain as to whether a pa- useful probabilistic guidance for diagnos- recommended in the absence of symp- tient’s complaint of is tic algorithms. tomatic or known CAD. Therefore, the attributable to myocardial ischemia (an role of coronary imaging in diabetic pa- angina equivalent) or simply to obesity 2. What are the implications of an tients is not to document the presence of and deconditioning. However, it is im- early diagnosis of coronary ischemia coronary atherosclerosis but to identify portant to recognize that, irrespective of or atherosclerosis? those with more extensive disease in cause, patients who are unable to exercise Several noninvasive imaging techniques whom further testing may be warranted are at increased cardiac risk. The inci- are evolving that enhance the anatomic in order to identify those with significant dence of inducible ischemia is increased diagnosis of CAD. Demonstration of ath- inducible myocardial ischemia. This is

DIABETES CARE, VOLUME 30, NUMBER 10, OCTOBER 2007 2731 Screening for CAD in diabetic patients based on the assumption (and the wide- to those patients with severe anatomic with full knowledge of the patient’s com- spread clinical practice) that patients with disease, where revascularization has par- plete cardiovascular risk profile. The up- severe myocardial ischemia involving a ticular benefits. However, extensive and dated American Heart Association large segment of the left ventricular myocar- expensive diagnostic testing to define the scientific statement (58) states that coro- dium are candidates for coronary angiogra- presence of CAD is not required before nary calcium testing is not valuable in in- phy and subsequent revascularization. The deciding to implement medical therapy dividuals at low Framingham risk but exact definition of “severe” ischemia is un- for established risk factors in most dia- may be useful as a screening tool in those known and has not been tested prospec- betic patients. There may be exceptions; at intermediate risk, which would include tively, but the available data (50) suggest an example is a patient with type 1 diabe- patients with diabetes. However, there that patients with ischemia involving 10% tes under 40 years of age who may have a was only limited support for coronary cal- or more of the left ventricle have a better 20- to 25-year history of diabetes but does cium testing of patients at intermediate outcome after myocardial revasculariza- not fall within current guidelines for ag- risk, with a class IIb recommendation tion compared with the results of medical gressive statin or aspirin therapy (10). (level of evidence B). Moreover, the therapy alone. Retrospective studies have American College of Cardiology appro- shown similar results in patients with di- 3. What tests, or sequence of tests, priateness criteria for cardiac CT (57) in- abetes (51). should be considered? With what dicate that the usefulness of screening The recently completed COURAGE frequencies should testing be done? asymptomatic intermediate-risk popula- (Clinical Outcomes Utilizing Revascular- The 1998 ADA Consensus Development tions with this technology is currently un- ization and Aggressive DruG Evaluation) Conference recommended exercise elec- known (56). trial (52) reported that among 2,287 pa- trocardiography to screen patients felt to Several studies have shown that the tients with stable angina, many of whom be high risk due to risk factor burden or coronary artery calcium score is valuable had multivessel or proximal left anterior presence of atherosclerosis at other sites. in identifying patients with a high likeli- descending coronary artery disease, in- Imaging for ischemia was recommended hood of inducible myocardial ischemia tensive medical treatment was as effective as the initial strategy only in patients with (59,60). These studies have consistently as percutaneous intervention combined abnormal resting electrocardiograms. observed