BACKGROUND: Carbon dioxide is a potent cerebral vasodilator. The change of carbon dioxide partial pressure may influence the intracranial pressure and the patients' neurological outcome. There are few reports about the influence of end-tidal CO2 (ETCO2), arterial CO2 (PaCO2) and its pressure difference P(a-ET)CO2 during a craniotomy on the Glasgow coma scale (GCS) score for evaluation of neurological status. In this study, authors tried to discover the influence of PaCO2, PETCO2, and P(a-ET)CO2 on neurological outcome. METHODS: The data of PaCO2 and PETCO2 and P(a-ET)CO2 during a craniotomy was saved. The correlations between each parameter, the GCS score and rCoBF were analyzed. To prevent a direct effect on carbon dioxide tension, blood pressure and body temperature were maintained within a normal range. At the same time, we inserted a probe of the thermal diffusion flowmetry monitor in the subdural space to monitor the regional cortical cerebral blood flow (rCoBF). All the data was saved simultaneously, at the moment of dura closure. RESULTS: There was a fair correlation between the PaCO2 and PETCO2. A low PaCO2 level correlated well with a good GCS score but, not with PETCO2. The mean P(a-ET)CO2 value was 4.4 +/- 3.1 mmHg. The high P(a-ET)CO2 level correlated well with a poor GCS score. High rCoBF correlated well with a good GCS score. However, the changes of PaCO2 and PETCO2 showed no correlations with the rCoBF. CONCLUSIONS: As a result, if we decrease the PaCO2 level by hyperventilation and increase the rCoBF level through proper management during anesthesia, we can improve the patients' neurological outcome.