Claim that antidepressants cause diabetes unproven

"Happy pills are linked to a higher risk of diabetes," is the headline in the Daily Mail.

The newspaper reports on a review that examined available evidence to see if there was a link between antidepressant use and type 2 diabetes.

While the review itself was thorough, the individual studies that were included in the review were of varying quality. They also used different methods, which makes it difficult to combine their results.

The review was also limited by the fact that it didn’t include any randomised controlled trials. These would be required to prove direct cause and effect (causality).

The researchers report an overall association between antidepressant use and diabetes, but they acknowledge that the studies gave mixed results and do not prove that antidepressant use causes diabetes.

Any association between antidepressant use and type 2 diabetes is likely to be extremely complex and involve multiple factors.

Overall, the limitations of this review make it difficult to draw meaningful conclusions. It can certainly not be used to prove causality. The researchers conclude with a recommendation that research into individual types of antidepressants, rather than antidepressants in general, be carried out.

If you are currently taking antidepressants you should not suddenly stop taking them as this could lead to a worsening of your symptoms as well as side effects.

Where did the story come from?

The review was carried out by researchers from the University of Southampton and was funded by the University of Southampton.

It was published in the peer-reviewed medical journal Diabetes Care.

The Daily Mail’s use of the term “Happy pills” trivialises the treatment of a range of life-threatening conditions. It fails to convey the complex effects that antidepressants have on the brain and nervous system.

BBC News and ITV News reported the story more sensitively and accurately, acknowledging that the review does not show that antidepressants cause diabetes, only that there is an association.

What kind of review was this?

This was a systematic review of all available studies that the researchers could identify that had collected data on adults who had taken antidepressants and had a recorded test or assessment for diabetes. It aimed to see if there was a link between taking antidepressants and developing diabetes, and whether antidepressants had an effect on control of blood sugar.

The researchers were only able to identify observational studies, which were of variable quality and methods (including variable populations, duration of follow-up, antidepressant use and measures of diabetes outcomes). It is not possible, based on these studies, to prove that the antidepressants caused diabetes. Randomised controlled trials would be required for this.

What did the research involve?

The researchers performed a comprehensive search of medical databases, such as The Cochrane Library and MEDLINES to identify appropriate studies.

Studies were judged to be appropriate if they involved:

  • adults aged 18 or over who had been prescribed antidepressants, and
  • who during the study period were either assessed for the development of diabetes, or
  • had a blood test to measure blood sugar levels.

A quality assessment for each study that was included was performed using tools appropriate to the study design.

A total of 1,638 articles were initially identified, and review of these articles brought the number down to three systematic reviews and 22 studies that were eligible for inclusion. Research designs included one case series and 21 observational studies (four cross-sectionalfive case-control and 12 cohort studies).

The studies were of such a wide variety that it was not possible to combine their results to perform any sort of useful meta-analysis. Instead, a narrative summary of each study was provided. 

What were the basic results?

The case series that was included in the review reported on 17 patients who were taking a variety of antidepressants. Increased blood sugar levels resolved in all patients after they stopped taking the antidepressants.

The four cross-sectional studies suggested a relationship between diabetes and depression, but this association was not explained by antidepressant use.

The five case-control studies showed an approximate doubling of diabetes risk in people receiving antidepressants.

There were inconsistent results from the 12 cohort studies. Several showed an increased risk of diabetes in those taking antidepressants, but not all of them. The most recent larger study showed a weaker association.

The three systematic reviews were largely based on small studies and were inconclusive.

How did the researchers interpret the results?

The researchers concluded that “the evidence suggests a link between antidepressant use and diabetes, but causality is not established. The strength of association in the larger most recent cohort studies is weak, which increases the chance that the finding occurs through residual confounding".


This systematic review of 22 observational studies looked at whether there was a link between taking antidepressants and developing diabetes.

The researchers reported an overall association between antidepressant use and diabetes, but acknowledge that the studies gave a mixed picture of results and do not prove that antidepressant use causes diabetes.

The review had several limitations:

  • Overall, the included observational studies were of variable design, methods and quality. There was a lack of any randomised controlled trials, which would be required to prove causality.
  • Not all of the studies accounted for the known risk factors for diabetes, or for other confounding factors that may be associated with both diabetes and antidepressant use (for example overweight and obesity).
  • The trials were difficult to compare as different measures were used to diagnose depression and diabetes (including self-reports in some cases rather than confirmed medical diagnoses). They also varied in terms of the type of antidepressants used, the length of time patients were on antidepressants and the length of follow-up.
  • Many of the trials had too few participants to show any reliable statistically significant conclusions.

It is possible that the link may be due to weight gain that can occur with some antidepressants, but it is difficult to know whether it was excess weight that directly led to the development of diabetes, or if any weight gain was due to antidepressant use. Also, the authors report that some studies adjusted for body weight and still observed an association between diabetes and antidepressant use.

Overall, the variable methods, quality and results of the studies included in this review make it difficult to draw any firm conclusions about the possible link between diabetes and antidepressant use.

It is not advisable to stop taking antidepressants abruptly, or without discussing it with your GP or the doctor in charge of your care. Most antidepressants need to be slowly reduced to prevent withdrawal side effects.

As altered glucose metabolism is a recognised potential side effect of some antidepressants, your blood glucose level should be checked before you begin taking antidepressants and it is recommended that this is repeated every six months during treatment. Your doctor can provide you with more information.

The authors conclude that more research is required but, in the meantime, we already know that a healthy lifestyle can prevent the onset of diabetes and is also beneficial for all types of mental illness. Read more about exercise, mood and stress.

NHS Attribution