Author + information
- †Icahn School of Medicine at Mount Sinai, New York, New York
- ‡University of Minnesota, Minneapolis, Minnesota
- ↵∗Reprint requests and correspondence:
Dr. Harvey Hecht, Icahn School of Medicine at Mount Sinai Cardiology, 1111 Amsterdam Avenue, New York, New York 10025.
Insanity: doing the same thing over and over again and expecting different results.
—Albert Einstein (1)
Risk stratification using historical elements (age, smoking and family history), clinical examination (hypertension) or biomarkers (lipids and glucose, etc.) predates the availability of high-quality imaging and was the main vehicle for cardiovascular risk prediction. It was soon realized that these markers, even in combination, were at best modest in their efficacy. Moreover, the bulk of events still occurred in the large population with low-risk scores, suggesting the need to develop better paradigms of risk prediction. Imaging has allowed us to visualize directly and indirectly atheromatous plaque and can unveil the consummate biomarker. It was but natural that imaging markers including the coronary artery calcium score (CAC) would be very effective tools for risk stratification as supported by multiple studies. CAC score seems to be far more facile than any other available marker for noninvasively predicting risk in both asymptomatic and symptomatic subjects. It allows reclassification of risk derived from conventional risk factor testing (net reclassification index). The question then naturally arises whether conventional risk factor based stratification schemes exert their limited predictive ability through: 1) just identifying the likelihood of coronary plaque burden; 2) better correlating with plaques that had a turbulent history (e.g., calcified plaque) or correlating with a tendency for plaque vulnerability; or 3) contributing to plaque formation, progression or rupture as in the case of low-density lipoprotein cholesterol, diabetes, smoking, and other factors. For all of these hypotheses, one could argue it would be better to just visualize the plaque and plaque progression itself via imaging (as with computed tomography imaging and CAC scoring) instead of guessing its presence or fate. Indeed, the finding of subclinical atherosclerosis through imaging is likely to upset the current logical notion that the results of population-wide risk factors can be extrapolated reliably to individuals for the institution of primary prevention.
Although the relationship between CAC and risk factors has been extensively investigated in the prediction of future sickness there is less information about how well they relate to each other in predicting health (vascular health)? The ability to remain true in sickness and in health is a much-cherished vow in the marital contract, but risk factor–based paradigms have shown less than optimal fidelity in sickness (see below), and, from a report in this issue of iJACC (2), in health as well.
Whelton et al. (2) tried to answer whether any risk profile can also reliably predict healthy aging of the coronaries. They evaluated 1,850 asymptomatic MESA (Multi-Ethnic Study of Atherosclerosis) participants with 0 CACs at baseline who underwent repeat CAC at a median of almost 10 years after the initial scan. Absence of CAC persisted in 55% of the cohort while 45% developed a median CAC score of 25. Healthy aging was defined as persistence of a 0 calcium score. A baseline atherosclerotic cardiovascular (ASCVD) risk score of <2.5% was associated with a greater probability of healthy aging compared with a score of ≥7.5%. Although the absence of risk factors was associated with healthy aging, there was no association of persistent 0 CAC with the Healthy Lifestyle variables of body mass index, physical activity, Mediterranean diet, and never smoking; no single traditional risk factor or combination of risk factors improved on the crude area under the curve of 0.65 obtained from the combination of age, gender, and ethnicity. During the decade-long follow-up, more than one-third of high-risk subjects (age ≥65 years, or those with ASCVD ≥7.5% or ≥3 traditional risk factors) remained free of coronary artery calcification. A similar proportion of low-risk subjects (age <55 years, ASCVD <2.5%, or those with no traditional risk factors) developed coronary calcification. Carotid intima media thickening and high-sensitivity C-reactive protein also played no role in predicting incident CAC. The findings are consistent with the earlier report from Min et al. (3), which evaluated 422 individuals with 0 CAC scores at baseline with annual scanning for 5 years. In the 106 who developed CAC >0, only age, diabetes, and smoking were associated with incident CAC, but no risk factors contributed to accelerated times to conversion. Among 621 with CAC scores of >0 at baseline, only the presence of CAC itself was associated with progression without any contribution from the presence or absence of traditional risk factors (3). This raises the interesting possibility that vascular health or the lack of it correlates only modestly with the presence or absence of a healthy or unhealthy risk factor profile. Because no identifiable low-risk phenotype could predict healthy arterial aging, we need to be looking elsewhere to predict the transition of healthy to unhealthy vasculature—imaging may fill in some of the gap in the future.
