High-sensitivity C-reactive protein: a useful marker for cardiovascular disease risk prediction and the metabolic syndrome.

Every year, 1.3 million Americans suffer a myocardial infarction with only one half of them exhibiting evidence of dyslipidemia. The identification of additional risk factors for cardiovascular disease is therefore of paramount importance. Of the examined novel biochemical markers, high-sensitivity C-reactive protein (hs-CRP) is the most promising. To date, 22 prospective epidemiologic studies have demonstrated that hs-CRP is a strong predictor of future vascular disease, 6 cohort studies have shown that hs-CRP measurements add prognostic value beyond that available from the Framingham risk score, and 8 cohort studies have confirmed that hs-CRP adds prognostic information in the metabolic syndrome and in the prediction of type 2 diabetes (1 ). These studies, done by various groups in Europe and the United States, examined middle-aged and elderly men and women, and some included different ethnic and racial groups. As a result of these findings and the fact that hs-CRP can be reliably and accurately measured by clinical laboratories, the American Heart Association (AHA) and the CDC issued joint guidelines about the implementation of hs-CRP measurement as part of the global risk assessment of cardiovascular disease (2 ). In this Journal, Levinson et al. (3 ) recently questioned the validity of using hs-CRP to predict future cardiovascular disease. These authors relied in their argument on data from the Women’s Health Study (WHS) (4 ) and the recently published article from the Reykjavik Study by Danesh et al. (5 ). We would like to point out some of the important issues regarding these two studies. Using the WHS cohort, we directly compared the ability of hs-CRP and LDL-cholesterol to predict future coronary events in 27 939 participants over a follow-up period of 8 years (4 ). Our data clearly showed that hs-CRP was superior to LDL-cholesterol in its ability to predict risk (fully adjusted relative risks, 2.3 for hs-CRP and 1.5 for LDL-cholesterol; both P 0.001). In addition, we demonstrated that hs-CRP provided additional prognostic information at all LDL-cholesterol concentrations and Framingham risk scores. After adjustment for all components of the Framingham risk score, hs-CRP remained a strong predictor of future risk. It is surprising that Levinson et al. (3 ), using these data, questioned the ability of hs-CRP, but not LDL-cholesterol, to predict future cardiovascular disease risk. We also examined 3597 of those women who have metabolic syndrome, as defined by the National Cholesterol Education Program, and have been followed for 8 years for first-ever cardiovascular events (6 ). The probability of cardiovascular event-free survival in these women was markedly affected by their hs-CRP values (Fig. 1). Baseline hs-CRP concentration was able to differentiate between low, moderate, and high risk of future cardiovascular events among those with metabolic syndrome, indicating the clinical benefit from its measurement in this patient population. Clearly, more aggressive interventional measures are warranted for those patients with increased hs-CRP. In the article by Danesh et al. (5 ), hs-CRP was measured in baseline samples from 2459 patients who developed a nonfatal myocardial infarction or died from coronary heart disease during a follow-up period of 30 years and in 3969 controls without coronary events. The authors concluded that hs-CRP is a moderate predictor of coronary heart disease risk and that the recommendation for its use should be revisited. However, these conclusions are not consistent with the presented data. The authors used the cutoff point of 2 mg/L rather than the 3 mg/L recommended by the AHA/CDC guidelines (2 ) and therefore may have underestimated the risk associated with hs-CRP. In addition, they adjusted not only for the Framingham risk score but also for diabetes, triglycerides, obesity, pulmonary function, social status, and other inflammatory markers that tend to reduce the predictive ability of hs-CRP. Even in this conservative analysis and estimation of increment in risk, an odds ratio for hs-CRP of 1.5 (a 50% increase in risk) is a significant finding from a clinical and a public health perspective (7 ). We reanalyzed our data from the WHS, using the cut points used by Danesh et al. (5 ). The odds ratios in both studies were almost identical for hs-CRP at a cut point of 2 mg/L, highest vs lowest quintile, and the 10-year risk estimate (Table 1). These odds ratios are also similar to those reported in the Monitoring Trends and Determinants in Cardiovascular Disease (MONICA) and the Atherosclerosis Risk in Communities (ARIC) studies (8 ). Furthermore, cholesterol concentrations in Iceland in 1967 were much higher than the current ones in the United 0 2 4 6 8