that the likelihood of ischemia with intensive medical treatment in pre- Over the course of the last several years, a in patients with a calcium score Ͻ100 is venting overall mortality or myocardial number of important developments in negligible, whereas those with a score of infarction. Results were similar in the cardiovascular imaging technology have Ն400 have a relatively high likelihood of approximate one-third of subjects with evolved, together with a growing database inducible ischemia. It should be noted diabetes. It may not be appropriate to ex- regarding the role of imaging in detecting that these studies evaluated patients un- trapolate the results of this trial in symp- CAD in asymptomatic patients with risk dergoing stress testing for clinical indica- tomatic patients (in which the event rate factors, including diabetes. For example, tions, in whom the likelihood of CAD for diabetic subjects approached 5% per we note recent studies of cardiac CT dem- would be higher than in the population year) to somewhat lower-risk asymptom- onstrating prognostic value of such imag- that might be selected for screening, and atic patients. The hypothesis that asymp- ing (54,55). Moreover, professional did not specifically study patients with di- tomatic patients with severe ischemia societies have updated their recommen- abetes. However, Anand et al. (17) stud- benefit from revascularization over and dations for the use of CT imaging (56) and ied asymptomatic patients with diabetes above aggressive medical management re- have developed appropriateness criteria and confirmed the higher incidence of in- mains unproven and is currently the sub- for the use of nuclear cardiology proce- ducible ischemia in patients with higher ject of several prospective randomized dures, cardiac CT, and cardiac magnetic calcium scores. Nearly one-third of those trials, some specifically targeted to the di- resonance (57). These new recommenda- patients had a calcium score Ͼ400, and abetic population. The results of these trials tions and appropriateness criteria have 28% of which had large ischemic defects. will either support or disprove the validity important implications for screening The decision to undertake coronary of testing for myocardial ischemia. asymptomatic patients with and without artery calcium screening should be based Are there other advantages to estab- diabetes. on sound clinical judgment and the test lishing the diagnosis of “ischemic heart The broad strategy for screening performed only if the results have the po- disease,” including definition of coronary asymptomatic patients remains uncer- tential to change patient management. atherosclerotic burden? Some would ar- tain, based on the very limited database There are populations, including the el- gue that compliance to medical therapy is and the lack of prospective clinical trials. derly and those with renal insufficiency, enhanced with definition of the presence If in the judgment of the clinician an with a very high prevalence of coronary of disease, although this has not been es- asymptomatic patient is a candidate for calcification but in whom calcium scores tablished by controlled studies. Negative CAD testing, it is reasonable to apply car- are less predictive of ischemia, which lim- implications of diagnosing CAD in the ab- diac CT for detection of coronary artery its the value of such testing. In such pa- sence of symptoms may include anxiety, calcification, using either electron beam tients, any screening test might be compromised insurability, false-positive or multislice technology, as the first step. considered inappropriate. Alternatively, and negative tests, and more. The cost im- The coronary calcium score is an excellent one might proceed directly to stress im- plications must also be considered given marker for the overall coronary athero- aging to assess myocardial ischemia. the constraints on health care resources sclerotic burden and identifies asymp- If coronary calcium testing is per- (53). tomatic individuals at higher risk for formed, it appears reasonable to proceed In summary, accurate and early diagno- inducible ischemia. The calcium score with further testing in diabetic patients sis of coronary atherosclerotic/ischemic may also identify those at risk of subse- with calcium scores Ͼ400, considering heart disease is likely of benefit primarily quent coronary events but should be used factors such as age and renal function.

