A large meta-analysis of longitudinal studies has found a two-way relationship between depression and obesity.

A total of 15 studies (N = 58,745) met the inclusion criteria and provided data in a form which could be used in the meta-analysis. Heterogeneity was low.

Seven of these studies made a clinical diagnosis of depression using a structured diagnostic interview, whereas the remaining eight studies identified subjects with elevated depressive symptoms e.g. a HADS-D score ≥8. Overweight was defined as a BMI of 25-29.99 and obesity as a BMI ≥30. The follow-up period ranged from 3 to 28 years.

Participants who were overweight or obese at baseline were at significantly increased risk of having clinical depression or elevated depressive symptoms at follow-up. The unadjusted odds ratios were 1.27 (95% CI: 1.07-1.51, P<0.01) and 1.55 (95% CI: 1.22-1.98, P<0.001) respectively.

Conversely, normal weight participants who were depressed at baseline had a significantly increased risk of being obese (but not of being overweight) at follow-up, odds ratio 1.58 (95% CI: 1.33-1.87, P<0.001).

Somatic symptoms, such as fatigue and sleep disturbance, may be caused by either obesity or depression. Symptom scores may therefore be misleading and the use of a diagnostic interview is to be preferred. Most of the participants (44,396) came from a single study which was flawed both as a result of using symptom scores and also due to a failure to exclude subjects with elevated depression symptoms at baseline.1 Removal of this study from the analysis, however, did not affect the results.

The presence of a significant association does not, of course, establish a causal relationship. Unadjusted data was used for the analysis as the studies did not adjust for the same confounders. The large Norwegian study found that the association between baseline obesity and depressive symptoms at follow-up was unchanged after controlling for a range of socioeconomic, lifestyle and psychological factors.1 However, neither this nor any of the other studies adjusted for diabetes which, when recognised, is known to increase the risk of depression.2

The authors suggest a number of biological mechanisms whereby obesity may cause depression.

One possible mediator is increased insulin resistance, but this is thought to increase serotonin levels3 - which may account for the paradoxical finding that obesity reduces the risk of suicide.1 I suspect that psychosocial mediators are more important, namely poor body image, a sense of failure and an awareness of social stigma.

The finding by the present study that the association between obesity and incident depression was significantly greater among Americans when compared with Europeans may support this view.

The increased risk of obesity in depressed patients, as with the increased risk of diabetes,2 is likely to be due in part to lifestyle factors e.g. lack of exercise and poor diet. It is easy to envisage a vicious cycle whereby depression leads to obesity, which in turn exacerbates low self-esteem and depression.

I think there are three lessons for practice here:

• When case finding among patients with chronic conditions, physical inactivity and obesity should suggest the possibility of depression.
• When treating mild to moderate depression in obese patients, referral to a group exercise programme is likely to be a better option than prescribing antidepressants, which may themselves cause further weight gain.

• When following up our depressed patients, there is a case for monitoring their BMI (and physical activity levels) as well as the PHQ-9.