Cohort studies, trials, and tribulations: systematic review and an evidence-based approach to arteriovenous malformation treatment

INTRODUCTION: There is uncertainty as to the best management of arteriovenous malformations of the brain (bAVM). However, the Spetzler-Martin grade (SMG) has been validated as an effective determinant of surgical risks. We performed a systematic review for the best evidence regarding the management of bAVM for series that incorporate an analysis based upon SMG.
EVIDENCE ACQUISITION: Medline, Embase, Scopus and Cochrane databases were searched for series between January 2000 and January 2018, with a minimum of 100 cases and that incorporated SMG stratification. From this primary search, series were selected for analysis that dichotomized outcomes at modified Rankin Scale (mRS) scores between 1 and 2 due to complications of treatment or reported favorable outcome (FO) (i.e. complete occlusion, no neurological deterioration and no post treatment hemorrhage). Case series that used a subset of the population other than SMG or had a prior history of hemorrhage were excluded. The series finally analyzed were explored for outcomes that reported: complications of treatment that led to a new permanent neurological deficit with mRS score >1 (adverse outcome); post treatment hemorrhage; occlusion rate; and FO. A comparison of treatment outcomes was made when more than one modality of treatment (surgery, radiosurgery, embolization or multiple treatment modalities) could be examined with results for specific Spetzler-Ponce class (SPC) A (i.e. SMG I and II), B (i.e. SMG III) or C (i.e. SMG IV and V).
EVIDENCE SYNTHESIS: The primary search produced 116 papers. After reviewing each publication and eliminating papers that had patient outcomes duplicated, 11 publications met the criteria for analysis (including: 5 exclusively surgery; 4 exclusively radiosurgery; 1 exclusively endovascular; and, 1 multi-modality). The following outcome comparisons analyzed were significant. For SPC A and B bAVM, there was a significantly higher rate of FO following treatment by surgery (98.6%; 95% CI: 97.5-99.2% and 76.4%; 95% CI: 70.0-81.7%, respectively) than radiosurgery (70.8%; 95% CI: 66.8-74.6% and 61.0%; 95% CI: 56.0-65.8%, respectively)(P<0.01). For SPC A and B bAVM, there were significantly fewer unobliterated bAVM following treatment by surgery (0.5%; 95% CI: 0.2-1.4% and 3.0%; 95% CI: 1.4-5.8%, respectively) than radiosurgery (23.9%; 95% CI: 20.4-27.8% and 30.9%; 95% CI: 27.9-34.0%, respectively) or embolization (7.6%; 95% CI: 4.3-12.9% SPC A) (P<0.01). Adverse outcomes from treatment were significantly higher for surgery (15.6%; 95% CI: 11.8-20.0%) than radiosurgery (3.3%; 95% CI: 2.3-4.8%) for SPC B (P<0.01) but not SPC A bAVM. No analysis of SPC C was possible.
CONCLUSIONS: Surgery remains, in general, the best choice for treating SPC A bAVM. For SPC B bAVM the decision as to best treatment should hinge on the likelihood of obliteration by radiosurgery. In cases where obliteration rate is expected to be high, radiosurgery should be the preferred treatment. There is insufficient information to make a recommendation from this analysis with regards the role of embolization for cure. There is no satisfactory standardized treatment for SPC C bAVM and treatment must remain individualized.