Although the findings from Whelton et al. (2) are interesting, they obviously should not detract from the benefits of practicing a healthy lifestyle, because the study was not designed to evaluate the effects of lifestyle and behavioral components on plaque development and progression. We should also note that the study population was extremely low risk, with 10-year hard event rates of 0.2% and 1.3% in the persistent 0 CAC and incident CAC groups, respectively, which renders differences in plaque evolution in response to any intervention very unlikely. Moreover, in an analysis of the entire eligible MESA population of 6,229 participants, the same group demonstrated slower CAC plaque progression and lower mortality in those with higher healthy lifestyle scores, albeit with a shorter scan interval and follow-up (4). The results must not be misinterpreted as a failure of healthy lifestyle, but must force us to think about additional and possibly better biomarkers for vascular risk.
Their report exemplifies the need for a balance between analyzing data from a population wide study and application of the results to individuals. It is seemingly easy to conclude from the statistical outcomes of a cohort study that a younger age, female gender, and lack of traditional risk factors are likely to be associated with the persistence of healthy arterial aging in an individual patient. This study, however, provides some data to the contrary and thus raises questions about clinical applicability, in view of the heterogeneity of disease development in individual subjects, wherein traditional risk factor-based prediction might be wrong in one-third of subjects on either side of the risk factor spectrum. Of course, risk factors have other pathogenetic aspects beyond just correlation with CAC (even when used as an index of vascular health) and it would be unreasonable to believe that events they predict happen only in those with development of CAC. The study did not measure serial changes in risk factors and their associations with serial changes in CAC which could present a different picture. It is also important to remember that the ultimate goals are first to predict and then, more importantly, to prevent events, for which CAC determined vascular age is an interesting intermediate step, with final proof yet to be provided.
The results of the current study mirror the observations made in substantially larger studies: the inability of risk factors to robustly predict outcomes in primary prevention and the superiority of the demonstration of subclinical atherosclerosis to the mere presence of risk factors. In more than one-half million patients presenting with a first myocardial infarction, almost 50% had ≤1 risk factor and only 20% had ≥3 risk factors (5). The net reclassification indices of the Framingham Risk Score by CAC from 3 major prospective population based trials (MESA , Heinz-Nixdorf , and Rotterdam ), ranged from 10% to 15% in the low-risk, 50% to 65% in the intermediate-risk, and 35% in the high-risk groups, with 19% to 25% for the entire population. In the intermediate-risk group, risk prediction by risk factors seems worse than a flip of the coin, questioning the robustness of population-based observation when applied to an individual on a one-on-one basis. In the Whelton et al. (2) study, one-third of the entire 0 CAC population were in the ASCVD ≥7.5% high-risk group and one-third of this group failed to develop CAC, contrary to their risk factor based prediction of high risk for events. If we believe, as much data show, that a CAC of 0 means very minimal risk for events, it is clear that risk factor based prediction has much to improve upon.
The authors correctly state that the prediction of individuals who will or will not continue to have healthy aging is elusive given the currently available risk factors and further work using genetic markers would be needed to perfect predictive values for individual application. Genetic factors may not only allow for the identification of the unhealthy phenotype, but would also explain the likelihood of the response to medical therapy and aggressive risk factor control (9). It is expected that the individual genetic fabric would reconcile the translation of population-based risk factor profile to disease development or healthy aging.
The goal of primary prevention is to produce a match made, not in heaven after an unpredicted fatal event, but on earth while the patient is still living. A multitude of data consistently shows that imaging the plaque itself (e.g., CAC) is the most powerful biomarker for both sickness and health. It may be time to acknowledge that the marriage of CAC to individual risk prediction, faithful in both vascular health and future sickness, is more likely to provide an accurate assessment of the future than risk factors.
↵∗ Editorials published in JACC: Cardiovascular Imaging reflect the views of the authors and do not necessarily represent the views of JACC: Cardiovascular Imaging or the American College of Cardiology.
Dr. Narula has received support from GE Healthcare and Philips Healthcare in the form of equipment to the institution. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- 2015 American College of Cardiology Foundation
- ↵Einstein A. Available at: https://www.quora.com/Did-Einstein-really-define-insanity-as-doing-the-same-thing-over-and-over-again-and-expecting-different-results. Accessed November 4, 2015.
- Whelton S.P.,
- Silverman M.G.,
- McEvoy J.W.,
- et al.
- Min J.K.,
- Lin F.Y.,
- Gidseg D.S.,
- et al.
- Erbel R.,
- Möhlenkamp S.,
- Moebus S.,
- et al.
- Elias-Smale S.E.,
- Proenc R.V.,
- Koller M.T.,
- et al.