Microsurgical treatment of arteriovenous malformations: analysis and comparison with stereotactic radiosurgery

Full access

Object. To compare microsurgical and stereotactic radiosurgical treatment of arteriovenous malformations (AVMs), the authors analyzed a prospective series of 72 consecutive patients who were treated microsurgically for cerebral AVMs by one neurosurgeon. The authors then compared the results of microsurgical treatment with published results of stereotactic radiosurgical treatment of small AVMs.

Methods. Patients were categorized by age, gender, presentation, and preoperative neurological status. The AVMs were categorized by size, location, presence of deep venous drainage, and Spetzler—Martin grade. Outcome was assessed for angiographic obliteration, hemorrhage following treatment, presence of a new, persistent postoperative neurological deficit, and Glasgow Outcome Scale (GOS) score. Ordinal logistic regression was used to model the GOS score and to predict new postoperative deficits. Generalized estimating equations were used to compare published results of microsurgical and stereotactic radiosurgical treatment of AVMs. Kaplan—Meier event-free survival plots were generated to compare the two modalities with respect to hemorrhage following treatment.

Overall, six patients (8.3%) exhibited a new persistent neurological deficit postoperatively. Sixty-five patients (90.3%) had a GOS score of 5. Three patients were moderately disabled and four patients were severely disabled. No patient was observed to be in a vegetative state and there were no treatment-related deaths. Seventy-one patients (98.6%) underwent intra- or postoperative angiography. Total excision of the AVM was angiographically confirmed in 70 patients (98.6% of those who underwent angiography). To date no patient has suffered from hemorrhage since the microsurgical treatment. When analysis was confined to patients whose AVMs were smaller than 3 cm in maximum diameter, the authors found a 100% angiographic obliteration rate, no new postoperative neurological deficit, and a good recovery in all patients. An analysis of all patients with Spetzler—Martin Grades I to III resulted in a 100% rate of angiographic obliteration, one patient with a new postoperative neurological deficit, and good recovery in 93% of the patients.

Size of the AVM, preoperative neurological status, and patient age are associated with GOS score (for all, p <0.02). The Spetzler—Martin grading system as well as each component of this system are associated with the development of a new postoperative neurological deficit (for all, p < 0.01). For the entire series there were fewer postoperative hemorrhages and deaths than those mentioned in published series of small AVMs treated with stereotactic radiosurgery. When these patients and published series of patients with microsurgically treated AVMs classified as Grade I to III were compared with similar patients treated radiosurgically there were significantly fewer postoperative hemorrhages (odds ratio = 0.210, p = 0.001), fewer deaths (odds ratio = 0.659, p = 0.019), fewer new posttreatment neurological deficits (odds ratio = 0.464, p = 0.013), and a higher incidence of obliteration (odds ratio = 28.2, p = 0.001) for the microsurgical group. Lifetable analysis confirms the statistically significant difference in hemorrhage-free survival time between the two groups (p = 0.002).

Conclusions. Based on this analysis, microsurgical treatment of Grades I to III AVMs is superior to stereotactic radiosurgery.

Abstract

Object. To compare microsurgical and stereotactic radiosurgical treatment of arteriovenous malformations (AVMs), the authors analyzed a prospective series of 72 consecutive patients who were treated microsurgically for cerebral AVMs by one neurosurgeon. The authors then compared the results of microsurgical treatment with published results of stereotactic radiosurgical treatment of small AVMs.

Methods. Patients were categorized by age, gender, presentation, and preoperative neurological status. The AVMs were categorized by size, location, presence of deep venous drainage, and Spetzler—Martin grade. Outcome was assessed for angiographic obliteration, hemorrhage following treatment, presence of a new, persistent postoperative neurological deficit, and Glasgow Outcome Scale (GOS) score. Ordinal logistic regression was used to model the GOS score and to predict new postoperative deficits. Generalized estimating equations were used to compare published results of microsurgical and stereotactic radiosurgical treatment of AVMs. Kaplan—Meier event-free survival plots were generated to compare the two modalities with respect to hemorrhage following treatment.

Overall, six patients (8.3%) exhibited a new persistent neurological deficit postoperatively. Sixty-five patients (90.3%) had a GOS score of 5. Three patients were moderately disabled and four patients were severely disabled. No patient was observed to be in a vegetative state and there were no treatment-related deaths. Seventy-one patients (98.6%) underwent intra- or postoperative angiography. Total excision of the AVM was angiographically confirmed in 70 patients (98.6% of those who underwent angiography). To date no patient has suffered from hemorrhage since the microsurgical treatment. When analysis was confined to patients whose AVMs were smaller than 3 cm in maximum diameter, the authors found a 100% angiographic obliteration rate, no new postoperative neurological deficit, and a good recovery in all patients. An analysis of all patients with Spetzler—Martin Grades I to III resulted in a 100% rate of angiographic obliteration, one patient with a new postoperative neurological deficit, and good recovery in 93% of the patients.

