Aneurysms Increase the Risk of Rebleeding After Stereotactic Radiosurgery for Hemorrhagic Arteriovenous Malformations
Background and Purpose—The purpose of this study was to define the risk of rebleeding after stereotactic radiosurgery (SRS) for hemorrhagic arteriovenous malformations with or without associated intracranial aneurysms.
Methods—Between 1987 and 2006, we performed Gamma Knife SRS on 996 patients with brain arteriovenous malformations; 407 patients had sustained an arteriovenous malformation hemorrhage. Sixty-four patients (16%) underwent prior embolization and 84 (21%) underwent prior surgical resection. The median target volume was 2.3 mL (range, 0.1–20.7 mL). The median margin dose was 20 Gy (range, 13.5–27 Gy).
Results—The overall rate of total obliteration defined by angiography or MRI was 56%, 77%, 80%, and 82% at 3, 4, 5, and 10 years, respectively. Before obliteration, 33 patients (8%) sustained an additional hemorrhage after SRS. The overall annual hemorrhage rate until obliteration after SRS was 1.3%. The presence of a patent aneurysm was significantly associated with an increased rehemorrhage risk after SRS (annual hemorrhage rate, 6.4%) compared with patients with a clipped or embolized aneurysm (annual hemorrhage rate, 0.8%; P=0.033).
Conclusions—When an aneurysm is identified in patients with arteriovenous malformations selected for SRS, additional endovascular or surgical strategies should be considered to reduce the risk of bleeding during the latency interval.
Stereotactic radiosurgery (SRS) has been widely used to manage patients with intracranial arteriovenous malformations (AVMs).1–4 The goal of treatment in cerebral AVMs is elimination of intracranial hemorrhage risk and prevention of new neurological deficits. In the absence of treatment, the overall risk of a spontaneous hemorrhage from an intracranial AVM appears to range from 2% to 5% per year.5,6 After an initial hemorrhage, early rebleeding rates have been reported to vary between 6% and 18% per year.5,7 We have confirmed that the presence of an aneurysm significantly increases the risk of a bleeding event after SRS for an AVM.1 The primary disadvantage of SRS lies in the risk of bleeding during the latency interval between SRS and obliteration. Total obliteration appears to reduce the cumulative residual lifetime risk of hemorrhage to ≤1%.8 The present report evaluated the risk of rebleeding in patients with hemorrhagic AVM with or without associated arterial aneurysms.
Methods and Materials
Between August 1987 and December 2006, 996 patients with AVM underwent single-stage SRS using the Leksell Gamma Knife (Elekta Instruments, Norcross, GA) at our center. We identified 407 patients (216 males and 191 females) whose AVM was diagnosed after they presented with one or more hemorrhages. Forty-five patients (11%) had an associated aneurysm. Locations of the aneurysms included intranidal (N=6), immediately prenidal (N=5), and circle of Willis (N=34). Before SRS, 7 patients with aneurysms of the circle of Willis underwent embolization and 13 patients underwent microsurgical neck clipping. No patient with an intranidal or immediately prenidal aneurysm underwent embolization or neck clipping. Twenty-five patients had patent aneurysms (intranidal: N=6, immediately prenidal: N=5, circle of Willis: N=14) at the time of SRS. In the entire cohort of patients with AVM with associated aneurysms, 34 patients had a prior hemorrhage from their AVM and 11 bled from their aneurysm (10 then underwent clipping and one underwent embolization). The detailed patient characteristics are shown in Table 1.
Our radiosurgical technique has been described in detail in previous reports.1 The margin SRS dose included the entire AVM nidus volume, defined as the shunt between the afferent arteries and the draining veins. Successfully embolized volumes were not included in the SRS target volume. SRS was performed with a Model U, B, C, or 4-C Leksell Gamma Knife (Elekta Inc, Atlanta GA). Intranidal aneurysms were included in the radiosurgical volume. Prenidal and circle of Willis aneurysms were not treated by radiosurgery.
