Do gene-environment interactions influence fasting plasma lipids? A study of twins

The aims of this study were to determine the influence of smoking, alcohol consumption, physical activity and hormone replacement therapy (HRT) on lipids, independently of genetic factors, and to detect whether gene-environment interactions influence these associations.

Fasting plasma total cholesterol, LDL cholesterol, HDL cholesterol, triglycerides, apolipoproteins AI and B and lipoprotein(a) were measured in 685 female twins (96 monozygotic, 230 dizygotic pairs and 33 singletons).

Smokers had higher triglyceride and lower HDL cholesterol levels than never-smokers (P < 0.001). After controlling for genetic influences, smoking accounted for 0.35 mmol L(-1) and 0.22 mmol L(-1) differences in triglyceride and HDL cholesterol levels, respectively (P < 0.005), remaining significant after excluding alcohol-discordant twin pairs. In a gene-environment interaction analysis, the association between smoking and triglycerides was exaggerated in subjects at high genetic risk of hypertriglyceridaemia (interaction P=0.04). All levels of alcohol consumption were associated with higher HDL cholesterol levels than abstinence, but only moderate alcohol consumers had lower LDL cholesterol and triglyceride levels. In monozygotic twins concordant for smoking, an alcohol intake > 10 units week(-1) accounted for a 0.32 mmol L(-1) difference in LDL cholesterol, independently of genetic effects (P=0.04). In postmenopausal women, those using HRT had 0.54 mmol L(-1) lower LDL cholesterol and 0.21 micromol L(-1) lower lipoprotein(a) levels than nonusers (P < 0.001 and P=0.04, respectively); these differences were attenuated after accounting for genetic effects in monozygotic twins. Although physically active subjects had higher levels of HDL cholesterol than nonactive subjects, this was nonsignificant after adjusting for genetic factors.

Smoking-induced aberrations in HDL cholesterol and triglycerides and alcohol-related differences in LDL cholesterol were independent of genetic influences. The association between smoking and hypertriglyceridaemia was accentuated in high genetic risk individuals.