Drinking less than a couple of glasses of water each day may predispose to hyperglycaemia, a new study suggests.
Over time, adults who drank no more than half a litre of water each day were more likely to have blood glucose levels rise into the pre-diabetes range, compared to people who drank more water.
The findings show a correlation between water intake and blood sugar, but do not prove cause-and-effect, said senior researcher Dr Lise Bankir, of the French national research institute INSERM.
Still, it is plausible, Dr Bankir said - and vasopressin is the potential missing link.
Vasopressin levels rise during dehydration, causing the kidneys to conserve water. But research suggests that higher vasopressin levels may act on receptors in the liver, causing it to produce and release glucose.
"There are good arguments to suggest that there could be a real cause-and-effect relationship in the association we have found," Dr Bankir said, "but this is not proof."
The findings are based on a longitudinal study of 3,615 French adults, ages 30 to 65, with normoglycemia at baseline. About 19% said they drank less than half a litre (500ml) of water each day.
Risk according to water intake
Over the next nine years, 565 study participants developed hyperglycaemia and 202 developed type 2 diabetes.
When the researchers looked at the participants' risk according to water intake, they found that people who drank at least 500ml of water per day were 28% less likely to develop hyperglycaemia than those who drank less.
Dr Bankir and her colleagues accounted for intake of sugary drinks and alcohol, as well as people's body weight at the start of the study, their reported exercise levels and other health factors. And the link between low water intake and high blood sugar persisted.
However, they could not control for everything, including generally healthy or less-healthy eating habits. "Healthier behaviours correlating with higher water drinking could account for the observed association," the researchers write.
There was no significant association of water intake with diabetes risk, but Dr Bankir believes the number of diabetes cases in the study was "too small to get a significant result." A larger study might have been able to detect a statistically significant link, she said.
The study was published online in Diabetes Care.
(Reuters Health, Amy Norton, October 2011)