Comparison of C-Reactive Protein and Low-Density Lipoprotein Cholesterol Levels in the Prediction of First Cardiovascular Events - Comparison of C-Reactive Protein and Low-Density Lipoprotein Cholesterol Levels in the Prediction of First Cardiovascular Events
The goal of this study was to compare the relative value of high sensitivity CRP and LDL cholesterol (LDL-C) in their association with cardiovascular events in "apparently" healthy women, and to determine whether the CRP provides added prognostic information after adjustment for all the components of the Framingham risk score.
Assays of LDL-C and CRP were obtained on 27,929 women participating in the Women's Health Study (WHS), a primary prevention trial designed to evaluate the utility of ASA and vitamin E in which women over 45 years of age were enrolled between 1992 and 1995. Participants were divided into quintiles of CRP and LDL-C. The discriminative value of predictive models was calculated on the basis of a minimal follow-up of 6 years.
Mean age was 54.7 years, 25% of patients had hypertension, 12% were smokers, 2.5% had diabetes, mean BMI was 25.9 kg/m2, and 44% were users of HRT. The approximate median CRP was 1.5 mg/L and LDL-C 125 mg/dL. In non-users of HRT the first quintile of CRP was less than 0.05 mg/L, the third quintile greater than 1.0 and less than 2.0 mg/dL, and the fifth quintile greater than 4 mg/dL. The first quintile of LDL-C was about less than 100 mg/dL, the third quintile 115-130 mg/dL, and fifth quintile greater than 150 mg/dL. The age adjusted relative risk for the first CV event increased linearly from 1.5 in the second quintile to 3.6 in the fifth quintile and major risk factor adjusted relative risk range from 1.4 in the second to 2.3 in the fifth quintile. The risk factor adjusted relative risk range for LDL-C was not significant until the fourth quintile and reached 1.5 in the fifth quintile. The area under the ROC curve for predicting CV events (sensitivity vs 1-specificity) for the risk factor adjusted RR was 0.81 for each biomarker. The distribution of CRP and LDL-C levels were only minimally correlated (r=0.08) suggesting each was detecting a different high-risk group. Using the median values as cutoffs, compared to low CRP + low LDL-C patients with low LDL-C + high CRP or a high LDL-C + low CRP had a 50% greater risk of events. The risk attributable to a high CRP + high LDL-C was higher in non-users of HRT. Increasing levels of CRP from less than 1 mg/L to greater than 3 mg/L increased the relative risk for CV events using the Framingham estimate of 10 year risk of events and cut points of LDL-C used in the ATP III guidelines.
Among patients participating in the Women's Health Study (WHS), C-reactive protein level is a stronger predictor of cardiovascular events than the LDL-C level and adds to the prognostic information derived from the Framingham risk score. The data relating CRP to subsequent risk is compelling, particularly in women. Clinicians should consider using the CRP for assessing CV risk in middle aged and older men and women, whose Global Risk Score is less thank 2% per year. As suggested by the authors, a statin trial in men and women with an LDL-C less than 130mg/dL and CRP above the median will be needed to help guide therapies.
Ridker PM, Rifai N, Rose L, et al. Comparison of C-Reactive Protein and Low-Density Lipoprotein Cholesterol Levels in the Prediction of First Cardiovascular Events. N Engl J Med. November 14, 2002;347:1557-65.
Keywords: C-Reactive Protein, Vitamin E, Body Mass Index, Cholesterol, LDL, Hydroxymethylglutaryl-CoA Reductase Inhibitors, Middle Aged, Women's Health, Risk Assessment, Hypertension, Diabetes Mellitus, Primary Prevention
< Back to Listings