BACKGROUND The present study examined the relation between fish consumption and mortality from all causes, ischemic heart disease, and stroke. METHODS The fish consumption data in 1961-1963, 1979-1981, and 1989-1991 and mortality data, age-standardized to 45-74 years, mean of the latest available 3 years, mostly around 1992-1993, in 36 countries, were obtained from the Food and Agriculture Organization and the World Health Organization, respectively. RESULTS There exists an inverse univariate correlation between log fish consumption in 1961-1963, 1979-1981, and 1989-1991 and log all-cause (P < 0.01 to < 0.001) and ischemic heart disease (P < 0.05 to < 0.01) mortality in both sexes. An inverse univariate correlation between log fish consumption and log stroke mortality was found only for the period 1961-1963 in both sexes (P < 0.05). Log fish consumption was independently, significantly, and inversely associated with log all-cause (all P < 0.001), ischemic heart disease (P < 0.01 to < 0.001), and stroke (P < 0.05 to < 0.001) mortality in all three time periods in both sexes, after adjusting for confounding factors. These associations remained significant even after exclusion of Iceland and Japan, countries with the highest amount of fish consumption and the lowest all-cause mortality rate. CONCLUSIONS Fish consumption is associated with a reduced risk from all-cause, ischemic heart disease, and stroke mortality at the population level.