Clinical, Environmental, and Genetic Risk Factors for Substance Use Disorders: Characterizing Combined Effects across Multiple Cohorts

Abstract

Importance: Substance use disorders (SUDs) incur serious social and personal costs. Screening techniques that identify persons at risk before problems develop can improve prevention efforts. Objective: To examine whether models that include polygenic scores (PGS) and a clinical/ environmental/risk index (CERI) are able to identify individuals as having a lifetime SUD. Design: We tested the predictive power of PGS and CERI for lifetime diagnosis of DSM–IV substance dependence using four longitudinal cohorts. Setting: The study included four samples: 1) the National Longitudinal Study of Adolescent to Adult Health (Add Health); 2) the Avon Longitudinal Study of Parents and Children (ALSPAC); 3) the Collaborative Study on the Genetics of Alcoholism (COGA); and 4) the Finnish Twin Cohort Study (FinnTwin12) for a combined sample of N = 15,134. Participants: Participants in Add Health (NEUR = 4,855; NAFR = 1,605) and COGA (NEUR = 1,878; NAFR = 870) included individuals of both European and African ancestries. Participants in ALSPAC (NEUR = 4,733) and FinnTwin12 (NEUR = 1,193) were limited to individuals of European ancestries. Exposures: A clinical/environmental risk index (CERI) composed of ten items and PGS for phenotypes with strong genetic overlap with SUDs (drinks per week, problematic alcohol use, externalizing problems, major depressive disorder, and schizophrenia). Main Outcomes: Meeting lifetime criteria for DSM–IV: 1) alcohol dependence, 2) drug dependence, and 3) any substance dependence (alcohol, other drug, or nicotine). Results: In the models containing the five PGS and CERI, the CERI was associated with all three outcomes (ORs = 1.35 – 1.64). PGS for problematic alcohol use was associated with alcohol dependence (OR = 1.14), PGS for externalizing was associated with drug dependence (OR = 1.14) and both were associated with any substance dependence (ORs = 1.11 – 1.19). Including the five PGS, CERI, and covariates explained 6% – 13% of the variance in SUDs. Those in the top 10% of CERI and PGS had relative risk ratios of 3.82 – 9.13 for each SUD relative to the bottom 90%. Conclusions and Relevance: Measures of clinical, environmental, and genetic, risk demonstrate modest ability to distinguish between affected and unaffected individuals for alcohol, drug, and any substance use disorders in young adulthood. These tools will continue to advance as we identify additional risk factors that can be incorporated into clinical practice and deliver on the goal of precision medicine.