Background There is some evidence to suggest that obesity is a risk factor for the development of depression, although this is not a universal finding. This discordance might be ascribed to the existence of a ‘healthy obese phenotype’ – that is, obesity in the absence of the associated burden of cardio-metabolic risk factors. We examined whether the association of obesity with depressive symptoms is dependent on the individual’s metabolic health.

Methods Participants were 3851 men and women (aged 63.0 ± 8.9 yrs, 45.1% men) from the English Longitudinal Study of Ageing, a prospective study of community dwelling older adults. Obesity was defined as body mass index ≥ 30 kg/m². Based on blood pressure, HDL-cholesterol, triglycerides, glycated haemoglobin, and C-reactive protein, participants were classified as ‘metabolically healthy’ (0 or 1 metabolic abnormality) or ‘unhealthy’ (≥ 2 metabolic abnormalities). Depressive symptoms were assessed at baseline and at 2 years follow up using the 8-item Centre of Epidemiological Studies Depression (CES-D) scale.

Results Obesity prevalence was 27.5%, but 34.3% of this group was categorized as metabolically healthy at baseline. Relative to non-obese healthy participants, after adjustment for baseline CES-D score and other covariates, the metabolically unhealthy obese participants had elevated risk of depressive symptoms at follow-up (odds ratio [OR] = 1.50, 95% CI, 1.05 – 2.15), although the metabolically healthy obese did not (OR=1.38, 95% CI, 0.88 – 2.17). We repeated the main analysis after excluding 451 participants with existing depressive symptoms (CES-D≥4) at baseline. There were 238 incident cases of depression at follow up, and in comparison with healthy non-obese participants only the metabolically unhealthy obese participants had elevated odds of incident depression (OR=1.56, 95% CI, 1.09 – 2.22), but not their metabolically healthy obese counterparts (OR=1.45, 95% CI, 0.92 – 2.30) nor unhealthy non-obese participants (OR=1.38, 95% CI, 0.98 – 1.94). In further analysis we examined the associations between individual metabolic risk factors and depression. There was a dose-response association between the number of metabolic risk factors and risk of depression, although the risk only became significant in participants with more than one risk factor. Adverse triglycerides, impaired glycaemic control, and low grade inflammation were associated with depression at follow-up in models adjusted for age, sex and baseline CES-D score.

Conclusion The association between obesity and risk of depressive symptoms appears to be partly dependent on metabolic health, although further work is required to confirm these findings.