Abstract
OBJECTIVE: To assess whether body mass index (BMI) modifies the associations of lipids with coronary heart disease (CHD). METHODS: In the Asia Pacific Cohort Studies Collaboration, total cholesterol (TC), high density lipoprotein cholesterol (HDLC) and triglycerides (TG) were measured for 333,297, 71,777 and 84,015 participants, respectively. All participants had measured BMI, categorized into underweight, normal, high-normal, overweight and obese, using standard definitions. For each BMI subgroup the effects of lipids on CHD were estimated per 1 standard deviation (SD) increase using Cox proportional hazard models, stratified by study and sex, adjusted for age and smoking. They were compared across the BMI groups, testing for interactions. RESULTS: In the analyses for TC, HDLC and TG, there were 3121, 714 and 808 CHD events during a mean follow-up of 6.7 years. The risk of CHD increased monotonically with increasing TC and decreasing HDLC in all BMI subgroups without evidence of heterogeneity (p for interaction > 0.4). In contrast, the hazard ratio for CHD for a one SD increase in log-transformed TG increased from 1.07 (95%CI 0.72-1.59) in underweight, 1.26 (1.10-1.44) in normal weight, 1.27 (1.08-1.49) in high-normal weight, 1.37 (1.22-1.55) in overweight, to 1.61(1.30-1.99) in obesity (p = 0.01 for interaction trend). These associations were attenuated (p = 0.07 for interaction) but remained significant in the overweight and obese after further adjustment for TC and HDLC. CONCLUSIONS: Greater excess body weight exacerbated the effects of TG, but not TC or HDLC, on CHD, suggesting that additional effort is required to reduce TG in the overweight and obese.