Size of the AVM, preoperative neurological status, and patient age are associated with GOS score (for all, p <0.02). The Spetzler—Martin grading system as well as each component of this system are associated with the development of a new postoperative neurological deficit (for all, p < 0.01). For the entire series there were fewer postoperative hemorrhages and deaths than those mentioned in published series of small AVMs treated with stereotactic radiosurgery. When these patients and published series of patients with microsurgically treated AVMs classified as Grade I to III were compared with similar patients treated radiosurgically there were significantly fewer postoperative hemorrhages (odds ratio = 0.210, p = 0.001), fewer deaths (odds ratio = 0.659, p = 0.019), fewer new posttreatment neurological deficits (odds ratio = 0.464, p = 0.013), and a higher incidence of obliteration (odds ratio = 28.2, p = 0.001) for the microsurgical group. Lifetable analysis confirms the statistically significant difference in hemorrhage-free survival time between the two groups (p = 0.002).

Conclusions. Based on this analysis, microsurgical treatment of Grades I to III AVMs is superior to stereotactic radiosurgery.

Arteriovenous malformations (AVMs) have traditionally been treated microsurgically. With the advent of stereotactic radiosurgery, considerable controversy has developed over what constitutes appropriate therapy. There are strong proponents of each modality, and no randomized, prospective study exists to test the two treatments against each other.

We report the results of a detailed statistical analysis of our consecutive series of AVMs treated microsurgically. Analysis of this series and other published series was undertaken to help clarify the choice between microsurgery and stereotactic radiosurgery for patients who harbor Spetzler—Martin20 Grades I to III AVMs.

Clinical Material and Methods
Patient Population

Between July 1985 and October 1996, 72 consecutive patients underwent microsurgical excision of an AVM, which was performed by one of the authors (R.E.H.). Data regarding patient demographics, medical history, presentation, neurological examination, and radiographic lesion characteristics were recorded prospectively. Patients with cavernous or venous malformations, dural arteriovenous fistulas, vein of Galen malformations, or other arteriovenous fistulas were not included.

The patients ranged in age from 2 to 81 years (mean 37.7 ± 18.3 years, median 39 years). Forty-three patients (60%) were male. Forty-five patients (63%) presented with hemorrhage with or without seizure, 13 (18%) with seizure alone, and 14 (19%) with other clinical syndromes. Neurological deficits were found preoperatively in 39 patients (54%). Table 1 shows the distribution of Spetzler—Martin grades20 among our patient population. Thirty-five patients (49%) had lesions in the left cerebral hemisphere and 31 patients (43%) had lesions in the right. Three patients had lesions of the corpus callosum. Three had lesions of the thalamus, brainstem, and/or basal ganglia. One patient had undergone clip ligation of a feeding vessel, which was performed by another surgeon several years before excision of the AVM.

TABLE 1

Distribution of Spetzler—Martin grading and its components in 72 patients surgically treated for AVMs

Grades & ComponentsNo. of Patients (%)
 I13 (18) 
 II13 (18) 
 III28 (39) 
 IV15 (21) 
 V3 (4) 
size
<3 cm19 (26) 
3–6 cm48 (67) 
>6 cm5 (7) 
eloquent location42 (58) 
deep venous drainage present27 (38) 

Operative Technique

All patients underwent microsurgical excision. Eight patients underwent preoperative embolization (one with Grade III, five with Grade IV, and two with Grade V AVMs). One patient whose AVM was classified as Grade III also underwent intraoperative embolization. Four patients required stereotactic localization of deep lesions. Three patients underwent two-stage resections because at the conclusion of the initial surgical procedure the surgeon believed there was residual AVM. No patient required a second-stage operation because of unexpected residual AVM. One patient refused the second-stage operation.

Postoperative Care

Postoperatively, the patients were given routine neurological care and evaluation. Outcome measures included presence of a new, persistent neurological deficit, the patient's score on the Glasgow Outcome Scale (GOS),11 hemorrhage from lesion postoperatively, and angiographic obliteration of the lesion. Seventy-one patients (98.6%) underwent postoperative or intraoperative angiography. One patient with a Grade V AVM refused postoperative angiography.

Statistical Analysis

Contingency table and logistic regression and ordinal logistic regression analyses were used to model the effects of preoperative factors on response measures: GOS score, presence of new neurological deficit, and adverse outcome (defined as a GOS score < 5 or presence of a new postoperative neurological deficit). The model fit was tested using changes in deviation and was assessed by including interaction terms, plotting residuals, and using probits rather than logits. Graphic techniques were used to assess covariate influence.

