Heavy episodic drinking and daily cigarette smoking in adolescence and early adulthood showed no clear association with psychological distress in midlife, according to a multicohort study published in the International Journal of Mental Health and Addiction. The research drew on four prospective cohorts from Finland and Sweden, following participants for periods ranging from 19 to 33 years, and found results that were broadly consistent across all four datasets. The finding stands in contrast to much of the earlier literature, which has generally reported stronger longitudinal links between early substance use and later mental health problems.
The study, led by Noora Berg of the Finnish Institute for Health and Welfare with colleagues from Finnish and Swedish research institutions, examined both the separate and combined associations of two substance use behaviors with midlife psychological distress. The cohorts were the TAM study from Tampere, Finland (N = 1334 in the analytical sample), FinnTwin16 (N = 4409), the Northern Finland Birth Cohort 1966 (N = 7147), and the Swedish Individual Development and Adaptation program (N = 514). Substance use was measured in adolescence at ages 14 or 16 and again in early adulthood at ages 22, 24, 25, or 26, while psychological distress was assessed decades later at ages 35, 42, 46, or 49 depending on the cohort.
The researchers defined heavy episodic drinking, abbreviated HED, as self-perceived drunkenness or the consumption of at least 60 grams of pure alcohol on a single occasion. Smoking was classified as daily cigarette use, distinguishing it from never, occasional, or former smoking. Psychological distress was measured with the 12-item General Health Questionnaire, a widely used screening instrument that asks about recent mental health states such as the ability to make decisions or feelings of unhappiness and depression. Items were scored on a binary basis, producing a total score from 0 to 12, and participants scoring 3 or more were classified as having high psychological distress.
Beyond the individual associations, the study addressed whether the timing and persistence of substance use mattered. Participants were classified into four categories for each behavior: use in neither period, adolescence only, early adulthood only, or both periods. The researchers also created joint exposure categories within each life phase, grouping participants as using neither substance, HED only, smoking only, or both. To test whether the combined effect of drinking and smoking exceeded the sum of their individual effects, they calculated the Relative Excess Risk due to Interaction, a measure of additive interaction, with confidence intervals computed by the method suggested by Knol and colleagues.
The analytical strategy was a two-step individual participant meta-analysis. Aligned logistic regression analyses, or generalized mixed models in the twin cohort to account for family-level clustering, were first run separately within each cohort, and the resulting estimates were then pooled using random-effects meta-analysis. Models were adjusted for parental socioeconomic position and for baseline psychological symptoms in adolescence or early adulthood, depending on the life stage analyzed. Women and men were examined separately, and sensitivity analyses repeated the main models using the GHQ-12 as a continuous score and with a stricter cut-off of 3 or 4.
The pooled results showed no clear evidence of associations between adolescent or early adulthood heavy episodic drinking or smoking and midlife psychological distress, and this pattern held for both women and men. Among women, one adjusted analysis in the TAM cohort found that early adulthood daily smoking was associated with less midlife distress (odds ratio 0.65, 95% confidence interval 0.43 to 0.99), an unexpected inverse association that appeared only among women who already had high psychological symptoms in early adulthood and did not persist when the stricter distress cut-off was used. Among men, no clear associations emerged in the main binary analyses, although continuous-score sensitivity analyses in the Northern Finland Birth Cohort showed small positive associations for adolescent drinking and smoking.
Heterogeneity across cohorts was generally low to moderate, and the authors noted that some elevated heterogeneity estimates likely reflected random variation rather than true differences, given the small number of cohorts. One notable exception concerned the Swedish IDA cohort, which included women only. In that cohort, adolescence-limited and early adulthood-limited heavy episodic drinking were associated with midlife distress in the main analyses (for example, an odds ratio of 2.91, 95% confidence interval 1.08 to 7.86 for adolescence-limited HED), and combined use of both substances was also associated with distress, but several of these associations did not survive the stricter sensitivity analyses. The authors cautioned that these cohort-specific findings limit generalizability.
The additive interaction analyses returned null results: engaging in heavy episodic drinking or smoking in both adolescence and early adulthood did not produce a greater association with distress than the sum of the separate age-phase associations, and co-occurring drinking and smoking within a life phase did not show interaction effects beyond additivity in either women or men. The authors had hypothesized stronger associations when substance use began in adolescence and persisted into early adulthood, and a greater-than-additive risk when both behaviors co-occurred, but the data did not support these expectations, apart from the elevated associations observed among Swedish women.
The authors offer several explanations for the discrepancy with prior research. Much of the earlier evidence relies on diagnostic outcomes, capturing the more severe end of the symptom and substance use spectrum, whereas this study used subclinical measures of distress. Previous studies of subclinical outcomes have also typically followed participants only into early adulthood, so shorter intervals between predictor and outcome may account for the stronger associations reported elsewhere. The authors emphasize that their findings do not prove that no association exists; rather, any association persisting into midlife appears weaker than those seen over shorter follow-ups and may attenuate over the life course. They also note that publication bias may have inflated the prominence of statistically significant results in the existing literature.
Methodological limitations temper the conclusions. Harmonizing measures across cohorts required dichotomizing variables, which may have reduced sensitivity to small effects, and the HED definitions differed somewhat between cohorts, relying on self-perceived drunkenness in the Finnish cohorts and on drinking frequency plus quantity in the Swedish cohort. The available control variables were limited and varied across cohorts, attrition bias remains possible despite evidence that adolescent substance use and distress were not linked to later non-participation in the TAM cohort, and the Northern Finland cohort measured adolescence two years earlier than the others. The cohorts were born between the 1950s and 1970s in two Nordic countries, so the findings may not generalize to contemporary adolescents, among whom alcohol use and smoking have declined while nicotine pouches and e-cigarettes have become common. Nevertheless, the harmonized analyses across independent datasets, the long follow-up windows, and the consistency of the null pattern across all four cohorts strengthen confidence that early heavy drinking and daily smoking are unlikely to be strong predictors of subclinical psychological distress decades later, and the authors call for further research using diverse measures and longitudinal designs to clarify these long-term dynamics.
Subject of Research: Long-term associations of adolescent and early adulthood alcohol use and cigarette smoking with midlife psychological distress
Article Title: The Individual and Joint Associations of Alcohol Use and Cigarette Smoking in Adolescence and Young Adulthood with Psychological Distress in Midlife – A Multicohort Study
Article References: Berg, N., Grundström, J., Aaltonen, S., Latvala, A., Miettunen, J., & Kiviruusu, O. (2026). The Individual and Joint Associations of Alcohol Use and Cigarette Smoking in Adolescence and Young Adulthood with Psychological Distress in Midlife – A Multicohort Study. International Journal of Mental Health and Addiction. https://doi.org/10.1007/s11469-026-01754-9
Image Credits: AI Generated
DOI: 10.1007/s11469-026-01754-9
Keywords: heavy episodic drinking, cigarette smoking, psychological distress, adolescence, early adulthood, midlife, prospective cohorts, GHQ-12, meta-analysis, Finland, Sweden, substance use
News Source: Glenn Wilkins. (October 29, 2026). Teen Drinking and Smoking Show No Clear Link to Midlife Distress in Four Decades-Long Cohorts. Scienmag.
