Objective: To determine the relative contribution of environmental and genetic influences on the joint distribution of heavy smoking, heavy alcohol use and heavy coffee drinking.
Method: Multivariate structural equation modeling in a large cohort of male twins (N = 2,220 monozygotic and 2,373 dizygotic twin pairs; mean age = 62.1 years) from the National Academy of Sciences-National Research Council's World War II Twin Registry.
Results: The best-fitting model identified two independent (i.e., uncorrelated) sets of genetic and environmental latent factors, with one set underlying joint heavy smoking and heavy alcohol use and the other set underlying joint heavy smoking and heavy coffee drinking (chi 2 = 14,13,22 df, p > .80). Heavy alcohol use and heavy coffee drinking were uncorrelated in this sample. While common genetic factors accounted for 35% to 78% of the total genetic variance in heavy substance use, a substantial amount of genetic variance remained specific to each of the three substances.
Conclusions: Several hypotheses involving genetic and environmental factors are presented to account for the independent clustering of heavy smoking and heavy alcohol use and of heavy smoking and heavy coffee drinking.