Treatment of arteriovenous malformations.

We read with great interest the paper by Dawson et al. (1] on combined treatment of arteriovenous malformations (AVMs) by embolization and stereotactic radiosurgery, and we would like to congratulate them for their efforts in the therapeutic approach to these lesions. Although we agree with the concept of using a combined approach, we do not recommend the same technical procedure. As discussed previously (2-4], embolization with particles is used optimally in the preoperative stage and mainly in neoplastic conditions that have a capillary barrier (3, 4]. Furthermore, permanent occlusion of the affected vessel cannot be expected with particles, and recanalization, even after dense packing with this material , is an almost constant finding on long-term follow-up (4]. Stable occlusion is being obtained now by using liquid agents such as N-butyl-2-cyanoacrylate (NBCA). These can be used in arteriovenous fistulas or malformations, and their use, although perhaps more challenging than the use of particles, is associated with low morbidity when administered by an experienced team. If complementary treatment is mandatory and radiosurgery is chosen, the remaining target must be as small as possible , preferably less than 2.5 em (5 , 6]. It is easy to understand that this "bull's-eye" must be a stable remnant. Any revascularization of the supposed previously totally embolized territories would lead to incomplete treatment with all risks that this carries, especially in AVMs that have bled (cases 3, 6, and 7 in the study by Dawson et al. (1]). Furthermore, embolization with particles is transitory, so changes in the angioarchitecture of the AVM, and even worsening of the lesion, as happened in case 6, can be expected. In conclusion , we stress that the combination of endovascular techniques and radiosurgery is a method of choice that should be considered for treatment of cerebral AVMs. We would only like to add that, on the basis of 15 years ' experience with particle embolization and 10 years ' experience with isobutyl-2-cyanoacrylate and NBCA, we think that embolization should be performed with definitive materials (NBCA) in order to obtain the best results, especially if the patient has intracranial bleeding and complete anatomic cure is mandatory. The facts that particles are easier to use and are used more widely do not make them a good agent. So far, in similar series , glue embolization has given the best results. Georges Rodesch Pierre Lasjaunias H6pital Bicetre, Universite Paris Sud 94275 Kremlin Bicetre Cedex, Paris, France 1023