Co-exposure to cadmium, cobalt, lead and other heavy metals occurs in many occupational settings, such as pigment and batteries production, galvanization and recycling of electric tools. However, little is known about interactions between several heavy metals. In the present study we determined DNA single strand break (DNA-SSB) induction and repair capacity for 8-oxoguanine in mononuclear blood cells of 78 individuals co-exposed to cadmium (range of concentrations in air: 0.05-138.00 micro g/m(3)), cobalt (range: 0-10 micro g/m(3)) and lead (range: 0-125 micro g/m(3)). Exposure to heavy metals was determined in air, blood and urine. Non-parametric correlation analysis showed a correlation between cadmium concentrations in air with DNA-SSB (P = 0.001, R = 0.371). Surprisingly, cobalt air concentrations correlated even better (P < 0.001, R = 0.401), whereas lead did not correlate with DNA-SSB. Logistic regression analysis including 11 possible parameters of influence resulted in a model showing that cobalt in air, cadmium in air, cadmium in blood and lead in blood influence the level of DNA-SSB. The positive result with cobalt was surprising, since exposure levels were much lower compared with the TRK-value of 100 micro g/m(3). To examine, whether the positive result with cobalt is stable, we applied several logistic regression models with two blocks, where all factors except cobalt were considered preferentially. All strategies resulted in the model described above. Logistic regression analysis considering also all possible interactions between the relevant parameters of influence finally resulted in the following model: Odds ratio = 1.286(Co in air) x 1.040(Cd in air) x 3.111(Cd in blood) x 0.861(Pb in air) x 1.023(Co in air x Pb in air). This model correctly predicts an increased level of DNA-SSB in 91% of the subjects in our study. One conclusion from this model is the existence of more than multiplicative effects for co-exposures of cadmium, cobalt and lead. For instance increasing lead air concentrations from 1.6 to 50 micro g/m(3) in the presence of constant exposures to cobalt and cadmium (8 micro g/m(3) and 3.8 micro g/m(3)) leads to an almost 5-fold increase in the odds ratio, although lead alone does not increase DNA-SSB. The mechanism behind these interactions might be repair inhibition of oxidative DNA damage, since a decrease in repair capacity will increase susceptibility to reactive oxygen species generated by cadmium or cobalt. Indeed, repair of 8-oxoguanine decreased with increasing exposures and inversely correlated with the level of DNA-SSB (P = 0.001, R = -0.427). Protein expression patterns of individuals exposed to cobalt concentrations of approximately 10 micro g/m(3) were compared with those of unexposed individuals using two-dimensional gel electrophoresis. Qualitative and apparent quantitative alterations in protein expression were selective and certainly occurred in <0.1% of all proteins. In conclusion, the hazard due to cobalt exposure - that has been classified only as IIB by the IARC - seems to be underestimated, especially when individuals are co-exposed to cadmium or lead. Co-exposure may cause genotoxic effects, even if the concentrations of individual heavy metals do not exceed TRK-values.