The median target volume was 2.3 mL (range, 0.1–20.7 mL). The median maximal diameter of the AVM nidus was 1.9 cm (range, 0.5–4.8 cm). The median prescription dose delivered to the nidus margin was 20 Gy (range, 13.5–27 Gy). The median maximum dose was 40 Gy (range, 22.2–50 Gy). The median number of isocenters was 2 (range, 1–12).
After radiosurgery, patients were instructed to have clinical and imaging assessments at 6-, 12-, 24-, and 36-month intervals. At the end of 3 years, if MRI suggested complete obliteration, a follow-up angiogram was requested. When obliteration of the AVM was suggested by MRI, cerebral angiography was requested. Complete AVM obliteration was defined as inability to detect a residual nidus at the time of follow-up angiography, contrast-enhanced T1- and T2-weighted MRI. Complete angiographic AVM obliteration was defined as disappearance of the nidus and absence of early venous drainage. At any time when a new neurological symptom or sign developed, the patient had CT and/or MRI to rule out hemorrhage or adverse radiation effects. This retrospective study was approved by the University of Pittsburgh Institutional Review Board.
Kaplan-Meier survival analysis was carried out to calculate rates of hemorrhage in patients with or without associated aneurysms. Patients were censored on loss of follow-up or at the time of hemorrhage. As such, rates of hemorrhage reflect a single hemorrhage experienced by a patient even if he or she had more than one hemorrhage. Annual hemorrhage rates were calculated based on years of follow-up and total number of hemorrhages. The log-rank test was used to assess differences in survival curves and Cox regression was used to assess hazard ratios in multivariate analysis.
Case Matching to Patients With AVM With and Without Coexisting Aneurysm
Of the 407 patients with hemorrhagic AVM who underwent single-stage SRS, 362 patients had no coexisting untreated aneurysm and were therefore eligible to be case–control matches in a comparison to 25 patients with hemorrhagic AVM who had patent coexisting aneurysms. Propensity score matching was performed between the aneurysm case cohort and the control cohort with without coexisting aneurysms.9 The propensity score was calculated by fitting a logistic regression model using the following 9 variables: sex, age, prior hemorrhage history, AVM location (deep versus superficial), target volume, margin dose, Spetzler-Martin grade, Pollock-Flickinger score, and follow-up duration. We used a nearest neighbor 1:2 matching algorithm based on the propensity score. A caliper width of 0.6 times the sample SD of the propensity score was chosen. After propensity score matching, the Mann-Whitney U test for continuous data and the Fisher exact test for categorical data were used to compare both groups.
At the time of assessment, 363 patients were alive and 44 had died. Seventeen patients died from a brain hemorrhage, one patient died from untreated symptoms of adverse radiation effects, and 27 patients died from causes unrelated to their AVM. The median follow-up after SRS was 66 months (range, 2–274 months). The median follow-up after imaging defined obliteration was 36 months (range, 1–259 months). All living patients had at least 1-year imaging and clinical assessment.
Response to Radiosurgery
AVM obliteration was documented by MRI in 255 patients and by angiography in 188 patients. The total obliteration rates after SRS (based on either angiography or MRI criteria) were 56%, 77%, 80%, and 82% at 3, 4, 5, and 10 years, respectively. The median time until MRI documentation of total obliteration was 32 months (95% CI, 28.9–35.1 month). The total obliteration rates based on angiography alone were 46%, 65%, 68%, and 70% at 3, 4, 5, and 10 years, respectively, at a median time of 37 months (95% CI, 35.2–38.6 months). Obliteration rates calculated by angiography alone from this series are biased (artificially lowered) by exclusion of patients with MR-defined obliteration who declined angiography. The exact time of obliteration remains unknown because the timing of documentation by imaging varies.
Hemorrhage Before and After Radiosurgery
A total of 407 patients had 486 hemorrhagic events during 14 298 risk-years before SRS. Assuming that patients were at risk for hemorrhage since their birth, the annual hemorrhage rate between birth and SRS was 3.4%. Fifty-five patients had ≥2 hemorrhages (a total of 79 hemorrhages during 478.4 risk-years); the annual rate of repeat hemorrhage between the time of the initial hemorrhage and SRS was 16.5% in this subset of patients with multiple bleeding events.
