As part of a large-scale epidemiological study, occupational isocyanate exposure was assessed in spray-painting environments. The aim was to assess which compounds contribute to isocyanate exposure in car body repair shops and industrial painting companies, and to identify tasks with high risk of isocyanate exposure. Mainly personal task-based samples (n = 566) were collected from 24 car body repair shops and five industrial painting companies using impingers with DBA in toluene. Samples were analysed by LC-MS for isocyanate monomers, oligomers and products of thermal degradation. From the 23 analysed compounds, 20 were detected. Exploratory factor analysis resulted in a HDI, TDI and MDI factor with the thermal degradation products divided over the TDI and MDI factors. The HDI factor mainly consisted of HDI oligomers and was dominant in frequency and exposure levels in both industries. Spray painting of PU lacquers resulted in the highest exposures for the HDI factor (<LOD-2643 microg/m(3) NCO), with no significant difference between the industries. Exposure variability during PU spray painting was large with a variability over time of (ww)S(2) = 9.1 compared with between-worker variability of (bw)S(2) = 1.6. Lower level exposure to the HDI factor was found during other painting-related tasks and even tasks without direct exposure to paint. Exposure to the TDI factor was found more regularly in car body repair shops than in industrial painting companies. Exposure levels were low (<LOD-5 microg/m(3) NCO) compared with the HDI factor and no clear contrast in levels between the tasks was observed. Exposure to the MDI factor was found incidentally during spraying and welding in car body repair shops (<LOD-0.5 microg/m(3) NCO). The results indicate that paint is the most important source and major contributor of isocyanate exposure in both industries with highest exposures during PU spraying. However, since respiratory protection is less extensively used during other tasks, lower level exposure during these other tasks may significantly contribute to the internal dose.