Comparison With Other Studies

The literature was reviewed for series of patients with AVMs treated with microsurgery or stereotactic radiosurgery. Each series was evaluated for “definite” and “probable” obliteration rates, presence of new neurological deficit following treatment, hemorrhage following treatment, and death. The definite obliteration rate was defined as the number of angiographically documented lesion obliterations divided by the total patient population. The probable obliteration rate included those patients with angiographic and nonangiographic evidence of AVM obliteration. In one microsurgical series,10 a number of patients did not undergo postoperative angiography because the surgeon was certain that no residual AVM was present. These patients were considered to have probable AVM obliterations. In stereotactic radiosurgery series, patients with angiographic or magnetic resonance (MR) imaging evidence of obliteration were considered to have probable AVM obliteration. The delineation between definite and probable obliteration is necessary because in some stereotactic radiosurgery studies patients were selected for angiography based on the appearance of AVM obliteration on MR imaging.6,18,26 Thus, angiographic obliteration rates may be falsely elevated. The pessimistic assessment of a definite obliteration rate for stereotactic radiosurgery series is probably counterbalanced by the optimistic assessment of probable obliteration. The true obliteration rate likely lies between the two.

We excluded series of patients treated solely by endovascular therapy and series in which particle beam therapy was used. Three radiosurgical and three microsurgical series (in addition to ours) were selected to provide adequate information for analysis. Series were chosen to represent modern state-of-the-art treatment protocols with adequate lengths of follow-up study. Many stereotactic radiosurgery series were reviewed. Three were identified that provided patient data sufficient to allow our analysis. These series represent the experiences of three of the preeminent stereotactic radiosurgery centers in North America with published AVM data.6,18,26 Several radiosurgical series were not used in the comparison because information regarding one or more aspects of diagnosis, treatment, follow up, or analysis were unclear or not provided.

The comparisons were limited to patients with small AVMs. Microsurgical series were restricted to patients with AVMs classified as Spetzler—Martin Grades I to III or, in the series of Sisti, et al.,19 patients with AVMs smaller than 3 cm. Stereotactic radiosurgery series were restricted to patients with AVMs classified as Grade I or II or having volume less than 10 cm3.

Generalized estimating equations were used to determine the pooled odds ratios. Kaplan—Meier12 estimates of hemorrhage-free survival time were calculated to compare patients with Grades I to III AVMs from the stereotactic radiosurgery series conducted by Friedman and colleagues5 with those from the microsurgical series conducted by Heros and associates10 and our own series.

Results
Dartmouth—Hitchcock Microsurgical Series

All patients survived surgical treatment. No patient experienced a hemorrhage from the AVM following surgical treatment. Six patients (8.3%) developed new, persistent neurological deficits after surgery. Five (83%) of these patients had visual deficits (hemianopsia in four cases and quadrantanopsia in one). Two patients (33%) exhibited motor deficits (one case of spastic leg and one case of hemiparesis in a patient with a hemianopsia). One patient with hemianopsia also developed dyslexia.

Good recovery was defined by using the GOS in 65 patients (90.3%). Three patients (4.2%) were moderately disabled and four patients (5.5%) were severely disabled. No patients were in a vegetative state or dead. Complete resection of the AVM was documented by angiography in 70 patients (97%). One patient has known residual AVM and another patient who refused angiography may have residual AVM.

When the analysis was confined to patients whose AVMs were less than 3 cm in maximum diameter, there was a 100% angiographic obliteration rate. In this group of patients no new postoperative neurological deficits were identified. Additionally, all patients made a good recovery. Confining the analysis to patients with AVMs classified as Spetzler—Martin Grades I to III resulted in a 100% angiographic obliteration rate, one patient with a new neurological deficit, and 92.5% of patients reaching the highest GOS score. Patients with Grades I to III AVMs comprised 75% of our series. However, they accounted for only 17% of the postoperative neurological deficits.

Table 2 details the results of a logistic regression analysis of the effects of preoperatively determined characteristics on the three outcome measures. Overall, the AVM grade and each of its components were correlated with new postoperative neurological deficits. Size, preoperative neurological deficit, and age were predictors of GOS score. Higher AVM grade, larger size of the AVM, and advanced patient age predicted adverse outcome. Multivariate modeling determined that new, persistent postoperative deficits were predicted best by Spetzler—Martin grade (p = 0.0004, pseudo-r2 = 0.318). Other combinations of AVM location or size, or presence of deep draining veins, did not significantly enhance the predictive capacity of the model. Statistically significant predictors of GOS were patient age, AVM size, and presence of preoperative neurological deficit (p = 0.0006, pseudo-r2 = 0.330). Adverse outcome could be moderately well predicted by AVM grade and patient age (p = 0.0126, pseudo-r2 = 0.153).

