Time trends in the incidence of cervical adenocarcinoma and adenosquamous cell carcinomas during the period 1973-1991 were examined using data provided by 60 populationbased cancer registries from 32 defined populations in 25 countries. Three components of the incidence trend were studied: age, calendar period of diagnosis and birth cohort. Cumulative incidence rates per 1,000 for 2 groups with age ranges 25-49 and 50-74 years were calculated from the model that best described the incidence data. There was a significant increase in the cumulative incidence of cervical adenocarcinomas in women born in the mid-1930s and in successive cohorts thereafter in some populations in the United States (whites and Hispanic women), Australia, New Zealand (non-Maori), England, Scotland, Denmark, Slovenia, Slovakia and Japan (Osaka) and among Chinese women in Singapore, with a general decline in the incidence in women born in earlier periods. In Sweden and Slovenia there is a suggestion of an increasing trend in both age groups. A decrease in incidence in both age groups was apparent in Finland, France and Italy. There were no changes in incidence in 24 registries covering other European, Asian and black populations in the United States. Part of the increase may be attributable to an increasing prevalence of human papillomavirus infection, and part to improvements in screening. Int. J. Cancer 75:536-545, 1998.1998 Wiley-Liss, Inc.In most Cancer Registries, close to three-fourths of cervical cancer cases are reported as squamous cell carcinomas. Adenocarcinoma and adenosquamous cell carcinomas represent 10-15% and cervical cancer with ''other or without'' specified histology represent the remaining 10-15%. During the last decade, several studies have reported increasing incidence rates of cervical adenocarcinomas, particularly for young women, in countries such as the United States (Peters et al., 1986;Schwartz and Weiss, 1986;Devesa et al., 1989;Zheng et al., 1996), England and Wales (Chilvers et al., 1987), Norway (Eide, 1987) and Sweden (Bergström et al., personal communication). The reasons for the increases are poorly understood. It is possible that some of the observed increase is due to trends in the proportion of cases histologically verified, changes in disease classification over time and increased detection of early asymptomatic lesions. However, the magnitude of the changes observed are generally too great to be entirely ascribed to such effects.The purpose of our work was to examine, on a worldwide basis, time trends in the incidence of cervical adenocarcinomas, especially in young women. The analyses include incidence data provided by 60 Cancer Registries for 32 defined populations in 25 countries over a median 18-year period (range, 10-30 years). The data permit comparison of time trends in different geographic areas as well as for different ethnic groups within a given country. The analyses of the secular trends in the incidence of squamous cell carcinomas will be presented in a separate report.
MATERIAL AND M...