Breast cancer incidence and its possible spatial association with pesticide application in two counties of England

Background. This study examines the spatial distribution of breast cancer incidence in Lincolnshire and Leicestershire, and its association with the application of selected pesticides. Methods. The Black-White (BW) join-count statistic and Moran I coefficient were used to investigate the localized distribution of breast cancer. Linear regression techniques were applied to examine the association between the breast cancer incidence rate and pesticide application. Results. The results from the BW join-count test and the Moran I coefficient test showed no overall evidence of localized distribution, 'clusters', of breast cancer. The regression analyses showed no sign of spatial association between breast cancer and application of pesticides in the urban areas as expected. The findings, however, did reveal a spatial association between the breast cancer incidence rates and the application of three selected pesticides [Aldicarb, Atrazine and Lindane (thought to potentially mimic oestrogen)] in rural Leicestershire. No such association was seen in Lincolnshire. Conclusions. Pesticides vary significantly in their spatial application. Overall, consistent associations between breast cancer incidence rates and the selected pesticides were not found, although the ecological study design has limitations and these are discussed. This approach is able to rule out strong associations; assessment of smaller risk would require a large and expensive study in individuals. The results of this study, although derived in the UK, have significance to the debate concerning the possibility of environmental compounds that mimic oestrogen and their consequences for human health. (C) 2004 Published by Elsevier Ltd on behalf of The Royal Institute of Public Health.