TABLE 2

Preoperative variables compared with outcome for 72 patients who underwent microsurgical excision of an AVM*

Outcome Coefficient (p value)
Preop VariableGOS ScoreNew Deficit PostopAdverse Outcome
patient
 age−0.053 (0.0162)−0.009 (0.6923)−0.035 (0.0483)
 gender0.781 (0.3375)−0.431 (0.6152)−0.069 (0.9145)
 presentation0.721 (0.2089)−0.650 (0.1884)0.136 (0.7378)
 preop deficit−2.226 (0.0128)0.633 (0.4717)−0.961 (0.1374)
AVM
 S—M grade−0.626 (0.0848)−2.065 (0.0002)−0.874 (0.0071)
 size−1.895 (0.0115)−2.226 (0.0099)−1.705 (0.0089)
 location−0.163 (0.8318)−11.745 (0.0088)−0.898 (0.1883)
 venous drainage−0.644 (0.3949)−2.372 (0.0127)−1.106 (0.0852)

S—M = Spetzler—Martin.

Selected Published Series

There are numerous reports of stereotactic radiosurgery of AVMs; three of these series6,18,26 were chosen for analysis. Three microsurgical series,7,10,19 in addition to our own, were analyzed as well. Table 3 details the characteristics of these series and the results of the analysis. Microsurgical treatment was statistically significantly more likely to provide definite or probable AVM obliteration. Microsurgically treated patients were significantly less likely to die, develop new postoperative neurological deficits, or experience hemorrhage postoperatively. Product-limit estimates of hemorrhage-free survival time statistically favored microsurgery over stereotactic radiosurgery (Fig. 1).

TABLE 3

Grade I to III AVMs treated with microsurgical excision or stereotactic radiosurgery*

No. of Patients (%)
Authors & YearTreatment MethodTotalDefinite ObliterationProbable ObliterationHemorrhageDeathNew Deficit
Heros, et al., 1990micro9179 (87) 91 (100) 00NR
Sisti, et al., 1993micro6763 (94) 63 (94) 001 (1.5)
Hamilton & Spetzler, 1994micro7171 (100) 71 (100) 000
Pollock, et al., 1994gamma6527 (42) 43 (66) 5 (7.7)2 (3.1)1 (1.5)
Friedman, et al., 1995LINAC5737 (65) 40 (70) 1 (1.8)01 (1.8)
Yamamoto, et al., 1995gamma9032 (36) 67 (74) 6 (6.7)3 (3.3)5 (5.6)
present seriesmicro5454 (100) 54 (100) 001 (1.9)
odds ratio21.7 28.2 0.2100.6590.464
 95% confidence interval5.40–87.0 4.27–185 0.123–0.3580.466–0.9330.254–0.851
p value0.001 0.001 0.0010.0190.013

Gamma = gamma knife radiosurgery; LINAC = linear accelerator radiosurgery; micro = microsurgery; NR = not reported.

Fig. 1.
Fig. 1.

Graph displaying Kaplan—Meier hemorrhage-free survival estimates following microsurgical extirpation or stereotactic radiosurgery of Spetzler—Martin Grade I to III AVMs. The microsurgical data were obtained from the current series and that of Heros, et al.10 The stereotactic radiosurgical data were obtained from the study by Friedman, et al.5

Discussion

We have reported a series of microsurgically treated AVMs with results similar to other modern series.7,10,19,20,23 Regression analysis demonstrates that patient age, lesion size, and preoperative neurological status are predictive of functional outcome. Adverse outcome from AVM microsurgery (as defined by a GOS score < 5 or the presence of a new postoperative neurological deficit) is predicted by AVM grade and patient age. Spetzler—Martin AVM grading and each of its components (size, location, and presence of deep venous drainage) are all predictive of new postoperative neurological deficit. Multivariate models show that grade alone is virtually as good a predictor as combinations of variables. Hamilton and Spetzler7 previously showed that the Spetzler—Martin grading system correlates well with transient and permanent postoperative neurological deficits.

A natural split is demonstrated in our population and in others.7,10,19,20 Patients with AVM Grades I to III have minimal long-term morbidity rates following resection. Conversely, higher-grade lesions carry an increasingly higher rate of morbidity. The implication is that a binary grading scale may be as informative as the multilevel Spetzler—Martin grading system. This principle is reflected in the observation that surgical decision making usually has an inflection point between Grades III and IV.4 Advanced age of the patient, although a factor in surgical decision making, should not, in and of itself, be considered a contraindication to surgical treatment of an AVM.8

Numerous reports on stereotactic radiosurgery have found that smaller AVMs respond better than larger ones, with respect to both lesion obliteration and irradiationinduced complications.1,3,5,6,13,16–18,21,22,26 Thus, a comparison of microsurgery and stereotactic radiosurgery in the treatment of small or low-grade AVMs should maximize the outcome statistics for both populations. In our analysis, microsurgery is far more effective if definite or probable AVM obliteration is the endpoint. Microsurgically treated patients were significantly less likely to suffer a hemorrhage, die, or develop new neurological deficits following treatment.