Twenty-nine patients had a single AVM hemorrhage after SRS. Three patients had 2 hemorrhages, the second of which was fatal in all 3 patients. Twelve patients died after a single AVM hemorrhage at a median of 22 months (range, 2–148 months) after SRS. The mortality rate due to AVM rehemorrhage after SRS was 4.2%. In 2997.2 patient-years of estimated hemorrhage risk (the interval from the date of SRS to the date of total obliteration on angiography or the date of last follow-up imaging showing a residual AVM), we confirmed 38 hemorrhages, which corresponds to an annual hemorrhage rate of 1.3%. In the first year (393.6 patient-years with 407 patients), the second year (355.1 patient-years with 376 patients), the third year (318.1 patient-years with 338 patients), 3 to 5 years (497.8 patient-years with 295 patients), 5 to 10 years (756.6 patient-years with 213 patients), and 10 to 20 years (644.0 patient-years with 108 patients) after SRS, 14, 4, 5, 4, 5, and 6 hemorrhages occurred, respectively. The annual rehemorrhage rates in patients without obliteration, in the interval of 0 to 1 year, was 3.6%, 1 to 2 years was 1.1%, 2 to 3 years was 1.6%, 3 to 5 years was 0.8%, 5 to 10 years was 0.7%, and 10 to 20 years was 0.9%. Based on Kaplan-Meier analysis (excluding the second hemorrhage in the patient who had >2 hemorrhages), the cumulative rate of AVM rehemorrhage after SRS was 3.2%, 4.3%, 5.5%, 6.6%, and 9.7% at 1, 2, 3, 5, and 10 years, respectively (Figure 1). No patient bled after documentation of AVM obliteration using either MRI or angiography. The detailed outcomes of univariate and multivariate analyses of rehemorrhage rates after SRS are shown in Table 2.
Hemorrhage Risks in Patients With or Without Aneurysms
Table 3 shows the outcomes of patients with aneurysms in addition to their AVM. In the group of patients without aneurysms (N=362), 2721.8 patient-years of estimated hemorrhage risk, we confirmed 27 hemorrhages in 23 patients (3 hemorrhages in one patient, 2 hemorrhages in 2 patients), which corresponds to an annual hemorrhage rate of 1.0%. Based on Kaplan-Meier analysis (excluding the second hemorrhage in the patient who had >2 hemorrhages), the cumulative rate of AVM rehemorrhage after SRS was 2.8%, 4.4%, and 5.2% at 1, 3, and 5 years, respectively. One of 20 patients who had their aneurysm clipped or embolized before SRS rebled from their AVM after SRS. In contrast, 9 of 25 patients with patent aneurysms before SRS rebled after SRS. Six patients with a patent aneurysm rebled from their AVM based on CT or MRI. Three patients died after rebleeding but the source of hemorrhage (AVM or aneurysm) could not be determined. For patients with AVM with clipped or embolized aneurysms, the rate of rehemorrhage from either their AVM or aneurysm after SRS was 0%, 6.2%, and 6.2% at 1, 3, and 5 years, respectively. In 118.4 patient-years of estimated hemorrhage risk (the interval from the date of SRS to the date of total obliteration on angiography or the date of last follow-up imaging showing a residual AVM), we confirmed one hemorrhage, which corresponds to an annual hemorrhage rate of 0.8%.
In the group of patients with AVM with patent aneurysms, the rate of rehemorrhage after SRS from either AVM or aneurysm rupture was 12.2%, 21.5%, and 26.7% at 1, 3, and 5 years, respectively. In 154.8 patient-years of estimated hemorrhage risk, we confirmed 10 hemorrhages in 9 patients, which correspond to an annual hemorrhage rate of 6.4%. The presence of a patent aneurysm was significantly associated with an increased rehemorrhage risk after SRS compared with patients with clipped or embolized aneurysms (P=0.033; Figure 1).
Comparison of Outcomes to Patients With AVM With and Without Coexisting Aneurysm
We evaluated variables that might affect outcomes between patients with patent aneurysms (N=25) and those without aneurysms (N=362). Patient age (P<0.0005), Pollock-Flickinger score (P=0.018), number of prior embolization (P=0.016), and Spetzler-Martin grade (P=0.022) were significantly different between patients with AVM with patent aneurysms and those without an aneurysm.
