Temporal changes in serum S-100 protein levels were compared between patients with ischemic stroke, transient ischemic attack (TIA) and traumatic brain injury (TBI). In addition, S-100 levels were correlated with clinical severity and outcome. Measurements were done with a LIA-mat Sangtec(R) 100 using an automated immunoluminometric assay. Serum S-100 was measured in 21 stroke patients, 18 TIA patients and ten TBI patients on days 1 (0-24 h), 2, 3, 4, 5 or 6 and 8 or 9. In a control group of 28 healthy volunteers one measurement was done. For the stroke and TIA patients, National Institutes of Health Stroke Scale (NIHSS) scores were obtained on admission and on day 10. For the TBI patients, Glasgow Coma Scale (GCS) scores were obtained on admission and Glasgow Outcome Scale (GOS) scores were obtained after 6 months:. Changes in serum S-100 levels over the first 3 days were significantly different between stroke and TBI patients (P = 0.014) and between stroke and TIA patients (P = 0.006). Peak concentrations of S-100 were most often observed on day 3 or 4 after stroke and on day 1 or 2 after TBI. In the stroke patients individual S-100 peak levels correlated well with the NIHSS score on admission (r = 0.58 P = 0.014) and the change in NIHSS score between day 10 and day 1 (r = 0.65, P = 0.005). In the TBI patients a good correlation between individual peak levels of S-100 and the GCS score on admission (r = -0.81, P = 0.010) and the GOS score 6 months after the trauma was found (r = -0.87, P = 0.004). We conclude that there is a significant difference in temporal changes of S-100 levels between ischemic stroke and TBI patients. This suggests different pathophysiological mechanisms. The results of this study further confirm that peak levels of serum S-100 correlate with neurological deficit resulting from either stroke or TBI. (C) 2000 Elsevier Science B.V. All rights reserved.