Overestimation of Cardiovascular Events Risk by the Atherosclerotic Cardiovascular Disease Risk Score

Study Aim:
To study how the AHA/ACC atherosclerotic cardiovascular disease (ASCVD) risk score performs among four different race/ethnic groups (White, Black, Chinese, and Hispanic) using data from the Multi-Ethnic Study of Atherosclerosis (MESA) cohort. Some authors of the ASCVD risk estimation guideline group previously stated that the overestimation is attributable to the inclusion of Chinese and Hispanics, since the derivation cohort for the ASCVD risk estimator does not include adults of these ethnicities. Another suggestion is that MESA is not representative because it is a highly treated population (with aspirin, anti-hypertensive agents, and statins).

Methods:
The investigators included 6,441 participants aged between 45-79 years from MESA, a prospective community-based cohort. MESA participants were compared with National Health and Nutrition Examination Survey (NHANES) 2013-2014 data to evaluate applicability of the cohort to the US population. The ASCVD risk score was calculated using baseline MESA data from 2000-2002, and expected and observed cardiovascular events in each race/ethnicity and sex were compared in 10-year follow-up. To ascertain the impact of each risk factor on the difference between observed events and expected events based on ASCVD, univariable and multivariable absolute risk regression was performed.

Results:
The MESA participants (mean age 61 years; 53% women) were similar to contemporary NHANES data except for age (62 vs. 56 years), ethnic diversity (whites 39% vs. 69%), and HDL cholesterol (51 mg/dL vs. 54 mg/dL). The calculated ASCVD risk scores and the proportion of participants on anti-hypertensive treatment and lipid lowering therapy in MESA were similar to the national representative NHANES sample. In comparison, the AHA-ACC ASCVD derivation cohort participants (which included individuals from several decades ago) compared with the NHANES data had significant differences in gender distribution (53% vs. 56%), ethnic diversity (whites 83% vs. 69%), total cholesterol (194 mg/dL vs. 215 mg/dL), smoking (26% vs. 19%), diabetes (9% vs. 13%), and baseline anti-hypertensive medication use (21% vs. 37%).

Among MESA participants, the ASCVD risk score overestimated cardiovascular events in both women (predicted events 9%; observed events 4%) and men (predicted events 14%; observed events 7%), with the risk score predicting approximately twice the number of observed events. The overestimation was seen in all four race/ethnic groups (discordance range from 67% in Hispanic men up to 314% in Chinese men).

In multivariable analyses, higher age, Chinese ethnicity, systolic blood pressure, alcohol use, diabetes, exercise, lipid-lowering therapy, and aspirin therapy were associated with risk overestimation (p < 0.05). Estimation of the ASCVD Cox model in MESA leads to significantly different coefficients for each of the variables included in the risk score.

Conclusions:
The investigators conclude that the ASCVD risk score significantly overestimates cardiovascular events in all four race/ethnic groups and in both genders in the MESA study. This study confirms that that the ASCVD risk score often overestimates an older adult's contemporary risk for myocardial infarction (MI) or stroke. Thus, it is important that clinicians and patients enter into a detailed risk discussion before embarking on pharmacologic therapy.

Perspective:
This study elaborates on previous reports that the ASCVD risk score applied to other cohorts may overestimate cardiovascular events.1-3 Several concerns had been raised of previous reports of potential overestimation:

  • Lack of active surveillance for cardiovascular events in addition to self-reported hospitalizations.
  • Inclusion of race/ethnic minorities, as overestimation may only be observed in Chinese and Hispanics who were not included in the pooled cohorts used to derive the ASCVD risk score.
  • More intensive preventive pharmacotherapy such as use of statin therapy, aspirin, anti-hypertensive agents, especially due to the availability of additional risk stratification data, e.g., coronary artery calcium, which was not available in prior cohort studies.
  • Increased use of revascularization procedures for unstable angina which may reduce hard MI events.

The authors tried to address several of these previous concerns. The MESA study classifies cardiovascular events through several methods in addition to participant reports (death certificates, hospitalization records, autopsy, and from physicians, relatives or friends for out-of-hospital deaths).4 The investigators also adjudicated possible missed events by crosslinking with the CMS database with no meaningful change in their results.

This study reports that the overestimation of events by the ASCVD risk score was present in White, Black, Chinese and Hispanics and not limited to specific ethnicities. Importantly, the authors were able to demonstrate that the MESA study population and natural history was similar to the wider American population by comparing data with NHANES. An alternative possibility is that the derivation cohort participants may not adequately/fully represent the modern American population, helping to explain the discordance seen between the observed and expected events in their multicultural cohort.

One other criticism of the MESA study can be that the participants underwent coronary artery calcium scoring which may have led to both individual and physician behavior change resulting in lower number of events. However, this study reported limited baseline testing data to the MESA participants. Additionally, although it is still possible that lifestyle advice and changes by physicians and participants were more aggressively pursued than in the general population, the authors do demonstrate that anti-hypertensive and lipid lowering treatment were comparable between the MESA study and NHANES data.