The demographic and clinical information of the patients with AVM with patent aneurysms (cohort group: N=25) was matched to patients without aneurysms (control group: N=50) as shown in Table 4. We found no difference in the rate of total obliteration after SRS in patients with AVM with patent aneurysms compared with those without aneurysms (P=0.368).
In the control group of patients with AVM without aneurysms, postradiosurgery cumulative rehemorrhage rates were 2.3%, 4.7%, and 7.4% at 1, 3, and 5 years, respectively. For patients with coexisting aneurysms, postradiosurgery cumulative rehemorrhage rates were 12.2%, 21.5%, and 26.7% at 1, 3, and 5 years, respectively (Figure 1). The presence of a patent aneurysm in patients with hemorrhagic AVM was significantly associated with a higher rate of rehemorrhage (log-rank: P=0.003; Cox proportional hazard: P=0.008; hazard ratio, 4.9; 95% CI, 1.52–16.1).
Publications generally report that the hemorrhage rate for untreated AVMs varies from 2% to 5% annually.5,6 After an initial hemorrhage, rebleeding rates increase to 6% to 18% per year.5,7 da Cost et al10 reported that a hemorrhagic presentation was a significant independent predictor of future bleeding (7.5% annual hemorrhage rate; hazard ratio, 2.15), whereas associated aneurysms, deep venous drainage, and prior embolization were not significantly associated with a higher risk of hemorrhage.
Recurrent Hemorrhage After SRS
Maruyama et al3 in a retrospective analysis involving 500 patients found a 54% reduction in the expected bleeding rate during the latency period. Our analysis found that the annual bleeding risk after SRS between the first year and the second year was reduced from 3.9% to 0.6%. In the present series, no patient sustained a hemorrhage after total obliteration was confirmed by either MR or angiography. Karlsson et al2 in a study of postradiosurgery hemorrhage rates reported that the risk for hemorrhage decreased during the latency interval. Yen et al4 reported that the hemorrhage rate in patients with a hemorrhage history (calculated between the time of diagnosis and SRS) was 10.4% but was reduced to 2.8% after SRS until total obliteration of AVM was confirmed. In the present series, factors associated with an increased hemorrhage rate after SRS included lower margin dose and the presence of a patent aneurysm (Table 3).
Aneurysms and AVMs
The incidence of a coexisting aneurysm with AVMs varies from 2.7% to 16.7% in the literature.11,12 In the present series, 11% of patients had a coexisting aneurysm. Locations of the aneurysms included intranidal (N=5), immediately prenidal (N=6), and circle of Willis (N=14). Twenty circle of Willis aneurysms were treated by clipping or embolization before SRS. We found that patients with hemorrhagic AVM with treated aneurysms were significantly less likely to rehemorrhage after SRS (Figure 1). Aneurysm development is more often noted in patients with larger volume AVMs and those with higher flows. However, in the present series, neither AVM volume (P=0.247) nor maximum diameter (P=0.269) was associated with the presence of a coexisting aneurysm. The risk of hemorrhage in patients with AVM with an aneurysm is reported as 7% to 10% annually, a risk significantly higher than the risk of a hemorrhage event from an AVM without a coexisting aneurysm.13,14 The annual hemorrhage rate in our patients with patent aneurysms was 6.4% compared with a bleed rate of 1.0% in patients without an aneurysm. Our findings mirror previously published results.1 Parkhutik et al15 also noted that the presence of an aneurysm increased the risk of postradiosurgical bleeding >5-fold. The present study indicates that the rate of brain hemorrhage 5 years after SRS was 26.7% in the presence of a patent aneurysm compared with 5.2% in patients without an aneurysm. The presence of a patent aneurysm was significantly associated with an increased rehemorrhage risk (in 6 patients hemorrhages were related to the AVM and in 3 patients we could not determine if the hemorrhage was related to the AVM or to the aneurysm). This result indicates that aneurysm management reduces combined bleeding risks during the latency interval after SRS. We believe that aneurysm treatment using surgical or endovascular techniques may reduce the combined bleeding risks during the latency interval. A prospective trial might provide stronger evidence for reduction of bleeding risk after more aggressive management of AVM-associated aneurysms. Accrual of patients for such a trial might be a significant limiting factor. Only intranidal aneurysms and flow-related aneurysms immediately proximal to the AVM involute as the AVM obliterates after SRS (Table 3).
