CLINICAL STUDIES

ROUTINE CEREBRAL ANGIOGRAPHY AFTER SURGERY FOR

SACCULAR ANEURYSMS: IS IT WORTH IT?

Riku P. Kivisaari, M.D.Department of Neurosurgery,Helsinki University CentralHospital, Helsinki, Finland

Matti Porras, M.D., Ph.D.Department of DiagnosticRadiology, Helsinki UniversityCentral Hospital, Helsinki, Finland

Juha Öhman, M.D., Ph.D.Department of Neurosurgery,Helsinki University CentralHospital, Helsinki, Finland

Jari Siironen, M.D., Ph.D.Department of Neurosurgery,Helsinki University CentralHospital, Helsinki, Finland

Keisuke Ishii, M.D., Ph.D.Department of Neurosurgery,Helsinki University CentralHospital, Helsinki, Finland

Juha Hernesniemi, M.D.,Ph.D.Department of Neurosurgery,Helsinki University CentralHospital, Helsinki, Finland

Reprint requests:Riku Kivisaari, M.D.,Department of Neurosurgery,Helsinki University CentralHospital, Topeliuksenkatu 5,00260 Helsinki, Finland.Email: riku.kivisaari@hus.fi

Received, January 7, 2004.

Accepted, May 6, 2004.

OBJECTIVE: The objective of this study was to determine whether an angiographicallyproven rate of saccular intracranial aneurysm occlusion after surgical clipping suggeststhat postoperative angiography should continue to be used routinely or should besupplanted by intraoperative angiography. These data also should establish a basis forcomparing surgery with new endovascular methods of treatment.METHODS: During a 3.5-year period, a consecutive series of 622 patients (955aneurysms, 808 of which were surgically clipped) who underwent postoperativeangiography were studied retrospectively. This series comprised 493 ruptured and 315unruptured aneurysms.RESULTS: Complete aneurysm closure was achieved in 88% of aneurysms, a neckremnant was discovered in 9%, and a fundus remnant was revealed in 3%. Of 493ruptured aneurysms, 86% were completely occluded. Of 315 unruptured aneurysms,91% were completely occluded. The results for clipping of complex aneurysms, i.e.,posterior circulation or large to giant aneurysms, were significantly inferior to those forsmall and anterior circulation aneurysms. In one-third of the large and giant aneu-rysms, a part of the base was left intentionally because of calcifications or strong wallor to prevent occlusion of any branches. In the series, a significant 5% complicationrate of major vessel occlusion was detected.CONCLUSION: Our retrospective analysis revealed that ruptured, posterior circula-tion, and large/giant aneurysms are more prone to incomplete clipping. Therefore,these aneurysms require postoperative if not intraoperative evaluation with angiogra-phy. Many clippings of anterior circulation aneurysms experience unexpected failures,which suggests that intraoperative angiography could be beneficial. This series, whichhas no selection bias, can be used as a basis to compare the results of other seriesreporting surgical or endovascular treatment.

KEY WORDS: Angiography, Intracranial aneurysm, Subarachnoid hemorrhage, Surgery

Neurosurgery 55:1015-1024, 2004 DOI: 10.1227/01.NEU.0000141043.07303.60 www.neurosurgery-online.com

Dandy (3) pioneered aneurysm surgeryby clipping the base of the aneurysm,and Yasargil (32) introduced the sys-

temic use of microsurgery to cure these fatalsacs. When an aneurysm is clipped, the base ofthe aneurysm is closed, and the walls are ap-posed for a continuous endothelial lining. Ifthe clip is perfectly positioned, this completelyoccludes the aneurysm and leaves the parentvessels intact. The complete closure of thebase of an intracranial aneurysm is crucial forthe prevention of rebleeding of a rupturedaneurysm or the subsequent growth and rup-ture of an unruptured aneurysm (4, 9, 12, 14,15, 26). During the past 10 years, endovascular

therapy with coils has been used increasinglyto occlude aneurysms (21, 31). In experiencedhands, aneurysms can be clipped or coiledwith low morbidity and mortality rates and ahigh short-term success rate (7, 10, 12, 31, 32).At many centers, control angiograms are sel-dom performed. The difficulty in predictingthe presence of a residual or unclipped aneu-rysm suggests that all patients should un-dergo intra- or postoperative catheter angiog-raphy, even if the aneurysm is opened and/orcoagulated, which is how we regularly handleaneurysms after clipping (1, 19, 24, 30). An-giography is a reliable method to evaluate thecompleteness of the closure of an aneurysm,

NEUROSURGERY VOLUME 55 | NUMBER 5 | NOVEMBER 2004 | 1015

but it is not completely without complications, especially inelderly patients or if performed by less-experienced practitio-ners (1, 22, 30).

