Hypothetical interventions on alcohol consumption and risk of mortality in healthy women: emulated target trials from the Norwegian women and health study.

We applied target trial emulation and the parametric g-formula to evaluate the effects of sustained alcohol consumption on 23-year mortality. Healthy women aged ≥ 35 years from the Norwegian Women and Health Study who completed questionnaires in 1998-2003 were eligible for the emulated trials (A and B). Trial B included only women reporting moderate alcohol consumption before baseline. In total, 57 964 and 14 393 women met the eligibility criteria for Trials A and B, respectively. The observed 23-year all-cause mortality risk was 6.97%. In Trial A, 23-year mortality risk differences (RDs) versus no intervention were 0.29% (95% CI, 0.02%, 0.58%) for no alcohol consumption, 0.12% (95% CI, 0.07%, 0.20%) for low consumption (>0 to 2 drinks/week), and - 0.44% (95% CI, -0.66%, -0.25%) for moderate consumption (>2 to 7 drinks/week). The lower mortality associated with moderate drinking was observed only among women aged ≥ 50 years. In Trial B, the 23-year mortality RD of reducing to low/no consumption (≤2 drinks/week) versus maintaining moderate consumption was 1.07% (95% CI, 0.10%, 1.66%). The results suggest a slight decrease in mortality from moderate alcohol consumption, particularly among women aged ≥ 50 years. In addition, the results suggest slightly increased mortality when reducing alcohol consumption from moderate levels.