Twelve studies were included: six nested case-control and six case-control. There were slightly over 10,000 participants (range 127 to 3,461) including 4,516 cancer cases. Follow-up in the nested studies ranged from four to 19 years. All studies were judged to be of good quality (scored between eight and 10 out of 11 according to the NHS criteria).
For associations between total prostate cancer and the four specific types of fatty acid only DPA showed a statistically significant association and this favoured higher levels of DPA (OR 0.76, 95% CI 0.60 to 0.96; six studies; I²=0%). Neither long-chain acids (DHA, DPA and EPA) nor fish oils (DHA and EPA) were significantly associated with total prostate cancer. There was no significant association when all four types of fatty acid were included (OR 1.0, 95% CI 0.90 to 1.11; I²=26%).
There was no significant association between omega-3 and advanced prostate cancer.
An association between fish oils (DHA and EPA) and high-grade prostate cancer favoured lower levels of fish oil (OR 1.38, 95% CI 1.05 to 1.82; I²=18%) but there was no significant association for any individual fatty acid or for all fatty acids combined.
Subgroup analyses showed no obvious differences between case-control and nested case-control studies. There was no evidence of publication bias.