2732 DIABETES CARE, VOLUME 30, NUMBER 10, OCTOBER 2007 Bax and Associates

The threshold of 400 for further testing is tries would be particularly helpful in cap- tween the National Heart, Blood, and consistent with the 2006 American Col- turing general population risk data (65). Lung Institute, the National Institute of lege of Cardiology appropriateness criteria There is a significant need to be inclusive Diabetes and Digestive and Kidney Dis- (57). Further testing could be performed in the populations assessed, so that ade- eases, the Centers for Medicare and Med- using single photon emission tomogra- quate representation of both sexes and all icaid Services, and the industry would phy to assess myocardial perfusion or racial groups is captured. Such data are help facilitate the application of newer stress echocardiography to assess isch- not readily available in the U.S., where, technologies to diabetes cardiovascular emic wall motion abnormalities. Cardiac for the present, we would likely rely on research. Beyond that, the establishment magnetic resonance is now able to assess analyses of populations studied during of an international clinical research net- perfusion, wall motion, or both during clinical trials. Integration of the efforts at work devoted to the study of cardiovas- pharmacologic stress but is less widely the National Heart, Blood, and Lung In- cular disease in diabetes would help available and in most situations more stitute and the National Institute of Dia- further our understanding. The need for expensive. betes and Digestive and Kidney Diseases such a network is clear with the current The timing of serial stress imaging to facilitate analyses of this sort could po- worldwide epidemic of obesity and type 2 studies in patients with initially normal tentially fill much of this need. While it diabetes. stress tests is also uncertain. The issue of would be desirable to prospectively study the “warranty” period of a normal study a large cohort of individuals with type 2 Closing comments was addressed by Hachamovich et al. (61) diabetes without CVD symptoms to test Although the CAD asymptomatic patient in a general population using myocardial such predictive models, a more intense with diabetes is by definition at least at perfusion imaging and Elhendy et al. (62) analysis of pooled data from completed intermediate risk for CVD events, it is dif- in patients with diabetes with exercise studies offers the attractive advantage of ficult to support routine CAD screening echocardiography, both suggesting that being available in a much more timely for these patients. As previous recom- the cardiac event rate is low within 2 years fashion. mendations for stratifying diabetic pa- after normal stress myocardial perfusion The BARI 2D (Bypass Angioplasty Re- tients based upon the number of risk or echo studies but that events may sub- vascularization Investigation 2 Diabetes) factors have not proven effective, the sequently occur, presumably due to pro- trial, focused entirely on patients with question remains whether there are indi- gressive atherosclerosis. type 2 diabetes, will provide a next step in viduals with diabetes in whom coronary outcomes. In this trial, patients with doc- artery imaging would seem particularly 4. What further research is needed umented CAD have been randomized to appropriate. Presumably, the motivation to evaluate the effectiveness of these