Target Trial Emulation of Alcohol Intake Interventions in Young Adults and Risk of CVD over 30 years: Data from the CARDIA study

Structured Abstract Aim This study uses a target trial emulation framework to estimate how different hypothetical sustained alcohol intake patterns affect 30-year cardiovascular disease risk in young adults. Methods This target trial emulation used longitudinal data from the CARDIA study, a multi-site cohort of Black and White young adults aged 18–30 at baseline (1985–1986), followed for 35 years. The exposure of this study consisted of self-reported daily alcohol intake categorized by sex into abstainer, light, moderate, heavy, and very heavy use categories. The emulated intervention compared hypothetical scenarios in which participants maintained consistent levels of alcohol consumption over time (assigned each of the 5 categories above) to estimate their long-term cardiovascular risk. Using the parametric g-formula to account for time-varying confounding and treatment–confounder feedback, we then estimated the 30-year incident risk of the composite cardiovascular outcome. Results Over 30 years, the overall CVD risk among all participants was 8.7%. Abstaining was associated with a slightly lower risk (8.3%; risk difference [RD] –0.4%; 95% CI, –0.6 to 0.1), while light (8.8%; RD 0.2%; 95% CI, –0.1 to 0.2) and moderate (9.4%; RD 0.7%; 95% CI, –0.1 to 0.9) drinking showed minimal risk differences. Heavy (RD 1.3%; 95% CI, –0.2 to 1.8) and very heavy drinking (RD 1.9%; 95% CI, –0.3 to 2.9) suggested modestly increased risks, though confidence intervals included no effect. Conclusion This hypothetical intervention indicates no meaningful difference in 30-year CVD risk among abstainers, light, or moderate drinkers in young adults. Heavy alcohol intake may modestly raise risk, but estimates are imprecise with a range from no effect to a modest increase. By addressing key biases and leveraging comprehensive longitudinal data, these findings offer valuable insights for clinical and public health guidance.