The outcomes of 151 patients with arteriovenous malformations (AVMs) treated either surgically or conservatively are presented. In terms of long-term survival rate and follow-up results, the patients who underwent total excision had more favorable results than did those treated conservatively. Patient age was the most important clinical factors, with the preoperative level of consciousness being the second most important. A grading system was formulated on the basis of the angiographical factors by using multiregression analysis; the size of the AVM influenced the surgical outcome twice as much as did the location and the pattern of arterial feeding. Thus, a large AVM (greater than or equal to 4 cm) was given 2 points; a small AVM (less than 4 cm) was given 0 points; a deep AVM was given 1 point; a superficial AVM was given 0 points; an AVM supplied by three of more artery systems was given 1 point; and an AVM supplied by one or two artery systems was given 0 points. AVMs were categorized into 5 grades from Grade 0 to 4 by the summation of these points. Predicted Karnofsky scale after surgery was calculated by this grading system and the following equation: predicted Karnofsky scale = 87.2 - 5.6 x Grade. Grade 0 and 1 AVMs showed high rates of total excision (Grade 0, 94%; Grade 1, 82%) and of satisfactory outcome (Grade 0, 90%; Grade 1, 82%) and were classified as "easy" lesions. Grade 2 AVMs are lesions classified as "moderate" and had a total excision rate of 76% and a satisfactory outcome rate of 71%.(ABSTRACT TRUNCATED AT 250 WORDS)