There is strong public belief that polyunsaturated fats protect against and ameliorate depression and anxiety.

To assess effects of increasing omega-3, omega-6 or total polyunsaturated fat on prevention and treatment of depression and anxiety symptoms.

We searched widely (Central, Medline and EMBASE to April 2017, trial registers to September 2016, ongoing trials updated to August 2019), including trials of adults with or without depression or anxiety, randomised to increased omega-3, omega-6 or total polyunsaturated fat for ≥24 weeks, excluding multifactorial interventions. Inclusion, data extraction and risk of bias were assessed independently in duplicate, and authors contacted for further data. We used random-effects meta-analysis, sensitivity analyses, subgrouping and Grading of Recommendations, Assessment, Development and Evaluations (GRADE) assessment.

We included 31 trials assessing effects of long-chain omega-3 (n = 41 470), one of alpha-linolenic acid (n = 4837), one of total polyunsaturated fat (n = 4997) and none of omega-6. Meta-analysis suggested that increasing long-chain omega-3 probably has little or no effect on risk of depression symptoms (risk ratio 1.01, 95% CI 0.92–1.10, I2 = 0%, median dose 0.95 g/d, duration 12 months) or anxiety symptoms (standardised mean difference 0.15, 95% CI 0.05–0.26, I2 = 0%, median dose 1.1 g/d, duration 6 months; both moderate-quality evidence). Evidence of effects on depression severity and remission in existing depression were unclear (very-low-quality evidence). Results did not differ by risk of bias, omega-3 dose, duration or nutrients replaced. Increasing alpha-linolenic acid by 2 g/d may increase risk of depression symptoms very slightly over 40 months (number needed to harm, 1000).

Long-chain omega-3 supplementation probably has little or no effect in preventing depression or anxiety symptoms.

L.H. and A.A. were funded to attend the World Health Organization Nutrition Guidance Expert Advisory Group (NUGAG) Subgroup on Diet and Health meetings and present review results. The authors report no other conflicts of interest.