There is uncertainty about the association between circulating concentrations of adiponectin and coronary heart disease (CHD) risk. We report new data from a prospective study in the context of a meta-analysis of previously published prospective studies. We measured baseline adiponectin levels in stored serum samples of 589 men with fatal CHD or nonfatal myocardial infarction and in 1231 controls nested within a prospective study of 5661 men (aged 40 to 59 years) recruited during 1978-1980, as well as in paired samples obtained 4 years apart from 221 of these participants. Baseline adiponectin concentrations correlated (P < 0.0001) positively with HDL cholesterol (r = 0.33) and inversely with C-reactive protein (r = -0.11) and BMI (r = -0.21), and the year-to-year consistency of adiponectin values was comparable to those of blood pressure and total cholesterol levels. No significant difference between median adiponectin levels at baseline was observed between cases and controls (10.2 versus 10.8 microg/mL; P = 0.5), despite the fact that body mass index, HDL, and C-reactive protein were all significant predictors of events in this cohort. The odds ratio for CHD was 0.89 (95% CI, 0.67 to 1.18) in a comparison of men in the top third of adiponectin concentrations compared with those in the bottom third, similar to a meta-analysis (including the present study) of 7 prospective studies involving a total of 1318 CHD cases (odds ratio, 0.84 [95% CI, 0.70 to 1.01]). In contrast to the strong associations previously reported between adiponectin levels and risk of type 2 diabetes, any association with CHD risk is comparatively moderate and requires further investigation.
