aNeurosurgical Operative Unit, Padova Hospital, 35100 Padova, ItalybDepartment of Enviromental Medicine and Public Health, University of Padova, 35100 Padova, Italy
Spetzler-Martin grade I-II AVMs consists of smallsuperficial AVMs, small deep noneloquent AVMs, andmiddle-sized superficial noneloquent AVMs. Epidemiologi-cally, they represent about half of the ruptured AVMsreferring to neurosurgical units [15]. If there is a general
consensus about their surgical operability [13,21,30,32,36],the timing of low-grade ruptured AVMs surgical removal isstill controversial [7,11,14]. We reviewed our recentexperience on early surgical treatment of ruptured gradeI-II AVMs to describe its related risks and benefits.
2. Patients and methods
A consecutive series of 27 patients with hemorrhagicgrade I-II AVMs surgically treated by a single neurosurgeon(GP) during the period January 2004-September 2008 within6 days after bleeding were analyzed. All patients showed a
1 F 30 14 Parietal 1 II 52 M 61 8 Frontal 1 II 53 F 37 11 Temporal 1 II 54 M 46 + 14 Cingulate
gyrus3 I 5
5 M 71 + 13 Temporalmesial
6 II 5
6 M 32 7 Rolandic 1 II 47 M 57 8 Temporal 1 I 48 F 17 + 15 Frontal 1 II 59 F 68 + 3 Frontal 1 I 110 M 25 15 Temporal 1 I 511 M 65 10 Cingulate
gyrus1 II 4
12 F 47 + 7 Frontal 1 II 413 M 63 7 Paracallosal 2 II 514 F 15 15 Parietal 5 I 515 F 32 + 3 Frontal 1 I 516 F 27 + 8 Occipital 1 I 517 M 27 14 Frontal 1 I 518 F 28 + 4 Cerebellar 1 II 519 F 30 15 Temporal 3 I 520 M 12 15 Frontal 4 I 521 M 45 + 8 Frontal 1 II 522 F 50 + 4 CP angle 1 II 123 M 31 5 Frontal 1 II 524 F 36 8 Frontal 1 II 425 F 38 + 7 Basal
ganglia1 II 3
26 F 50 3 Cerebellar 1 II 327 M 74 8 Cingulate
gyrus1 I 4
F indicates female; M, male.
663G. Pavesi et al. / Surgical Neurology 72 (2009) 662–667
cerebral AVM on DSA at admission, and surgical removalwas controlled by postoperative DSA. The average length offollow-up is 22 months (3-48 months). There were 14 femaleand 13 male patients, with a mean age of 41.3 years (range12-74 years). Patients demographics and their clinical andradiological features were obtained from the medicalrecords. Neurological grade on admission was rated usingthe GCS; neurological outcome was assessed using GOS(5 = good recovery, 4 = moderate disability, 3 = severedisability, 2 = vegetative state, 1 = dead) according tooutpatient evaluation. The presence or absence of anisocoria
Table 2Correlation between preoperative clinical findings and functional outcome (GOS)
GOS: 1 death, 2 vegetative state, 3 severe disability, 4 moderate disability, 5 good
before surgery and the GCS on admission were correlatedwith GOS using the exact Fisher exact test.
3. Results
There were 27 patients with SM grade I and II AVMsoperated in the acute stage of bleeding. Table 1 summarizesthe most relevant clinical findings. Most patients, 21 (78%),were operated within the first day of bleeding, whereas theothers were operated within the sixth day of bleeding, mainlybecause of delayed diagnosis. Before surgery, 11 patients(41%) showed a GCS greater than 8, and 16 patients (59%) aGCS of 8 or less. Some patients with a GCS score of 8 or lesswere admitted with a pharmacologically induced coma,following protective sedation and intubation performed inemergency department.
