CLINICAL STUDIES

BYPASS SURGERY FOR COMPLEX BRAIN ANEURYSMS: AN ASSESSMENT OF INTRACRANIALINTRACRANIAL BYPASS
Nader Sanai, M.D.
Department of Neurological Surgery, University of California at San Francisco, San Francisco, California

Zsolt Zador, M.D.
Department of Neurological Surgery, University of California at San Francisco, San Francisco, California

Michael T. Lawton, M.D.
Department of Neurological Surgery, University of California at San Francisco, San Francisco, California Reprint requests: Michael T. Lawton, M.D., Department of Neurological Surgery, University of California at San Francisco, 505 Parnassus Avenue, M779, Box 0520, San Francisco, CA 94143-0520. Email: lawtonm@neurosurg.ucsf.edu Received, September 21, 2008. Accepted, March 17, 2009. Copyright © 2009 by the Congress of Neurological Surgeons

OBJECTIVE: Bypass surgery for brain aneurysms is evolving from extracranial-intracranial (EC-IC) to intracranial-intracranial (IC-IC) bypasses that reanastomose parent arteries, revascularize efferent branches with in situ donor arteries or reimplantation, and reconstruct bifurcated anatomy with grafts that are entirely intracranial. We compared results with these newer IC-IC bypasses to conventional EC-IC bypasses. METHODS: During a 10-year period, 82 patients underwent bypass surgery as part of their aneurysm management. A quarter of the patients presented with ruptured aneurysms and two-thirds presented with compressive symptoms from unruptured aneurysms. Most aneurysms (82%) had non-saccular morphology and 56% were giant sized. Common locations included the cavernous internal carotid artery (23%), middle cerebral artery (20%), and posteroinferior cerebellar artery (12%). RESULTS: Forty-seven patients (57%) received EC-IC bypasses and 35 patients (43%) received IC-IC bypasses, including 9 in situ bypasses, 6 reimplantations, 11 reanastomoses, and 9 intracranial grafts. Aneurysm obliteration rates were comparable in ECIC and IC-IC bypass groups (97.9% and 97.1%, respectively), as were bypass patency rates (94% and 89%, respectively). Three patients died (surgical mortality, 3.7%), and 4 patients were permanently worse as a result of bypass occlusions (neurological morbidity, 4.9%). At late follow-up (mean duration, 41 months), good outcomes (Glasgow Outcome Scale score 5 or 4) were measured in 68 patients (90%) overall, and were similar in EC-IC and IC-IC bypass groups (91% and 89%, respectively). Changes in Glasgow Outcome Scale score were slightly more favorable with IC-IC bypass (6% worse or dead after IC-IC bypass versus 14% with EC-IC bypass). CONCLUSION: IC-IC bypasses compare favorably to EC-IC bypasses in terms of aneurysm obliteration rates, bypass patency rates, and neurological outcomes. IC-IC bypasses can be more technically challenging to perform, but they do not require harvest of extracranial donor arteries, spare patients a neck incision, shorten interposition grafts, are protected inside the cranium, use caliber-matched donor and recipient arteries, and are not associated with ischemic complications during temporary arterial occlusions. IC-IC bypass can replace conventional EC-IC bypass with more anatomic reconstructions for selected aneurysms involving the middle cerebral artery, posteroinferior cerebellar artery, anterior cerebral artery, and basilar apex.
KEY WORDS: Aneurysm, Clipping, Extracranial-to-intracranial bypass, In situ bypass, Intracranial-to-intracranial bypass, Reanastomosis, Reimplantation, Revascularization
Neurosurgery 65:670–683, 2009
DOI: 10.1227/01.NEU.0000348557.11968.F1

www.neurosurgery-online.com

ABBREVIATIONS: ACA, anterior cerebral artery; ATA, anterior temporal artery; BTO, balloon test occlusion; EC-IC, extracranial-intracranial; GOS, Glasgow Outcome Scale; ICA, internal carotid artery; IC-IC, intracranial-intracranial; MCA, middle cerebral artery; OA, occipital artery; PCA, posterior cerebral artery; PICA, posteroinferior cerebellar artery; SCA, superior cerebellar artery; STA, superficial temporal artery; VA, vertebral artery

670 | VOLUME 65 | NUMBER 4 | OCTOBER 2009

www.neurosurgery-online.com

INTRACRANIAL-INTRACRANIAL BYPASS FOR COMPLEX ANEURYSMS

D

espite its well-publicized failure to benefit patients with ischemic stroke in the extracranial-intracranial (EC-IC) Bypass Trial, EC-IC bypass surgery has been essential in the management of brain aneurysms that are too complex for conventional clipping or endovascular coiling (2, 4, 8, 11, 22, 27). Revascularization of a territory distal to an unclippable, giant, dolichoectatic, or thrombotic aneurysm enables the aneurysm to be occluded without risk of ischemic complications, or the parent artery’s blood flow to be reversed or reduced safely. Superficial temporal artery (STA)-to-middle cerebral artery (MCA) bypass was the prototype, and subsequently an array of bypasses was developed with the same concept of redirecting extracranial blood flow from scalp arteries or cervical carotid arteries to the brain, either directly with 1 anastomosis or with interposition grafts and 2 anastomoses. In recent years, innovative bypasses have been introduced anecdotally that revascularize intracranial arteries with other intracranial arteries, without extracranial donor arteries (6, 7, 9, 15, 16, 19). These intracranialintracranial (IC-IC) bypasses are simple, elegant, and more anatomic than their EC-IC counterparts. IC-IC bypasses require no harvest of extracranial donors, spare patients a neck incision, shorten any interposition grafts, are protected within the cranium, and use caliber-matched donor and recipient arteries. These advantages of IC-IC bypasses appeal to experienced bypass surgeons, and the use of these bypasses has increased noticeably. Sekhar et al. (26) performed at least 11 IC-IC bypasses in an overall experience with 119 bypasses in 115 patients. Lawton et al. (15) performed 28 IC-IC bypasses (44%) in an overall experience with 63 bypasses in 61 patients. The development of IC-IC bypasses represents an evolution of bypass surgery for brain aneurysms. However, IC-IC bypass results have not been compared with conventional EC-IC bypass results in a large clinical experience. We have embraced IC-IC bypasses in our aneurysm practice, and therefore reviewed our experience with bypass surgery for complex aneurysms in 82 patients. We hypothesized that IC-IC bypasses would compare favorably to EC-IC bypasses in terms of efficacy, safety, and patient outcomes, and that these data would support the trend toward intracranial arterial reconstruction.

Medical records, radiographic studies, operative reports, intraoperative photographs, neurological course, and clinical follow-up evaluations were reviewed retrospectively. Neurological outcomes were assessed using the Glasgow Outcome Scale (GOS). A clinical nurse or clinician not directly involved in the care of these patients performed all outcome assessments preoperatively, early postoperatively (6 weeks), and at late follow-up.

Patients
During a 10-year period between November 1997 and November 2007, 1984 aneurysms were treated microsurgically in 1578 patients by the senior author (MTL). Of these patients, 82 (5%) underwent cerebral revascularization as part of their aneurysm management; 50 patients were women and 32 men, with a mean age of 53 years (age range, 12–78 years) (Table 1). Twenty-one patients presented with subarachnoid hemorrhage (26%). Hunt and Hess grade III was the most common clinical grade (48%), and 4 patients presented with poor Hunt and Hess grade. Fifty-six patients (68%) presented with unruptured aneurysms and neurological symptoms, with cranial neuropathy or hemiparesis from mass effect present in 38 patients (68%). Eight patients (14%) presented with transient ischemic attacks or stroke in association with thrombotic aneurysms. Five patients presented with incidental, unruptured aneurysms (6%).