immediate revascularization combined for such testing would be the clinical sus- recommendations? with aggressive medical management or picion that the individual is at high risk Several considerations suggest that better to a program of aggressive medical manage- for having a CVD event in the short term. approaches to patient selection for CAD ment with delayed or no revascularization. Further development and testing of dia- screening would be of considerable value. The trial will provide a perspective on the betes-specific “risk engines” may be help- CARDS (Collaborative Atorvastatin Dia- benefits of revascularization in the era of ful in identifying these subjects. In betes Study) included individuals with modern medical therapy and important patients deemed on clinical grounds to be type 2 diabetes and no history of cardio- information for developing strategies for at particularly high risk, coronary artery vascular disease. In that study, the active the treatment of asymptomatic patients. calcium scoring may be the reasonable treatment group, in whom LDL choles- Proper trials testing the new imaging mo- first test, with subsequent functional im- terol was lowered to ϳ80 mg/dl, still had dalities, such as those called for in recent aging performed if the calcium scoring in- a 1.5% per year major cardiovascular American College of Cardiology consen- dicates a substantial atherosclerotic event rate. While other, non-LDL risk fac- sus recommendations, are a particularly burden. tors were not optimally treated in this important area of investigation. The re- study, its findings serve to emphasize that sults of these trials will have a major im- type 2 diabetic patients being treated ac- pact on the implementation of the References cording to intensive treatment guidelines recommendations outlined in the current 1. Hogan P, Dall T, Nikolov P: Economic likely have a residual “intermediate risk” article and will provide a basis for further costs of diabetes in the U.S. in 2002. Dia- betes Care 26:917–932, 2003 (1–2% per year) for cardiac events. In the revisions in the future. 2. American Diabetes Association: Consen- recently completed COURAGE trial Future studies should include inno- sus development conference on the diag- (63,64), overall event rates among dia- vative new approaches to characterization nosis of coronary heart disease in people betic patients were nearly 5% per year, of plaque structure and stability, consider with diabetes: 10–11 February 1998, Mi- again indicating a significant residual risk use of biomarkers, and perhaps incorpo- ami, Florida. Diabetes Care in intensively medically treated patients rate genetic testing in combination with 21:1551–1559, 1998 as well as those undergoing percutaneous imaging strategies. These trials should 3. Beckman JA, Creager MA, Libby P: Diabe- intervention. These findings would seem specifically include strong representation tes and atherosclerosis: epidemiology, to highlight the need for improved meth- of women and minorities who bear a dis- pathophysiology, and management. JAMA ods to stratify residual risk within popu- proportionate burden of disease and mor- 287:2570–2581, 2002 4. Colhoun HM, Betteridge DJ, Durrington lations undergoing intensive medical tality from diabetes. It will be critical as PN, Hitman GA, Neil HA, Livingstone SJ, management. part of future trials to determine whether Thomason MJ, Mackness MI, Charlton- As a first step, development and test- screening for CAD is cost effective and Menys V, Fuller JH: Primary prevention of ing of improved risk prediction models helps to improve patient adherence to cardiovascular disease with atorvastatin against data available from national regis- risk-reducing therapies. Partnership be- in type 2 diabetes in the Collaborative

DIABETES CARE, VOLUME 30, NUMBER 10, OCTOBER 2007 2733 Screening for CAD in diabetic patients