In addition, the investigators identified several risk factors that are associated with risk overestimation. The risk factors were higher age (most important), Chinese ethnicity, systolic blood pressure, alcohol use, diabetes, exercise, lipid-lowering therapy, and aspirin therapy. In MESA the relationship of risk factors to cardiovascular events was significantly different than in the ASCVD derivation cohort. These findings are consistent with other reports, for example recent published data from the Framingham Heart Study show that the prevalence of risk factors have changed over the decades.5

The use of invasive therapy has increased over the past 15 years. However, in the Women's Health Study (after accounting for revascularization) there was still an overestimation of events by the ASCVD risk score.6 In addition, although the number of acute myocardial infarctions have decreased according to NHANES data driven by a decrease in mainly STEMI, the number of NSTEMIs have remained stable between 2002-2011.7,8 In addition, including revascularizations of unstable angina in the MESA study as events would still likely lead to overestimation by the ASCVD risk score.9

This report adds to the ongoing concerns for the use of 2013 ASCVD risk score. However, other studies comparing the ASCVD risk score based approach versus an approach using the enrollment criteria of major trials or a hybrid approach have reported that the ASCVD risk score based approach performs the best in identifying patients to treat.10 Thus, although the ASCVD risk score often overestimates risk, currently it provides a useful guide to undertake a primary prevention strategy with clinicians and patients engaging in a discussion of the treatment plan as recommended by the AHA/ACC guidelines.11-13

These discussions of management strategies particularly for individuals with an ASCVD risk score of 5-15% could also be guided by the measurement of coronary artery calcium and lipoprotein(a).14-16 In addition, clinicians should note that the ASCVD risk estimator does highlight (when you select "Other") that the model may significantly under- or overestimate in other race/ethnic groups. The development of a better risk prediction tool needs to use newer risk factor and outcomes data, and include data from multiple races/ethnic groups that are representative of the American population.

References

  1. Ridker PM, Cook NR. Statins: new American guidelines for prevention of cardiovascular disease. Lancet 2013;382:1762-5.
  2. DeFilippis AP, Young R, Carrubba CJ, et al. An analysis of calibration and discrimination among multiple cardiovascular risk scores in a modern multiethnic cohort. Ann Intern Med 2015;162:266-75.
  3. Muntner P, Colantonio LD, Cushman M, et al. Validation of the atherosclerotic cardiovascular disease Pooled Cohort risk equations. JAMA 2014;311:1406-15.
  4. Bild DE, Bluemke DA, Burke GL, et al. Multi-Ethnic Study of Atherosclerosis: objectives and design. Am J Epidemiol 2002;156:871-81.
  5. Schnabel RB, Yin X, Gona P,et al. 50 year trends in atrial fibrillation prevalence, incidence, risk factors, and mortality in the Framingham Heart Study: a cohort study. Lancet 2015;386:154-62.
  6. Cook NR, Ridker PM. Further insight into the cardiovascular risk calculator: the roles of statins, revascularizations, and underascertainment in the Women's Health Study. JAMA Intern Med 2014;174:1964-71.
  7. Yoon SS, Dillon CF, Illoh K, Carroll M. Trends in the prevalence of coronary heart disease in the U.S.: National Health and Nutrition Examination Survey, 2001-2012. Am J Prev Med 2016. [Epub ahead of print]
  8. Khera S, Kolte D, Aronow WS, et al. Non-ST-elevation myocardial infarction in the United States: contemporary trends in incidence, utilization of the early invasive strategy, and in-hospital outcomes. J Am Heart Assoc 2014;3:e00095.
  9. Silverman MG, Harkness JR, Blankstein R, et al. Baseline subclinical atherosclerosis burden and distribution are associated with frequency and mode of future coronary revascularization: multi-ethnic study of atherosclerosis. JACC Cardiovasc Imaging 2014;7:476-86.
  10. Mortensen MB, Afzal S, Nordestgaard BG, Falk E. Primary prevention with statins: ACC/AHA risk-based approach versus trial-based approaches to guide statin therapy. J Am Coll Cardiol 2015;66:2699-709.
  11. Martin SS, Sperling LS, Blaha MJ, et al. Clinician-patient risk discussion for atherosclerotic cardiovascular disease prevention: importance to implementation of the 2013 ACC/AHA Guidelines. J Am Coll Cardiol 2015;65:1361-8.
  12. Navar AM, Stone NJ, Martin SS. What to say and how to say it: effective communication for cardiovascular disease prevention. Curr Opin Cardiol 2016;31:537-44.
  13. Stone NJ, Robinson JG, Lichtenstein AH, et al. 2013 ACC/AHA guideline on the treatment of blood cholesterol to reduce atherosclerotic cardiovascular risk in adults: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines. Circulation 2014;129:S1-45.
  14. Detrano R, Guerci AD, Carr JJ, et al. Coronary calcium as a predictor of coronary events in four racial or ethnic groups. N Engl J Med 2008;358:1336-45.
  15. Polonsky TS, McClelland RL, Jorgensen NW, et al. Coronary artery calcium score and risk classification for coronary heart disease prediction. JAMA 2010;303:1610-6.
  16. Emerging Risk Factors C, Erqou S, Kaptoge S, et al. Lipoprotein(a) concentration and the risk of coronary heart disease, stroke, and nonvascular mortality. JAMA 2009;302:412-23.

Clinical Topics: Diabetes and Cardiometabolic Disease, Clinical Topic Collection: Dyslipidemia, Prevention, Advanced Lipid Testing, Lipid Metabolism, Nonstatins, Novel Agents, Statins, Smoking

Keywords: Angina, Unstable, Antihypertensive Agents, Atherosclerosis, Blood Pressure, Cholesterol, HDL, Cohort Studies, Coronary Vessels, Diabetes Mellitus, Hydroxymethylglutaryl-CoA Reductase Inhibitors, Lipoprotein(a), Myocardial Infarction, Primary Prevention, Risk Factors, Smoking, Stroke, Dyslipidemias


< Back to Listings