Obliteration of the AVM is not a certain consequence of stereotactic radiosurgery.1,3,5,6,12,16–18,21,22,26 Friedman, et al.,6 reported lesion obliteration demonstrated on angiography in 37 (65%) of 57 patients with smaller AVMs (mean follow-up time 33 months). Forty (70%) of the 57 patients probably had their lesions obliterated. Pollock, et al.,18 documented angiographic obliteration of only 27 (42%) of 65 AVMs with Spetzler—Martin Grade I or II in a series that had a follow-up duration longer than 24 months. The Mayo group26 was able to prove obliteration in 32 (36%) of 90 AVMs with volumes of 10 cm3 or less, whereas 74% of the AVMs were probably obliterated (follow up 12–60 months). Colombo, et al.,3 reported probable obliteration rates of 53% and 90% at 1 and 2 years, respectively, in patients with AVMs smaller than 25 mm. Betti and associates1 found a 9 to 69% cure rate in their population of patients with AVMs smaller than 25 mm, depending on the dose of radiation. Heffez and colleagues (unpublished data) found the median time to obliteration of AVMs smaller than 35 mm to be 2.04 years. In contrast, microsurgical series, including our own, demonstrate nearly 100% obliteration rates immediately.7,10,19,23

Because AVM obliteration following stereotactic radiosurgery is not immediate, the interval between treatment and obliteration is of concern. With respect to the propensity to hemorrhage, the natural history of the AVM is not improved until the lesion is obliterated.3,5,9,16,18,21,22,24 Pollock, et al.,18 reported that 7.7% of patients with Grade I or II AVMs experienced hemorrhage within the first 8 months following stereotactic radiosurgery. Although these authors quoted a yearly risk of hemorrhage of 3.7%, for a period of time the risk of AVM hemorrhage following stereotactic radiosurgery may exceed the risk of hemorrhage from untreated lesions.2,15 Yamamoto and coworkers26 reported posttreatment hemorrhage in 6.6% of their patients who had AVMs with a volume less than 10 cm3. Betti and associates1 reported a 12.5% rate of post—stereotactic radiosurgery hemorrhage in their series overall. Friedman and colleagues5,6 reported a 3.6% incidence of post—stereotactic radiosurgery hemorrhage in patients with Grade I to III AVMs. Steiner, et al.,21 did not discuss the rate of hemorrhage that occurred during the first 2 years following stereotactic radiosurgery. However, patients in whom the lesion was not obliterated after 2 years showed a 1.9 to 6.5% yearly hemorrhage risk, depending on the analysis.

Hemorrhage that occurs during the latency period is not a benign event. In the series of Yamamoto and coworkers,26 50% of patients who experienced such a hemorrhage died. Similarly, in the series of Pollock, et al.,18 40% of the patients with Grade I or II AVMs who experienced hemorrhage during the latency period died. Only 17% of patients with Grades I to III AVMs who bled following stereotactic radiosurgery died in the series of Friedman, et al.5 Nussbaum and colleagues14 report that the lifetime cost of radiosurgery is predominantly the result of the treatment of delayed morbidity from hemorrhage. This is one of the reasons that, in their analysis, AVM microsurgery was found to be cost effective relative to stereotactic radiosurgery.

The mortality rate associated with microsurgical extirpation of small AVMs approaches 0%.4,7,10,20,23 On the other hand, in some stereotactic radiosurgery series mortality rates are reported to be 3 to 5%.1,17,21,26 Additionally, information on long-term radiation-related complications has been included in many series. Colombo, et al.,3 noted a 5% rate of radionecrosis arising between 2 and 18 months after stereotactic radiosurgery. Almost half of these patients had persistent neurological deficits. Two percent of the patients included in the series of Friedman, et al.,6 experienced transient delayed complications that were believed to be due to stereotactic radiosurgery. Another 1.3% developed permanent neurological deficits attributable to the procedure. Yamamoto, et al.,25 found a 7.5% incidence of delayed radiation-related morbidity in a selected group of patients. Cyst formation and delayed oncogenesis in patients who undergo radiosurgery are other potential concerns.