Early rebleeding, within 4 hours from the first episode,occurred in 2 patients: both patients showed an initial GCS of14 and a small hematoma on admission CT, beforeneurological worsening; after surgery, at follow-up, 1 patientshowed a moderate disability and the other died. Earlyrebleeding rate was 7.4%.
Associated aneurysms at afferent vessels of the AVMwere observed and treated in 4 cases (15%). At follow-up,17 patients showed a GOS grade 5 (62.9%), 6 patients agrade 4 (22.2%), 2 patients a grade 3 (7.4%), and 2 patientsa grade 1 (7.4%).
No significant correlation of GOS with anisocoria andGCS was found (Table 2).
In 11 patients (40.7%), anisocoria was observed beforesurgery. After surgery, pupils returned isocoric in 9 (81.8%)of 11 patients.
All patients with preoperative GCS greater than 8 showedfavorable results (91%good recovery, 9%moderate disability).
4. Discussion
The timing of surgical treatment of bleeding AVMs in theliterature is far different from that of aneurysms and lessdebated. Early surgery for ruptured AVMs with associatedhematomas is reported only in few articles with small series[11,14,18,27,28,34]. All the reported series of early surgeryon ruptured AVMs concern patients with massive hemato-
Fisherexact test
4 (MD) 5 (GR)
1 75 10 P ≤ .16315 71 10 P ≤ .10506 (22.2%) 17 (63.0%)
recovery.
664 G. Pavesi et al. / Surgical Neurology 72 (2009) 662–667
mas and rapid clinical deterioration. To our knowledge,within these series, even fewer are the cases that report acutesurgical resection of bleeding AVMs in patients without life-threatening hematomas and presenting with a good GCSscore. Puzzilli et al [28] reported 5 patients with GCS 10 to15, who were operated early. Kuhmonen et al [14] reportedthe largest experience on early surgery for ruptured AVMswith 12 patients with HH grade I-II out of a series of 49patients operated within 4 days after hemorrhage: theyshowed that outcome was predicted by the primary severityof bleed rather than SM grade, location of AVM, and size ofintracerebral hematoma.
Grade I-II AVMs are homogeneous from the surgicaltechnique point of view: they include small- or middle-sizedand superficially seated AVMs, and they generally have alow surgical risk, related with their surgical accessibility. Weexcluded higher SM AVMs grades from our early surgicalmanagement because direct surgical risk necessarilyincreases in this less uniform AVM population, andcombined or alternative treatment options make theirmanagement decision more complex [16]. Two main reasonsusually justify a nonaggressive management of brain AVMsin the first days from hemorrhage: (1) AVMs early
Fig. 1. Preoperative CT (A) and carotid angiography (B) showing a small hematomday from having hemorrhage and complaining only of headache; surgery was perfoevacuation and AVM removal; she was discharged neurologically intact 4 days af
rebleeding is not considered as frequent and as catastrophiclike in ruptured aneurysms context; (2) operating an AVMduring the acute phase of bleeding entails management of aswelling brain with a related additional difficulty to obtainsurgical clearness. Our observations suggest that bothreasons are, in the case of grade I-II AVMs, misleading iftaken as exclusion criteria for early surgery. Our small cohortshowed an acute rebleeding rate before treatment of 7.4%within a short window of observational time, given the earlysurgical therapy to which patients were submitted. This‘ultra-early’ rebleeding rate represents around half of the firstyear rebleeding risk that, according to population-based andlarge observational studies data for bleeding AVMs, rangesfrom 4% to 18% [2,4,8,23,19]. However, acute rebleedingrate has not been estimated with care, and early timing oftreatment for bleeding AVMs is usually considered only forpatients presenting with large hematomas or with AVM-related aneurysms [12,22]. In the present experience,rebleeding led to a fatal outcome in 1 patient and to aminor deficit in another patient. Although large observa-tional studies of patients with untreated AVM show a lowclinical impact of AVM hemorrhage and a low risk of earlyrehemorrhage [2,6], other authors point out the clinical
a surrounding a mesial parietal AVM of a 15-year-old girl admitted on fourthrmed the following day. Immediate postoperative findings (C-D) confirm clotter.
