Estimation of respirable dust exposure among coal miners in South Africa.

The use of retrospective occupational hygiene data for epidemiologic studies is useful in determining exposure-outcome relationships, but the potential for exposure misclassification is high. Although dust sampling in the South African coal industry has been a legal requirement for several decades, these historical data are not readily adequate for estimating past exposures. This study describes the respirable coal mine dust levels in three South African coal mines over time. Each of the participating mining operations had well-documented dust sampling information that was used to describe historical trends in dust exposure. Investigator-collected personal dust samples were taken using standardized techniques from the face, backbye (underground jobs not at the coal face), and surface from 50 miners at each mine, repeated over three sampling cycles. Job histories and exposure information was obtained from a sample of 684 current miners and 188 ex-miners. Linear models were developed to estimate the exposure levels associated with work in each mine, exposure zone, and over time using a combination of operator-collected historical data and investigator-collected samples. The estimated levels were then combined with work history information to calculate cumulative exposure metrics for the miner cohort. The mean historical and investigator-collected respirable dust levels were within international norms and South African standards. Silica content of the dust samples was also below the 5% regulatory action level. Mean respirable dust concentrations at the face, based on investigator-collected samples, were 0.9 mg/m(3), 1.3 mg/m(3), and 1.9 mg/m(3) at Mines 1, 2, and 3, respectively. The operator-collected samples showed considerable variability across exposure zones, mines, and time, with the annual means at the face ranging from 0.4 mg/m(3) to 2.9 mg/m(3). Statistically significant findings were found between operator- and investigator-collected dust samples. Model-based arithmetic mean dust estimates at the face were 1.2 mg/m(3), 2.0 mg/m(3), and 0.9 mg/m(3) for Mines 1, 2, and 3, respectively. Using these levels, the mean cumulative exposure for the cohort was 56.8 mg-years/m(3). Current miners had a mean cumulative exposure of 66.5 mg-years/m(3), compared with ex-miners of 26.8 mg-years/m(3). Improvements in dust management or the use of different sampling equipment could account for the significant differences seen between operator- and investigator-collected data. Regression modeling for estimating mean dust levels over time using combined historical and investigator-collected data seems a reasonable method and useful in constructing models to describe cumulative exposures in a cohort of current and ex-miners.