Aneurysm Characteristics
The most common locations were cavernous ICA (19 aneurysms, 23%), MCA (16 aneurysms, 20%), and posteroinferior cerebellar artery (PICA) (10 aneurysms, 12%) (Table 2). Forty-six aneurysms (56%) were giant in size, and 67 (82%) had fusiform or dolichoectatic morphology. Thirty-one patients (38%) had thrombotic aneurysms. Eight aneurysms (10%) had been treated endovascularly with coils, of which 3 were incompletely treated and 5 were recurrent. Fifteen patients (19%) had multiple aneurysms, with 26 other aneurysms diagnosed.

Indications
Bypasses were performed only when conventional clipping failed, as a result of complex anatomy, large or giant size, dolichoectatic morphology, intraluminal thrombus, or atherosclerotic tissue at the neck. Balloon test occlusion (BTO) was used to select 26 patients, all of whom had cavernous or supraclinoid ICA aneurysms, for aneurysm management with a bypass. Ten patients failed the test with balloon inflation alone, and 16 patients failed with additional hypotensive challenge (lowering mean arterial pressure with nitroprusside drip by 20 mm Hg, or 25% of mean arterial pressure, whichever was greater). Failed BTO was used as an indication for bypass. High-flow bypass was used in patients who failed BTO immediately, and low-flow bypass was used in patients who failed BTO after hypotensive challenge. The decision to perform a bypass with aneurysms in other locations was based on the aneurysm’s unclippability and patients’ angiographic anatomy (presence or absence of collateral circulation from the circle of Willis or leptomeningeal connections).

PATIENTS AND METHODS
Study Design
The study was approved by the Institutional Review Board and conducted in compliance with Health Insurance Portability and Accountability Act regulations. The prospectively maintained database for the Cerebrovascular Service at the University of California–San Francisco was searched for patients who had bypass surgery as part of their aneurysm treatment. Patients were divided into 2 groups according to type of bypass. EC-IC bypass involved donor arteries from external carotid artery branches (STA and occipital artery [OA]), cervical carotid arteries (common carotid artery, internal carotid artery [ICA], and external carotid artery), or other extracranial arteries (e.g., subclavian artery). IC-IC bypass involved intracranial donor arteries, and were further categorized as in situ bypass (adjacent donor artery), reimplantation (aneurysm branch artery onto parent artery), reanastomosis (primary repair of parent artery), and intracranial bypass graft (graft interposed between donor and recipient arteries).

Surgical Technique

In Situ Bypass
In situ bypass requires donor and recipient arteries that lie parallel and in close proximity to one another. Four sites have this anatomy: MCA branches (M2 and M3 segments) and anterior temporal artery (ATA) as they course through the sylvian fissure (Fig. 1); bilateral anterior cerebral arteries (ACA) as they course through the interhemi-

NEUROSURGERY

VOLUME 65 | NUMBER 4 | OCTOBER 2009 | 671

SANAI ET AL.

TABLE 1. Study demographicsa Total
Total no. of patients Median age (range), y Female sex Patients with ruptured aneurysm 82 53 (12–78) 50 (61%)

TABLE 2. Aneurysm demographicsa EC-IC bypass
47 (57%) 57 (12–77) 35 (74%)

IC-IC bypass
35 (43%) 45 (15–78) 15 (43%)

Aneurysm location
Internal carotid artery

Total

EC-IC bypass
17 (36%) 9 (19%) 3 (6%) 2 (4%) 7 (15%) 0 (0%) 0 (0%) 0 (0%) 2 (4%) 2 (4%) 1 (2%) 4 (9%)

IC-IC bypass
2 (6%) 0 (0%) 0 (0%) 0 (0%) 9 (26%) 3 (9%) 2 (6%) 1 (3%) 0 (0%) 0 (0%) 1 (3%) 4 (11%)

Cavernous Supraclinoid Ophthalmic Terminus

19 (23%) 9 (11%) 3 (4%) 2 (2%) 16 (20%) 3 (4%) 2 (2%) 1 (1%) 2 (2%) 2 (2%) 2 (2%) 8 (10%)

All Grade I Grade II Grade III Grade IV Grade V
Patients with unruptured aneurysm (symptomatic)

21 (26%) 6 (29%) 1 (5%) 10 (48%) 4 (19%) 0 (0%)

8 (17%) 2 (25%) 1 (13%) 3 (38%) 2 (25%) 0 (0%)

13 (37%) 4 (31%) 0 (0%) 7 (54%) 2 (15%) 0 (0%)

Middle cerebral artery

Total
Anterior cerebral artery

Anterior communicating artery Pericallosal Distal ACA
Basilar artery

All Headache Cranial neuropathy Hemiparesis Seizure TIA (stroke) Hydrocephalus Incomplete coiling or recurrence
Patients with unruptured aneurysm (incidental)

56 (68%) 22 (39%) 29 (52%) 9 (16%) 3 (5%) 8 (14%) 4 (7%) 7 (13%)

37 (79%) 15 (41%) 22 (59%) 6 (16%) 2 (5%) 4 (11%) 2 (5%) 2 (5%)

19 (54%) 7 (37%) 7 (37%) 3 (16%) 1 (5%) 4 (21%) 2 (11%) 5 (26%)

Basilar bifurcation Posterior cerebral artery Superior cerebellar artery Basilar trunk
Posterior inferior cerebellar artery

PICA Vertebral artery
Total
a

10 (12%) 3 (4%) 82 (100%)

0 (0%) 0 (0%) 47 (100%)

10 (29%) 3 (9%) 35 (100%)

All
Aneurysm characteristics

5 (6%) 23.8 (4–60) 15 (18%) 67 (82%) 46 (56%) 31 (38%) 4.5

2 (4%) 28.5 (6–60) 10 (21%) 37 (79%) 34 (72%) 18 (38% 4.7

3 (9%) 17.2 (4–49) 5 (14%) 30 (86%) 12 (34%) 13 (37%) 4.4

EC-IC, extracranial-intracranial; IC-IC, intracranial-intracranial; ACA, anterior cerebral artery; PICA, posteroinferior cerebellar artery.

Mean diameter (range), mm Saccular Fusiform, dolichoectatic Giant Thrombotic
Outcome measure

Mean preoperative GOS score
a

and a stacked straight clip closing the fenestration). In cases in which clip reconstruction fails, the neck can be clipped to exclude the aneurysm, preserve the parent artery, and sacrifice the branch artery. The occluded branch artery can then be reconstituted with reimplantation onto the parent artery. Alternatively, the branch artery can be reimplanted to an adjacent donor artery that is not the parent artery, as long as that donor artery lies near the branch. Similar to in situ bypasses, this favorable anatomy occurs with MCA, ACA, and PICA aneurysms. Reimplantation requires 1 end-to-side anastomosis.

EC-IC, extracranial-intracranial; IC-IC, intracranial-intracranial; TIA, transient ischemic attack; GOS, Glasgow Outcome Scale.

Reanastomosis
Reanastomosis requires trapping the aneurysm, completely detaching afferent and efferent arteries, and reconnecting cut ends with an end-toend anastomosis. This technique works well with fusiform aneurysms that are small or medium in size. Saccular aneurysms at bifurcations are difficult to reconstruct with primary reanastomosis because the second efferent branch must either be reimplanted or bypassed with an extracranial donor artery. Large and giant aneurysms may be difficult to reanastomose because ends of parent artery are separated after excising an aneurysm. Mobilizing the ends of afferent and efferent arteries may enable the first stitch to pull them together with minimal tension. If the gap in the parent artery is too long and the tension too great, the suture will tear through the artery wall as it is tightened and ruin the repair. Some large aneurysms in PICA and MCA territories have a redundant parent artery that will allow primary reanastomosis despite their size. Reanastomosis requires 1 end-to-end anastomosis.

spheric fissure over the corpus callosum (A3 and A4 segments) (Fig. 2); posterior cerebral artery (PCA) (P2 and P3 segments) and superior cerebellar artery (SCA) as they course through the ambient cistern around the cerebral peduncle (Fig. 3); and bilateral PICAs as they course through the cisterna magna to meet behind the medulla underneath the cerebellar tonsils (Fig. 4). In situ bypasses require 1 side-toside anastomosis.