The Case-Matching Analysis
In our comparison of patients with coexisting aneurysms to patients without aneurysms, we found additional factors that affected outcomes, including patient age, the Pollock-Flickinger score, the number of prior embolization procedures, and the Spetzler-Martin grade. A case-matching study was important to reduce potential bias. In the case-matching study, patients with AVM with coexisting aneurysms were still found to have a higher rehemorrhage rate compared with patients with AVM without aneurysms (P=0.008; hazard ratio, 4.9; 95% CI, 1.5–16.1). This analysis indicates that patients with a coexisting aneurysm had a 5 times increased risk of either AVM or aneurysm hemorrhage compared with patients without an aneurysm or with patients who had successfully clipped or embolized aneurysms. (P=0.033; Figure 1). All patients who underwent clipping or embolization of their aneurysm before SRS had aneurysms located in the circle of Willis. No patient with a prenidal or an intranidal aneurysm underwent clipping or embolization. Intranidal and immediate prenidal aneurysms completely involuted after successful SRS. These aneurysms were often difficult to access by endovascular techniques and were believed to have a high risk for either clipping or embolization. We recommend that circle of Willis aneurysms that have bled undergo embolization or microsurgical neck clipping. We also recommend that patients with circle of Willis aneurysms found in conjunction with a hemorrhagic AVM undergo embolization or microsurgical neck clipping (Figure 2).
Weaknesses of the Current Study
Although angiography is the “gold standard” for obliteration, 25% of the patients in this series did not undergo angiography after their follow-up MRI revealed obliteration. Although the absence of follow-up angiography in these patients may lead to selection bias, we agree with Pollock et al16 who reported that only 3% of patients with MR-defined obliteration had a residual nidus demonstrated by angiography.
The present study is not a prospective randomized study but retrospective outcome analysis that includes a matched cohort of patients with AVM with or without aneurysms. We performed a propensity score case-match control study to reduce bias. In this study, we excluded patients who were lost to follow-up, a technique that might produce bias. During our 23-year AVM experience, increasing experience with dose–volume relationships, conformality and selectivity of treatment planning, and reliance on both angiographic and then MRI data gradually expanded our knowledge and improved SRS techniques and outcomes.
SRS is an important option for properly selected patients with AVMs. When a circle of Willis aneurysm is identified in such patients, additional endovascular or surgical strategies should be considered to reduce the risk of bleeding during the latency interval. Pre- or intranidal aneurysms will obliterate when the AVM obliterates. Embolization after SRS may or may not reduce the risk of rebleeding during the latency interval after radiosurgery. A prospective trial to assess this hypothesis is currently under evaluation by the North American Gamma Knife Consortium.
Sources of Funding
The work described in this report was funded by a research grant to Dr Kano from AB Elekta, Stockholm, Sweden, and the Osaka Medical Research Foundation for Incurable Diseases.
Drs Lunsford and Kondziolka are consultants for AB Elekta and Dr Lunsford is a stockholder.
This study was reported at the Annual Meeting of the American Association of Neurological Surgeons, Miami, FL, April 17, 2012.
- Received May 17, 2012.
- Revision received June 28, 2012.
- Accepted July 18, 2012.
- © 2012 American Heart Association, Inc.
- Yen CP,
- Sheehan JP,
- Schwyzer L,
- Schlesinger D
- Pollock BE,
- Flickinger JC,
- Lunsford LD,
- Bissonette DJ,
- Kondziolka D
- da Costa L,
- Wallace MC,
- Ter Brugge KG,
- O'Kelly C,
- Willinsky RA,
- Tymianski M
- Parkhutik V,
- Lago A,
- Tembl JI,
- Vazquez JF,
- Aparici F,
- Mainar E,
- et al