Occupational exposures, in addition to smoking, may contribute to development and progression of COPD. SUBJECTS: Data are for 5734 subjects in the Lung Health Study (LHS), a study of smoking cessation interventions for persons with early COPD (defined as FEV1/FVC ratio < 0.70 and FEV1 55% - 90% predicted). Each had baseline and up to 5 annual follow-up spirometry tests. METHODS: FEV1% predicted post-bronchodilator at each follow-up year was modeled as determined by both baseline characteristics and time-dependent variables using a repeated-measures general linear model with first-order autoregression. Baseline characteristics included age, airway responsiveness by methacholine challenge, baseline FEVI, and self-reported dust exposure, fume exposure, and mask use. Time-dependent variables' included smoking (smoking status, cigarettes/day, or smoking intensity as a categorical variable) and self-reported dust exposure, fume exposure, and mask use. RESULTS: Exposure to "fume" (both as a baseline characteristic and as a time-dependent characteristic) showed relationships to FEV1 in many models. Significant results were found in men, but not in women. Subjects reporting fume exposure at baseline had 0.44% lower FEV1% predicted than those not reporting (p < 0.06), and fume exposure at any follow-up year was associated with a 0.21% lower FEV1% predicted (p < 0.05) when adjusted for age, baseline FEV1% predicted, smoking status, and baseline cigarettes! day. No significant effects of dust exposure or mask use were observed. CONCLUSIONS: There is an indication that self-reported fume exposure adversely affects FEV1 in persons with early COPD. The effect may be additive to that of smoking and be present even if smoking ceases.