The risk of hemorrhage after radiosurgery for cerebral arteriovenous malformations

Author: Reviewer: Vasthi Christensen, MD
Content Contributor: The Abramson Cancer Center of the University of Pennsylvania
Last Reviewed: April 10, 2005

Authors: Maruyama K, Kawahara N, Shin M, Tago M, Kishimoto J, Kurita H, Kawamoto S, Morita A, Kirino T.
Source:New England Journal of Medicine, 352(2); 146-153, 2005.


  • Arteriovenous malformations (AVMs) are defined by the presence of arteriovenous shunting through a nidus of tangling abnormal arteries and veins linked by fistulas.
  • AVMs affect 0.01-0.50% of the population
  • AVMs clinically present as hemorrhage, seizures, progressive neurologic deficit, or headache.
  • Surgery is the best known treatment for AVMs; the nidus is obliterated by ligation of feeding arteries and veins.
  • Angiography shows that stereotactic radiosurgery obliterates many AVMs (in 80-95% of patients) after a latency period of a few years.
  • Hemorrhage has been reported to occur in 2-5% of patients per year during the latency period.
  • It has been unclear whether – and to what extent – the risk is reduced during this latency period as compared with the risk before radiosurgery, or for example, after obliteration.

Materials and Methods

  • To address these questions, a retrospective study was performed at the University of Tokyo involving 500 patients who were treated with stereotactic radiosurgery (gamma knife) for cerebral AVMs between 1990 and 2003.
  • Selection criteria included a small AVM (<3cm) located in a critical or eloquent area of the brain (surgically inaccesable).
  • Surgery was recommended for amenable locations of malformations, except for patient with coexisting medical conditions.
  • Radiosurgery was performed within 3 months of evaluation.
  • Dose applied to the margin of each malformation was designed to be at least 20Gy with the use of 50% isodose lines, but doses were occasionally reduced depending on the volume and location of malformations.
  • Follow up: CT/MRI and clinical evaluation every 6 months
  • Primary end point was first hemorrhage after date of diagnosis.
    • Rates of hemorrhage were assesed during 3 periods: before radiosurgery, latency period, and after angiographic obliteration.
    • In the overall analysis, patients were divided into two groups: those initially presenting with hemorrhage and those without hemorrhage at presentation.


  • Median observation period was 7.8 years.
  • 310 patients with hemorrhage at presentation, 190 presented without hemorrhage.
  • The maximal dose was 40 Gy and the median dose to the margins was 21 Gy, but radiation doses were less than 20Gy in 35 patients
  • Cumulative rates of obliteration were 81% at 4 years and 91% at 6 years.
  • Radiation side effects: transient radiation-induced neurologic deterioration found in 5% of patients, with 1.5% having persistent neurologic deterioration.
  • Hemorrhage occurred before radiosurgery in 8.4% of patients, 5% of patient in the latency period, and 2.4% after obliteration.
  • The risk of hemorrhage was reduced by 54% during the latency period (HR=0.46, p=0.006) and by 88% after proven obliteration when compared with the risk of hemorrhage before radiosurgery.
  • However, the risk of hemorrhage among the 190 patients without hemorrhage at presentation did not significantly decrease from the value before radiosurgery.
  • The authors show that among patients presenting with hemorrhage, the rates of subsequent hemorrhage before radiosurgery is similar over 3 years.

Author's Conclusions

  • The risk of hemorrhage from AVMs is significantly decreased after radiosurgery
  • This is an important endpoint because morbidity after rupture of an AVM is 53-81% and mortality after rupture is 10-17%.

Scientific Implications

There did not appear to be a benefit for radiosugery for cerebral AVMs for patients who presented without hemorrhage, although the risk tended to decrease during the latency period after obliteration. Perhaps the decreased rate of hemorrhage is due to a natural decline in the rate of recurrent hemorrhage. In other words, a natural decline in the rate of recurrent bleeding has been reported within one year of rupture of AVMs. Is the benefit seen in this trial due to radiosurgery of the fact that they have already hemorrhaged? The authors go on to disprove this later in their article, emphasizing that rates of hemorrhage among their patients remained stable over the untreated years. However, they admit that the number of untreated patients is small.

This is a retrospective study without a control group. They show a decreased risk of hemorrhage after radiosurgery, but only for patients who presented with hemorrhage. Perhaps radiosurgery for AVMs should be reserved for those patients who present initially with hemorrhage and withheld for those who present without hemorrhage. Furture studies may include treatment of AVMs with heavy-charged particles such as protons.

Related Blog Posts

December 8, 2023

The Gift of Time

by Carolyn Vachani, MSN, RN, AOCN

January 23, 2023

News on the Passing of the Lymphedema Treatment Act!

by OncoLink Team

September 16, 2022

Wine Racks and the Mystery of Survival

by Rodney Warner, JD