We have tried to provide a fair evaluation of the published results for microsurgical and stereotactic radiosurgical treatment of AVMs. However, we clearly have a surgical bias. An issue that arises in the comparison of microsurgery and stereotactic radiosurgery for AVMs is that prognostication of outcome from the former is usually based on diameter-dependent analysis, whereas that of the latter is more appropriately analyzed using lesion volume. Additionally, many of the patients included in stereotactic radiosurgery series were selected because of the depth or predicted surgical difficulty of their lesions. These two points raise some concern for the comparison of the two modalities. However, whereas the location of the lesion may affect the likelihood of new, post—stereotactic radiosurgery neurological deficits, it should not dramatically influence the biological response following the procedure, except where dosing was limited by adjacent structures.

Another criticism is that our model does not rigorously treat adjuvant therapies such as embolization. We believe, however, that the model is still valid for several reasons. Embolization is usually reserved for larger, higher-grade lesions. In the studies included in our statistical analysis embolization was not used to a great degree in the subset of patients under comparison. Use of AVM embolization may change the morbidity rate of subsequent microsurgical or stereotactic radiosurgery treatment of that AVM. For the purpose of this analysis, any incidence of morbidity from AVM embolization is attributed to the other treatment modalities.

Our analysis of these patients focused on long-term clinical outcome following treatment and did not include data on transient postoperative complications. Such complications are, of course, more frequent than permanent neurological deficits following excision or irradiation of AVMs and will add to the costs associated with treatment of these lesions. When transient postoperative complications have been included in a cost analysis, microsurgery has still emerged as cost effective relative to stereotactic radiosurgery.14

Concluding that microsurgical excision of AVMs is superior to stereotactic radiosurgery assumes that the surgery can be performed with outcomes similar to those in the studies used for our statistical analysis. The same concern also exists for stereotactic radiosurgery: that is, patient evaluation and treatment planning need to be conducted with the expertise necessary to match the results of stereotactic radiosurgery performed at major centers.

One might speculate that a well-designed, prospective, randomized trial in which stereotactic radiosurgical and microsurgical excision of Spetzler—Martin Grades I to III AVMs were compared would provide more definitive data regarding the relative efficacies and risk profiles of the two treatment modalities. Statistically, it is very unlikely, however, that such a study would change the conclusions concerning AVM obliteration and hemorrhage following treatment. The difference in these outcomes measures is so great that any type of design bias (such as series selection, patient selection, or endpoint evaluation) is an unlikely explanation. For example, assume that an unknown, confounding variable exists that increases the odds of AVM obliteration 10 times. If this variable were present in 80% of the surgical group and only 30% of the radiosurgical group in our analysis, its random distribution in a prospective study would still result in a considerable benefit of microsurgery compared to stereotactic radiosurgery. A randomized trial might confirm or refute a benefit of surgery with respect to decreased incidence of mortality or a reduction in permanent neurological deficits because the odds ratios in our study are much closer to one for these outcome events. Such a trial would require approximately 2000 patients to show a reduction in morbidity rate from 5 to 2.5%.

Conclusion

The series presented is similar to other recent series of microsurgical treatment for AVM. Prognostic inference can be made based on the Spetzler—Martin grading, preoperative neurological status, and patient age. Based on our review and analysis, we conclude that microsurgery is superior to stereotactic radiosurgery in the treatment of smaller, surgically accessible AVMs when obliterating the AVM, reducing the risk of hemorrhage, reducing permanent neurological morbidity, and reducing mortality are the desired outcomes.

References

  • 1.

    Betti OOMunari CRosler R: Stereotactic radiosurgery with the linear accelerator: treatment of arteriovenous malformations. Neurosurgery 24:3113211989Neurosurgery 24:

  • 2.

    Brown RD JrWiebers DOForbes Get al: The natural history of unruptured intracranial arteriovenous malformations. J Neurosurg 68:3523571988J Neurosurg 68:

  • 3.

    Colombo FPozza FChierego Get al: Linear accelerator radiosurgery of cerebral arteriovenous malformations: an update. Neurosurgery 34:14211994Neurosurgery 34:

  • 4.

    Fisher WS III: Therapy of AVMs: a decision analysis. Clin Neurosurg 42:2943121995Fisher WS III: Therapy of AVMs: a decision analysis. Clin Neurosurg 42:

  • 5.

    Friedman WABlatt DLBova FJet al: The risk of hemorrhage after radiosurgery for arteriovenous malformations. J Neurosurg 84:9129191996J Neurosurg 84:

  • 6.

    Friedman WABova FJMendenhall WM: Linear accelerator radiosurgery for arteriovenous malformations: the relationship of size to outcome. J Neurosurg 82:1801891995J Neurosurg 82:

  • 7.

    Hamilton MGSpetzler RF: The prospective application of a grading system for arteriovenous malformations. Neurosurgery 34:271994Neurosurgery 34:

  • 8.

    Harbaugh KSHarbaugh RE: Arteriovenous malformations in elderly patients. Neurosurgery 35:5795841994Neurosurgery 35:

  • 9.