We conducted this study because of the high incidence(16–20 per 100,000 persons per year) of subarachnoid hemor-rhage (SAH) in Finland (9, 11, 27, 28), our defined catchmentarea comprising 2 million people. As our institution is the onlyneurosurgical center, there is no referral bias, and we haveextensive experience with more than 10,000 patients with ce-rebral aneurysms. Until recently, routine intra-arterial angiog-raphy was performed during the early postoperative periodfor all patients who underwent aneurysm clipping at ourinstitution; subsequently, computed tomographic angiogra-phy replaced digital subtraction angiography (DSA).

As many neurosurgical centers do not perform postopera-tive controls, we analyzed the completeness of the closure ofintracranial aneurysms treated by experienced surgeons. Thisanalysis also should establish a basis for comparing surgerywith endovascular therapy and the results of intraoperativeangiography.

PATIENTS AND METHODS

Patients

Helsinki is the only neurosurgical unit serving a populationof 2 million in Southern Finland. There are no admissionbiases, because all of the patients, including those in poorcondition or moribund, are admitted. Of 932 patients withSAH or an unruptured saccular intracerebral aneurysmtreated at our hospital between August 1, 1998, and December31, 2001, a consecutive series of 622 patients (41% men, 59%women; age range, 23–81 yr) was studied retrospectively.Excluded were 32 patients with fusiform aneurysm, 103 pa-tients who had nonaneurysmal SAH, and 44 patients whounderwent endovascular treatment or surgery other than clip-ping (e.g., trapping or bypass). In addition, 77 patients with nocontrol digital subtraction angiogram were excluded. The rea-sons for not performing control DSA included severe calcifi-cations at the aortic arch and its branches, a complication inobtaining the diagnostic digital subtraction angiogram, or per-fect visualization of the anatomy at operation. All patientswith saccular aneurysm were treated, with the exception of 54moribund or rapidly deteriorating or extremely old (�85 yr)patients. Patients with large hematomas also were treated. Theaneurysms were operated on by 12 surgeons; the senior author(JH) performed most of the aneurysm operations (68%).

The 622 patients presented with 955 aneurysms. Of these,808 aneurysms were clipped, and the results were monitoredpostoperatively by angiography. A total of 121 patients (19%)had no history of SAH, i.e., they had incidental aneurysms. Atotal of 493 of the aneurysms (61%) were ruptured causingSAH; 315 (39%) of the 808 aneurysms were unruptured. Thedemographic features of the patients with SAH are listed inTable 1, and those of the patients without SAH are listed inTable 2.

TABLE 1. Patients with subarachnoid hemorrhage operated onfor ruptured aneurysmsa

Clinical characteristics No. of patients (%)

SexMale 210 (43%)Female 283 (57%)

Total 493 (100%)

Age (yr)Average 53Range 23–81

Patients with multiple aneurysms 152 (31%)WFNS grade

1 208 (42%)2 89 (18%)3 27 (5%)4 84 (17%)5 85 (17%)Total 493 (100%)

Aneurysm locationICA 98 (20%)MCA 143 (29%)AComA 174 (35%)Peric 31 (6%)VBA 47 (10%)Total 493 (100%)

Aneurysm sizeSmall (�7 mm) 218 (44%)Medium (7–14 mm) 238 (48%)Large (15–24 mm) 20 (4%)Giant (�24 mm) 4 (1%)NA 13 (3%)Total 493 (100%)

Timing of surgery�24 h 288 (58%)�72 h 400 (81%)3–10 d 63 (13%)�10 d 30 (6%)Total 493 (100%)

GOS score (at 3 mo)GR 229 (46%)MD 100 (20%)SD 94 (19%)VS 20 (4%)Dead 50 (10%)Total 493 (100%)

a WFNS, World Federation of Neurosurgical Societies scale (2); ICA, inter-nal carotid artery; MCA, middle cerebral artery; AComA, anterior commu-nicating artery; Peric, pericallosal artery; VBA, vertebrobasilar arteries; NA,not available; GOS, Glasgow Outcome Scale (13); GR, good recovery;MD, moderate disability; SD, severe disability; VS, vegetative state.