Reimplantation
Complex aneurysms with branches that originate from the aneurysm base can often be reconstructed with tandem clipping techniques that preserve branch arteries (a fenestrated clip encircling the branch origin

672 | VOLUME 65 | NUMBER 4 | OCTOBER 2009

www.neurosurgery-online.com

INTRACRANIAL-INTRACRANIAL BYPASS FOR COMPLEX ANEURYSMS

A

A

B

C

B

C

D

E

D

E

FIGURE 1. Intracranial-intracranial (IC-IC) bypass techniques for middle cerebral artery (MCA) aneurysms. A, overview of MCA anatomy. B, in situ bypass (anterior temporal artery [ATA]-MCA). C, reanastomosis (MCA). D, reimplantation (MCA-MCA). E, intracranial bypass graft (anterior cerebral artery [ACA]-MCA double reimplantation). ICA, internal carotid artery; ST, superior trunk; IT, inferior trunk; RAG, radial artery graft.

FIGURE 2. IC-IC bypass techniques for anterior cerebral artery (ACA) aneurysms. A, overview of MCA anatomy. B, in situ bypass (A3-A3). C, reanastomosis (ACA). D, reimplantation (pericallosal-callosomarginal). E, intracranial bypass graft (ACA double implantation). ICA, internal carotid artery; MCA, middle cerebral artery; PcaA, pericallosal artery; CmaA, callosomarginal artery; L, left; and R, right; RAG, radial artery graft.

Intracranial Bypass With Grafts
Bypasses with interposition grafts connect donor and recipient arteries that are entirely intracranial, differentiating them from EC-IC bypasses that use extracranial donors. In contrast to EC-IC bypasses with saphenous vein grafts spanning from the neck to the sylvian fissure, intracranial bypass grafts are shorter and radial artery grafts are sufficiently long. Radial artery grafts are preferred over saphenous vein grafts because they are composed of arterial tissue, have higher long-term patency rates, and match the caliber of intracranial arteries. A preoperative Allen test with Doppler ultrasound ensures adequate perfusion of the hand with ulnar artery and a competent palmar arch. Intraoperatively, the forearm is accessed for harvest more easily than the thigh, particularly when the patient is positioned laterally or prone for posterior circulation aneurysms. Vasospasm in radial artery grafts has been described, but can be avoided by using pressure distension to

dilate the graft before implantation, and by bathing the graft in a mixture of nitroprusside and heparin. Unlike other IC-IC techniques, intracranial bypass grafts require at least 2 anastomoses, which can be end-to-side, end-to-end, or side-to-side. Anastomoses are planned to minimize brain ischemia during the time that intracranial arteries are temporarily occluded and sutured.

RESULTS
Aneurysm Treatment
Forty-seven patients (57%) received EC-IC bypasses and 35 patients (43%) received IC-IC bypasses. EC-IC bypasses included 16 low-flow bypasses with STA donors in 15 patients and an OA donor in 1 patient (Table 3). High-flow EC-IC

NEUROSURGERY

VOLUME 65 | NUMBER 4 | OCTOBER 2009 | 673

SANAI ET AL.

A

B

A

C

D

B

C

E

F

D

E

FIGURE 3. IC-IC bypass techniques for basilar artery apex aneurysms. A, overview of basilar artery apex anatomy. B, in situ bypass (superior cerebellar artery [SCA]–posterior cerebral artery [PCA]). C, reanastomosis (PCA). D, reimplantation (anterior temporal artery [ATA]-SCA). E, intracranial bypass graft (vertebral artery [V A]-SCA bypass). F, intracranial bypass (middle cerebral artery [MCA]-PCA bypass). ICA, internal carotid artery; ACA, anterior cerebral artery; BA, basilar artery; CN III, oculomotor nerve.

FIGURE 4. IC-IC bypass techniques for posterior inferior cerebellar artery (PICA) aneurysms. A, overview of PICA anatomy. B, in situ bypass (PICA-PICA). C, reanastomosis (PICA). D, reimplantation (PICA– vertebral artery [V A]). E, intracranial bypass graft (V A-PICA bypass). BA, basilar artery; L, left; R, right; RAG, radial artery graft.

bypasses were performed in 31 patients using saphenous vein grafts in 27 patients and radial artery grafts in 4 patients. IC-IC bypasses consisted of in situ bypasses in 9 patients (26%), reimplantation in 6 patients (17%), reanastomosis in 11 patients (31%), and intracranial bypass grafts in 9 patients (26%) (Table 4). Unlike extracranial bypass grafts, intracranial bypass grafts used the radial artery more frequently than the saphenous vein (6 patients versus 3 patients, respectively). Seven patients had complex bypass configurations (Table 5), of which 5 were categorized as IC-IC bypass patients and 2 as EC-IC bypass patients. Aneurysm location influenced bypass design. Overall, 31 of 33 patients with ICA aneurysms underwent EC-IC bypass, which was easier than IC-IC bypass with the petrous-to-supraclinoid ICA bypass. Similarly, 9 of 14 aneurysms involving the basilar artery apex or trunk were managed with EC-IC bypass

because of their deep location and limited proximal donor sites. IC-IC bypasses for basilar artery aneurysms included MCAPCA bypass with radial artery grafts (2 patients), vertebral artery (VA)-SCA bypass (1 radial artery graft and 1 saphenous vein graft), and ATA-SCA reimplantation (1 patient). In contrast to ICA and basilar artery aneurysms, ACA and PICA aneurysms were revascularized exclusively with IC-IC bypasses. The distal ACA territory was far removed from extracranial donor arteries, making IC-IC bypasses more appealing. IC-IC bypass options were numerous with PICA aneurysms and eliminated the tedious dissection required to harvest the OA. Bypasses for MCA aneurysms were split between EC-IC (7 patients) and IC-IC bypasses (9 patients). Approach depended on aneurysm location. ICA aneurysms were approached through a pterional craniotomy (29 patients),

674 | VOLUME 65 | NUMBER 4 | OCTOBER 2009

www.neurosurgery-online.com

INTRACRANIAL-INTRACRANIAL BYPASS FOR COMPLEX ANEURYSMS

TABLE 3. Extracranial-intracranial bypassesa Technique Low-flow bypass No. of patients 9 3 3 0 1 0 6 16 5 3 0 1 0 Graft Flow Anastomosis 1 1 1 1 1 1 2 2 2 2 2 2 2 Technique E-S E-S E-S E-S E-S E-S E-S E-S, E-E E-S, E-E E-S E-S E-S, E-E E-S, E-E

TABLE 4. Intracranial-intracranial bypassesa Technique In situ bypass No. of patients 1 1 2 0 5 1 1 1 3 5 1 5 2 0 1 1 1 2 2 Graft Flow Anastomosis 1 1 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 Technique S-S S-S S-S S-S S-S E-S E-S E-S E-S E-E E-E E-E E-S E-S E-S E-S E-S E-S E-S

STA-MCA STA-PCA STA-SCA STA-AICA OA-PCA OA-PICA
High-flow bypass

No No No No No No Yes Yes Yes Yes Yes Yes Yes

Low Low Low Low Low Low High High High High High High High

ATA-MCA MCA-MCA ACA-ACA PCA-SCA PICA-PICA
Reimplantation

No No No No No No No No No No No No Yes Yes Yes Yes Yes Yes Yes

Low Low Low Low Low Low Low Low Low Low Low Low High High High High High High High

MCA-MCA PC-CM ATA-SCA PICA-VA
Reanastomosis

CCA-MCA ECA-MCA ICA-MCA Subclavian-MCA CCA-PCA/SCA ECA-PCA/SCA ICA-PCA/SCA

MCA ACA PICA
IC bypass graft

a STA, superficial temporal artery; MCA, middle cerebral artery; E-S, end-to-side; PCA, posterior cerebral artery; AICA, anterior inferior cerebellar artery; OA, occipital artery; PICA, posterior inferior cerebellar artery; CCA, common carotid artery; ECA, external carotid artery; E-E, end-to-end; ICA, internal carotid artery; SCA, superior cerebellar artery.

