Genetic and environmental contributions to marijuana use development: Twin analyses across three independent data sets

• In Event: 1-125 - Genetically-Informed Studies of Adolescent Substance Use and Psychopathology

Integrative Statement

The prevalence of marijuana (MJ) use among those 18 years of age or younger has increased by roughly 50% in the last 15 years (Kerr et al., 2018). Likewise, several US states, as well as Canada have legalized recreational MJ use for adults. Early initiation and regular use of MJ during the critical window of adolescence has been associated with several adverse outcomes in emerging adulthood including increased risk for psychopathology, poor educational attainment, and diminished cognitive functioning (Volkow et al., 2014). However, some genetically-informed studies of adolescent MJ use, using twins, do not provide support that such adverse outcomes are causally related to use (e.g., Jackson et al., 2016). Rather, there may be common genetic and shared environmental liabilities. While many studies have examined developmental patterns of MJ use, there lacks systematic, genetically-informed research on the contributions to, patterns of, and outcomes related to adolescent MJ use. Our study compared growth patterns of MJ use from early adolescence to early adulthood across three large longitudinal studies of twins and non-twins.

Data were compiled across three data sets from the Minnesota Twin Family Study (MTFS), the Colorado Twin Study (CTS), and the National Longitudinal Study of Adolescent to Adult Health (Add Health) for a total samples size of N = 26447. As our interest is in early adolescence to early adulthood MJ use, we included data from those aged ~14 to ~24 years. Each study administered MJ use frequency items. Using these items, we harmonized data across studies forming one variable of ‘the number of days of MJ use in the last month’. Linear latent growth models (LGM) were run independently in each data set. Given our interest in age-related development, and with differing wave structures across studies, we chose to use age as our time metric. The mean age of first use centered around mid-adolescence (M = 16.5, SD = 2.8). Therefore, the LGM intercept was centered at 16.5 years. Intercept mean and variance estimates can be interpreted in terms of the number of days of past month MJ use at age 16.5. Slope mean and variance estimates can be interpreted in terms of the yearly growth in past month MJ use (total age span of ~10 years). Across studies, the mean number of days of past month MJ use at age 16.5 ranged from 0.80 days to 1.5 days. The mean growth in past month MJ use (across ages 14 to 24), ranged from 2.7 days to 3.4 days. Multiple group, likelihood ratio difference tests indicated no significant differences, across studies, in terms of slope means, slope/intercept variances/covariances, and sex effects. Generalized additive mixed models (GAMM) replicated the LGM results. Biometric LGM analyses were also conducted. Heritability estimates for the intercept ranged from 30% to 80%; for the slope they ranged from 40% to 90%. Non-shared environmental effects were consistently low (< under 30%), suggesting that casual associations with adult outcomes, using discordant twin designs, are not likely to be found. Associations with outcomes and implications will be discussed.

Authors

• Eric Lee Thibodeau, Department of Psychology, University of Minnesota, Minneapolis, MN
  Presenting Author

• Stephanie Zellers, University of Minnesota
  Non-Presenting Author

• Frank D Mann, University of Minnesota
  Non-Presenting Author

• Irene Elkins, University of Minnesota
  Non-Presenting Author

• Robin P Corley, University of Colorado
  Non-Presenting Author

• John Hewitt, University of Colorado
  Non-Presenting Author

• Michael Stallings, University of Colorado
  Non-Presenting Author

• Robert Krueger, University of Minnesota
  Non-Presenting Author

• Robert Iacono, University of Minnesota
  Non-Presenting Author

• Matt McGue, University of Minnesota
  Non-Presenting Author

• Scott Vrieze, University of Minnesota
  Non-Presenting Author