KIVISAARI ET AL.

1016 | VOLUME 55 | NUMBER 5 | NOVEMBER 2004 www.neurosurgery-online.com

Angiography

Patients were assessed preoperatively by use of computedtomographic angiography or catheter DSA; both carotid arteriesand at least one vertebral artery were assessed. PostoperativeDSA was performed with an Integris V3000 (1024 � 1024 matrix;Philips Medical Systems, Best, The Netherlands), within 24 hoursafter the operation. At least four different projections per vesselwere obtained. All angiographic studies were analyzed by anexperienced neuroradiologist (MP) who has assessed nearly10,000 intracranial aneurysms. The analysis focused on: 1) thecompleteness of the closure of the aneurysm; and 2) the occlusion

of any major artery or branchor perforator. The closure ofthe aneurysm was consid-ered incomplete if even asmall neck remnant (�1mm) or fundus of the aneu-rysm was reliably visible(Figs. 1–3). Incomplete clo-sures were divided into twocategories: 1) neck remnantonly; and 2) residual fillingof the fundus of the aneu-rysm. The size of the aneu-rysm was approximated inmillimeters by comparisonwith the intracavernous partof the internal carotid artery(5 mm) and the middle partof the basilar artery (3 mm).

RESULTS

Complete closure of aneu-rysm was achieved in 711(88%) of the 808 aneurysms(Table 3). Approximatelytwo-thirds (59) of the 97 in-complete closures were un-expected, and one-third (38) of the incomplete closures wereplanned to save a major vessel, because of calcifications, or owingto a thick aneurysm wall. The frequency of unexpected incom-

FIGURE 1. Digital subtraction angio-grams of a 58-year-old man. A, rup-tured aneurysm of the basilar artery,World Federation of Neurosurgical So-cieties Grade 1. The patient underwentsurgery on Day 2. B, to save a smallperforator, a small neck remnant was leftat operation. The patient died secondaryto rebleeding 10 days after the operation.

TABLE 2. Patients without subarachnoid hemorrhage operatedon for unruptured aneurysmsa

Clinical characteristics No. of patients

SexMale 81 (36%)Female 141 (64%)

Total 222 (100%)

Incidental 121

Age (yr)Average 52Range 28–81

Multiple aneurysms 144 (65%)Aneurysm location

ICA 64 (20%)MCA 163 (51%)AComA 44 (14%)Peric 14 (4%)VBA 30 (9%)Total 315 (100%)

Aneurysm sizeSmall (�7 mm) 215 (68%)Medium (7–14 mm) 78 (25%)Large (15–24 mm) 8 (3%)Giant (�24 mm) 7 (2%)NA 7 (2%)Total 315 (100%)

GOS score (incidental at 3 mo)GR 98 (81%)MD 13 (11%)SD 7 (6%)VS 1 (1%)Dead 2 (2%)Total 121 (100%)

a ICA, internal carotid artery; MCA, middle cerebral artery; AComA, anteriorcommunicating artery; Peric, pericallosal artery; VBA, vertebrobasilar arteries;NA, not available; GOS, Glasgow Outcome Scale (13); GR, good recovery;MD, moderate disability; SD, severe disability; VS, vegetative state.

FIGURE 2. Computed tomographic (A) and digital subtraction (B) angio-grams of a 57-year-old woman. A, unruptured giant aneurysm of the leftmiddle cerebral artery bifurcation. Only part of the aneurysm is filled withcontrast medium. Immediately after the operation, the patient experiencedslight hemiparesis, which subsided. B, complete closure of the aneurysm.