Petrous–supraclinoid ICA ICA-MCA ACA-MCA MCA-ACA ACA-ACA MCA-PCA VA-SCA
a

and an orbitozygomatic craniotomy was used for additional exposure with giant aneurysms (4 patients). Similarly, a pterional craniotomy was adequate for MCA aneurysms (14 patients), and an orbitozygomatic craniotomy was used for giant aneurysms (2 patients). ACA aneurysms were exposed through bifrontal craniotomies, with the midline of the head positioned parallel to the floor and angled up 45 degrees to allow gravity to retract the dependent hemisphere. All bypasses for basilar apex aneurysms were performed through orbitozygomatic craniotomies. A VASCA bypass for 2 basilar trunk aneurysms was performed through a combined far lateral–subtemporal craniotomy. One distal OA-PCA bypass was performed through a torcular craniotomy. PICA bypasses were performed through far lateral craniotomies, although the PICA-PICA bypass did not require resection of occipital condyle or much lateral exposure when the aneurysm was not accessed, as with a staged endovascular occlusion (3 patients). Brain relaxation was achieved with mannitol (1 g/kg) and cerebrospinal fluid drainage through a ventriculostomy, fenestration in the lamina terminalis, or dissection into a subarachnoid cistern. During the anastomosis when parent arteries were temporarily occluded, mild hypothermia and barbiturateinduced electroencephalographic burst suppression were used to increase tolerance to ischemia. The average intracranial cross-clamp time was 46 minutes (range, 32–63 minutes) for EC-IC bypass and 46 minutes (range, 26–76 minutes) for IC-IC bypass. Changes in somatosensory evoked potentials or the

ATA, anterior temporal artery; MCA, middle cerebral artery; S-S, side-to-side; ACA, anterior cerebral artery; PCA, posterior cerebral artery; SCA, superior cerebellar artery; PICA, posterior inferior cerebellar artery; E-S, end-to-side; PC, pericallosal artery; CM, callosomarginal artery; VA, vertebral artery; E-E, end-to-end; ICA, internal carotid artery.

electroencephalogram were rare, and were managed by increasing blood pressure with pressor agents. Heparin irrigation was used liberally in the surgical field during the anastomosis, but systemic heparin was not used. Aneurysm occlusion was performed during surgery in 54 patients and consisted of 27 aneurysm trappings (33%), 16 proximal occlusions (20%), 6 distal occlusions (7%), and 5 aneurysm clippings (6%) (Table 6). The remaining 28 patients (34%) underwent staged endovascular aneurysm occlusion. Twenty-two of these patients were in the EC-IC bypass group, reflecting the large number of ICA aneurysms. In contrast, half of the aneurysms in the IC-IC bypass group were trapped during surgery, reflecting the accessibility of these more distally located aneurysms. The small number of aneurysms clipped directly reflects the nonsaccular morphology of these aneurysms. Endovascular staging was typically performed 2 to 3 days after the bypass procedure. Patients were started on aspirin (350 mg/d) immediately after surgery.

NEUROSURGERY

VOLUME 65 | NUMBER 4 | OCTOBER 2009 | 675

SANAI ET AL.

TABLE 5. Complex bypass techniquesa EC-IC bypass group Double reimplantation IC-IC bypass group 0 1 1 1 1

TABLE 6. Operative and clinical outcomes after extracranialintracranial and intracranial-intracranial bypassesa Total
Bypassed aneurysms Aneurysm occlusion 82 5 (6%) 27 (33%) 16 (20%) 6 (7%) 28 (34%) 80 (98%) 75 (91%) 82 3 (4%) 4 (5%)

EC-IC bypass
47 2 (4%) 10 (21%) 11 (23%) 2 (4%) 22 (47%) 46 (98%) 44 (94%) 47 3 (6%) 1 (2%)

IC-IC bypass
35 3 (9%) 17 (49%) 5 (14%) 4 (11%) 6 (17%) 34 (97%) 31 (89%) 35 0 (0%) 3 (9%)

ECA-MCA-MCA ACA-MCA-MCA ACA-PC-CM
IC-IC EC-IC

1 0 0 0 0

Clipping Trapping Proximal occlusion Distal occlusion Endovascular occlusion
Aneurysm obliteration Bypass patency No. of patients

MCA-MCA reanastomosis + STA-MCA ATA-MCA + STA-MCA
Reanastomosis with interposition STA

MCA-STA-MCA
Double EC-IC

0 1

1 0

Surgical mortality Transient neurological morbidity Late outcome

STA-MCA, double barrel
a

EC-IC, extracranial-intracranial; IC-IC, intracranial-intracranial; ECA, external carotid artery; MCA, middle cerebral artery; ACA, anterior cerebral artery; PC, pericallosal artery; CM, callosomarginal artery; STA, superficial temporal artery; ATA, anterior temporal artery.

GOS score 5 GOS score 4 GOS score 3 GOS score 2 Dead Total Lost
Change in GOS score at late follow-up

59 (78%) 9 (12%) 4 (5%) 0 (0%) 4 (5%) 76 3

34 (81%) 4 (10%) 1 (2%) 0 (0%) 3 (7%) 42 2

25 (74%) 5 (15%) 3 (9%) 0 (0%) 1 (3%) 34 1

Aneurysm Outcomes
Angiography was performed after surgery in all patients to evaluate bypass patency and aneurysm exclusion. Overall, 80 of 82 aneurysms were obliterated angiographically (97.6% obliteration rate). All 5 clipped and 27 trapped aneurysms were completely excluded. Of the 16 aneurysms that were proximally occluded, 15 were angiographically occluded postoperatively. Of the 6 aneurysms that were distally occluded, 5 were angiographically occluded postoperatively. Two fusiform aneurysms were filling angiographically after surgery, but had new intraluminal thrombosis, smaller angiographic size, and reduced flow. The 28 aneurysms that were treated with staged endovascular therapy were completely occluded, 16 of them with coils and 12 with proximal balloons. All 6 aneurysms with IC-IC bypasses were treated with direct coil occlusion, whereas only 9 of 21 aneurysms with EC-IC bypasses were treated with direct coil occlusion. Aneurysm obliteration was comparable in the EC-IC and IC-IC bypass groups (97.9% and 97.1%, respectively). Overall, 75 of 82 bypasses were patent on postoperative angiography (91%). Three EC-IC bypasses and 4 IC-IC bypasses occluded, with comparable patency rates in the 2 groups (94% and 89%, respectively). Intraoperative events predicted later occlusion in 4 cases. Two bypass grafts from the cervical carotid to MCA became limp at the end of the cases and both bypasses were revised (1 proximally, 1 proximally and distally). One patient with a gunshot-related dissecting ACA aneurysm had damaged parent arteries that were reanastomosed after aneurysm excision. The initial repair occluded, the parent artery was excised back to more normal tissue, and the anastomosis was revised with increased tension. One MCA thrombosed after clipping of a large M1 segment aneurysm; the aneurysm was excised, the M1 segment was

Improved Unchanged Worse Dead
a

15 (20%) 53 (70%) 4 (5%) 4 (5%)

8 (19%) 28 (67%) 3 (7%) 3 (7%)

7 (21%) 25 (74%) 1 (3%) 1 (3%)

EC-IC, extracranial-intracranial; IC-IC, intracranial-intracranial; GOS, Glasgow Outcome Scale.

thrombectomized, and reanastomosis restored MCA flow. Despite immediate revisions in these 4 cases, the bypasses occluded postoperatively. The remaining 3 bypass occlusions were unexpected. One patient had a saphenous vein with significant varicosities; 1 MCA reanastomosis required an STA interposition graft to bridge the gap in the parent artery; and an allograft saphenous vein used in the remaining patient occluded.

