The Impact of Peer Substance Use and Polygenic Risk on Trajectories of Heavy Episodic Drinking Across Adolescence and Emerging Adulthood

James J. Li*, Seung Bin Cho, Jessica E. Salvatore, Howard J. Edenberg, Arpana Agrawal, David B. Chorlian, Bernice Porjesz, Victor Hesselbrock, Danielle M. Dick, COGA Investigators

*Corresponding author for this work

Research output: Contribution to journalArticleAcademicpeer-review


Background: Heavy episodic drinking is developmentally normative among adolescents and young adults, but is linked to adverse consequences in later life, such as drug and alcohol dependence. Genetic and peer influences are robust predictors of heavy episodic drinking in youth, but little is known about the interplay between polygenic risk and peer influences as they impact developmental patterns of heavy episodic drinking. Methods: Data were from a multisite prospective study of alcohol use among adolescents and young adults with genome-wide association data (n = 412). Generalized linear mixed models were used to characterize the initial status and slopes of heavy episodic drinking between age 15 and 28. Polygenic risk scores (PRS) were derived from a separate genome-wide association study for alcohol dependence and examined for their interaction with substance use among the adolescents’ closest friends in predicting the initial status and slopes of heavy episodic drinking. Results: Close friend substance use was a robust predictor of adolescent heavy episodic drinking, even after controlling for parental knowledge and peer substance use in the school. PRS were predictive of the initial status and early patterns of heavy episodic drinking in males, but not in females. No interaction was detected between PRS and close friend substance use for heavy episodic drinking trajectories in either males or females. Conclusions: Although substance use among close friends and genetic influences play an important role in predicting heavy episodic drinking trajectories, particularly during the late adolescent to early adult years, we found no evidence of interaction between these influences after controlling for other social processes, such as parental knowledge and broader substance use among other peers outside of close friends. The use of longitudinal models and accounting for multiple social influences may be crucial for future studies focused on uncovering gene–environment interplay. Clinical implications are also discussed.

Original languageEnglish
Pages (from-to)65-75
Number of pages11
JournalAlcoholism: Clinical and Experimental Research
Issue number1
Publication statusPublished - 1 Jan 2017
Externally publishedYes

Cite this