Fish

In the umbrella review, each 100 g a day increase in fish intake was associated with a relative risk (the risk in one group relative to another) of 0.92 for all-cause mortality, with a 95% confidence interval (the range the true value probably lies in) of 0.87 to 0.97 [1]. The association with cardiovascular mortality (death from heart and blood vessel disease) was stronger, a relative risk of 0.75 per 100 g a day [1].

A meta-analysis (a study that pools the results of earlier studies) of twelve prospective studies with 672,389 participants and 57,641 deaths found a relative risk of 0.94 for the highest compared with the lowest fish intake [2]. The association was nonlinear (it did not change in a straight line with intake): compared with never eating fish, 60 g a day was associated with a relative risk of 0.88 [2]. A review that separated fish types found lower total mortality with fatty fish, a relative risk of 0.97, and no such association with lean fish [3].

All of these findings come from observational studies, which track the diets people already eat rather than assigning them. Fish eaters differ from others in many ways, including income, other foods and exercise, and statistical adjustment cannot fully remove these differences. The umbrella review rated the evidence for all-cause mortality as moderate quality. The stronger association with heart disease deaths is consistent with the role of dietary fat quality in cardiovascular disease, and fish is a feature of several healthy dietary patterns.