BACKGROUND: Time trends of cancer of the gastric cardia differ between populations and the reasons are not fully understood. The object of this study was to investigate the occurrence of cancer of the gastric cardia in descriptive relation to age at death, calendar period, birth cohort and gender in the Netherlands between 1969 and 1994. METHODS: Data on the number of people with cancer of the gastric cardia as the underlying cause of death from 1969 to 1994 were obtained from annual publications by the National Causes of Death Registry of Statistics Netherlands. To estimate the separate effects of age, calendar period and birth cohort on the trend in mortality, a simultaneous analysis of these factors was performed using a log-linear Poisson model. RESULTS: In 1969, the mortality rates from cancer of the gastric cardia for males and females per 100,000 people were 2.1 and 1.1; in 1994 the mortality rates were 1.5 and 0.7, respectively. Examination of the time trend suggested that mortality for cancer of the gastric cardia may reflect a period phenomenon, although a cohort effect may have also contributed to the observed time trend. Furthermore, more males than females died from cancer of the gastric cardia. The difference was most striking in the younger age categories. CONCLUSION: In this Dutch population, the age-period-cohort-gender analysis indicated that the mortality rates decreased after the period 1975-1979 which might be explained by a decrease in exposure to risk factor(s) or an increase in exposure to protective factor(s).
BACKGROUND: Time trends of cancer of the gastric cardia differ between populations and the reasons are not fully understood. The object of this study was to investigate the occurrence of cancer of the gastric cardia in descriptive relation to age at death, calendar period, birth cohort and gender in the Netherlands between 1969 and 1994. METHODS: Data on the number of people with cancer of the gastric cardia as the underlying cause of death from 1969 to 1994 were obtained from annual publications by the National Causes of Death Registry of Statistics Netherlands. To estimate the separate effects of age, calendar period and birth cohort on the trend in mortality, a simultaneous analysis of these factors was performed using a log-linear Poisson model. RESULTS: In 1969, the mortality rates from cancer of the gastric cardia for males and females per 100,000 people were 2.1 and 1.1; in 1994 the mortality rates were 1.5 and 0.7, respectively. Examination of the time trend suggested that mortality for cancer of the gastric cardia may reflect a period phenomenon, although a cohort effect may have also contributed to the observed time trend. Furthermore, more males than females died from cancer of the gastric cardia. The difference was most striking in the younger age categories. CONCLUSION: In this Dutch population, the age-period-cohort-gender analysis indicated that the mortality rates decreased after the period 1975-1979 which might be explained by a decrease in exposure to risk factor(s) or an increase in exposure to protective factor(s).