Drinkers with obesity-related organ problems reported almost six times the alcohol: 11.1 g a day against 1.9 g. Then researchers looked at who was in each group

In 2025 a Lancet commission drew a new line through obesity. On one side is preclinical obesity: excess body fat with organs and daily function still intact. On the other is clinical obesity, where the fat is already linked to organ dysfunction or limits on everyday activities. The obvious next question is whether people on either side eat and drink differently. A study published on 3 November 2026 put that to 135 adults arriving for their first visit at an obesity clinic in Tomar, Portugal.

The number that jumps out concerns the people who drank. Among them, those with clinical obesity reported a median of 11.1 grams of alcohol a day, against 1.9 grams with preclinical obesity (p = 0.004). For scale, a US standard drink holds about 14 grams, so that is most of one drink a day against a few sips. Alcohol also made up 3% of their energy intake rather than 0.6%. Read alone, it invites a tidy story in which drink helps push obesity over the line.

Now look at who was in each group. The share who drank at all barely differed: 23 of 48 people with preclinical obesity and 42 of 87 with clinical obesity, about 48% each by our arithmetic from the paper's counts. Across everyone, including non-drinkers, the median was zero grams in both groups. But the clinical group was 12 years older on average (56 against 44), somewhat heavier (mean BMI 39.0 against 36.8) and held twice the proportion of men, 38% against 19%. That last detail matters, because men who drank reported a median of 21.5 grams a day and women who drank 2.4.

Splitting by sex shows why the pooled figure misleads. Women who drank averaged 2.2 grams a day with preclinical obesity and 8.3 with clinical obesity. Among men the means ran the other way, 36.6 and 30.8. When the authors adjusted for age, sex and BMI, alcohol intake was no longer linked to clinical obesity (adjusted odds ratio 0.990, 95% CI 0.969 to 1.011). The same thing happened with body shape: alcohol correlated with waist-to-hip ratio and body fat percentage until the same adjustment, and then neither link held.

Diet told a duller and perhaps more important story. Healthy Eating Index scores were almost identical, 49.3 and 49.5, and the authors' grading gave both totals an F. Energy intake (2,205 and 2,174 kcal) and fibre (16.7 and 17.0 g) were close too. Median scores for whole grains and sodium were zero in both groups. The one difference was a small one in protein, 18.5% of energy against 17.2%, which also showed per kilogram of current body weight but not per kilogram of ideal body weight. Whatever separates the two kinds of obesity, this snapshot didn't find it on the plate.

None of this clears alcohol. The comparison rests on 65 drinkers at one clinic, with intake recalled on a questionnaire, and people tend to under-report drinking. A single snapshot can't show which came first, and physical activity, medication and income weren't accounted for. Organ problems came from medical records rather than study blood tests, and liver disease wasn't assessed at all, so the study cannot speak to the place where alcohol and obesity may matter most. Nearly one man in four (23.8%) drank above the study's limit of 30 grams a day, and the authors still argue for alcohol screening across obesity care. What the paper does show is how easily a crude gap can come from who happens to be in each group.