Review findings on the role of dietary patterns in preventing depression are inconsistent, possibly due to variation in assessment of dietary exposure and depression. We studied the association between dietary patterns and depressive symptoms in six population-based cohorts and meta-analysed the findings using a standardised approach that defined dietary exposure, depression assessment and covariates.

Included were cross-sectional data from 23 026 participants in six cohorts: InCHIANTI (Italy), LASA, NESDA, HELIUS (the Netherlands), ALSWH (Australia) and Whitehall II (UK). Analysis of incidence was based on three cohorts with repeated measures of depressive symptoms at 5–6 years of follow-up in 10 721 participants: Whitehall II, InCHIANTI, ALSWH. Three a priori dietary patterns, Mediterranean diet score (MDS), Alternative Healthy Eating Index (AHEI-2010), and the Dietary Approaches to Stop Hypertension (DASH) diet were investigated in relation to depressive symptoms. Analyses at the cohort-level adjusted for a fixed set of confounders, meta-analysis used a random-effects model.

Cross-sectional and prospective analyses showed statistically significant inverse associations of the three dietary patterns with depressive symptoms (continuous and dichotomous). In cross-sectional analysis, the association of diet with depressive symptoms using a cut-off yielded an adjusted OR of 0.87 (95% confidence interval 0.84–0.91) for MDS, 0.93 (0.88–0.98) for AHEI-2010, and 0.94 (0.87–1.01) for DASH. Similar associations were observed prospectively: 0.88 (0.80–0.96) for MDS; 0.95 (0.84–1.06) for AHEI-2010; 0.90 (0.84–0.97) for DASH.

Population-scale observational evidence indicates that adults following a healthy dietary pattern have fewer depressive symptoms and lower risk of developing depressive symptoms.

Our aim was to evaluate the relationship between adherence to different Dietary Approaches to Stop Hypertension (DASH) diet indices and the risk of depression.

In a prospective study we assessed 14051 participants of a dynamic (permanently ongoing recruitment) prospective cohort (the Seguimiento Universidad de Navarra (SUN) Project), initially free of depression. At baseline, a validated FFQ was used to assess adherence to four previously proposed DASH indices (Dixon, Mellen, Fung and Günther). To define the outcome we applied two definitions of depression: a less conservative definition including only self-reported physician-diagnosed depression (410 incident cases) and a more conservative definition that required both clinical diagnosis of depression and use of antidepressants (113 incident cases). Cox regression and restricted cubic splines analyses were performed.

After a median follow-up period of 8 years, the multiple-adjusted model showed an inverse association with the Fung DASH score (hazard ratio (HR)=0·76; 95 % CI 0·61, 0·94) when we used the less conservative definition of depression, and also under the more conservative definition (HR=0·63; 95 % CI 0·41, 0·95). We observed a weak inverse association with the Mellen DASH score, but no statistically significant association was found for the other definitions. The restricted cubic splines analyses suggested that these associations were non-linear (U-shaped).

Moderate adherence to the DASH diet as operationalized by Fung and Mellen was related to lower depression risk. Since these associations were non-linear, additional prospective studies are required before the results can be generalized and clinical recommendations can be given.