Abstract
Studying genetic contributions to substance initiation is crucial for identifying at-risk individuals and developing targeted prevention strategies. Investigating these factors during adolescence is vital, as this period is critical for brain development and represents an age of experimentation and initiation of substance use. Here we generate polygenic scores (PGSs), using data from the PGS catalog, across a range of substance use related traits to assess PGS in predicting i) measures of impulsivity taken from the UPPS-P questionnaire and ii) self-reported use of nicotine/tobacco, cannabis, alcohol and caffeine in early-mid adolescence. Repeat cross-sectional analyses across age bands (ages 9-10, 11-13, and 13-15) were conducted using the longitudinal Adolescent Brain Cognitive Development (ABCD) Study(®) (total N=8,753; 55% female). Due to the large contribution of European-like (EUR-like) individuals in discovery samples, we performed ancestry stratified analysis in EUR-like (n=5,225), African (AFR-like; n=637) and admixed (MIX-like; n=2,891) groups reflecting genetic similarity to continental ancestry groups. In the EUR-like group, PGS related to nicotine/tobacco were associated with greater impulsivity across all subscales of the UPPS-P at all ages. Analyses across ages 9-15 years old revealed PGS-impulsivity associations that: a) grew as the sample aged (e.g. Smoking Status PGS with Lack of Perseverance: 9-10 years-old: β=0.065, 11-13 years-old: β=0.11, 13-15 years-old: β=0.12) and b) others that diminished as the sample aged (e.g. Alcohol Consumption PGS with Sensation Seeking: 9-10 years-old: β=0.070, 11-13 years-old: β=0.062, 13-15 years-old: β=0.03). Evaluating the performance of PGS against self-reported substance use, PGS of nicotine/tobacco traits were associated with regular consumption of caffeine across ages. At ages 13-15, PGS of smoking traits were associated with cannabis and tobacco exposure (e.g., Smoking Initiation PGS and self-reported cannabis use, ΔR(2)=0.0094), in addition to weekly caffeine consumption. Across ages, nicotine/tobacco and alcohol PGS and regular energy drink consumption associations grew over time (e.g., Smoking Status PGS: 9-10 years-old: β=0.088, 11-13 years-old: β=0.24, 13-15 years-old: β=0.29). As with impulsivity, some PGS associations decreased over time (Alcohol Consumption PGS and self-reported alcohol use: 9-10 years-old: β=0.12, 11-13 years-old: β=0.11, 13-15 years-old: β=0.083). Replication of our EUR-like results in AFR-like and MIX-like sub-samples revealed a significant attenuation of effects, underscoring the importance of collecting genetic studies in larger ancestrally diverse cohorts. Our results highlight the dynamic relationship between genetic risk factors of substance use, trait impulsivity, and self-reported substance initiation throughout adolescence. Further, evidence here indicates caffeine consumption represents an early risk factor for problematic substance use in later life. Results support PGSs, in conjunction with larger phenotypic profiles, for identification of prevention efforts.