Patient Outcomes
Three patients died in the perioperative period (surgical mortality, 3.7%), all with basilar trunk aneurysms. These patients underwent uncomplicated bypass procedures to revascularize the basilar apex (STA-SCA, STA-PCA, and external carotid artery–SCA), but subsequent endovascular therapy resulted in aneurysm rerupture during coiling, basilar artery thrombosis after bilateral VA occlusions, and intracerebral hemorrhage while taking heparin after bilateral VA occlusions. Permanent neurological morbidity was observed in 4 patients (4.9%), all related to bypass occlusions. These patients had MCA

676 | VOLUME 65 | NUMBER 4 | OCTOBER 2009

www.neurosurgery-online.com

INTRACRANIAL-INTRACRANIAL BYPASS FOR COMPLEX ANEURYSMS

strokes after occlusion of high-flow EC-IC bypasses in 3 patients and an MCA reanastomosis in 1 patient. Two other bypass occlusions (after aneurysm excision and reanastomosis) did not cause any permanent neurological deficits, and 1 bypass occlusion caused only transient neurological deficits. In addition to this patient with transient neurological deficits related to bypass occlusion, 3 patients had postoperative epidural hematomas with deficits that resolved completely (transient neurological morbidity, 4.9%). Excluding the 3 surgical mortalities and 3 additional patients lost to follow-up, final neurological outcomes were assessed in 76 patients (93%). The mean duration of follow-up was 41 months (range, 1–125 months) and did not differ significantly between EC-IC and IC-IC bypass groups (38.6 and 42.7 months, respectively). Four patients died after hospital discharge, 3 from complications in rehabilitation and 1 from delayed growth of a basilar trunk aneurysm with resulting brainstem compression. Three of the 4 late deaths were in the EC-IC bypass group. Good outcomes (GOS score 5 or 4) were measured in 68 patients (90%) overall, and were similar in EC-IC and IC-IC bypass groups (91% and 89%, respectively) (Table 6). At late follow-up, 15 patients (20%) were improved and 53 (70%) were unchanged, relative to preoperative neurological condition, excluding lost patients. Changes in outcome by GOS score were slightly more favorable in the IC-IC bypass group than the EC-IC bypass group (6% versus 14% worse or dead in IC-IC versus EC-IC bypass group, respectively). The mean final GOS scores reflected a similar trend, with a mean GOS score of 4.3 in the EC-IC bypass group and 4.6 in the IC-IC bypass group. Relative to preoperative neurological condition, mean GOS score decreased 0.30 in the EC-IC bypass group and increased 0.21 in the IC-IC bypass group.

DISCUSSION
Evolution of Bypass Surgery for Brain Aneurysms
Bypass surgery for brain aneurysms began with the introduction and popularization of the STA-MCA bypass by Yasargil (29). ¸ This simple bypass protected patients from ischemic complications after deliberate arterial occlusion during the treatment of MCA and some ICA aneurysms. Bypass surgery evolved with EC-IC bypasses that used other extracranial donor arteries and interposition grafts connected to proximal donor sites in the neck (1, 3, 5, 10, 12–15, 17, 18, 20, 21, 23, 25, 28, 30). Even though these second-generation EC-IC bypasses yield excellent results, bypass surgery for aneurysms is evolving further as IC-IC bypasses eliminate extracranial donor arteries and reconstruct cerebral circulation in ways that resemble normal vascular anatomy. In our clinical experience, this third generation of IC-IC bypasses compared favorably to traditional EC-IC bypasses. Aneurysm obliteration rates, bypass patency rates, and neurological outcomes (late GOS score and change in GOS score) were similar to EC-IC bypass patients, supporting a progression toward vascular reconstruction that is entirely intracranial. EC-IC bypasses are technically easier to perform than IC-IC bypasses. For example, STA-MCA bypass requires 1 end-to-side

anastomosis that is usually straightforward because the donor artery is large and mobilizes to enable visualization of both suture lines. In contrast, in situ bypass between 2 MCA branches requires a more challenging side-to-side anastomosis between arteries with limited mobility. Similarly, an external carotid artery-MCA bypass requires a proximal anastomosis that can be performed in a superficial cervical site with no cerebral ischemia, whereas an A1 ACA-MCA intracranial bypass graft requires a proximal anastomosis in a narrow surgical corridor that is even deeper than the distal anastomosis to MCA (Fig. 5). Although cross-clamping the A1 ACA does not produce ischemia in patients with a competent anterior communicating artery, temporary clips on a major intracranial artery induce some time pressure. Therefore, IC-IC bypasses add a degree of difficulty. Based on our experience, we think the extra effort with IC-IC bypasses is justified. First, the caliber of scalp arteries is variable and sometimes too diminutive to revascularize an efferent artery. Although scalp arteries can dilate over time, they might not meet the demand immediately. Deep bypasses to midline or paramedian arteries can require 8 cm or more of scalp artery, and it may be too small at the anastomotic depth to be safe. In contrast, in situ bypass, reanastomosis, and reimplantation techniques use donor arteries that match or exceed the caliber of recipient arteries. Second, EC-IC bypasses that use cervical carotid artery require long interposition grafts at the limit of the radial artery graft. Consequently, saphenous veins were used more frequently than radial arteries, introducing caliber mismatches between graft and intracranial artery. Longer grafts are also associated with lower long-term patency rates. In contrast, intracranial bypass grafts are shorter, and enabled frequent use of radial artery grafts. Their smaller caliber closely resembles that of intracranial arteries and enhances the anastomosis. Although late patency rates were not measured in this study, shorter grafts with arterial composition are more likely to remain patent. Third, IC-IC bypasses eliminate neck incisions, reduce invasiveness, and improve cosmesis. Intracranial bypasses are less vulnerable than EC-IC bypasses to neck torsion, injury, and occlusion with external compression. Fourth, IC-IC bypasses eliminate the harvest of an extracranial donor artery, saving time and tedious effort. Intracranial donor arteries reside in the surgical field and require minimal preparation. Finally, temporarily occluding an intracranial artery for bypass is well tolerated in the territories of most IC-IC bypasses. In situ bypasses and reimplantations require temporary occlusion of 2 intracranial arteries to perform the anastomosis, instead of just 1 recipient artery with EC-IC bypass. However, neurophysiological changes were rarely encountered during these occlusions and always resolved with a boost in arterial pressure. We did not observe any neurological morbidity in this patient series related to temporarily occluding an intracranial artery during anastomosis, or related to an intracranial donor artery that would not have been involved in an EC-IC bypass. These advantages of IC-IC bypass justify their use. Although they are more technically challenging to perform, they are within the expertise of experienced bypass neurosurgeons, and result in elegant bypasses that we think represent the next generation of bypass surgery for aneurysms.

NEUROSURGERY

VOLUME 65 | NUMBER 4 | OCTOBER 2009 | 677

SANAI ET AL.

