News · Nutrition & Diet
Heavy drinking and obesity both raised diabetes risk, separately
A Diabetologia study followed 7,817 Korean adults for 16 years and found obesity and heavy drinking each raised diabetes risk, but combined less than adding them up would predict.
Based on a peer-reviewed cohort study in Diabetologia
- Researchers followed 7,817 Korean adults aged 40 and over from 2001-02 until 2017-18, publishing in Diabetologia.
- Nobody had diabetes at the start.
- Obesity tracked with 91% higher diabetes risk, or about 22 extra cases per 1,000 person-years.
- Heavy drinking, defined as 30 g a day or more, tracked with 23% higher risk, about 15 extra cases per 1,000 person-years.
- Low-to-moderate drinking showed weaker associations that appeared only on one of the two statistical scales tested.
- Measuring exposures repeatedly over time produced stronger associations than measuring once at the start.
- Combining obesity and drinking produced less added risk than simply summing the two would predict.
- Observational, in a single Asian cohort. Drinking is self-reported and alcohol metabolism varies by population.
Obesity and heavy drinking are both established diabetes risks, usually studied one at a time. A study in Diabetologia followed nearly eight thousand Korean adults for sixteen years to see what happens when someone carries both.
Obesity and high alcohol consumption were consistently associated with incident diabetes, with stronger associations observed in the time-varying models.
Two questions in one design
The study set out to test something more specific than whether either exposure matters. The aim was to examine whether the independent and joint effects of obesity and alcohol consumption on diabetes risk differ when exposures are assessed at baseline or as time-varying variables.
That second half is the methodological point. Most cohort studies classify someone once, at enrolment, and carry that label for twenty years. People do not cooperate with that: they gain weight, they start drinking more, they stop.
Sixteen years of updates
In total 7,817 participants aged 40 and over without diabetes at baseline were followed from 2001-2002 until 2017-2018.
The analysis ran two ways: with exposures fixed at baseline, and with them updated through follow-up. Comparing the two is what reveals how much a single snapshot costs.
What each exposure carried
Both mattered, and obesity mattered more. Obesity raised the risk by about 91%, or 21.9 additional cases per 1000 person-years.
High alcohol consumption, defined as 30 g a day or more, raised it by about 23%, or 14.6 additional cases per 1000 person-years.
Reporting the absolute figures alongside the relative ones is worth noticing. Twenty-two extra cases per thousand person-years is a number a reader can hold; a percentage increase on an unstated baseline is not.
Moderate drinking stayed ambiguous. Associations for low-to-moderate alcohol consumption were weaker, and held up only on one of the two scales tested. A result that depends on which scale you pick is a result to describe as uncertain.
The interaction, which goes the unexpected way
If two risks each raise your odds, the intuitive assumption is that carrying both is the sum of the two. It was not.
No multiplicative interactions were observed. But antagonistic additive interactions were observed between obesity and current alcohol consumption.
Antagonistic here means the combination produced less excess risk than adding the separate excesses together would predict. Both exposures are still harmful. Their arithmetic is simply not additive, which matters for anyone estimating population burden by stacking risk factors.
What one cohort cannot show
Nobody was assigned to drink or to gain weight, and heavy drinkers differ from light drinkers in diet, smoking, income and much else.
Alcohol intake is also self-reported, and self-reported drinking is reliably an underestimate, which would tend to push the true thresholds higher than the ones printed here.
The setting matters too: this is a single Asian cohort, and both diabetes risk at a given body size and alcohol metabolism differ across populations. The direction of these findings is likely to travel further than the numbers are.
People also ask
What does it mean that the combination was less than the sum?
It means the two risks do not simply stack. Someone who is both obese and drinking heavily had higher risk than either alone, but lower than you would get by adding the separate excesses together. Researchers call that an antagonistic additive interaction. It does not mean drinking is protective for anyone; it means the arithmetic of combining risks is not straightforward.
How much is 30 grams of alcohol a day?
Roughly two to three standard drinks, depending on the country's definition. It is above most national guidelines, which is why the study treats it as the high-consumption category rather than as ordinary drinking.
Why does time-varying measurement matter?
Because people change. Someone classified as a light drinker at enrolment in 2001 may be drinking heavily by 2010, and a study that only measures once will attribute their outcome to the wrong category. Updating exposures through follow-up produced stronger associations here, which suggests single-baseline studies have been understating both effects.
Does moderate drinking raise diabetes risk?
This study is equivocal, which is itself informative. Associations for low-to-moderate consumption were weaker and reached significance only on the additive scale, not the multiplicative one. When a result depends on which statistical scale you choose, the honest description is uncertain rather than positive or negative.
Does a Korean cohort apply elsewhere?
Partly. Body composition at a given weight, diabetes risk at a given body mass index, and alcohol metabolism all differ across populations, and East Asian populations carry variants affecting how alcohol is processed. The qualitative finding that both exposures matter is likely to travel; the specific numbers are less certain to.