S-100b is thought to be a screening marker of hypoxic brain damage in patients with cardiac arrest. However, the time-dependent occurrence and relevance of increased S-100b serum levels in out-of-hospital patients with cardiopulmonary resuscitation (CPR) is still discussed. The purpose of our study was to evaluate the diagnostic utility of S-100b measurements in comparison to that of adhesion molecules sE-selectin and sP-selectin in patients with CPR. Sixteen out-of-hospital patients (median age 69.6 years; range 59.2-82.2 years) suffering from cardiac arrest due to ventricular fibrillation, asystole, or electromechanical dissociation were recruited prospectively. Blood samples were drawn on scene after the return of spontaneous circulation (ROSC) and 12 hours after successful CPR. The reference group consisted of 10 patients with isolated severe head trauma (SHT) (Glasgow Coma Score </ or =8), and the control group comprised 20 healthy volunteers. Serum concentrations of S-100b, determined by immunoluminometric assay, were compared with serum levels of sE-selectin and sP-selectin measured by an enzyme-linked immunosorbent assay and correlated with the patients' survival. In the CPR group, S-100b serum levels (2.37 ng/ml; 1.37-4.09 ng/ml) at study entry (11.6 minutes after arriving on scene) did not significantly differ from those of SHT patients (2.88 ng/ml; 1.78-8.81 ng/ml). Both groups showed significant differences from the healthy controls (0.04 ng/ml; 0.01-0.82 ng/ml). At 12 hours after CPR the serum levels had decreased to 0.41 ng/ml (0.24-0.51 ng/ml) but continued to be significantly elevated compared to that of the control group. sE-selectin values in serum increased from 56.00 ng/ml (38.50-85.50 ng/ml) on scene to 79.00 ng/ml (52.00-127.00 ng/ml) after 12 hours (p < 0.05). The first measurements differed significantly from serum levels of the control group (22.50 ng/ml; 14.00-34.00 ng/ml) and from those of the SHT group (45.00 ng/ml; 39.00-63.75 ng/ml). At 12 hours after study entry the sE-selectin values were not significantly different from those of the SHT group (51.50 ng/ml; 39.00-95.88 ng/ml). sP-selectin serum levels increased slightly from 199.50 ng/ml (184.25-227.25 ng/ml) to 247.00 ng/ml (206.50-354.75 ng/ml). First and second measurements did not reveal any significant differences in either the SHT group or the healthy controls. When correlated with survival, S-100b measurements exhibited constantly high serum levels for patients, decreasing within the first 24 hours, whereas they decreased significantly in patients with longer survival. sP-selectin values on scene slightly increased in cases of survivals less than 24 hours after CPR. sE-selectin serum levels always remained within normal levels and revealed no significance later on. In contrast to the endothelium-derived adhesion molecules sE-selectin and sP-selectin, comparison of measurements of specific neuroprotein S-100b early after cardiac arrest and 12 hours later seem to provide an indication of the severity of hypoxic brain damage and the prognosis after CPR. Further investigations are required to better understand the CPR-related mechanisms of blood-brain barrier damage.