A1 ACA-MCA and MCA-PCA bypasses has made them preferred choices. Selecting a bypass from among the 4 IC-IC techniques depends on aneurysm anatomy, suitability of the donor artery, depth of the surgical field, and type of anastomosis. Fusiform aneurysms lend themselves to reanastomosis because they are often distally located away from bifurcations, with 1 afferent and 1 efferent artery. End-to-end E F D repair requires aneurysm excision back to healthy arterial tissue on both ends, and joining ends without tension. Mobilizing the redundant artery and resecting aneurysm can bring the arteries together. End-to-end anastomosis is the easiest anastomosis: forceps tips in the lumen enable visualization of translucent arterial walls and guidance of the needle through its bites; the numJ G H I ber of bites needed to complete the anastomosis is less; and arteries rotate to visualize both suture lines. In contrast to fusiform aneurysms, saccular aneu rysms with multiple efferent arteries require other reconstructive techniques. In situ bypass and reimplantation revascularize 1 efferent artery when the other can be preserved with clipFIGURE 5. Intracranial bypass graft with double reimplantation. A, this ruptured right middle cerebral artery ping. For example, the ACA(MCA) aneurysm was coiled and recurred 6 months later (right internal carotid artery [ICA] angiogram, anterior ACA bypass works when clipoblique view). B, intraoperatively, the 2 M2 MCA trunks originated from the aneurysm (An) base and could not be ping or coiling an anterior kept open with direct clipping. Note the strand of coil in the lumen of the temporal M2 trunk (black arrow). C, the communicating artery aneuA1 segment of the anterior cerebral artery (ACA) was used as the donor for a radial artery graft (RAG) that was rysm sacrifices 1 A2 ACA. The sutured with an end-to-side anastomosis (D and E). The frontal M2 trunk was reimplanted onto the RAG with a sideother patent A2 ACA supplies to-side anastomosis, (F) shown after suturing the deep suture line intraluminally and (G) after completing the superficial suture line. H, the end of the RAG was looped to the temporal M2 trunk and sewn with an end-to-side anasthe distal bypass and restores tomosis. Postoperative angiography (right ICA injection, lateral [I] and anteroposterior [J] views) confirmed patency flow to the opposite ACA. of the bypass and filling of both MCA trunks (red arrows indicate anastomoses). Side-to-side anastomosis is probably the most difficult Bypass Selection anastomosis because the deep suture line is sewn inside the lumen. After approximating the 2 arteries with sutures at each Most current EC-IC bypasses convert to an IC-IC bypass preend of the arteriotomies, the first bite transitions the needle from sented in this clinical experience. ACA and PICA territories outside the lumen where the knot is tied, to inside the lumen were particularly amenable to intracranial reconstruction, and where running bites are taken. Bites are taken between 2 outer even though MCA and basilar apex territories were divided layers of arterial wall, keeping track of 4 translucent layers. The between EC-IC and IC-IC techniques, growing experience with

A

B

C

678 | VOLUME 65 | NUMBER 4 | OCTOBER 2009

www.neurosurgery-online.com

INTRACRANIAL-INTRACRANIAL BYPASS FOR COMPLEX ANEURYSMS

A

B

C

D

E

F

G

last bite transitions the needle again from inside to outside the lumen to tie the knot. The second suture line is performed from outside the lumen and is much easier. The arteriotomy length should be 3 times the diameter of the arteries in order to communicate generously between arteries. Consequently, side-to-side anastomoses require more bites than other anastomoses. This difficult anastomosis should be avoided in deep, narrow surgical corridors, but can be performed in the sylvian fissure, cisterna magna, and interhemispheric fissure. Reimplantation also salvages a branch artery compromised by aneurysm clipping with an end-to-side anastomosis to the parent artery, the other efferent artery, or an uninvolved bystander. PICAVA reimplantation was the most frequent location for this technique, but it works well in MCA and ACA territories (pericallosal-to-callosomarginal reimplantation) (Fig. 6). These recipient reimplanta-

FIGURE 6. Recipient reimplantation. A, this thrombotic ACA aneurysm, seen on axial T1-weighted magnetic resonance imaging scans, originated at the bifurcation of the pericallosal (PC) and callosomarH I ginal (CM) arteries. B, right ICA angiogram, lateral view. C and D, the aneurysm (An) was exposed in the interhemispheric fissure through a bifrontal craniotomy, using gravity to retract the right hemisphere (right hemisphere down; left hemisphere up; nose facing to the right). Attempts to clip reconstruct the neck were unsuccessful; intraluminal thrombus caused the clips to occlude the pericallosal artery. E, rather than opening the aneurysm, removing thrombus, and attempting to reconstruct a neck, the pericallosal artery was clip occluded, transected, and mobilized to the callosomarginal artery. An end-to-side PC-CM anastomosis was performed: back wall (F), front wall (G), and after completion (H). I, the CM artery supplied blood flow to the entire distal ACA territory. Distal clip occlusion of the aneurysm resulted in its complete thrombosis.

NEUROSURGERY

VOLUME 65 | NUMBER 4 | OCTOBER 2009 | 679

SANAI ET AL.

and a spatulated end of the reimplanted recipient to cover the arteriotomy. Simple continuous sutures are placed loosely and tightened after all bites have been taken. The site of reimplantation is selected to slacken the reimplanted artery and allow it to be shifted from side to side to visualize both suture lines. Complex reconstructions E F D are required when multiple efferent arteries are compromised by clipping. For example, the double reimplantation technique rebuilds a bifurcation with 3 anastomoses (Fig. 5). A radial artery graft is first connected proximally to a donor artery to ready the bypass graft. The first efferent artery is reimplanted on the live graft and blood flow is restorI H G ed immediately. The second efferent artery is reimplanted distally on the graft, allowing the graft to supply the first reimplanted artery during this second reimplantation. Placement of a temporary clip distal to the first and proximal to the second anastomosis redirects blood flow to the reimplanted trunk while keeping FIGURE 7. Donor reimplantation. A, a 55-year-old woman presented with subarachnoid hemorrhage from this multhe other surgical site dry. This tilobulated left superior cerebellar artery (SCA) aneurysm, seen on rotational angiogram with 3-dimensional reconsuccessive reimplantation of struction. Only the superior lobule could be coiled, leaving residual neck to preserve the SCA origin. She was branch arteries minimizes isreferred for surgery to protect her from rehemorrhage. After inspecting the anatomy intraoperatively, it seemed chemia, with temporary ocunlikely that the aneurysm could be clipped without occluding the SCA and likely that a bypass would be needed to clusion times for each of the preserve it. B, a prominent anterior temporal artery (ATA) was found in the sylvian fissure, and (C) it had sufficient efferent arteries equal to the length to reach the SCA. D, ATA was transected distally and reimplanted onto the SCA with an end-to-side anastotime needed to complete 1 mosis. E, after bypass patency was confirmed, (F) the aneurysm (An) neck was dissected and clipped (G). H and I, anastomosis. A double reimindocyanine green videography confirmed good flow in the ATA-SCA bypass, as did the postoperative angiogram (left plantation technique adapts to ICA injection, lateral view, with opacification of the left SCA [arrows]). Note the course of the SCA over the cerebellar vermis (red asterisk). The patient tolerated ATA sacrifice without neurological sequela. MCA, middle cerebral triple reimplantation for triartery; ICA, internal carotid artery; PCA, posterior cerebral artery; CN3, oculomotor nerve. furcated anatomy. Other intracranial bypass grafts replenish cerebral blood flow with fewer anastomoses (Fig. 8). For examtions connect the proximal end of a branch to the side of donor, ple, the MCA-PCA bypass revascularizes quadrifurcated but donor reimplantations can also connect the distal end of a anatomy of the basilar apex with a single deep anastomosis to branch to a recipient artery to rededicate the branch artery to the P2 PCA and a superficial anastomosis to an MCA trunk that supplying a new vascular territory. For example, ATA supplies is already exposed en route to the PCA site. Intracranial bypass a silent vascular territory, and when reimplanted onto SCA, can grafts such as the MCA-PCA bypass do not fully reconstruct supply the SCA or basilar apex (Fig. 7). Reimplanted arteries can arterial anatomy and might not enable complete exclusion of the therefore donate or receive blood flow. The end-to-side anastoaneurysm, but might reverse flow or create more benign hemomosis is identical to STA-MCA bypass, with a generous arteridynamics inside the aneurysm. otomy in the donor (at least 2 times the diameter of the artery)

A

B

C

680 | VOLUME 65 | NUMBER 4 | OCTOBER 2009

www.neurosurgery-online.com

INTRACRANIAL-INTRACRANIAL BYPASS FOR COMPLEX ANEURYSMS

A

B

C

PCA aneurysm), or serendipitous options (e.g., ATA-SCA bypass). We select the bypass that facilitates an eu rysm occlusion, restores nor mal blood flow, and is technically most feasible.

