Associations among dust exposure, smoking habits, and demographic factors and longitudinal changes of lung function were assessed among male steel workers. Cohort descriptive data analysis was conducted in 541 steel workers who had performed spirometry at least twice between 1982 and 1991 (mean follow-up, 6.1 years). The annual change (slope) in FVC, FEV1, FEV1/FVC%, and in body weight was determined by simple linear regression. The Pearson correlation coefficient between weight change and spirometry changes was calculated. Comparisons were also done in 75 pairs of steel workers matched by age, height, initial FEV1, and smoking status, but whose FEV1 declines differed by > or = 60 mL/yr. The FEV1 and FVC declined an average of 44 and 50 mL/yr, respectively, for the cohort as a whole. The FEV1 and FVC declined 52 and 54 mL/yr for current smokers, 43 and 53 mL/yr for ex-smokers, and 36 and 43 mL/yr for nonsmokers, respectively. Increasing weight was highly correlated with accelerated decline in lung function (p<0.0001). In the matched pairs, mean slopes for FVC, FEV1, and FEV1/FVC ratio were -96 mL/yr, -95 mL/yr, and -0.40%/yr for the rapid decliners; and +5 mL/yr, +10 mL/yr, and +0.10%/yr for their partners (p<0.0001). Matched pair comparisons showed that the rapid decliners averaged a 4.313 kg weight gain, while their partners gained 1.044 kg during the follow-up period. The slope of weight gain was 0.708 kg/yr for rapid decliners and 0.191 kg/yr for comparison workers (p<0.0036). Weight gain, in addition to aging and cigarette smoking, was found to be associated with the longitudinal rate of decline in FVC, FEV1, and FEV1/FVC ratio.