When I first saw the sample size - 1,087,803 participants - I figured this was worth reading closely. Most alcohol-and-liver studies work with thousands. This one enrolled over a million South Korean men and followed them for 11 years. A dataset that size tends to settle arguments.

The argument it settles: whether alcohol guidelines built on Western populations apply to East Asian men with fatty liver. Short answer, no. The reason traces back to a single gene.

The setup

Researchers pulled men aged 40 and older with MASLD from South Korea's 2011-2012 national health examinations. MASLD - metabolic dysfunction-associated steatotic liver disease - is the current umbrella term for fatty liver tied to metabolic factors like obesity and insulin resistance, not to drinking itself. That matters. The cohort started with men whose livers were already vulnerable before alcohol entered the picture.

Alcohol intake was sorted into five categories: none, under 70 g/week, 70-140 g/week, 140-210 g/week, and 210 g or more per week. A standard US drink holds roughly 14 g of pure alcohol, so 70 g/week is about 5 drinks spread over seven days.

Primary outcome: liver-related events (LREs) - newly diagnosed hepatocellular carcinoma, cirrhosis with or without decompensation, and liver-related death. Median follow-up 11 years.

The numbers

26,742 participants (2.5%) developed an LRE during follow-up.

The dose-response came out J-shaped. The lowest tier - under 70 g/week - showed no significant elevated risk compared with not drinking at all. Above that floor, risk climbed:

  • 70-140 g/week. No significant increase across the full cohort. But in vulnerable subgroups - men with diabetes, BMI under 25, or abnormal alanine aminotransferase levels - risk already rose here.
  • 140-210 g/week. Adjusted hazard ratio 1.10 (95% CI: 1.05-1.14).
  • 210+ g/week. Adjusted hazard ratio 1.30 (95% CI: 1.25-1.34).

Spline analysis confirmed the nonlinear shape. Risk doesn't climb evenly with each additional drink, it accelerates past the threshold.

Why the line sits lower here

The mechanism runs through ALDH2 (aldehyde dehydrogenase 2, a gene that helps the body break down alcohol). Many East Asians carry a variant that reduces ALDH2 activity. When ALDH2 function is impaired, acetaldehyde - the toxic intermediate in alcohol metabolism - accumulates faster at lower doses. Standard guidelines were calibrated on populations where this variant is rare.

The J-shape matters, though not in the obvious way. Even in this vulnerable population, the lowest drinking tier showed no significant elevated risk. The finding is about where the threshold sits, not about whether one exists.

The subgroup data is the part worth taking seriously. Men with diabetes or elevated liver enzymes saw their risk curve shift at 70-140 g/week. Earlier than the general cohort, and well below where most clinical guidelines flag concern.

Limits. It's male-only. ALDH2 variant frequency varies within East Asian populations, and the researchers inferred genetic risk from ethnicity rather than direct genotyping. These numbers apply to men with MASLD, not to every East Asian drinker. Still, a million men followed for 11 years is enough to trust the general shape of what they found.

The practical question is whether the reference ranges your doctor uses were calibrated for your biology. For East Asian men with a fatty liver diagnosis, this study says probably not.

Source: Clinical Gastroenterology and Hepatology, DOI