A controlled randomized study of endoscopic evacuation versus medical treatment was performed in 100 patients with spontaneous supratentorial intracerebral (subcortical, putaminal, and thalamic) hematomas. Patients with aneurysms, arteriovenous malformations, brain tumors, or head injuries were excluded. Criteria for inclusion were as follows: patients' age between 30 and 80 years; a hematoma volume of more than 10 cu cm; the presence of neurological or consciousness impairment; the appropriateness of surgery from a medical and anesthesiological point of view; and the initiation of treatment within 48 hours after hemorrhage. The criteria of randomization were the location, size, and side of the hematoma as well as the patient's age, state of consciousness, and history of hypertension. Evaluation of outcome was performed 6 months after hemorrhage. Surgical patients with subcortical hematomas showed a significantly lower mortality rate (30%) than their medically treated counterparts (70%, p less than 0.05). Moreover, 40% of these patients had a good outcome with no or only a minimal deficit versus 25% in the medically treated group; the difference was statistically significant for operated patients with no postoperative deficit (p less than 0.01). Surgical patients with hematomas smaller than 50 cu cm made a significantly better functional recovery than did patients of the medically treated group, but had a comparable mortality rate. By contrast, patients with larger hematomas showed significantly lower mortality rates after operation but had no better functional recovery than the medically treated group. This effect from surgery was limited to patients in a preoperatively alert or somnolent state; stuporous or comatose patients had no better outcome after surgery. The outcome of surgical patients with putaminal or thalamic hemorrhage was no better than for those with medical treatment; however, there was a trend toward better quality of survival and chance of survival in the operated group.