Atorvastatin Diabetes Study (CARDS): tion of silent myocardial ischemia in 24. Golomb BA, Dang TT, Criqui MH: Pe- multicentre randomised placebo-con- asymptomatic diabetic subjects: the DIAD ripheral arterial disease: morbidity and trolled trial. Lancet 364:685–696, 2004 study. Diabetes Care 27:1954–1961, mortality implications. Circulation 114: 5. MRC/BHF Heart Protection Study of cho- 2004 688–699, 2006 lesterol lowering with simvastatin in 15. Schuijf JD, Pundziute G, Jukema JW, 25. Mann JF, Gerstein HC, Pogue J, Bosch J, 20,536 high-risk individuals: a random- Lamb HJ, van der Hoeven BL, de Roos A, Yusuf S: Renal insufficiency as a predictor ised placebo-controlled trial. Lancet 360: van der Wall EE, Bax JJ: Diagnostic accu- of cardiovascular outcomes and the im- 7–22, 2002 racy of 64-slice multislice computed to- pact of ramipril: the HOPE randomized 6. Haffner SM, Alexander CM, Cook TJ, Boc- mography in the noninvasive evaluation trial. Ann Intern Med 134:629–636, 2001 cuzzi SJ, Musliner TA, Pedersen TR, Kjek- of significant coronary artery disease. Am J 26. Doobay AV, Anand SS: Sensitivity and shus J, Pyorala K: Reduced coronary Cardiol 98:145–148, 2006 specificity of the ankle-brachial index to events in simvastatin-treated patients 16. Mazzone T: The role of electron beam predict future cardiovascular outcomes: a with coronary heart disease and diabetes computed tomography for measuring systematic review. Arterioscler Thromb or impaired fasting glucose levels: sub- coronary artery atherosclerosis. Curr Diab Vasc Biol 25:1463–1469, 2005 group analyses in the Scandinavian Sim- Rep 4:20–25, 2004 27. Gerstein HC, Mann JF, Yi Q, Zinman B, vastatin Survival Study. Arch Intern Med 17. Anand DV, Lim E, Hopkins D, Corder R, Dinneen SF, Hoogwerf B, Halle JP, Young 159:2661–2667, 1999 Shaw LJ, Sharp P, Lipkin D, Lahiri A: Risk J, Rashkow A, Joyce C, Nawaz S, Yusuf S: 7. Effects of ramipril on cardiovascular and stratification in uncomplicated type 2 di- Albuminuria and risk of cardiovascular microvascular outcomes in people with abetes: prospective evaluation of the com- events, death, and heart failure in diabetic diabetes mellitus: results of the HOPE bined use of coronary artery calcium and nondiabetic individuals. JAMA 286: study and MICRO-HOPE substudy: Heart imaging and selective myocardial perfu- 421–426, 2001 Outcomes Prevention Evaluation Study sion scintigraphy. Eur Heart J 27:713– 28. Grimm RH Jr, Svendsen KH, Kasiske B, Investigators. Lancet 355:253–259, 2000 721, 2006 Keane WF, Wahi MM: Proteinuria is a risk [see comments] 18. Schuijf JD, Bax JJ, Shaw LJ, de Roos A, factor for mortality over 10 years of fol- 8. Pyorala K, Pedersen TR, Kjekshus J, Lamb HJ, van der Wall EE, Wijns W: low-up: MRFIT Research Group: Multiple Faergeman O, Olsson AG, Thorgeirsson Meta-analysis of comparative diagnostic Risk Factor Intervention Trial. Kidney Int G: Cholesterol lowering with simvastatin performance of magnetic resonance imag- Suppl 63:S10–S14, 1997 improves prognosis of diabetic patients ing and multislice computed tomography 29. Rutter MK, McComb JM, Brady S, Mar- with coronary heart disease: a subgroup for noninvasive coronary angiography. shall SM: Silent myocardial ischemia and analysis of the Scandinavian Simvastatin Am Heart J 151:404–411, 2006 microalbuminuria in asymptomatic sub- Survival Study (4S). Diabetes Care 20: 19. Soedamah-Muthu SS, Fuller JH, Mulnier jects with non-insulin-dependent diabe- 614–620, 1997 HE, Raleigh VS, Lawrenson RA, Colhoun tes mellitus. Am J Cardiol 83:27–31, 1999 9. Bhatt DL, Marso SP, Hirsch AT, Ringleb HM: All-cause mortality rates in patients 30. Vinik AI, Maser RE, Mitchell BD, Freeman PA, Hacke W, Topol EJ: Amplified benefit with type 1 diabetes mellitus compared R: Diabetic autonomic neuropathy. Dia- of clopidogrel versus aspirin in patients with a non-diabetic population from the betes Care 26:1553–1579, 2003 with diabetes mellitus. Am J Cardiol 90: UK general practice research database, 31. Valensi P, Sachs