Purpose: Although high consumption of fish may be one of the contributing factors for Japanese longevity, no epidemiological study using Japanese data has tested this hypothesis.
Subjects and methods: The relationship between fish consumption and all-cause as well as cause-specific mortality was analyzed using the database of NIPPON DATA80. At baseline in 1980, history, physical, and blood biochemical measurement and a nutritional survey by the food-frequency method were performed in randomly selected community-based subjects aged 30 years and over in Japan. After exclusion of subjects with significant comorbidities at baseline, we followed 3945 men and 4934 women for 19 years. Men and women were analyzed comprehensively. Age- and sex-adjusted and multivariate adjusted relative risk for all-cause or cause-specific mortality was calculated using a Cox proportional hazards model with delayed entry.
Results: During 19 years of followup, there were 1745 deaths. Subjects were divided into 5 groups according to fish consumption frequency. The multivariate Cox analyses showed that relative risks for subjects who ate fish more than twice daily compared with those of subjects who ate 1 to 2 times weekly were 0.99 (95% confidence intervals: 0.77-1.27) for all-cause, 1.26 (0.70-2.29) for stroke, 0.92 (0.20-4.23) for cerebral hemorrhage, 1.09 (0.48-2.43) for cerebral infarction, and 0.91 (0.35-2.35) for coronary heart disease mortality.
Conclusion: Our results did not provide evidence in support of the fish hypothesis, perhaps because the majority of the Japanese subjects in the study ate fish more than the threshold level shown to be beneficial in the previous studies.