Carol A. Prescott1 & Kenneth S. Kendler1
Structured interviews were conducted with 1,937 adult female twins (including 569 MZ and 394 DZ pairs) participating in a longitudinal study of psychopathology. Assessment included lifetime history of alcohol, tobacco and marijuana use; density (quantity x frequency) at maximum use; and problematic use as indicated by Fagerstrom tolerance scores (tobacco) and DSM-IV diagnoses of abuse and dependence and symptom counts (alcohol and marijuana). Lifetime use of alcohol was reported by 93% of subjects, tobacco by 67%, and marijuana by 48%. Diagnoses of abuse or dependence were obtained by 13.6% of the sample for alcohol and by 7.8% for marijuana. Univariate twin analyses of lifetime use, density of use, number of dependence symptoms, and abuse/dependence diagnoses yielded similar results for all three substances. Results for lifetime use vs abstinence were consistent with a combination of genetic and shared environmental factors. Analyses of density and problematic use indicated additive genetic factors were the major source of pair resemblance. The covariation among substances for density and problematic use was moderate, (r's=.3-.5). The sources of this covariation were investigated using two multivariate models: (1) a common factors model which structured the covariation as arising from common additive genetic, shared environmental and specific environmental sources but did not further structure the individual loadings; and (2) a more restrictive latent liability model which structured all covariation through a single latent factor, requiring that the variation due to each source be proportional across substances. For substance use density, number of symptoms, and diagnosis, the latent liability model could not be rejected. For all three phenotypes, the latent factor accounted for a moderate (20-40%) proportion of the variance for each substance, and was estimated as due primarily to additive genetic sources, with common environmental effects only for density. Unique variance was attributed to genetic and specific environmental sources.
Address: VIPBG, VCU, PO Box 980126, Richmond VA 23298-0126 USA; (804) 828-5968;, (804) 828- 1471 (fax); cprescot@hsc.vcu.edu; http://electro.psi.v cu.edu:80/~vipbg/
1Virginia Institute for Psychiatric & Behavioral Genetics, Department of Psychiatry, Medical College of Virginia of Virginia Commonwealth University, Richmond VA, USA. Support for data collection and analysis was provided by NIH grants AA/DA-09095, MH-40828. The Virginia Twin Registry is maintained by L. Corey, and is supported by NIH grants HD-26746 and NS-31564. We thank B. Brooke, L. Halberstadt, and D. Foley for their roles in data collection and L. Karkowski and C. Gardner for assistance with data analysis.