R-N, Harfouche B, 625–628, 2002 1992–1999. Diabetologia 49:660–666, Lormeau B, Paries J, Cosson E, Paycha F, 10. Buse J, Ginsberg H, Bakris G, Clark NG, 2006 Leutenegger M, Attali J-R: Predictive Costa F, Eckel RH, Fonseca V, Gerstein H, 20. De Lorenzo A, Lima RS, Siqueira-Filho value of cardiac autonomic neuropathy in Grundy SM, Nesto RW, Pignone MP, AG, Pantoja MR: Prevalence and prognos- diabetic patients with or without silent Plutzky J, Porte D, Redberg R, Stritzel KF, tic value of perfusion defects detected by myocardial ischemia. Diabetes Care 24: Stone NJ: Primary prevention of cardio- stress technetium-99m sestamibi myocar- 339–343, 2001 vascular diseasese in people with diabetes dial perfusion single-photon emission 32. Prevalence of unrecognized silent myo- mellitus. Circulation 115:114–126, 2007 computed tomography in asymptomatic cardial ischemia and its association with 11. Pitt B, Waters D, Brown WV, van Boven patients with diabetes mellitus and no atherosclerotic risk factors in noninsulin- AJ, Schwartz L, Title LM, Eisenberg D, known coronary artery disease. Am J Car- dependent diabetes mellitus: Milan Study Shurzinske L, McCormick LS: Aggressive diol 90:827–832, 2002 on Atherosclerosis and Diabetes (MiSAD) lipid-lowering therapy compared with 21. Giri S, Shaw LJ, Murthy DR, Travin MI, Group. Am J Cardiol 79:134–139, 1997 angioplasty in stable coronary artery dis- Miller DD, Hachamovitch R, Borges-Neto 33. Boulton AJM, Vinik AI, Arezzo JC, Bril V, ease: Atorvastatin versus Revasculariza- S, Berman DS, Waters DD, Heller GV: Im- Feldman EL, Freeman R, Malik RA, Maser tion Treatment Investigators. N Engl J Med pact of diabetes on the risk stratification RE, Sosenko JM, Ziegler D: Diabetic neu- 341:70–76, 1999 using stress single-photon emission com- ropathies: a statement by the American 12. Rajagopalan N, Miller TD, Hodge DO, puted tomography myocardial perfusion Diabetes Association (ADA Statement). Frye RL, Gibbons RJ: Identifying high- imaging in patients with symptoms sug- Diabetes Care 28:956–962, 2005 risk asymptomatic diabetic patients who gestive of coronary artery disease. Circu- 34. Klein BE, Klein R, McBride PE, Cruick- are candidates for screening stress single- lation 105:32–40, 2002 shanks KJ, Palta M, Knudtson MD, Moss photon emission computed tomography 22. Adler AI, Neil HA, Manley SE, Holman SE, Reinke JO: Cardiovascular disease, imaging. J Am Coll Cardiol 45:43–49, RR, Turner RC: Hyperglycemia and hy- mortality, and retinal microvascular char- 2005 perinsulinemia at diagnosis of diabetes acteristics in type 1 diabetes: Wisconsin 13. Scognamiglio R, Negut C, Ramondo A, and their association with subsequent car- epidemiologic study of diabetic retinopa- Tiengo A, Avogaro A: Detection of coro- diovascular disease in the United King- thy. Arch Intern Med 164:1917–1924, nary artery disease in asymptomatic pa- dom prospective diabetes study (UKPDS 2004 tients with type 2 diabetes mellitus. JAm 47). Am Heart J 138:S353–S359, 1999 35. Hiller R, Sperduto RD, Podgor MJ, Ferris Coll Cardiol 47:65–71, 2006 23. Hanefeld M, Schmechel H, Schwanebeck FLd, Wilson PW: Diabetic retinopathy 14. Wackers FJ, Young LH, Inzucchi SE, U, Lindner J: Predictors of coronary heart and cardiovascular disease in type II dia- Chyun DA, Davey JA, Barrett EJ, Taillefer disease and death in NIDDM: the Diabe- betics: the Framingham Heart Study and R, Wittlin SD, Heller GV, Filipchuk N, tes Intervention Study experience. Diabe- the Framingham Eye Study. Am J Epide- Engel S, Ratner RE, Iskandrian AE: Detec- tologia 40 (Suppl. 2):S123–S124, 1997 miol 128:402–409, 1988

2734 DIABETES CARE, VOLUME 30, NUMBER 10, OCTOBER 2007 Bax and Associates

