Association between Occupational Exposure to Textile Fibre Dusts and Lung Cancer in a Population-Based Case-Control Study in Montreal: A Preliminary Analysis Comparing Results from Three Analytical Methods

摘要

Objectives: To compare results estimated using two causal inference methods, inverse probability of exposure weighting (IPEW) and G-computation, to that estimated using the conventional multivariable logistic regression, on the association between occupational exposure to textile fibre dusts and lung cancer risk. Methods: A population-based case-control study on lung cancer was conducted from 1996 to 2001 in Montreal, Canada. Cases were individuals diagnosed with incident lung cancer and population controls were randomly selected from electoral lists and frequency-matched to cases by age, sex and electoral district. Questionnaires on lifetime occupational history, smoking and demographic characteristics were collected during in-person interview. Experts reviewed subjects' work history and assessed exposure to 294 agents, including textile fibre dusts. Odds ratios (OR) and their 95% confidence intervals (CIs) for the association between ever exposure to textile fibre dusts and lung cancer risk were estimated using three different methods: 1) IPEW, 2) G-computation, and 3) conventional multivariable regression. Results: The two causal inference methods produced higher point estimates (ORIPEW=1.17, 95% Cl=0.86-1.57; ORG-computation=l.ll, 95% Cl=0.80-1.49) compared to that estimated using the conventional multivariable logistic regression (OR=0.87, 95% Cl=0.68-l.ll). However, all three sets of OR results were close to the null value. Conclusion: The different methods provided rather similar results, albeit not identical. They are compatible with a null association between occupational exposure to textile fibre dusts and lung cancer.