Estimating impacts of reducing acrylonitrile exposure on lung cancer mortality in an occupational cohort with the parametric g-formula

Alexander Keil,Gregory Haber,Barry Graubard,Patricia A Stewart,Debra Silverman,Stella Koutros
DOI: https://doi.org/10.1136/oemed-2024-109532
IF: 4.9
2024-06-04
Occupational and Environmental Medicine
Abstract:Objectives To inform the potential human carcinogenicity of acrylonitrile, we estimate associations between acrylonitrile exposures and lung cancer mortality in US workers with the objectives of (1) assessing potential for healthy worker survivor bias and (2) adjusting for this bias while assessing the expected lung cancer mortality under different hypothetical occupational exposure limits on acrylonitrile exposure using the parametric g-formula. Methods We used data from a cohort of 25 460 workers at facilities making or using acrylonitrile in the USA. We estimated HRs to quantify associations between employment and lung cancer mortality, and exposure and leaving employment. Using the parametric g-formula, we estimated cumulative lung cancer mortality at hypothetical limits on acrylonitrile exposure. Results Recent and current employment was associated with lung cancer, and exposure was associated with leaving employment, indicating potential for healthy worker survivor bias. Relative to no intervention, reducing the historical exposure under limits of 2.0, 1.0 and 0.45 parts per million would have been expected to reduce lung cancer mortality by age 90 by 4.46 (95% CI 0.78 to 8.15), 5.03 (95% CI 0.96 to 9.11) and 6.45 (95% CI 2.35 to 10.58) deaths per 1000 workers, respectively. A larger lung cancer mortality reduction would be expected under elimination of exposure: 7.21 (95% CI 2.72 to 11.70) deaths per 1000 workers. Conclusions Healthy worker survivor bias likely led to underestimation of excess risk. Our results corroborate previous study findings of an excess hazard of lung cancer among the highest exposed workers.
public, environmental & occupational health
What problem does this paper attempt to address?