36. Akasaka T, Yoshida K, Hozumi T, Takagi 48. Eddy DM, Schlessinger L: Archimedes: a Consensus Document on Electron Beam T, Kaji S, Kawamoto T, Morioka S, Yo- trial-validated model of diabetes. Diabetes Computed Tomography). Circulation 115: shikawa J: Retinopathy identifies marked Care 26:3093–3101, 2003 402–426, 2007 restriction of coronary flow reserve in pa- 49. Goraya TY, Leibson CL, Palumbo PJ, 57. Hendel RC, Patel MR, Kramer CM, Poon tients with diabetes mellitus. J Am Coll Weston SA, Killian JM, Pfeifer EA, Jacob- M, Hendel RC, Carr JC, Gerstad NA, Gil- Cardiol 30:935–941, 1997 sen SJ, Frye RL, Roger VL: Coronary ath- lam LD, Hodgson JM, Kim RJ, Kramer 37. Intensive blood-glucose control with sul- erosclerosis in diabetes mellitus: a CM, Lesser JR, Martin ET, Messer JV, Red- phonylureas or insulin compared with population-based autopsy study. JAm berg RF, Rubin GD, Rumsfeld JS, Taylor conventional treatment and risk of com- Coll Cardiol 40:946–953, 2002 AJ, Weigold WG, Woodard PK, Brindis plications in patients with type 2 diabetes 50. Hachamovitch R, Hayes SW, Friedman RG, Hendel RC, Douglas PS, Peterson ED, (UKPDS 33): UK Prospective Diabetes JD, Cohen I, Berman DS: Comparison of Wolk MJ, Allen JM, Patel MR: ACCF/ Study (UKPDS) Group. Lancet 352:837– the short-term survival benefit associated ACR/SCCT/SCMR/ASNC/NASCI/SCAI/ 853, 1998 [published erratum appears in with revascularization compared with SIR 2006 appropriateness criteria for car- Lancet 354:602, 1999] medical therapy in patients with no prior diac computed tomography and cardiac 38. Laakso M: Hyperglycemia and cardiovas- coronary artery disease undergoing stress magnetic resonance imaging: a report of cular disease in type 2 diabetes. Diabetes myocardial perfusion single photon emis- the American College of Cardiology 48:937–942, 1999 sion computed tomography. Circulation Foundation Quality Strategic Directions 39. Nathan DM, Cleary PA, Backlund JY, Ge- 107:2900–2907, 2003 Committee Appropriateness Criteria nuth SM, Lachin JM, Orchard TJ, Raskin 51. Sorajja P, Chareonthaitawee P, Rajago- Working Group, American College of Ra- P, Zinman B: Intensive diabetes treatment palan N, Miller TD, Frye RL, Hodge DO, diology, Society of Cardiovascular Com- and cardiovascular disease in patients Gibbons RJ: Improved survival in asymp- puted Tomography, Society for with type 1 diabetes. N Engl J Med 353: tomatic diabetic patients with high-risk Cardiovascular Magnetic Resonance, 2643–2653, 2005 SPECT imaging treated with coronary ar- American Society of Nuclear Cardiology, 40. Abbott RD, Donahue RP, Kannel WB, tery bypass grafting. Circulation 112: North American Society for Cardiac Im- Wilson PW: The impact of diabetes on I311–316, 2005 aging, Society for Cardiovascular Angiog- survival following myocardial infarction 52. Boden WE, O’Rourke RA, Teo KK, Harti- raphy and Interventions, and Society of in men vs women: the Framingham gan PM, Maron DJ, Kostuk W, Knudtson Interventional Radiology. J Am Coll Car- Study. JAMA 260:3456–3460, 1988 M, Dada M, Casperson P, Harris CL, Sper- diol 48:1475–1497, 2006 41. Chaowalit N, Arruda AL, McCully RB, tus JA, Shaw L, Chaitman BR, Mancini 58. Budoff MJ, Achenbach S, Blumenthal RS, Bailey KR, Pellikka PA: Dobutamine stress GB, Berman DS, Weintraub WS: Design Carr JJ, Goldin JG, Greenland P, Guerci echocardiography in patients with diabe- and rationale of the Clinical Outcomes AD, Lima JA, Rader DJ, Rubin GD, Shaw tes mellitus: enhanced prognostic predic- Utilizing Revascularization and Aggres- LJ, Wiegers SE: Assessment of coronary tion using a simple risk score. J Am Coll sive DruG Evaluation (COURAGE) trial artery disease by cardiac computed to- Cardiol 47:1029–1036, 2006 Veterans Affairs Cooperative Studies Pro- mography: a scientific statement from the 42. Vanzetto G, Halimi S, Hammoud T, Fa- gram no. 424. Am Heart J 151:1173–1179, American Heart Association Committee gret D, Benhamou PY, Cordonnier D, 2006 on Cardiovascular Imaging and Interven- Denis B, Machecourt J: Prediction of car- 53. Diamond GA, Kaul S, Shah PK: Screen tion, Council on Cardiovascular Radiol- diovascular events in clinically selected testing cardiovascular prevention in ogy and