OBJECTIVE The nervous system is the major target of the toxic effect of manganese (Mn) and its compounds in welding fumes. In humans, inhalation is the most frequent route of Mn access, therefore, the respiratory tract and lungs are usually involved in the process of translocation of inhaled noxius agent by blood to the brain. This study was performed to assess whether it is possible to use neurophysiological tests for the detection of early effects of exposure to low Mn concentrations. It is also known that irritating welding fumes affect distal bronchioles of nonciliated, epithelial Clara cells, which secret anti-inflammatory and immunossupresive Clara cell protein (CC16) into the respiratory tract. The examination of usefulness of CC16 as early pulmonary biomarker for neurophysiological abnormal results of welding fumes exposure was performed. METHODS The study group comprised 59 welders employed at different workposts in a shipyard, matched for age and smoking habits with the control group composed of 23 mechanicians and electricians not exposed to welding fumes. Subjective neurological symptops (CNS), visual evoked potentials (VEP) and electroencephalography (EEG) were examined in welders and the relationships between Mn concentrations in the air, blood and urine as well as between cummulative exposure index (CEI) (Mn mg/m3 x years of exposure) were investigated. Effects of exposure were expressed in the form of biomarkers of the body burden, and CC16 as early pulmonary biomarker in welding exposure was examined by sensitive latex-immunoassay. RESULTS Abnormal results of VEP and EEG and the lowest CC16 levels were found in the youngest welders exposed to welding fumes. Those changes were related to the highest Mn airborne levels (xg > 0.3 mg/m3) and high blood Mn concentrations (approximately 14.0 microg/dL). The highest values of correlation coefficients were found only in welders characterized by abnormal neurophysiological results, VEP (r = 0.83) and VEP and VEP+EEG (r = 0.82). The multiple linear regression analysis from all analyzed subgroups, indicated that those with only abnormal neurophysiological tests, VEP and EEG, showed the highest values of partial correlation. It also revealed partial correlation cofficiants between Mn in the air, CEI (Mn mg/m3 x years) and CC16, Mn-B and Mn-U in VEP and VEP+EEG groups. It was found that the highest partial correlations were between the magnitude of exposure--Mn mg/m3, CEI and Mn-B concentration (R2 = 0.72, R2 = 0.66) as well as between CC16 pulmonary biomarker effects and Mn-B concentration (R2 = 0.51). CONCLUSIONS The subclinical effects revealed in neurological endpoints and abnormal results of neurophysiological tests, VEP and EEG, confirmed that those sensitive tests could be used for the detection of early effect of exposure to low manganese concentration. Inhibition of Clara cell protein secretion in younger welders not adapted to the Mn environment suppresses anti-inflammatory effect in the respiratory tract and probably enhances the absorption and thus the incidence of subclinical neurotoxic symptoms related to airborne Mn and Mn-B levels.