Multivariate assessment of lipid parameters as predictors of coronary heart disease among postmenopausal women: potential implications for clinical guidelines.

BACKGROUND

Over the past decade, lipid measurements have been significantly improved and standardized. We evaluated the usefulness of multiple plasma lipid parameters in predicting coronary heart diseases (CHD) among women.

METHODS AND RESULTS

Among 32,826 women from the Nurses' Health Study who provided blood samples at baseline, 234 CHD events were documented during 8 years of follow-up. In a nested study, these cases were matched to controls (1:2) for age, smoking, fasting status, and month of blood draw. We estimated the relative risk (RR) for each lipid parameter, adjusted for C-reactive protein, homocysteine, body mass index, family history, hypertension, diabetes, postmenopausal hormone use, physical activity, alcohol intake, and blood draw parameters. The RRs associated with an increase of approximately 1 SD (mg/dL) were as follows: HDL cholesterol (HDL-C) (RR=0.6 [0.5 to 0.8], SD=17), apolipoprotein B100 (apoB100) (RR=1.7 [1.4 to 2.1], SD=32), LDL cholesterol (LDL-C) (RR=1.4 [1.1 to 1.7], SD=36), total cholesterol (TC) (RR=1.4 [1.1 to 1.6], SD=40), and triglycerides (RR=1.3 [1.0 to 1.5], SD=80). Among the lipid indexes, the RRs were: apoB100/HDL-C (RR=1.7 [1.4 to 2.1], SD=1.0), TC/HDL-C (RR=1.6 [1.3 to 1.9], SD=1.3), LDL-C/HDL-C (RR=1.5 [1.3 to 1.9], SD=1.0), and non-HDL-C (RR=1.6 [1.3 to 1.9], SD=42 mg/dL). After simultaneous control for several lipid biomarkers, HDL-C was the primary contributor of the variation in multivariate models (P=0.01), followed by LDL-C (P=0.01), whereas triglycerides and apoB100 did not contribute further information. HDL-C-related ratios were the strongest contributors to predicting CHD (P<0.0001).

CONCLUSIONS

Lower levels of HDL-C may be a key discriminator of higher CHD events among postmenopausal women. HDL-C-related ratios (such as TC/HDL-C) provide a powerful predictive tool independently of other known CHD risk factors.