Incidence of digestive cancers and occupational exposure to asbestos

While the role of exposure to asbestos in the development of several cancers such as mesotheliomas and bronchopulmonary cancers is now well established, the possible relationship between digestive cancers, other than peritoneal mesotheliomas, and occupational exposure to asbestos is still controversial. The great majority of the studies are based on mortality data. The aim of the study was to analyse the relationship between digestive cancer incidence and occupational exposure to asbestos in a population of subjects for whom precise occupational exposure data and precise incidence data were available. The population consisted of salaried and retired workers from a company using asbestos to manufacture fireproof textiles and friction materials. There were 1454 men (79.9%) and 366 women (20.1%). A cumulative exposure index and a mean exposure concentration in fibres/ml for each subject were calculated with the aid of an in-house job-exposure matrix. The number of cases of digestive cancer observed was compared with the expected and Standardized Incidence Ratio (SIR) was estimated. Precise occupational exposure data allowed us to study the dose-response relationship between asbestos exposure and risk of digestive cancer using Cox model. Fifty-six digestive cancers occurred in the study population over the 18-year follow-up period for 48.4 expected (SIR = 1.16 [0.87-1.50]). Comparing with incidence in the county, SIR was not significant for any of the digestive localization, but for peritoneum. However, even after taking into account the potential confounders via the Cox model, there was a significant dose-response relationship between the occurrence of digestive cancers and the mean exposure concentration, even after exclusion of peritoneum cancers. Our study provides initial evidence suggesting a relationship between occupational exposure to asbestos and the risk of digestive cancer: first, it is a study of incidence although the risk evidenced is not significant; secondly, a dose-effect relationship is demonstrated in the whole population. However, these preliminary results require confirmation by more powerful studies focusing on larger series. (C) 2002 Lippincott Williams Wilkins.