RESUMO
PURPOSE: Previous research suggests several sociodemographic risk factors for the persistence of harmful alcohol use. However, the evidence is limited due to short follow-up times, retrospective reporting and samples comprising only people with alcohol dependence. We pooled data from six prospective cohort studies to systematically evaluate whether the sociodemographic risk factors differ between the incidence and persistence of harmful alcohol use. METHODS: Data were from six prospective cohort studies from the US, UK and Japan (n = 28,394). We conducted a two-stage meta-analysis to examine the associations of six sociodemographic risk factors (sex, age, presence of a partner, educational attainment, smoking and psychological distress) with the incidence and persistence of harmful alcohol use. Tests of heterogeneity were used to evaluate whether the associations differ between the incident and persistent use. RESULTS: Male sex, younger age, higher education, smoking, and psychological distress were associated with a greater risk of both the incidence and the persistence of harmful alcohol use in mutually adjusted models (ORs = 0.97-1.67). There were no differences in the associations of these risk factors with incident and persistent use, except that the association of psychological distress was greater with incident use compared to persistent use (p for heterogeneity < 0.05). CONCLUSIONS: These findings suggest that the incidence and persistence of harmful alcohol use share a similar set of sociodemographic risk factors in the general population.