Breast cancer age at diagnosis patterns in four Latin American Populations: A comparison with North American countries

Abstract

BackgroundIn the Latin America countries (LAC), one in five breast cancer (BC) cases occur in women younger than 45 years, almost twice the frequency seen in developed countries. Most BC cases in younger women are premenopausal and are generally more difficult to detect at early stages and to treat than postmenopausal cancers. We employ data from four high quality population-based registries located in LAC and assess the extent to which the higher frequency of BC occurring in younger women is due to a younger population structure, compared to that of developed countries. Next, we analyze secular and generational trends of incidence rates in search for additional explanations. MethodsUsing data from the International Agency for Research on cancer, between 1988 and 2007, the age distribution of BC incident cases for registries located in Brazil, Colombia, Costa Rica, Ecuador is compared to that of USA and Canadian registries, both before and after removing differences in population age structure. An age-period-cohort modelling of incidence rates is also conducted in all compared registries to identify secular and generational effects. ResultsBC incident cases in the LAC registries present, on average, at an earlier age than in the USA and Canadian registries and for 2003–2007, between 20 and 27% of cases occur in women aged 20–44. About two thirds of the difference in age distribution between LAC and USA registries is attributable to the younger age distribution in the LAC base populations. The USA registries show the highest age-specific BC incidence rates of all compared aggregated registries, at all ages. However, in all the LAC registries incidence rates are rapidly increasing, fueled by a strong birth cohort effect. This cohort effect may be explained by important reduction in fertility rates occurring during the second half of the 20th century, but also by a greater exposure to other risk factors for BC related to the adoption of life styles more prevalent in developed countries. ConclusionThe younger age at presentation of BC incident cases seen in the analyzed LAC registries, and possibly in many Latin American countries, is not only attributable to their relatively young population age structure but also to the low incidence rates in older women. As more recently born cohorts, with greater exposure to risk factors for postmenopausal BC, reach older age, incidence rates will be more similar to the rates seen in the USA and Canadian registries. There is a need for additional research to identify determinants of the higher BC rate among younger women in these countries.