Objectives To describe mortality in workers exposed to chrysotile asbestos, and determine exposure–response relationships between asbestos exposure and mortality from lung cancer and asbestosis by fitting and comparing different models.

Methods A prospective cohort consisting of 586 workers in an asbestos textile factory was followed from 1 January 1972 to 31 December 2006. A structured questionnaire was administered to collect personal information and exposure data. Paired concentration samples measured in the workshops were used to convert dust concentrations to fibre concentrations. Individual cumulative asbestos exposure was estimated as the product of fibre concentrations and duration of employment in each job and expressed as fibre-years/ml. The vital status of cohort members was followed annually. Poisson regression analysis was applied to fit log-linear, log-quadratic, power, additive relative risk and categorical models to estimate exposure–response relationships between cumulative fibre exposure and mortality from lung cancer and asbestosis.

Results Of the 226 deaths (14.6 per 1000 person-years) over the 35-year follow-up, 51 were from lung cancer (3.29 per 1000 person-years) and 37 from asbestosis (2.39 per 1000 person-years). A significant exposure–response relationships with either lung cancer or asbestosis (p<0.001) was observed in the final model. The power model with lagged 10 years was found to be the best model of those evaluated for both lung cancer (β coefficient=0.53) and asbestosis (β coefficient=0.63).

Conclusions The study confirmed strong associations between exposure to chrysotile asbestos and lung cancer and asbestosis, in which clear exposure–response relationships were observed.