Regression analysis of trends in mortality from hepatocellular carcinoma in Japan, 1972-2001

Abstract

Background: In Japan, higher mortality rates from hepatocellular carcinoma (HCC) have been observed than in any other industrialized country and the reported numbers of deaths due to HCC have increased significantly in recent years. We assessed whether there was a real increase in mortality from HCC and which factors contributed to the recent increasing trends of the number of deaths from HCC.

Methods: Poisson regression models were used to analyse panel data for the period 1972-2001.

Results: In both sexes, age-standardized mortality rates from HCC increased significantly over the past three decades. Among males the annual percentage increase in age-adjusted mortality rates was approximately 1-3% during the period 1972-96. On the other hand, female mortality trends were less clear. The estimated birth cohort effects suggested that there was a peak in the mortality risk among the cohorts born during the period 1927-36, which corresponded to those aged between 50-64 years during the period 1987-96. This was the period when a large increase of both age-specific mortality rates among older age groups and age-standardized mortality rates were observed particularly among males.

Conclusions: This study shows that an upward trend of mortality from HCC in Japan was primarily due to the effects among birth cohorts born around 1930, which are consistent with the pattern of past exposure to hepatitis C virus. The cohort effects have contributed to a large increase in mortality from HCC in recent years and the number of deaths from HCC is expected to increase by 2010-15.