Background and Aims Studies have linked adolescent alcohol use with adverse consequences in adulthood, yet it is unclear how strong the associations are and to what extent they may be due to confounding. Our aim was to estimate the strength of association between different patterns of adolescent drinking and longer‐term psychosocial harms taking into account individual, family and peer factors. Design Participant‐level data were integrated from four long‐running longitudinal studies: Australian Temperament Project, Christchurch Health and Development Study, Mater Hospital and University of Queensland Study of Pregnancy and Victorian Adolescent Health Cohort Study. Setting Australia and New Zealand. Participants Participants were assessed on multiple occasions between ages 13 and 30 years (from 1991 to 2012). Number of participants varied (up to n = 9453) by analysis. Measurements Three patterns of alcohol use (frequent, heavy episodic and problem drinking) were assessed prior to age 17. Thirty outcomes were assessed to age 30 spanning substance use and related problems, antisocial behaviour, sexual risk‐taking, accidents, socio‐economic functioning, mental health and partner relationships. Findings After covariate adjustment, weekly drinking prior to age 17 was associated with a two‐ to threefold increase in the odds of binge drinking [odds ratio (OR) = 2.14; 95% confidence interval (CI) = 1.57–2.90], drink driving (OR = 2.78; 95% CI = 1.84–4.19), alcohol‐related problems (OR = 3.04; 95% CI = 1.90–4.84) and alcohol dependence (OR = 3.30; 95% CI = 1.69–6.47) in adulthood. Frequency of drinking accounted for a greater proportion of the rate of most adverse outcomes than the other measures of alcohol use. Associations between frequent, heavy episodic and problem drinking in adolescence and most non‐alcohol outcomes were largely explained by shared risk factors for adolescent alcohol use and poor psychosocial functioning. Conclusions Frequency of adolescent drinking predicts substance use problems in adulthood as much as, and possibly more than, heavy episodic and problem drinking independent of individual, family and peer predictors of those outcomes.