Thyroid cancer incidence in women and proximity to industrial air pollution sources: a spatial analysis in a middle size city in Colombia

Abstract

Manizales, a mid-size city in Colombia, hosts industries like metallurgy, electrical, chemical, and rubber and plastic industrial facilities that have released into atmosphere some pollutants postulated as thyroid cancer (TC) promoters, such as dioxins and furans, dichloromethane, lead and copper. In this article we aim to detect clusters of TC and analyze their spatial association with industrial pollution. TC cases (2003-2010) were obtained from Manizales' Population-based Cancer Registry (PCR-Mz). Atmospheric emissions from industries were obtained from official reports of environmental authority. Data was spatially aggregated into census tracts and analyzed with Bayesian Besag-York and Mollie (BYM) models. Three exposure approaches were used: i) presence or absence of industries into census tracts, ii) sum of air discharges, and iii) an exposure index (EI) that considered the distance and orientation of the census tract regard to industries, average wind direction and speed, and population mobility. Models were fitted by exposure definition and sex, and included traffic and socioeconomic variables for adjustment. Using the Kulldorff's spatial exploration statistic we also performed point-data analyses in order to detect and localize clusters with individual data. Ecological regression models showed that, for women, smoothed standardized incidence ratio (sSIR) increase in 15% [95% credibility interval: 3-27%] and 63% [95% CI: 18-125%] per one standard deviation increase in EI for dichloromethane and PCDDs/Fs, respectively. Point-data analysis confirmed a cluster of female cases close to an industry emitting chlorinated solvents. These results suggest that dichloromethane and PCDDs/Fs emitted from industrial sources might be suspected as thyroid cancer risk factors.