Genetic and environmental influences on illicit drug use and tobacco use across birth cohorts KENNETH S. KENDLER 1, 2 , CHARLES GARDNER 1 ,KRISTEN C. JACOBSON 1 , MICHAEL C. NEALE 1, 2, 3 AND CAROL A. PRESCOTT 1, 3 1 Virginia Institute for Psychiatry and Behavioral Genetics and Departments of Psychiatry ; 2 Human Genetics ; and 3 Psychology, Medical College of Virginia Commonwealth University, Richmond, VA, USA ABSTRACT Background. The prevalence of use of many psychoactive substances has changed considerably in recent years. While genetic factors impact on overall risk for substance use, we know little about whether the etiological importance of these factors differs across birth cohorts. One theory, which postulates that heritability of deviant traits increases in permissive environments, predicts a positive relationship across cohorts between prevalence and heritability of substance use. Method. The lifetime history of use of tobacco, cannabis, cocaine, sedatives and stimulants were assessed in 4826 twins from male–male and female–female pairs born in Virginia from 1934 to 1974. Using empirical methods based on prevalence by birth year, these twins were divided into three cohorts for each substance (e.g. for cannabis 1934–1953, 1954–1968 and 1969–1974). Structural equation modeling was performed using the Mx software package. Results. Prevalence rates for psychoactive substance use differed substantially across cohorts, most markedly for cocaine, sedatives and stimulants, which were highest in the 1958–1963 cohort. However, for all substances, the best-fit model constrained estimates of the etiological role of genetic and environmental risk factors to be equal across both sex and cohort. Conclusions. We found no evidence in this sample for any systematic relationship between herita- bility and prevalence of psychoactive substance use – which should be a rough index of drug availability and/or acceptability. This sample had reasonable power to detect large changes in heritability across cohorts and at least moderate power to detect relatively small changes. INTRODUCTION Genetic studies in man commonly assume that the impact of genetic factors is the same across diverse environments. However, a large body of evidence suggests that many aspects of the environment can modify genetic effects in many organisms including humans (Boerwinkle & Hallman, 1993). Typically, such studies of genotype–environment interaction examine the impact of genetic factors on the response to discrete environmental risk factors such as industrial toxins (Perera, 1997), head trauma (Jordan et al. 1997), tobacco smoke (Khaw & Barrett-Connor, 1986), diet (Tikkanen et al. 1990), or stressful life events (Kendler et al. 1995). Examining changes in heritability across birth cohorts is an alternative approach to the study of gene–environment interaction. Such analyses plausibly assume that short-term changes in the magnitude of genetic effects in human populations reflect changing environ- mental exposures. Few risk factors have changed as dramatically in frequency in Western popu- lations in the 20th century as the use of certain * Address for correspondence : Kenneth S. Kendler, M.D., Department of Psychiatry, PO Box 980126, Richmond, VA 23298- 0126. (Email : kendler@hsc.vcu.edu) Psychological Medicine, 2005, 35, 1–8. f 2005 Cambridge University Press doi:10.1017/S0033291705004964 Printed in the United Kingdom 1