665G. Pavesi et al. / Surgical Neurology 72 (2009) 662–667
severity of ruptured AVMs' natural history [1,9,23]. Theearly rebleeding rate we observed, together with the lowsurgical risk associated with small/accessible AVMs[3,5,7,21,25,30,31], and with the long-term efficacy ofsurgical treatment compared with radiosurgery or emboliza-tion in low-grade AVMs [26,30], supports our propensity toadopt an early treatment for all patients with grade I-IIAVMs, regardless of their clinical grade on admission.Similarly to what happens for ruptured aneurysm treatment,where a general agreement is being established toward earlytreatment for HH grade I-III patients with small anteriorcirculation aneurysms, patients in good neurological condi-tions with low-grade AVMs should be approached with thesame curative intentions. In fact, for those 11 patients whoshowed a GCS score greater than 8 at admission, surgeryoffered immediate protection from rebleeding with lowsurgical risk (91% GR, 9% MD) (Fig. 1). On the other hand,regarding patients operated with a GCS score of 8 or less, itis intuitive and confirmed by our results (9 out of 11 ofpatients with preoperative anisocoria showed normal pupils
Fig. 2. Preoperative CT (A) and carotid angiography (B) showing a large hematomaold woman admitted with a GCS score of 5; surgery was urgently performed, andangiogram (D) confirms AVM removal. At follow-up, speech impairment and hem
diameter after surgery) that prompt clot evacuation avoidspoor outcome (GOS 1-2) in most patients with life-threatening hematoma on admission. The observation thata preoperative bad neurological grade and the presence ofanisocoria are not significantly correlated with the functionaloutcome indicates the efficacy of an aggressive treatment inthese patients.
However, if early surgery virtually eliminates the risk ofrebleeding and achieves a prompt brain relaxation, its useremains controversial because brain edema is most severe inthe first days after hemorrhage and this necessitates morebrain retraction, and the risk of intraoperative AVM ruptureis higher. Such arguments against early surgery are so widelyspread in the neurosurgical practice that it is commonlyjustified a multistaged surgical management with hemor-rhagic clot evacuation as first step treatment [10,11,20].However, such management is, in our opinion, obviouslyencumbered by uncontrollable intraoperative bleeding risk ifAVM removal strategy is not considered in the sameoperation. In the particular context of grade I-II AVMs, we
of the basal ganglia caused by rupture of a small lenticular AVM of a 38-year-a CT at 2 weeks (C) shows the residual area of brain damage; postoperativeiparesis are still present (GOS 3).
666 G. Pavesi et al. / Surgical Neurology 72 (2009) 662–667
found no additional difficulties rather than routine manage-ment of an acute hemorrhagic brain, whose relaxation, ifcisternal or ventricular cerebro-spinal fluid subtraction is notsufficient, is even helped by the possibility of evacuating thehematoma that almost invariably surrounds a part of themalformation, thus creating a dissection plane where asurgeon may start his or her microsurgical work [17,33].
In addition, it has been argued that DSA, in the presenceof an intracerebral hematoma, may not totally fill the AVMnidus, thus providing an incomplete imaging diagnosis [35].In the present series, we observed a satisfying correlationbetween DSA and surgical anatomy in all cases, within thelimits of the technique: actually, surgeons must be aware thatsome small feeding arteries in the deep white matter orrunning beneath the major draining veins cannot be easilyidentified on conventional angiograms, but they must beintraoperatively looked for to achieve complete hemostasisof the surgical field. All patients showed complete oblitera-tion of their AVM on postoperative DSA. Although it isreported that hidden AVMs compartments or perinidaldilated capillary network may cause recurrence after totalextirpation, these are autonomous anatomic entities, morefrequently observed in large AVMs with an importanthemodynamic stress rather than in low-grade AVMs [24,29].
