This work was undertaken to investigate the usefulness of diisocyanate-related protein adducts in blood samples as biomarkers of occupational exposure to toluene diisocyanate (TDI; 2,4- and 2,6-isomers) and 4,4'-methylenediphenyl diisocyanate (MDI). Quantification of adducts as toluene diamines (TDAs) and methylenedianiline (MDA) was performed on perfluoroacylated derivatives by gas chromatography-mass spectrometry (GC-MS/MS) in negative chemical ionisation mode. TDI-derived adducts were found in 77% of plasma and in 59% of globin samples from exposed workers manufacturing flexible polyurethane foam. The plasma levels ranged from 0.003 to 0.58 nmol mL(-1) and those in globin from 0.012 to 0.33 nmol g(-1). The 2,6-isomer amounted to about two-thirds of the sum concentration of TDA isomers. MDI-derived adducts were detected in 3.5% of plasma and in 7% of globin samples from exposed workers manufacturing rigid polyurethane foam. A good correlation was found between the sum of TDA isomers in urine and that in plasma. The relationship between globin adducts and urinary metabolites was ambiguous. Monitoring TDI-derived TDA in plasma thus appears to be an appropriate method for assessing occupational exposure. Contrary to TDI exposure, adducts in plasma or globin were not useful in assessing workers' exposure to MDI. An important outcome of the study was that no amine-related adducts were detected in globin samples from TDI- or MDI-exposed workers, alleviating concerns that TDI or MDI might pose a carcinogenic hazard. Further studies are nevertheless required to judge whether diisocyanates per se could be such a hazard.