Trends of stomach cancer mortality in Eastern Asia in 1950-2004: comparative study of Japan, Hong Kong and Singapore using age, period and cohort analysis.

To characterize the temporal trends of stomach cancer mortality in Eastern Asia and to better interpret the causes of the trends, we performed age, period and cohort analysis (APC analysis) on the mortality rates in Japan, Hong Kong and Singapore during 1950-2004, as well as the rates in the US as a control population. For the APC analysis, Holford's approach was used to avoid the identification problem. Age-standardized mortality rates (ASMR) decreased consistently in all four areas during the observation period in both males and females. Japan had the highest ASMR in both sexes, followed by Singapore, Hong Kong and the US, but the differences in ASMR among the four areas diminished with time. The results of APC analysis suggested that the decreasing mortality rates in Eastern Asia were caused by the combination of decreasing cohort effect since the end of the 1800s and decreasing period effect from the 1950s. The US showed similar results, but its decreases in the period and cohort effect preceded those of Eastern Asia. Possible causes for the decrease in the cohort effect include improvement in the socioeconomic conditions during childhood and a decrease in the prevalence of H. pylori infection, while possible causes for the decrease in the period effect include a decrease in dietary salt intake and improvements in cancer detection and treatment. These findings may help us to predict future changes in the mortality rates of stomach cancer.