Mortality occurred in 2 patients (7.4%): 1 patient died 10days after surgery of pulmonary embolism in the context of asevere neurological deterioration due to early rebleeding inthe preoperative stage; the other developed a brainstemischemia after transtentorial herniation after hemorrhage. Atfollow-up, overall disability was observed in 8 patients(29.6%). If we make a rough comparison between ourmorbidity/mortality figures with AVMs natural history data,indicating that 23% to 33% of patients have a new deficitafter AVMs rupture and 23% to 29% of patients die fromAVM rupture [1,23], we may conclude that early surgicaltreatment of grade I-II bleeding AVMs has reductedmortality in our patients, whereas little can be done inreducing hemorrhage-linked disability (Fig. 2).
5. Conclusions
Early surgical management of grade I-II hemorrhagicAVMs in the presence of microneurovascular skills notdifferent from those demanded by other AVMs surgeryor aneurysm clipping leads to favorable results, mini-mizing the risk of rebleeding, reducing the time ofhospital stay, and allowing a more rapid rehabilitativecourse whenever necessary.
References
[1] Brown Jr RD, Wiebers DO, Forbes G, et al. The natural history ofunruptured intracranial arteriovenous malformations. J Neurosurg1988;68:352-7.
[2] Choi JH, Mast H, Sciacca RR. Clinical outcome after recurrenthemorrhage in patients with untreated brain arteriovenous malforma-tion. Stroke 2006;37:1243-7.
[3] Deruty R, Pelissou-Guyotat I, Morel C, et al. Reflections on themanagement of cerebral arteriovenous malformations. Surg Neurol1998;50:245-56.
[4] Halim AX, Johnston SC, Singh V. Longitudinal risk of intracranialhemorrhage in patients with arteriovenous malformation of the brainwithin a defined population. Stroke 2004;35:1697-702.
[5] Hamilton MG, Spetzler RF. The prospective application of agrading system for arteriovenous malformations. Neurosurgery1994;34:2-7.
[6] Hartmann A, Mast H, Mohr JP. Morbidity of intracranial hemorrhagein patients with cerebral arteriovenous malformation. Stroke 1998;29:931-4.
[7] Hashimoto N, Nozaki K, Takagi Y, et al. Surgery of cerebralarteriovenous malformations. Neurosurgery 2007;61(1 Suppl):375-89.
[8] Hernesniemi J, Dashti R, Juvela S. Natural history of brainarteriovenous malformations: a long term follow-up study of risk ofhemorrhage in 238 patients. Neurosurgery 2008;63:823-31.
[9] Hillman J. Population-based analysis of arteriovenous malformationtreatment. J Neurosurg 2001;95(4):633-7.
[10] Iwama T, Yoshimura K, Imhof HG, et al. Emergency craniotomy forintraparenchymal massive hematoma after embolization of supraten-torial arteriovenous malformations. Neurosurgery 2003;53:1251-60.
[12] Meisel HJ, Mansmann U, Alvarez H, et al. Cerebral arteriovenousmalformations and associated aneurysms: analysis of 305 cases from aseries of 662 patients. Neurosurgery 2000;46(4):793-802.
[13] Kobayashi S, Hongo K, Bhagwati SN, et al. Management ofarteriovenous malformations: part II. Surg Neurol 1997;48:2-6.
[14] Kuhmonen J, Piippo A, Vaart K, et al. Early surgery for rupturedcerebral arteriovenous malformations. Acta Neurochir Suppl 2005;94:111-4.
[15] Laakso A, Dashti R, Seppanen J, et al. Long-term excess mortality in623 patients with brain arteriovenous malformations. Neurosurgery2008;63(2):244-55.
[16] Lawton MT. UCSF Brain Arteriovenous Malformation Study Project.Spetzler-Martin Grade III arteriovenous malformations: surgical resultsand a modification of the grading scale. Neurosurgery 2003;52:740-9.
