European Journal of Epidemiology. 2008 Apr 26 [Epub ahead of print] [Link]

Gasparrini A, Pizzo AM, Gorini G, Seniori Costantini A, Silvestri S, Ciapini C, Innocenti A, Berry G.

Public and Environmental Health Research Unit (PEHRU), London School of Hygiene & Tropical Medicine, Keppel Street, London, WC1E 7HT, UK,


Background: Several papers have reported state-wide projections of mesothelioma deaths, but few have computed these predictions in selected exposed groups. Objective To predict the future deaths attributable to asbestos in a cohort of railway rolling stock workers.

Methods: The future mortality of the 1,146 living workers has been computed in term of individual probability of dying for three different risks: baseline mortality, lung cancer excess, mesothelioma mortality. Lung cancer mortality attributable to asbestos was calculated assuming the excess risk as stable or with a decrease after a period of time since first exposure. Mesothelioma mortality was based on cumulative exposure and time since first exposure, with the inclusion of a term for clearance of asbestos fibres from the lung.

Results: The most likely range of the number of deaths attributable to asbestos in the period 2005–2050 was 15–30 for excess of lung cancer, and 23–35 for mesothelioma.

Conclusion: This study provides predictions of asbestos-related mortality even in a selected cohort of exposed subjects, using previous knowledge about exposure-response relationship. The inclusion of individual information in the projection model helps reduce misclassification and improves the results. The method could be extended in other selected cohorts.

Keywords: Asbestos – Predictions – Mesothelioma – Lung cancer