Thirty-six car painters and 115 control persons participated in a follow-up investigation 6 years after the initial study, including measurement of lung function (dynamic spirometry and nitrogen washout test) and estimation of exposure to diisocyanates based on individual working routines, use of respiratory protection equipment, and measurements. The mean exposure for the car painters was 0.0015 mg/m3 hexamethylenediisocyanate (HDI) and 0.09 mg/m3 hexamethylenediisocyanate-biurettrimer (HDI-BT), but frequently there were peak exposures exceeding 2.0 mg/m3 HDI-BT for at least 30 sec. Compared with smoking controls, the smoking car painters had greater yearly reduction in FVC (95 versus 38 ml), FEV1 (61 versus 28 ml), and VC (77 versus 30 ml). The nonsmoking car painters showed no differences in lung volumes compared with their nonsmoking controls. The impairment correlated well with the frequency of high peak exposures to HDI-BT, but not with the mean exposure to diisocyanates.