Occupational Exposure Reconstruction for Tungsten Carbide Manufacturing Workers

Objectives: The objectives of the study were to generate exposure estimates for a cohort of hardmetal workers to be used in an occupational epidemiological mortality study, to determine the sensitivity of the estimates to correction factors and task differences, and to analyze the influence of qualitative factors on measured agent concentrations. Methods: Industrial hygiene measurements were obtained from 21 hardmetal plants located in five countries. Job-exposure matrices were constructed for cobalt, tungsten, and nickel for the period 1952 – 2014. The matrices consisted of 69 job classes based on job title and process and the exposure estimates were calculated from the available measurements. Correction factors were applied to total aerosol and inhalable fraction measurements and closed- and open-face cassette measurements to assess estimate sensitivity. One job class was broken into tasks and exposure estimates calculated to assess task differentiation sensitivity. Measurements were coded for 10 qualitative factors and analyzed by analysis of variance to assess potential impact on measured concentrations. Results: The exposure level estimates were similar to or lower than those previously reported for the 1952 – 2014 study period. The estimates were not sensitive to total aerosol/inhalable fraction corrections up to a factor of five and were not sensitive to task differentiation. Among the qualitative factors assessed, American Governmental Conference of Industrial Hygienists (ACGIH) threshold limit value (TLV) time periods exerted the most influence on measured concentrations. Conclusions: The matrices constructed provide the necessary job and exposure differentiation required for the occupational epidemiological analyses. The insensitivity of the estimates to rather large correction factors suggests that incorporation of such modifiers would not have had a marked impact on the estimates obtained. The difficulty of including qualitative factors in exposure reconstructions was not resolved here. However, the analyses provided insight into some possible effects of these factors on measured concentration that may be useful for other studies in terms of evaluating the appropriateness and usefulness of including such factors.