    Heros RCKorosue K: Radiation treatment of cerebral arteriovenous malformations. N Engl J Med 323:1271291990N Engl J Med 323:

  • 10.

    Heros RCKorosue KDeibold PM: Surgical excision of cerebral arteriovenous malformations: late results. Neurosurgery 26:5705781990Neurosurgery 26:

  • 11.

    Jennett BBond M: Assessment of outcome after severe brain damage. A practical scale. Lancet 1:4804841975Lancet 1:

  • 12.

    Kaplan ELMeier P: Nonparametric estimation from incomplete observations. J Am Stat Assoc 53:4574811958J Am Stat Assoc 53:

  • 13.

    Lunsford LDKondziolka DFlickinger JCet al: Stereotactic radiosurgery for arteriovenous malformations of the brain. J Neurosurg 75:5125241991J Neurosurg 75:

  • 14.

    Nussbaum ESHeros RCCamarata PJ: Surgical treatment of intracranial arteriovenous malformations with an analysis of cost-effectiveness. Clin Neurosurg 42:3483691995Clin Neurosurg 42:

  • 15.

    Ondra SLTroupp HGeorge EDet al: The natural history of symptomatic arteriovenous malformations of the brain: a 24-year follow-up assessment. J Neurosurg 73:3873911990J Neurosurg 73:

  • 16.

    Pollock BEFlickinger JCLunsford LDet al: Hemorrhage risk after stereotactic radiosurgery of cerebral arteriovenous malformations. Neurosurgery 38:6526591996Neurosurgery 38:

  • 17.

    Pollock BEKondziolka DLunsford LDet al: Repeat stereotactic radiosurgery of arteriovenous malformations: factors associated with incomplete obliteration. Neurosurgery 38:3183241996Neurosurgery 38:

  • 18.

    Pollock BELunsford LDKondziolka Det al: Patient outcomes after stereotactic radiosurgery for “operable” arteriovenous malformations. Neurosurgery 35:181994Neurosurgery 35:

  • 19.

    Sisti MBKader AStein BM: Microsurgery for 67 intracranial arteriovenous malformations less than 3 cm in diameter. J Neurosurg 79:6536601993J Neurosurg 79:

  • 20.

    Spetzler RFMartin NA: A proposed grading system for arteriovenous malformations. J Neurosurg 65:4764831986J Neurosurg 65:

  • 21.

    Steiner LLindquist CAdler JRet al: Clinical outcome of radiosurgery for cerebral arteriovenous malformations. J Neurosurg 77:181992J Neurosurg 77:

  • 22.

    Steiner LLindquist CCail Wet al: Microsurgery and radiosurgery in brain arteriovenous malformations. J Neurosurg 79:6476521993 (Editorial)J Neurosurg 79:

  • 23.

    Sundt TM JrPiepgras DGStevens LN: Surgery for supratentorial arteriovenous malformations. Clin Neurosurg 37:491151991Clin Neurosurg 37:

  • 24.

    Wascher TMSpetzler RF: Radiosurgery of arteriovenous malformations. J Neurosurg 76:104510461992J Neurosurg 76:

  • 25.

    Yamamoto MJimbo MHara Met al: Gamma knife radiosurgery for arteriovenous malformations: long-term follow-up results focusing on complications occurring more than 5 years after irradiation. Neurosurgery 38:9069141996Neurosurgery 38:

  • 26.

    Yamamoto YCoffey RJNichols DAet al: Interim report on the radiosurgical treatment of cerebral arteriovenous malformations. The influence of size, dose, time, and technical factors on obliteration rate. J Neurosurg 83:8328371995J Neurosurg 83:

Article Information

Address for Dr. Pikus: University of Miami School of Medicine, Miami, Florida.Address reprint requests to: Robert E. Harbaugh, M.D., Department of Surgery (Neurosurgery), Dartmouth—Hitchcock Medical Center, 1 Medical Center Drive, Lebanon, New Hampshire 03756.

© AANS, except where prohibited by US copyright law.

Headings

Figures

  • View in gallery

    Graph displaying Kaplan—Meier hemorrhage-free survival estimates following microsurgical extirpation or stereotactic radiosurgery of Spetzler—Martin Grade I to III AVMs. The microsurgical data were obtained from the current series and that of Heros, et al.10 The stereotactic radiosurgical data were obtained from the study by Friedman, et al.5

References

1.

Betti OOMunari CRosler R: Stereotactic radiosurgery with the linear accelerator: treatment of arteriovenous malformations. Neurosurgery 24:3113211989Neurosurgery 24:

2.

Brown RD JrWiebers DOForbes Get al: The natural history of unruptured intracranial arteriovenous malformations. J Neurosurg 68:3523571988J Neurosurg 68:

3.

