People who use antidepressants may be at higher than average risk of getting type 2 diabetes, according to a new look at past studies. But exactly how the two are linked isn't clear.
Some antidepressants may be of greater concern than others. And some factors - such as higher doses and longer durations of use - seem to raise diabetes risks as well, a team of UK researchers reported.
But conflicting data point to a need for more research.
"A definitive association cannot be drawn," lead researcher Katharine Barnard, of the University of Southampton in the UK, told Reuters Health.
"There is clearly a link between antidepressant medication and type 2 diabetes," she wrote in an email. But so far, studies can't say whether the drugs actually cause diabetes.
"The results… are all over the place - study to study and drug to drug," Dr. Peter D. Kramer said. He is a clinical professor of psychiatry at Brown University in Providence, Rhode Island, and a spokesperson for the American Psychiatric Association.
"This study says yes, there's some reason to worry," said Kramer, who did not participate in the new research. "This is something to watch for and think about."
The number of antidepressant prescriptions written in the UK increased from 20.1 million in 1999 to 46.7 million in 2011, the researchers noted.
In light of that trend and studies showing the drugs may affect blood sugar, Barnard and her colleagues collected 25 years' worth of research to see whether there is a link between antidepressant use and diabetes risk.
They found one study that looked at 17 individual reports of blood sugar changes tied to the use of antidepressants.
Those included people with normal blood sugar levels who developed high blood sugar anywhere from three weeks to five months after starting antidepressants. Blood sugar levels returned to normal after people stopped taking the drugs.
The researchers also found 21 larger studies that ranged in size from 1,000 to more than 200,000 participants and yielded conflicting results.
For example, in one study of about 166,000 people with depression, researchers looked at 2,200 who were later diagnosed with diabetes. They found people who had used moderate to high doses of antidepressants for over two years were 84 percent more likely to get diabetes than those who hadn't used antidepressants recently.
In other studies the link between antidepressant use and diabetes was much weaker, or small enough that it could have been due to chance, according to findings published in the journal Diabetes Care.
Three past reviews on the topic found risks may vary by type of antidepressant. But results analyzed in those studies weren't always consistent either.
It's possible that weight gain associated with certain antidepressants could explain a higher diabetes risk, the researchers said. However, some studies took participants' weight into account and still found a link between the drugs and diabetes.
The studies included in the review were of varying quality, the UK team noted. They also used different methods to measure depression and diabetes - either asking participants directly or reviewing pharmacy and medical records.
And none of the studies were done using the gold standard design for medical research, known as a randomized controlled trial.
In that type of study, people would be randomly assigned to take antidepressants or not and followed to see who develops diabetes.
Kramer said the review suggests a need for caution.
"If you are on antidepressants, you should be checked for diabetes or glucose tolerance," he said. "And if you are gaining weight then discuss that with your doctor."
Marjorie Cypress, a nurse practitioner and spokesperson for the American Diabetes Association, stressed how important it is that people on antidepressants keep taking their medications.
"Some of these drugs can have some very bad side effects if you stop taking them suddenly," she told Reuters Health.
"If you have concerns, always talk to your healthcare provider," she said. She also noted that "keeping your weight down and exercising are probably the best things you can do to prevent (type 2) diabetes."