In adults with metabolic syndrome, the risk tied to sugary drinks didn't climb steadily. It seemed to start around 90 calories a day

Most advice about sugary drinks treats them as a straight line: more is worse. A team at Beijing Friendship Hospital wondered whether that holds for people whose metabolism is already under strain. So they turned to NHANES, the long-running US health survey, and split 26,885 adults aged 20 to 79 into two groups: 8,656 with metabolic syndrome, meaning at least three of five warning signs such as a large waist, high blood pressure and raised fasting glucose, and 18,229 without it.

Each person had described what they ate and drank over one 24-hour period at some point between 2003 and 2018, and those answers were linked to national death records up to the end of 2019. Over a median of about eight years, 1,434 people with metabolic syndrome and 1,601 without it died. After adjusting for age, sex, income, smoking, alcohol, body mass index, existing illnesses and more, the group with metabolic syndrome showed a bent curve rather than a straight one. Below about 90 calories a day from sugary drinks there was no clear link with death. Above it, each standard-deviation step up went with an 18% higher risk (hazard ratio 1.18). Sugar told a similar story past about 20 grams a day, roughly five teaspoons, though the bend itself narrowly missed statistical significance.

It is tempting to read that as a daily allowance. Please don't; the authors don't either. They call these potential inflection points and say they should not be used as clinical cut-offs. Intake came from a single day's recall, which can't show a person's usual habit or how it changed over the following years. The listed adjustments don't include physical activity or overall diet quality, both of which could travel with drink choices. And the study looked at death from any cause, so it can't say what the drinks might have had to do with any particular death.

Two details deserve a second look. The paper says there was no association in people without metabolic syndrome, but the estimate there (1.09 per standard deviation of calories, 95% CI 0.99 to 1.21) isn't far from the one for people with it across the whole intake range (1.14, 95% CI 1.05 to 1.23). The intervals overlap heavily, so 'smaller or less certain' is a fairer reading than 'no effect'. The other is weight. The 950 g/day inflection point sounds like a lot of soda, but the medians in the metabolic syndrome group were 867 g and only about 91 calories, and the methods list carbonated and bottled water among the drinks counted. Grams here may not mean what you'd assume, which makes calories and sugar the more useful numbers.

Within the metabolic syndrome group, the link looked stronger in women than in men: per standard deviation of sugar, a hazard ratio of 1.23 against 1.04, with a formal test for the difference (p = 0.02). Older adults looked higher too, but that gap did not pass the same test. None of this proves that cutting back would lengthen anyone's life, since no one was asked to change what they drank. Still, for someone already told they have metabolic syndrome, swapping a daily sweet drink for water or unsweetened tea is a low-cost thing to try and easy to raise at the next check-up.