Colombo FPozza FChierego Get al: Linear accelerator radiosurgery of cerebral arteriovenous malformations: an update. Neurosurgery 34:14211994Neurosurgery 34:

4.

Fisher WS III: Therapy of AVMs: a decision analysis. Clin Neurosurg 42:2943121995Fisher WS III: Therapy of AVMs: a decision analysis. Clin Neurosurg 42:

5.

Friedman WABlatt DLBova FJet al: The risk of hemorrhage after radiosurgery for arteriovenous malformations. J Neurosurg 84:9129191996J Neurosurg 84:

6.

Friedman WABova FJMendenhall WM: Linear accelerator radiosurgery for arteriovenous malformations: the relationship of size to outcome. J Neurosurg 82:1801891995J Neurosurg 82:

7.

Hamilton MGSpetzler RF: The prospective application of a grading system for arteriovenous malformations. Neurosurgery 34:271994Neurosurgery 34:

8.

Harbaugh KSHarbaugh RE: Arteriovenous malformations in elderly patients. Neurosurgery 35:5795841994Neurosurgery 35:

9.

Heros RCKorosue K: Radiation treatment of cerebral arteriovenous malformations. N Engl J Med 323:1271291990N Engl J Med 323:

10.

Heros RCKorosue KDeibold PM: Surgical excision of cerebral arteriovenous malformations: late results. Neurosurgery 26:5705781990Neurosurgery 26:

11.

Jennett BBond M: Assessment of outcome after severe brain damage. A practical scale. Lancet 1:4804841975Lancet 1:

12.

Kaplan ELMeier P: Nonparametric estimation from incomplete observations. J Am Stat Assoc 53:4574811958J Am Stat Assoc 53:

13.

Lunsford LDKondziolka DFlickinger JCet al: Stereotactic radiosurgery for arteriovenous malformations of the brain. J Neurosurg 75:5125241991J Neurosurg 75:

14.

Nussbaum ESHeros RCCamarata PJ: Surgical treatment of intracranial arteriovenous malformations with an analysis of cost-effectiveness. Clin Neurosurg 42:3483691995Clin Neurosurg 42:

15.

Ondra SLTroupp HGeorge EDet al: The natural history of symptomatic arteriovenous malformations of the brain: a 24-year follow-up assessment. J Neurosurg 73:3873911990J Neurosurg 73:

16.

Pollock BEFlickinger JCLunsford LDet al: Hemorrhage risk after stereotactic radiosurgery of cerebral arteriovenous malformations. Neurosurgery 38:6526591996Neurosurgery 38:

17.

Pollock BEKondziolka DLunsford LDet al: Repeat stereotactic radiosurgery of arteriovenous malformations: factors associated with incomplete obliteration. Neurosurgery 38:3183241996Neurosurgery 38:

18.

Pollock BELunsford LDKondziolka Det al: Patient outcomes after stereotactic radiosurgery for “operable” arteriovenous malformations. Neurosurgery 35:181994Neurosurgery 35:

19.

Sisti MBKader AStein BM: Microsurgery for 67 intracranial arteriovenous malformations less than 3 cm in diameter. J Neurosurg 79:6536601993J Neurosurg 79:

20.

Spetzler RFMartin NA: A proposed grading system for arteriovenous malformations. J Neurosurg 65:4764831986J Neurosurg 65:

21.

Steiner LLindquist CAdler JRet al: Clinical outcome of radiosurgery for cerebral arteriovenous malformations. J Neurosurg 77:181992J Neurosurg 77:

22.

Steiner LLindquist CCail Wet al: Microsurgery and radiosurgery in brain arteriovenous malformations. J Neurosurg 79:6476521993 (Editorial)J Neurosurg 79:

23.

Sundt TM JrPiepgras DGStevens LN: Surgery for supratentorial arteriovenous malformations. Clin Neurosurg 37:491151991Clin Neurosurg 37:

24.

Wascher TMSpetzler RF: Radiosurgery of arteriovenous malformations. J Neurosurg 76:104510461992J Neurosurg 76:

25.

Yamamoto MJimbo MHara Met al: Gamma knife radiosurgery for arteriovenous malformations: long-term follow-up results focusing on complications occurring more than 5 years after irradiation. Neurosurgery 38:9069141996Neurosurgery 38:

26.

Yamamoto YCoffey RJNichols DAet al: Interim report on the radiosurgical treatment of cerebral arteriovenous malformations. The influence of size, dose, time, and technical factors on obliteration rate. J Neurosurg 83:8328371995J Neurosurg 83:

TrendMD

Metrics

Metrics

All Time Past Year Past 30 Days
Abstract Views 3 3 3
Full Text Views 36 36 25
PDF Downloads 26 26 24
EPUB Downloads 0 0 0

PubMed

Google Scholar