Genetic and Environmental Vulnerabilities UnderlyingAdolescent Substance Use and Problem Use: Generalor Specific?Susan E. Young,1,3Soo Hyun Rhee,1,2Michael C. Stallings,1,2Robin P. Corley,1and John K. Hewitt1,2Received 9 Dec. 2005—Final 27 Feb. 2006Are genetic and environmental risks for adolescent substance use specific to individualsubstances or general across substance classes? We examined this question in 645 monozygotictwin pairs, 702 dizygotic twin pairs, 429 biological sibling pairs, and 96 adoptive (biologicallyunrelated) sibling pairs ascertained from community-based samples, and ranging in age from12 to 18 years. Substance use patterns and symptoms were assessed using structuredpsychiatric interviews. Biometrical model fitting was carried out using age- and sex-specificthresholds for (a) repeated use and (b) problem use, defined as one or more DSM-IVsymptoms of abuse or dependence. We hypothesized that problem use would be moreheritable than use in adolescence, and that both genetic and environmental risks underlyingtobacco, alcohol, and marijuana use and problem use would be significantly correlated.Results of univariate analyses suggested significant heritable factors for use and problem usefor all substances with the exception of alcohol use. Shared environmental factors wereimportant in all cases and special twin environmental factors were significant for tobacco use,tobacco problem use, and alcohol use. Multivariate analyses yielded significant geneticcorrelations between each of the substances (for both levels studied), and significant sharedenvironmental correlations among use variables only. Our results suggest that tobacco,alcohol, and marijuana problem use are mediated by common genetic influences, but sharedenvironmental influences may be more substance-specific for problem use.KEY WORDS: Adolescence; alcohol; heritability; marijuana; siblings; substance use; tobacco; twins.INTRODUCTIONLarge-scale epidemiological studies have shown thatby late adolescence about two-thirds of adolescentsreport having used tobacco, over 80% report expe-rience with alcohol, and approximately half-reportexperimentation with marijuana (CDC, 2000; John-ston et al., 2001; SAMSHA, 2001). Moreover,adolescents commonly report experimentation wi thmultiple substances rather than exhibiting substance-specific preferences. Although the developmentof substance use disorders (SUD) in adolescence isless prevalent, it is not rare (Kandel et al., 1997;Young et al., 2002), and early onset of use is awell-established risk factor for the progression fromuse to abuse and dependence (Bucholz et al., 2000;1Institute for Behavioral Genetics, University of Colorado, 447UCB, Boulder, CO, 80303, USA.2Department of Psychology, University of Colorado, 345 UCB,Boulder, CO, 80303, USA.3To whom correspondence should be addressed at Institute forBehavioral Genetics, University of Colorado, 447 UCB, Boulder,CO, 80303, USA. Tel.: +1-303-492 1235; Fax: +1-303-492 8063;e-mail: [email protected]Ó 2006 Springer Science+Business Media, Inc.Behavior Genetics (Ó 2006)DOI: 10.1007/s10519-006-9066-7Chen and Anthony, 2003; Grant and Dawson, 1998).It has been argued that this trajectory may be aby-product of an underlying pattern of disinhibitorypsychopathology rather than a direct risk (McGueet al., 2001).Although the evidence for a genetic risk under-lying substance use and the development of substanceuse problems has been converging among studies ofadult twins (Kaprio et al., 1987; Kendler et al., 1999;McGue et al., 1992; True et al., 1999; Tsuang et al.,1996), the extent to which adolescent substance useproblems are due to genetic factors is less wellunderstood, possibly due to the developmental natureof substance use disorders. A review of adolescenttwin and adoption studies conducted through 2002(Hopfer et al., 2003) summarizes research conductedon samples in the U.S. (Maes et al., 1999; McGueet al., 2000; Rhee et al., 2003c) and abroad (Koop-mans et al., 1997), which consistently show thatadolescent tobacco use is under strong genetic con-trol. In contrast, genetic influences on alcohol andmarijuana use (defined as experimentation in ado-lescence) appear to be much less important, based ona number of large-scale adolescent twin studies(Koopmans et al., 1999; Maes et al., 1999; McGueet al., 1996; 2000; Miles et al., 2002; Rose et al., 2001;Viken et al., 1999). Clinically significant substanceuse problems, such as symptoms of abuse or depen-dence, have been rarely studied in adolescents. In arecent study on a subset of the current Col oradoadolescent sample (Rhee et al., 2003c), heritabilityfor tobacco problem use (defined as one or moreDSM-IV dependence symptom) was significant, butvaried considerably by sex (h2=0.26 for males andh2=0.95 for females). McGue et al. (2000) reportedsignificant genetic influences on tobacco dependence,but not on marijuana abuse/dependence diagnoses.The search for susceptibility genes that mayunderlie adolescent substance use and substance usedisorders can be guided by knowledge about whethergeneral or substance-specific risks are operating.Although sparse, findings in the extant adolescentliterature on the etiology of comorbidity of use acrossmultiple substances suggest that genetic and envi-ronmental influences are common across substanceclasses (Han et al., 1999; Koopmans et al., 1999; Rheeet al., 2003a). However, most family studies of thecauses of comorbidity among substance use disordershave involved adults and have provided inconsistentconclusions regarding general versus substance-specific familial risk (Bierut et al., 1998; Meller et al.,1998; Merikangas et al., 1998; Tsuang et al., 1998).As noted in previous research (Rhee et al., 2003b), theconflicting results among the adult studies may bedue, in part, to differences in the analytical methodsused to examine substance-involved probands andtheir relatives, as well as differences in sample selec-tion. In an adolescent sibling study involving matchedclinical and control samples, Rhee et al. (2003a), useda biometrical model fitting approach to compare 13alternative hypotheses of the causes of comorbidity.Results suggest that the comorbidity of alcohol andillicit drug dependence was most likely due to a singleunderlying familial liability.The primary aim of the present study was toexamine the genetic and environmental covariationamong tobacco, alcohol, and marijuana use andproblem use in