Use of antidepressant drugs by patients at risk for type 2 diabetes is associated with a significantly increased risk of developing the disease, although the effect is not seen in patients who take metformin, according to a presentation at the annual meeting of the American Diabetes Association.

Previous studies have yielded conflicting reports on the role of depression in risk for diabetes. Researchers have concurred, however, that people who have diabetes and depression need better management for their psychiatric illness in order to optimize ability to control diabetes.

"In a sub-analysis of the Diabetes Prevention Program, which sought to reduce progression to type 2 diabetes by either lifestyle modification or the drug metformin, compared with placebo, in people at very high risk for diabetes, we found that those in the lifestyle and placebo arms of the trial who were taking antidepressant drugs when they started the study, or who took them frequently during the study, were two to three times more likely to develop diabetes during the study," said Richard R. Rubin, PhD, Associate Professor, Medicine and Pediatrics, The Johns Hopkins University School of Medicine, Baltimore, MD, who was a co-investigator in the study.

There was no increased risk of diabetes seen in those not taking antidepressants in the study, suggesting that the antidepressant medications predispose to diabetes. While it is well-known that some antipsychotic drugs may increase a person's risk for diabetes, this is the first indication that antidepressant drugs can increase the risk of diabetes. Although this finding occurred in an already high-risk population, it has serious public health implications.

Dr. Rubin noted that the association between antidepressants and diabetes has not been reported before, and the mechanism for this association is unknown. The effect remained after adjustment for all likely risk factors such as increased weight.

"We have to keep in mind that that this occurred in a population at very high risk for diabetes so we don't know whether the findings would apply to the general population," said Rubin, "but this should be explored further."