Across five FINRISK surveys (1992–2012, 34,756 Finnish adults linked to national registries), the hazard of major adverse liver outcomes per unit of alcohol intake barely moved: HR 1.28 (95% CI 1.19–1.38) in 1992 versus 1.33 (1.17–1.52) in 2012, interaction p=0.94. This held even as high waist-hip ratio tripled (11.0% to 34.7%) and diabetes nearly tripled (3.7% to 10.8%). Metabolic adjustment did not weaken the alcohol signal. The all-cause mortality association, by contrast, faded from HR 1.17 to 0.99.
The result cuts against a popular worry that rising obesity and diabetes make alcohol progressively more dangerous to the liver, and against the opposing suggestion that alcohol's liver effect is weakening. Neither appears true at the population level here. Practically, alcohol remains a dose-dependent liver hazard regardless of metabolic status, so adults should not read a healthy waistline as a licence to drink, nor assume that already-high metabolic risk makes alcohol irrelevant. Important caveats apply: intake was self-reported and likely underestimated, the 2012 estimate rests on fewer events and a wider interval, and a population-level absence of interaction does not rule out synergy in individuals with severe obesity or heavy drinking. The shrinking all-cause mortality association probably reflects changing competing risks and drinker profiles rather than any new safety. Overall, this is a confirmatory, reassuringly consistent finding that reinforces existing advice to limit alcohol for liver health.