Lung cancer risk assessment associated with exposure to hexavalent chromium: Results of pooled analysis of three cohorts.

Abstract

Quantitative risk assessments of inhaled hexavalent chromium (CrVI) have primarily relied on studies of two chromate production worker cohorts, involving male workers exposed to high concentrations of airborne CrVI that caused an exposure-dependent increase in lung cancer and elicited severe respiratory irritation. An additional larger cohort of CrVI-exposed aerospace workers, including women, and with lower intensity exposures, was recently updated with a longer follow-up period and reconstructed cumulative exposure estimates. A risk assessment was conducted using a pooled analysis of individual-level dose-response information from three cohorts, including the aerospace workers, to generate lung cancer inhalation unit risk estimates (IURs). IURs are estimates of the upper-bound of lifetime increased cancer risk from continuous inhalation exposure to 1 µg/m³ and are commonly used in quantitative risk assessment. Use of the pooled cohort broadens the range for dose-response analysis due to the comparatively low exposures to CrVI in the aerospace cohort; moreover, sex-specific risks could be expanded with the inclusion of female workers. Standardized logistic regression using person-year-specific weighted cumulative exposures was used to estimate the probability of lung cancer mortality, with varied weighting approaches. The primary model included male workers from all cohorts with a minimum of one year of CrVI exposure; secondary analyses examined the impacts of sex and short-term exposure. The best fit dose-response function was a Michaelis-Menten relationship. The IUR from the primary model was 0.01 (95%CI: 0.006-0.01) per µg/m³. Best estimates of the three IURs from secondary analyses, which removed the respiratory use adjustment factors, included females from the aerospace cohort and added workers with less than one year of exposure from the chromate production cohorts, were relatively consistent with the primary analyses, ranging from 0.008 to 0.03 per µg/m³. Use of this pooled IUR estimate allows for inclusion of a broader dose-response range and reduced uncertainty through use of a larger population, additional years of follow-up, and incorporation of broader geographical and cohort characteristics with improved relevancy to the general population.