Dolores Catelan1, Dario Consonni2, Annibale Biggeri3, Barbara Dallari2, Angela C Pesatori4, Luciano Riboldi2, Carolina Mensi2. 1. Department of Statistics Computer Science Applications "G. Parenti", University of Florence, Florence, Italy. Electronic address: dolores.catelan@unifi.it. 2. Occupational Health Unit, Fondazione IRCCS Ca' Granda-Ospedale Maggiore Policlinico, Milan, Italy. 3. Department of Statistics Computer Science Applications "G. Parenti", University of Florence, Florence, Italy. 4. Occupational Health Unit, Fondazione IRCCS Ca' Granda-Ospedale Maggiore Policlinico, Milan, Italy; Department of Clinical and Community Sciences, University of Milan, Milan, Italy.
Abstract
INTRODUCTION: Measuring and mapping the occurrence of malignant mesothelioma (MM) is a useful means to monitor the impact of past asbestos exposure and possibly identify previously unknown sources of asbestos exposure. OBJECTIVE: Our goal is to decompose the observed spatial pattern of incidence of MM in the Lombardy region (Italy) in gender-specific components linked to occupational exposure and a shared component linked to environmental exposure. MATERIALS AND METHODS: We selected from the Lombardy Region Mesothelioma Registry (RML) all incident cases of MM (pleura, peritoneum, pericardium, and tunica vaginalis testis) with first diagnosis in the period 2000-2016. We mapped at municipality level crude incidence rates and smoothed rates using the Besag York and Mollié model separately for men and women. We then decomposed the spatial pattern of MM in gender-specific occupational components and a shared environmental component using a multivariate hierarchical Bayesian model. RESULTS: We globally analyzed 6226 MM cases, 4048 (2897 classified as occupational asbestos exposure at interview) in men and 2178 (780 classified as occupational asbestos exposure at interview) in women. The geographical analysis showed a strong spatial pattern in the distribution of incidence rates in both genders. The multivariate hierarchical Bayesian model decomposed the spatial pattern in occupational and environmental components and consistently identified some known occupational and environmental hot spots. Other areas at high risk for MM occurrence were highlighted, contributing to better characterize environmental exposures from industrial sources and suggesting a role of natural sources in the Alpine region. CONCLUSION: The spatial pattern highlights areas at higher risk which are characterized by the presence of industrial sources - asbestos-cement, metallurgic, engineering, textile industries - and of natural sources in the Alpine region. The multivariate hierarchical Bayesian model was able to disentangle the geographical distribution of MM cases in two components interpreted as occupational and environmental.
INTRODUCTION: Measuring and mapping the occurrence of malignant mesothelioma (MM) is a useful means to monitor the impact of past asbestos exposure and possibly identify previously unknown sources of asbestos exposure. OBJECTIVE: Our goal is to decompose the observed spatial pattern of incidence of MM in the Lombardy region (Italy) in gender-specific components linked to occupational exposure and a shared component linked to environmental exposure. MATERIALS AND METHODS: We selected from the Lombardy Region Mesothelioma Registry (RML) all incident cases of MM (pleura, peritoneum, pericardium, and tunica vaginalis testis) with first diagnosis in the period 2000-2016. We mapped at municipality level crude incidence rates and smoothed rates using the Besag York and Mollié model separately for men and women. We then decomposed the spatial pattern of MM in gender-specific occupational components and a shared environmental component using a multivariate hierarchical Bayesian model. RESULTS: We globally analyzed 6226 MM cases, 4048 (2897 classified as occupational asbestos exposure at interview) in men and 2178 (780 classified as occupational asbestos exposure at interview) in women. The geographical analysis showed a strong spatial pattern in the distribution of incidence rates in both genders. The multivariate hierarchical Bayesian model decomposed the spatial pattern in occupational and environmental components and consistently identified some known occupational and environmental hot spots. Other areas at high risk for MM occurrence were highlighted, contributing to better characterize environmental exposures from industrial sources and suggesting a role of natural sources in the Alpine region. CONCLUSION: The spatial pattern highlights areas at higher risk which are characterized by the presence of industrial sources - asbestos-cement, metallurgic, engineering, textile industries - and of natural sources in the Alpine region. The multivariate hierarchical Bayesian model was able to disentangle the geographical distribution of MM cases in two components interpreted as occupational and environmental.