Dust-related dose-response decrements in lung function among coal miners have been reported in several studies, with varying magnitudes across populations. Few studies have compared differences between current and former coal miners. No studies on dose response relationships with lung function have been conducted in South African coal mines, one of the top three producers of coal internationally. The objectives of this study were (1) to describe the relationship between respirable dust exposure and lung function among current and former South African coal miners and to determine whether differential dust related effects were present between these employment categories; (2) to examine dust related dose response relationships, controlling for potential confounding by smoking and a history of tuberculosis (TB). Six hundred and eighty-four current and 188 ex-miners from three bituminous coal mines in Mpumalanga Province were studied. Interviews assessing work histories, smoking profiles and other risk factors were conducted. Work histories were also obtained from company records. Standardised spirometry was performed by trained technicians. Cumulative respirable dust exposure (CDE) estimates were constructed from company-collected sampling and measurements conducted by the researchers. Regression models examined the associations of CDE with per cent predicted FEV(1) and FVC, controlling for smoking, past history of TB and employment status. A statistically significant decline in FEV(1) of 1.1 and 2.2 ml/mg-year/m(3) was found in representative 40-year-old, 1.7-m tall current and former miners, respectively. Significant differences were found between the highest and medium exposure categories. Ex-miners had a lower mean per cent predicted lung function than current miners for each cumulative exposure category, suggesting a "healthy worker" effect. Past history of TB contributed to 21 and 14% declines in per cent predicted FEV(1) and FVC, respectively. Thus, in this cohort, a dose-related decline in lung function was associated with respirable dust exposure, with a magnitude of effect similar to that seen in other studies and important differences between current and former employees. A "healthy worker" effect may have attenuated the magnitude of this relationship. TB was a significant contributor to lung function loss.