Intervention, and Committee on high-risk NIDDM patients: prognostic asymptomatic diabetic patients. J Am Coll Cardiac Imaging, Council on Clinical value of exercise stress test and thallium- Cardiol 49:1915–1917, 2007 Cardiology. Circulation 114:1761–1791, 201 single-photon emission computed 54. Shaw LJ, Raggi P, Schisterman E, Berman 2006 tomography. Diabetes Care 22:19–26, DS, Callister TQ: Prognostic value of car- 59. Berman DS, Wong ND, Gransar H, 1999 diac risk factors and coronary artery cal- Miranda-Peats R, Dahlbeck J, Hayes SW, 43. Mortality after 16 years for participants cium screening for all-cause mortality. Friedman JD, Kang X, Polk D, Hachamo- randomized to the Multiple Risk Factor Radiology 228:826–833, 2003 vitch R, Shaw L, Rozanski A: Relationship Intervention Trial. Circulation 94:946– 55. Raggi P, Shaw LJ, Berman DS, Callister between stress-induced myocardial isch- 951, 1996 TQ: Prognostic value of coronary artery emia and atherosclerosis measured by 44. Gaede P, Vedel P, Larsen N, Jensen GV, calcium screening in subjects with and coronary calcium tomography. J Am Coll Parving HH, Pedersen O: Multifactorial without diabetes. J Am Coll Cardiol 43: Cardiol 44:923–930, 2004 intervention and cardiovascular disease in 1663–1669, 2004 60. He ZX, Hedrick TD, Pratt CM, Verani MS, patients with type 2 diabetes. N Engl J Med 56. Greenland P, Bonow RO, Brundage BH, Aquino V, Roberts R, Mahmarian JJ: Se- 348:383–393, 2003 Budoff MJ, Eisenberg MJ, Grundy SM, verity of coronary artery calcification by 45. Guzder RN, Gatling W, Mullee MA, Lauer MS, Post WS, Raggi P, Redberg RF, electron beam computed tomography Mehta RL, Byrne CD: Prognostic value of Rodgers GP, Shaw LJ, Taylor AJ, Wein- predicts silent myocardial ischemia. Cir- the Framingham cardiovascular risk traub WS, Harrington RA, Abrams J, culation 101:244–251, 2000 equation and the UKPDS risk engine for Anderson JL, Bates ER, Grines CL, Hlatky 61. Hachamovitch R, Hayes S, Friedman JD, coronary heart disease in newly diag- MA, Lichtenberg RC, Lindner JR, Pohost Cohen I, Shaw LJ, Germano G, Berman nosed type 2 diabetes: results from a GM, Schofield RS, Shubrooks SJ Jr, Stein DS: Determinants of risk and its temporal United Kingdom study. Diabet Med 22: JH, Tracy CM, Vogel RA, Wesley DJ: variation in patients with normal stress 554–562, 2005 ACCF/AHA 2007 clinical expert consen- myocardial perfusion scans: what is the 46. The Mount Hood 4 Modeling Group: sus document on coronary artery calcium warranty period of a normal scan? JAm Computer modeling of diabetes and its scoring by computed tomography in Coll Cardiol 41:1329–1340, 2003 complications: a report on the Fourth global cardiovascular risk assessment and 62. Elhendy A, Arruda AM, Mahoney DW, Mount Hood Challenge meeting. Diabetes in evaluation of patients with : a Pellikka PA: Prognostic stratification of Care 30:1638–1646, 2007 report of the American College of Cardi- diabetic patients by exercise echocardiog- 47. Eddy DM, Schlessinger L: Validation of ology Foundation Clinical Expert Con- raphy. J Am Coll Cardiol 37:1551–1557, the archimedes diabetes model. Diabetes sensus Task Force (ACCF/AHA Writing 2001 Care 26:3102–3110, 2003 Committee to Update the 2000 Expert 63. Boden WE, O’Rourke RA, Koon KT, Har-

DIABETES CARE, VOLUME 30, NUMBER 10, OCTOBER 2007 2735 Screening for CAD in diabetic patients

tigan PM, Maron DJ, Kostuk WJ, Knudt- Mancini GB, Weintraub WS: Optimal AD: Derivation and validation of a predic- son M, Dada M, Casperson P, Harris CL, medical therapy with or without PCI for tion score for major coronary heart dis- Chaitman BR, Shaw L, Gosselin G, Nawaz stable coronary disease. N Engl J Med 356: ease events in a U.K. type 2 diabetic S, Title LM, Gau G, Blaustein MD, Booth 1503–1516, 2007 population. Diabetes Care 29:1231–1236, DC, Bates ER, Spertus JA, Bergman DS, 64. Donnan PT, Donnelly L, New JP, Morris 2006

2736 DIABETES CARE, VOLUME 30, NUMBER 10, OCTOBER 2007