WITHDRAWN: Do Classes of Polysubstance Use in Adolescence Differentiate Growth in Substances Used in the Transition to Young Adulthood?

Introduction: Previous research has found that early onset and chronic substance use in adolescence predicts substance use problems during young adulthood. Past studies have differentiated classes of polysubstance use in adolescence; however, the associations of adolescent polysubstance use classes with longitudinal substance use trajectories from adolescence to young adulthood have not been studied. The current study examined substance use classes during adolescence and longitudinal trajectories of each substance used across the transition to young adulthood. Method: Data were collected biennially from 662 youth and followed ten years across six measurement assessments. Using baseline data (T1), Latent Class Analysis (LCA) was used to identify classes of polysubstance use (cigarette, alcohol, marijuana, and illicit drug use) during adolescence. Using an accelerated longitudinal design that treated age as the time variable, we then used T2 through T6 data to fit Latent Growth Models (LGM) for cigarette, alcohol, marijuana, and illicit drug use to examine longitudinal trajectories of each substance used by class. Using a model based approach accounted for uncertainty in modal class assignment from the LCA when fitting the LGMs. Results: A three-class model fit the data best and included a poly-use class, that had high probabilities of use of all substances, a co-use class, that had high probabilities of use of alcohol and marijuana, and a low-use class that had low probabilities of use of all substances. We then examined trajectories of each substance used by class. Strong continuity of substance use was found by class across 14 years. For example, individuals in the poly-use class continued to use all substances from 14 to 28 years; the co-use class had stable trajectories of alcohol and marijuana use, low levels of cigarette use, and decreasing levels of illicit drug use; and the low-use class were low in the use of all substances over time except for alcohol use, which peaked at 22 years of age and then decreased. Additionally, for some substances, higher average levels of use at age 14 were associated with changes in growth of other substances used over time. Conclusions: The simultaneous use of substances is common in adolescence, and young adult outcomes that are attributed to single drug types may be better explained by differences in polysubstance use. As such, efforts that only target a single drug type may be missing an important opportunity to reduce the use and subsequent consequences related to the use of multiple substances.