Limitations
Our conclusion that IC-IC bypasses can replace many EC-IC bypasses is based on a comparison between 2 groups of patients that were different D E F and highly selected. This conclusion applies to ACA, PICA, basilar apex, and many MCA aneurysms, but not to most ICA aneurysms. EC-IC by passes will remain the preferred choice for petrous, cavernous, and supraclinoid ICA aneurysms because they are easier and less risky than the petrous-to-supraclinoid ICA bypass. Therefore, the progresG H I sion from EC-IC to IC-IC bypass does not apply to all aneurysms. The STA-MCA bypass will remain a versatile technique, and we are not suggesting that it be abandoned. Seventeen bypasses were performed in this clinical experience, but several discussed here were not performed, such as the PCA-SCA and petrous ICA-MCA bypasses. Indications for these bypasses FIGURE 8. Intracranial bypass graft. A, this fusiform, giant basilar trunk aneurysm (left vertebral artery [V A]) angiogram, anterior oblique view) enlarged rapidly over a 2-year period, changing from a small, asymptomatic are few and the technical debump in the basilar trunk to a compressive mass with new hemiparesis, dysarthria, and gait instability. It was treated mands are high. Other in with a V A-to-superior cerebellar artery (SCA) bypass. B, a combined far lateral–temporal craniotomy provided a subtracranial bypasses, such as temporal view of the basilar artery (BA) apex. C and D, a radial artery graft (RAG) was anastomosed to the right Spetzler’s “figure 8 anastomoSCA. E, the proximal end of the graft was connected to the side of extradural V at the foramen magnum. F, the A sis,” (12a) were not a part of course of the V A-SCA bypass is shown intraoperatively (arrows) and (G) angiographically (right V injection, A our experience, but should be anteroposterior view, with anastomoses indicated by the red arrows). V A-SCA bypass and clip occlusion of the right included in the menu of IC-IC V resulted in thrombosis of the aneurysm lumen, shown preoperatively in blue and postoperatively in red on overA bypasses. Innovation will add laid volumetric images generated from contrast-enhanced magnetic resonance angiography. H, lateral view. I, anteroto this menu over time and posterior view. L, left; R, right; PCA, posterior cerebral artery. deepen our armamentarium of intracranial bypasses. This study did not include patients in whom a planned Bypass selection ultimately depends on an intraoperative bypass was aborted. On several occasions, we found that P3 assessment of the aneurysm and surrounding anatomy. We PCA aneurysms were difficult to revascularize with intracradevise a primary bypass strategy and several contingency nial reconstructive techniques because surgical corridors were strategies, with preparations for each (such as prepping a graft deep and narrow through all approaches (subtemporal, possite). There might be several viable options (e.g., PICA-PICA terior interhemispheric, and supratentorial-infraoccipital). bypass and PICA reimplantation), no options (e.g., P3 segment

NEUROSURGERY

VOLUME 65 | NUMBER 4 | OCTOBER 2009 | 681

SANAI ET AL.

These aneurysms are increasingly treated with endovascular techniques (24). Bypass with aneurysm occlusion is a good strategy for giant, dolichoectatic, thrombotic, or previously coiled aneurysms because it avoids the unpredictability of thrombectomy with clip reconstruction. It also avoids risky adjuncts such as hypothermic circulatory arrest. However, deliberate hemodynamic alteration with bypass and aneurysm occlusion can also be risky and unpredictable. Poor outcomes were encountered with flow reversal in basilar trunk aneurysms as a result of basilar artery thrombosis or occlusion of perforators. Other complications resulted from heparinization to decelerate aneurysm thrombosis, with subsequent intracranial hemorrhage. Bypass with incomplete aneurysm occlusion relies on some intraluminal aneurysm thrombosis, which has unavoidable dangers. This management of basilar trunk aneurysms might not be the best strategy for this difficult disease. We are hopeful that stents or other endovascular devices will offer reconstructive options without open surgery, but these therapies are not available presently and will need to be evaluated critically before they replace surgical bypass strategies. This experience demonstrates that old-fashioned microneurosurgery can be applied creatively to change the practice of bypass surgery for complex aneurysms. Only suture and meticulous technique are needed to move from the current generation of EC-IC bypasses to the next generation of IC-IC bypasses. New technology, such as ELANA (excimer laserassisted nonocclusive anastomosis) (14) or automated staplers, will not facilitate most IC-IC bypasses. Instead, continued evolution will require mastering the simple craft of arterial anastomosis in all of its variations, and expending some extra effort to create more elegant and anatomic bypasses.

Disclosure
The authors have no personal financial or institutional interest in any of the drugs, materials, or devices described in this article.

REFERENCES
1. Auguste KI, Quiñones-Hinojosa A, Lawton MT: The tandem bypass: Subclavian artery-to-middle cerebral artery bypass with dacron and saphenous vein grafts. Technical case report. Surg Neurol 56:164–169, 2001. 2. Ausman JI, Diaz FG: Critique of the extracranial-intracranial bypass study. Surg Neurol 26:218–221, 1986. 3. Barnett DW, Barrow DL, Joseph GJ: Combined extracranial-intracranial bypass and intraoperative balloon occlusion for the treatment of intracavernous and proximal carotid artery aneurysms. Neurosurgery 35:92–98, 1994. 4. Barnett HJ, Sackett D, Taylor DW, Haynes B, Peerless SJ, Meissner I, Hachinski V, Fox A: Are the results of the extracranial-intracranial bypass trial generalizable? N Engl J Med 316:820–824, 1987. 5. Baskaya MK, Kiehn MW, Ahmed AS, Ates O, Niemann DB: Alternative vascular graft for extracranial-intracranial bypass surgery: Descending branch of the lateral circumflex femoral artery. Neurosurg Focus 24:E8, 2008. 6. Bederson JB, Spetzler RF: Anastomosis of the anterior temporal artery to a secondary trunk of the middle cerebral artery for treatment of a giant M1 segment aneurysm. Case report. J Neurosurg 76:863–866, 1992. 7. Candon E, Marty-Ane C, Pieuchot P, Frerebeau P: Cervical-to-petrous internal carotid artery saphenous vein in situ bypass for the treatment of a high cervical dissecting aneurysm: Technical case report. Neurosurgery 39:863– 66, 1996.

