Previous studies reported controversial results concerning alterations of astrocytes in schizophrenia. Because S100B may be regarded as a marker for astrocytes, the objective of this study was to examine S100B serum concentrations in 30 patients with schizophrenia with a monoclonal two-site immunoluminometric assay that specifically detects S100B. An ANOVA revealed medication ( p < 0.005) and deficit vs. nondeficit syndrome ( p < 0.05) as factors that influenced S100B significantly. S100B was higher in schizophrenic patients treated with antipsychotic drugs for approximately 3 weeks (241.1 F 152.5 ng/l) in comparison with unmedicated patients (111.4 F 31.8 ng/l, p < 0.005), and healthy age-matched controls (112.8 F 53.4 ng/l, p < 0.001; Bonferroni corrected two-tailed Student's t-test). There was no difference of S100B between unmedicated patients and controls ( p>0.05). Patients with deficit (250.6 F 154.9 ng/l) had higher S100B levels than patients with nondeficit schizophrenia (146.7 F 107.2 ng/l, p < 0.05) or controls ( p < 0.005). S100B was positively correlated with the subscore 'thought disturbance' of the Brief Psychiatric Rating Scale ( p < 0.05). In summary, increased serum levels of S100B may indicate alterations of astrocytes during early treatment with antipsychotics and in deficit schizophrenia. Whether S100B is elevated due to injured astrocytes and a disrupted blood -brain barrier, or by active secretion of S100B by astrocytes, has to be clarified by further studies. D