The Doctor Weighs In

I recently came across an article in the April 23 issue of the Archives of Internal Medicine (vol. 167, pp. 802-807; 2007) describing a newly-discovered connection between depression and diabetes type 2 in older adults. The study enrolled 4,681 non-diabetic men and women over age 65, and followed them for 10 years. The participants filled out every year a questionnaire to measure their depressive symptoms, and every 2-4 years had their blood pressure measured. After removing confounding factors that are well known to increase the incidence of diabetes, like increased body mass index, alcohol consumption, and smoking, they discovered that even a single report of high depressive symptoms is highly associated with increased risk of diabetes type 2. In fact, there was a 60% increased chance of developing diabetes after reporting one incidence of high depression; this is significant by any measure.

Is it a good study?

I think it is an excellent study. The study group was ‘clean’, namely all diabetics and pre-diabetics were excluded from participation in the study. It was longitudinal, following the 4,581 participants for 10 years. The numbers were large enough and the observation period long. Furthermore, the authors accounted for all the obvious confounding factors, like obesity, smoking status, alcohol consumption, and anti depressant use. What is most convincing is that there was a relationship between the score of depression and the likelihood of developing diabetes, what we call in experimental medicine a ‘dose/ response’ relationship. The highest score had the highest increase in risk (60%).

How could depression cause diabetes?

There could be several explanations for this observation, and the answer is not really known. But if we look at the hormonal changes that accompany depression, we may find a possible link to diabetes. Cortisol is a stress hormone that is elevated in depressive states. Cortisol also happens to inhibit the secretion of insulin from the beta cells in the pancreas. Lower blood insulin would in turn lead to higher blood sugar, a hallmark of diabetes.

Caveat emptor

As always, when I speculate I feel obliged to caution that what ‘makes sense’, in biology it ain’t necessarily so. Biology is so complex, and largely still unknown, that any attempt to ‘dry lab’ it is an exercise in futility. Still, using speculation to form a testable hypothesis is how science works. For instance, we could test the cortisol hypothesis by measuring its blood levels in patients reporting an episode of depression and see if there is a correlation between it and the onset of diabetes. Or an experiment could be devised in which any participant reporting a depressive episode would be immediately treated with anti depressive drugs. Would that reduce the risk of developing diabetes?

I am sure many of you have other suggestions to study the depression/diabetes relationship. Send in your suggestions; we promise to publish and critique them.

Dov Michaeli, MD, Ph.D.