8. EC/IC Bypass Study Group: Failure of extracranial-intracranial arterial bypass to reduce the risk of ischemic stroke. Results of an international randomized trial. The EC/IC Bypass Study Group. N Engl J Med 313:1191– 1200, 1985. 9. Evans JJ, Sekhar LN, Rak R, Stimac D: Bypass grafting and revascularization in the management of posterior circulation aneurysms. Neurosurgery 55:1036–1049, 2004. 10. Friedman JA, Piepgras DG: Current neurosurgical indications for saphenous vein graft bypass. Neurosurg Focus 14:e1, 2003. 11. Goldring S, Zervas N, Langfitt T: The Extracranial-Intracranial Bypass Study. A report of the committee appointed by the American Association of Neurological Surgeons to examine the study. N Engl J Med 316:817–820, 1987. 12. Hadeishi H, Yasui N, Okamoto Y: Extracranial-intracranial high-flow bypass using the radial artery between the vertebral and middle cerebral arteries. Technical note. J Neurosurg 85:976–979, 1996. 12a. Hanel RA, Spetzler RF: Surgical treatment of complex intracranial aneurysms. Neurosurgery SHC Suppl 3:SHC1289–SHC1299, 2008. 13. Kato Y, Sano H, Imizu S, Yoneda M, Viral M, Nagata J, Kanno T: Surgical strategies for treatment of giant or large intracranial aneurysms: Our experience with 139 cases. Minim Invasive Neurosurg 46:339–343, 2003. 14. Langer DJ, Van Der Zwan A, Vajkoczy P, Kivipelto L, Van Doormaal TP, Tulleken CA: Excimer laser-assisted nonocclusive anastomosis. An emerging technology for use in the creation of intracranial-intracranial and extracranial-intracranial cerebral bypass. Neurosurg Focus 24:E6, 2008. 15. Lawton MT, Hamilton MG, Morcos JJ, Spetzler RF: Revascularization and aneurysm surgery: Current techniques, indications, and outcome. Neurosurgery 38:83–94, 1996. 16. Lemole GM Jr, Henn J, Javedan S, Deshmukh V, Spetzler RF: Cerebral revascularization performed using posterior inferior cerebellar artery-posterior inferior cerebellar artery bypass. Report of four cases and literature review. J Neurosurg 97:219–223, 2002. 17. Mohit AA, Sekhar LN, Natarajan SK, Britz GW, Ghodke B: High-flow bypass grafts in the management of complex intracranial aneurysms. Neurosurgery 60:ONS105–ONS123, 2007. 18. Morgan MK, Sekhon LH: Extracranial-intracranial saphenous vein bypass for carotid or vertebral artery dissections: A report of six cases. J Neurosurg 80:237–246, 1994. 19. Quiñones-Hinojosa A, Lawton MT: In situ bypass in the management of complex intracranial aneurysms: Technique application in 13 patients. Neurosurgery 57:140–145, 2005. 20. Quiñones-Hinojosa A, Du R, Lawton MT: Revascularization with saphenous vein bypasses for complex intracranial aneurysms. Skull Base 15:119–132, 2005. 21. Regli L, Piepgras DG, Hansen KK: Late patency of long saphenous vein bypass grafts to the anterior and posterior cerebral circulation. J Neurosurg 83:806–811, 1995. 22. Relman AS: The extracranial-intracranial arterial bypass study: What have we learned? N Engl J Med 316:809–810, 1987. 23. Rivet DJ, Wanebo JE, Roberts GA, Dacey RG Jr: Use of a side branch in a saphenous vein interposition graft for high-flow extracranial-intracranial bypass procedures. Technical note. J Neurosurg 103:186–187, 2005. 24. Sanai N, Tarapore P, Lee A, Lawton MT: The current role of microsurgery for posterior circulation aneurysms: A selective approach in the endovascular era. Neurosurgery 62:1236–1253, 2008. 25. Santoro A, Guidetti G, Dazzi M, Cantore G: Long saphenous-vein grafts for extracranial and intracranial internal carotid aneurysms amenable neither to clipping nor to endovascular treatment. J Neurosurg Sci 43:237–251, 1999. 26. Sekhar LN, Natarajan SK, Ellenbogen RG, Ghodke B: Cerebral revascularization for ischemia, aneurysms, and cranial base tumors. Neurosurgery 62 [Suppl 3]:1373–1410, 2008. 27. Sundt TM Jr: Was the international randomized trial of extracranial-intracranial arterial bypass representative of the population at risk? N Engl J Med 316:814–816, 1987. 28. Ustün ME, Büyükmumcu M, Ulku CH, Cicekcibasi AE, Arbag H: Radial artery graft for bypass of the maxillary to proximal middle cerebral artery: An anatomic and technical study. Neurosurgery 54:667–671, 2004.

682 | VOLUME 65 | NUMBER 4 | OCTOBER 2009

www.neurosurgery-online.com

INTRACRANIAL-INTRACRANIAL BYPASS FOR COMPLEX ANEURYSMS

29. Yasargil MG: Anastomosis between the superficial temporal artery and a ¸ branch of the middle cerebral artery, in Yasargil MG (ed): Microsurgery ¸ Applied to Neurosurgery. Stuttgart: Georg Thieme-Verlag, 1969, pp 105–115. 30. Zhang YJ, Barrow DL, Day AL: Extracranial-intracranial vein graft bypass for giant intracranial aneurysm surgery for pediatric patients: Two technical case reports. Neurosurgery 50:663–668, 2002.

COMMENTS

S

anai et al. have made an excellent contribution to the literature. They report a number of creative microsurgical solutions to difficult neurovascular pathologies. Clearly, the types of aneurysmal disease illustrated do not have acceptable endovascular options with existing technology. This type of creative approach represents a transition in the way microvascular surgeons conceptualize neurovascular anatomy. In the early days of microvascular surgery, the focus was on recreating normal anatomy by complex clip reconstructions even if prolonged temporary arterial occlusion was required. It became obvious that perfect reconstructions were not always possible or posed substantial ischemic risks. The use of extracranial-intracranial (EC-IC) bypass grafts with the superficial temporal artery, saphenous vein, or radial artery clearly advanced the field. This contribution may eliminate the need for cervical dissections for proximal anastomoses in some cases and, in many circumstances, may obviate the need for graft harvest. The tradeoff for using an intracranial interposition graft is the requirement to put 2 different vascular territories at risk during the creation of the bypass. Perhaps nontraditional anastomoses, such as the excimer laser-assisted nonocclusive anastomosis technique, will reduce this risk in selected cases. This article will be a valuable reference to cerebrovascular and cranial base surgeons as they search for solutions in complex vascular cases. Christopher S. Eddleman Bernard R. Bendok H. Hunt Batjer Chicago, Illinois

morbidity and mortality. The comparison of these 2 groups is not entirely unbiased, as the aneurysm location and morphology will dictate the available options for revascularization. Nevertheless, the comparison does demonstrate the utility of these innovative approaches and emphasizes the need to be versatile and creative in managing complex intracranial aneurysms. The use of IC-IC bypasses is particularly useful in cases where an unanticipated need for revascularization occurs during the course of an operation. IC-IC procedures do not require as much preoperative planning as EC-IC bypasses, owing to the need for preparation of the extracranial donor vessel in the latter. As the population of intracranial aneurysms presenting for surgical management becomes increasingly complex, the technical skills and innovation of cerebrovascular surgeons will be increasingly challenged. This excellent article, which reviews these elegant vascularization options, clearly indicates that this group is meeting the challenge. Daniel L. Barrow Atlanta, Georgia

T

T

his article represents yet another outstanding contribution from the cerebrovascular surgery group at the University of California, San Francisco. The authors report a retrospective review of a decade of experience in managing patients with complex intracranial aneurysms using either EC-IC or intracranial-intracranial (IC-IC) bypasses. In doing so, they report on a number of bypass techniques that were previously undescribed, the subject of case reports, or reported in very small case series. In an interesting comparison of the outcome of patients undergoing EC-IC versus IC-IC bypass, it is clear that the results are quite similar and associated with a low risk of operative

he authors report their experience with IC-IC bypass techniques for the treatment of complex aneurysms that cannot be coiled or clipped without revascularization. The data show that IC-IC bypasses perform comparably to EC-IC bypasses in terms of patient safety, outcome, and efficacy. However, both groups were highly selected, and the data cannot be used to support “progression toward vascular reconstruction that is entirely intracranial” unless the indications and favorable anatomic factors for IC-IC bypass are present. In some instances, the flow provided by an IC-IC bypass may be inadequate, as in the treatment of basilar trunk aneurysms with flow reversal. Conversely, this technique appears particularly attractive for middle cerebral artery, anterior cerebral artery, and posteroinferior cerebellar artery-toposteroinferior cerebellar artery bypasses. In our experience, IC-IC bypasses have been similarly useful, especially in challenging cases in which the patient’s anatomy is appropriate. However, with both donor and recipient territories at risk in the rare event of bypass occlusion or with extended cross-clamping, our first preference is for clip reconstruction of the aneurysm versus suitable EC-IC bypass. Finally, the authors’ experience reinforces that a combined approach is often needed when an aneurysm cannot be treated by endovascular or standard microsurgical techniques alone. The authors have presented a thorough, clear description of their impressive experience with IC-IC bypass, including relevant technical notes, indications, and rationales for the different challenging cases depicted. Rasha Germain Robert F. Spetzler Phoenix, Arizona

NEUROSURGERY

VOLUME 65 | NUMBER 4 | OCTOBER 2009 | 683

Sign up to vote on this title
UsefulNot useful