Ann Work Expo Health. 2026 May 12;70(4):wxag036. doi: 10.1093/annweh/wxag036.
ABSTRACT
INTRODUCTION: The cohort of chrysotile cohort workers from Quebec was established in the 1960s. It remains one of the most influential investigations into health risk of chrysotile asbestos. The cohort principally relies on measurements of asbestos-containing dust via impingers (as million particles per cubic foot of inhaled air, mpcf) for exposure assessment, though counts of fibers (fibers per cubic centimeter of inhaled air, longer than 5 µm with aspect ratio greater than 3:1 f/cm3) are more toxicologically relevant.
OBJECTIVE: To develop an empirical model that predicts exposure to fibers as a function of dust levels and workplace in chrysotile asbestos mining operations in Quebec. The model is intended to be used in re-analysis of epidemiologic data that accounts for measurement error in exposure.
METHODS: We obtained a copy of 623 individual parallel measurements of dust and fibers collected in Quebec in the 1970s and their contextual information. We fitted mixed-effects linear models that predict fiber concentrations as a function of the counts of dust particles, with random effect of sampling site. To evaluate the general model performance, we conducted a 10-fold cross-validation.
RESULTS: Fiber concentrations ranged from 0.06 to 307 f/cm3. Dust counts ranged from 0.04 to 9.12 mpcf. The average fiber-to-dust ratio of (f/cm3)/mpcf) was 12 (SD 16, median 7), ranging from 0.07 to 227. We estimated a positive association between logarithms of fiber and dust counts, which is not materially affected by adjustment for workplaces. The model with the logarithms of dust levels per se explained 24% of between-site variance and 12% of within-site variance in the logarithms of fiber concentrations. The average cross-validated R2 was 68%.
DISCUSSION: We confirmed observations that the ratio of fibers to dust counts of chrysotile asbestos is not constant but depends on a variety of work characteristics, including the number of dust particles. Our results are consistent with similar analyses conducted by others. However, we could not access all side-by-side impinger and fiber measurements that existed based on published reports, and available measurements do not cover all the times and workplaces where members of the cohort of Quebec millers and miners were exposed to chrysotile. Our models of exposure to fibers have a general form that adheres to multiplicative Berkson-type error. We cannot rule out the dependence of this error on risk of health outcomes.
CONCLUSION: We conclude that it is possible to calibrate impinger counts of dust to the fiber concentrations in the air for chrysotile asbestos mined and processed in Quebec in the 1970s and quantify the associated uncertainty.
PMID:42166758 | DOI:10.1093/annweh/wxag036
