ObjectiveAsbestos causes mesothelioma (MM) and cancer of lung, ovary, larynx and possibly other organs. Risk reduction after cessation of exposure and at long latency is debated. For MM, the predicted increase of rates with time since first exposure (TSFE) might be attenuated by fibre clearance. The ban of asbestos use in Italy gives an opportunity to measure long term risk in formerly exposed workers.MethodsIn a pool of 43 Italian asbestos cohorts (asbestos-cement, rolling stock, shipbuilding and other industries), standardised mortality ratios (SMRs) were computed for the major causes, with consideration of time factors (period, duration, latency). The functional relation between pleural cancer mortality rates and TSFE was evaluated using Poisson regression, including a term for fibre clearance.ResultsThe study includes 54,409 subjects (6,054 women): 55.3% alive, 43.0% died (cause known for 94%; 789 pleural cancers), and 1.7% lost to follow-up. Mortality was increased for all causes (SMR = 1.06; p < 0.001), all malignancies (SMR = 1.18; p < 0.01), pleural and peritoneal malignancies (SMRs = 14.66 and 6.95; both p < 0.01), lung (SMR = 1.26; p < 0.01) and ovarian cancer (SMR = 1.39; p < 0.05). Pleural cancer rate increased during the first 40 years of TSFE and reached a plateau thereafter. The model including asbestos elimination fitted the data much better than the traditional model (p = 0.0004; estimated asbestos elimination rate of 5% per year).DiscussionOur results provide information on the risk after asbestos exposure, even for rarer diseases. Increased risk for ovarian cancer is confirmed. Results agree with the hypothesis that pleural cancer risk, rather than increasing indefinitely, reaches a plateau at long latency, coherently with asbestos clearance from the lungs. As a consequence, early exposures to asbestos cannot be considered to be the exclusive determinant of the long term development of pleural cancer, with implications also for the risk related to non occupational exposure and for asbestos removal.