The efficacy of cisternal drainage in association with early aneurysmal surgery remains highly controversial. The authors attempted to clarify the indications for this procedure and the proper timing of drainage removal in a series of 205 patients with no evidence of intracerebral hematoma who underwent surgical obliteration of bleeding aneurysms within 72 hours after subarachnoid hemorrhage. The 136 patients in whom cisternal drainage was performed constituted Group A and the remaining 69 patients Group B. The acid-base balance and lactate concentration were measured serially in cisternal cerebrospinal fluid (CSF) and arterial blood of 33 patients. Subarachnoid blood demonstrated by computed tomography (CT) was graded according to the system of Fisher et al. The outcome at 6 months did not differ significantly between Groups A and B among patients of preoperative CT grade 2 and clinical grades I-II. However, among patients of CT grade 3 or clinical grade III, those in Group A had better outcomes. The rates of symptomatic vasospasm in Groups A and B were 39.7% and 40.6%, respectively. However, persistent vasospasm was more frequent in Group B (11.8% vs. 26.1 %), particularly in CT grade 3 patients. Ventricular enlargement was more prevalent in Group A (33% vs. 17%). CSF pH higher than arterial pH after the 7th postoperative day was associated with a poor outcome and was fairly well correlated with a rise in CSF HCO, ₃^- and a fall in CSF PCO₂. CSF lactate increased with clinical deterioration but was not well correlated with preoperative CT findings, total volume of CSF outflow, or prognosis. These results suggest that cisternal drainage is indicated in cases of preoperative CT grade 3 or clinical grade III, and that the drainage tube should be removed as soon as possible if the CSF pH falls below the arterial pH and CSF lactate decreases toward the normal level.