[17] Lawton MT, Du R, Tran MN, et al. Effect of presenting hemorrhage onoutcome after microsurgical resection of brain arteriovenous mal-formations. Neurosurgery 2005;56:485-93.
[18] Liang X, Liu Z, Wang H, et al. Surgical strategy for cerebralarteriovenous malformation with acute surgery. Neural Regen Res2007;2(7):416-20.
[19] Mast H, Young WL, Koennecke HC. Risk of spontaneous haemor-rhage after diagnosis of cerebral arteriovenous malformation. Lancet1997;350:1065-8.
[20] Meyer PG, Orliaguet GA, Zerah M, et al. Emergency management ofdeeply comatose children with acute rupture of cerebral arteriovenousmalformations. Can J Anesth 2000;47(8):758-66.
[21] Morgan MK, Rochford AM, Tsahtsarlis A. Surgical risks associatedwith the management of Grade I and II brain arteriovenousmalformations. Neurosurgery 2004;54(4):832-9.
[22] Nataf F, Schlienger M, Bayram M, et al. Microsurgery or radiosurgeryfor cerebral arteriovenous malformations? A study of two paired series.Neurosurgery 2007;61:39-50.
[23] Ondra SL, Troupp H, George ED, et al. The natural history ofsymptomatic arteriovenous malformations of the brain: a 24-yearfollow-up assessment. J Neurosurg 1990;73:387-91.
[24] Pellettieri L, Svendsen P, Wikholm G, et al. Hidden compartments inAVMs—a new concept. Acta Radiol 1997;38(1):2-7.
[25] Pik JH, Morgan MK. Microsurgery for small arteriovenous malforma-tions of the brain: results in 110 consecutive patients. Neurosurgery2000;47(3):571-7.
[26] Pikus HJ, Beach ML, Harbaugh RE. Microsurgical treatment ofarteriovenous malformations: analysis and comparison with stereo-tactic radiosurgery. J Neurosurg 1998;88(4):641-6.
667G. Pavesi et al. / Surgical Neurology 72 (2009) 662–667
[27] Purdy PD, Batjer HH, Sampson D. Management of hemorrhagiccomplications from pre-operative embolization of arteriovenousmalformations. J Neurosurg 1991;74:205-11.
[28] Puzzilli F, Mastronardi L, Ruggeri A, et al. Early surgical treatment ofintracerebral hemorrhages caused by AVM: our experience in 10 cases.Neurosurg Rev 1998;21:87-92.
[29] Sato S, Kodama N, Sasaki T, et al. Perinidal dilated capillary networksin cerebral arteriovenous malformations. Neurosurgery 2004;54:163-8.
[30] Schaller C, Schramm J. Microsurgical results for small arteriovenousmalformations accessible for radiosurgical or embolization treatment.Neurosurgery 1997;40(4):664-74.
[31] Sisti MB, Kader A, Stein BM. Microsurgery for 67 intracranialarteriovenous malformations less than 3 cm in diameter. J Neurosurg1993;79:653-60.
[32] Spetzler RF, Martin NA. A proposed grading system for arteriovenousmalformations. J Neurosurg 1986;65:476-83.
[33] Steudel WI, Hopp G, Hacker H. The significance of the size of anintracerebral hematoma for treatment and prognosis. Adv Neurosurg1983;11:99-104.
[34] Steudel WI, Lorenz R, Berkefeld J. Acute operation on arteriovenousmalformation with hematomas—report on six cases. Neurochirurgia(Stuttg) 1992;35(1):26-30.
[35] Willinsky RA, Fitzgerald M, TerBrugge K, et al. Delayed angiographyin investigation of intracerebral hematomas caused by small arter-iovenous malformations. Neuroradiology 1993;35(4):307-11.
[36] Zhao J, Wang S, Li J, et al. Clinical characteristics and surgical resultsof patients with cerebral arteriovenous malformations. Surg Neurol2005;63(2):156-61.
