In 6,755 MESA participants aged 45-84 followed without baseline cardiovascular disease, re-analysing alcohol intake with occasional drinkers as the reference group and reallocating former drinkers by past drinking patterns reshaped the results. Moderate drinking's protective links to myocardial infarction (HR 0.64, 95% CI 0.38-1.06) and cardiovascular mortality (HR 0.81, 0.53-1.25) lost significance. Light drinking (2-7 drinks/week) rose to higher risk of MI (HR 1.46), stroke (1.54), MACE (1.33) and heart failure (1.42), and coronary artery calcium above 300 showed a dose-response with intake.
The J-shaped curve has long rested on lifetime abstainers as the comparison group, a group that often includes people who stopped drinking because of illness. This paper goes a step further than the usual fix of dropping former drinkers, arguing that doing so leaves survivorship bias among current drinkers, and its 'intention-to-treat' reallocation is a clever attempt to address that. The pattern now matches Mendelian randomization findings and the 2023 meta-analytic work by Stockwell and colleagues, which also found no clear protection at low intake. Limits matter: this is observational, alcohol was self-reported at a single point, and the reallocation rests on assumptions about why people quit. The moderate-drinking intervals remain wide, so the light-drinking harm is more striking than the null result at moderate intake. Editorially, this is confirmatory rather than paradigm-shifting, but it strengthens the case that adults should not start or continue drinking for heart health.