In Reply: Risk for Hemorrhage the First 2 Years After Gamma Knife Surgery for Arteriovenus Malformations: An Update

To the Editor: We are thankful for the insightful comments by Mehkri et al1 to our recently published article,2 giving us an opportunity to explain what at first glance may seem to be contradictory conclusions drawn in the present study as compared with our earlier studies. As pointed out, the results in the recent article are similar to those published in 2001,3 the study hereafter being referred to as the KLS study. Both studies concluded that the lowest dose to the arteriovenous malformation (AVM) nidus, the AVM volume, and the age of the patient all are independent factors related to the risk for hemorrhage the first 2 years after Gamma Knife surgery (GKS). The consensus 20 years ago was that AVMs are congenital. An imperative consequence thereof is that the risk for hemorrhage must increase with age. We quantified the relation between risk and age in an earlier study, assuming that AVMs are congenital.4 This risk/age relation did fit well with the age dependence found in the KLS study. It was thus concluded that the age dependence is a consequence of the natural course of AVMs and not caused by the treatment. We were in the KLS study unable to conclude whether the AVM volume dependence was treatment-induced or due to the natural course of AVMs. There is today strong evidence contradicting the dogma that AVMs are congenital.5 Instead, they are deemed developmental. Using this new information, we could, as the authors point out, in another study show that the risk for hemorrhage in untreated AVMs is both age-independent and AVM volume–independent.6 Thus, the higher hemorrhage risk after GKS for larger AVMs and older patients is not due to the natural course but treatment-induced. We completely agree with Mehkri et al that AVMs are a heterogeneous group and that management of patients with AVM must be individualized. Expertise in microsurgery, radiosurgery, and embolization is necessary to optimize the management for each individual patient. This statement directly contradicts the conclusion from the ARUBA study7 that all unruptured AVMs should be left untreated. Indeed, our study shows that the risk of AVM hemorrhage for small AVMs treated with high doses are lower than that for untreated patients already within 6 months after GKS. There are numerous other studies published contradicting the findings in the ARUBA study, one being the NASSAU study,8 allowing us to override the ARUBA conclusion. Thus, the management of unruptured AVMs must be individualized, just as for ruptured ones. We also agree with Mehkri et al that the clinical outcome after an AVM hemorrhage is a crucial piece of information for AVM management. Is the clinical outcome after a hemorrhage in an earlier unruptured AVM “often mild, with bleeding often mainly confined to the brain arteriovenous malformation itself,” as stated by Mohr et al?7 No. To shed light on this topic, we analyzed the clinical outcome for all 383 patients who hemorrhaged after GKS.9 The only factor related to the outcome was age. Neither prior hemorrhage, AVM volume, or location had any impact on the outcome. The mortality rate was independent of whether the AVM had hemorrhaged or not before GKS, contradicting one of the assumptions on which the ARUBA study was based. The 21% mortality rate we found is comparable with the earlier published studies analyzing postradiosurgery hemorrhages and prospective studies of the natural course.8 We fail to understand the relation between our article and the statements in the last paragraph of the letter. Headache and seizures are usually not hemorrhage-related but caused by the AVM itself. Thus, headache may improve, and the likelihood of seizure is likely to decrease once an AVM is obliterated, independent of treatment modality. Whether the symptoms improve or not after failed radiosurgery is controversial. Independent thereof, the potential symptomatic benefit failed radiosurgery may result in should, in our opinion, not be a part of the decision-making. Instead, additional treatments should be offered so that the aim of complete AVM obliteration can be achieved.