Markers for Prediction of Cardiovascular Disease Risk

Abstract

To the Editor: The article by Dr Yeboah and colleagues compared the ability of several risk markers to improve prediction of coronary heart disease (CHD) and cardiovascular disease (CVD) among individuals at intermediate risk. The authors reported that coronary artery calcium (CAC) provided superior reclassification compared with other risk markers and recommended CAC as a tool for refining cardiovascular risk prediction in individuals at intermediate risk. While the added predictive ability of CAC in CHD risk prediction was substantial and confirmed previous findings, we believe that supporting CAC as a candidate for CVD screening based on the results is less grounded. While it is straightforward to define an intermediaterisk group for CHD, this is not the case for CVD because accepted thresholds are lacking. In the study by Yeboah et al, the authors included persons at intermediate risk for CHD and therefore the results might not necessarily apply to persons at intermediate risk for CVD. While CAC has been shown to accurately predict CHD in different populations, CAC has not been proven to be a useful predictor for stroke. In the study by Yeboah et al, addition of CAC to the Framingham risk score provided overall net reclassification improvement of 0.66 for CHD and 0.47 for CVD risk prediction. The percentage net correct reclassification in the group without events hardly changed after adding a non–CAC-related outcome such as stroke (40.4% for CHD and 36.0% for CVD). However, the percentage net correct reclassification for those with events, which reflects the ability to identify persons who will benefit from intensive treatment, dropped from 25.5% for CHD and 10.6% for CVD. The percentage net correct reclassification for persons with CHD events of 25.5% in the current study and 24.0% previously reported imply that adding CAC to risk prediction models moves a substantial proportion of persons initially at intermediate risk to the high-risk group, in which they qualify for more aggressive treatment. This supports the incorporation of CAC in CHD risk assessment. However, whether 10.6% net correct reclassification of persons with CVD events provided by CAC is sufficiently large to warrant recommending CAC as a screening tool for CVD is doubtful. The general trend in developing new guidelines on cardiovascular risk prevention is moving toward focusing on broader CVD risk rather than on CHD risk only. However, before considering new markers for CVD risk prediction, all components of this broad outcome should be considered and limitations for stroke risk prediction should be recognized.