The prognostic value of the Spetzler's grading system is studied in a series of 52 AVMs treated by a combined management, using one or several of the 3 available techniques: surgical resection, endovascular embolization, radiosurgery. The symptoms at the time of treatment were haemorrhage 50%, seizures 31%, headache and deficit 19%. Three grade groups were considered: I and II (31%), III (33%), IV and V (36%). Overall, AVMs were managed as follows: resection alone 25%, embolization plus resection 23%, embolization alone 23%, radiosurgery with various combinations 29%. According to the grade groups, the most frequently used technique was resection alone for grade I-II AVMs (44%), radiosurgery for grade III AVMs (41%) and embolization alone for grade IV-V AVMs (42%). The clinical outcome was evaluated in terms of deterioration due to treatment. The best results were obtained in grade I-II AVMs (81% with no deterioration) then in grade III AVMs (65%) and in grade IV-V (58%). However, when we consider the outcome in terms of favourable results (no or only minor deterioration) we obtained a similar outcome for grade I-II and grade III AVMs (94% each), and only 79% for grade IV-V malformations. The angiographic outcome showed a better eradication rate in grade III AVMs (88% complete eradication), than in grade I-II AVMs (75%) and in grade IV-V (47%). Our conclusion is that the Spetzler's grading system in this series was well correlated with both the clinical and the angiographic outcome. However, we found no real difference between grade I-II and grade III AVMs. So, in terms of prognostic value, the grade I, II, and III AVMs could be considered together as low-grade malformations, with a better prognosis than the high-grade malformations (grade IV and V).