CEREBRAL ANGIOGRAPHY AFTER SURGERY FOR SACCULAR ANEURYSMS

NEUROSURGERY VOLUME 55 | NUMBER 5 | NOVEMBER 2004 | 1017

plete closure was extremelyhigh at the following loca-tions: anterior communicat-ing artery (84% of all incom-plete closures wereunexpected), internal carotidartery (73% of all incompleteclosures were unexpected),and pericallosal arteries(71% of all incomplete clo-sures were unexpected) (Ta-ble 4). Of the 493 rupturedaneurysms, 424 (86%) werecompletely occluded, and 46of the 69 incomplete closureswere unexpected. Of the 94unruptured aneurysms thatwere operated on during thesame operation with a ruptured aneurysm, 83 (88%) were com-pletely occluded, and 6 of the 11 incomplete closures were un-expected. A total of 221 unruptured aneurysms were in patientswithout SAH or after a 2- to 3-month recovery period after SAH.Of the 221 aneurysms, 204 (92%) were completely occluded, and7 of the 17 incomplete closures were unexpected. Overall, therewere 315 unruptured aneurysms, of which 287 (91%) were com-pletely occluded. The difference in the number of completelyclosed aneurysms between ruptured and all unruptured aneu-rysms is significant (P � 0.003, �2 test) (Table 3). Neither theWorld Federation of Neurosurgical Societies grade (2) at admis-sion of the patients with SAH or the timing of surgery influencedthe success rate of complete closure (Table 5).

The closure percentage (expected and unexpected) of theruptured vertebrobasilar aneurysms (74%) was significantlylower than that of ruptured middle cerebral, carotid, or ante-

rior communicating artery aneurysms. In addition, the closurepercentage of the ruptured aneurysms of the middle cerebralartery (92%) was significantly higher than that of rupturedpericallosal or anterior communicating artery aneurysms,whereas other differences were not statistically significant.The unruptured vertebrobasilar aneurysms also were signifi-cantly less often completely secured (77%) than the unrup-tured aneurysms of the carotid (97%) and middle cerebralarteries (92%) (Table 6).

The size of the aneurysm had a significant impact on closurerate (Table 6). Of the 20 large (15–24mm) ruptured aneurysms,only 12 (60%) were completely occluded, whereas 92% of thesmall (2–7 mm) and 84% of the medium-size (8–14 mm) un-ruptured aneurysms were completely occluded. Of the unrup-tured aneurysms, 95% of small aneurysms were completelyoccluded, whereas only 85% of medium, 50% of large, and57% of giant aneurysms were completely occluded. Twenty-four (41%) of the 59 aneurysms with unexpected incompleteclosure were reoperated, and 7 (12%) were treated with endo-vascular coiling (Fig. 4). After these treatments, four aneu-rysms still had neck remnants.

There were 44 major vessel (or branch of major vessel)occlusions; 32 of these were unexpected. Furthermore, thefrequency of unexpected major vessel occlusion was high atthese locations: internal carotid artery (75%), middle cerebralartery (79%), and pericallosal arteries (83%) (Table 7). Majorvessel occlusion led to reoperation in 14 patients. Two occlu-sions were treated with extracranial-intracranial bypass oper-ations. In 12 occlusions, repositioning of the aneurysm clipwas performed, and good angiographic results were achievedin 9 patients.

Overall, control DSA lead to reoperation or coiling in 45patients. A flowchart illustrating excluded patients and themajor results of control DSA are shown in Figure 4. Despite thefar greater number of large, giant, and posterior circulationaneurysms, there were fewer fundus remnants in the aneu-rysms operated on by the senior author (JH), with an experi-ence of more than 2500 aneurysms treated, as compared witha group of 11 other surgeons (2.4% versus 6.1%; P � 0.0192).The same comparison of neck remnants was 10.1% versus11.7% (not significant), as giant and posterior circulation an-eurysms were operated on by the senior author. Outcomes at3 months are listed in Table 8.

DISCUSSION

An incompletely treated aneurysm may regrow and lead torecurrent symptoms of hemorrhage or mass effect. The re-bleeding rate of aneurysms with residual necks is between 3.5and 28%, (5–8, 12). In long-term follow-up of patients withextremely complex aneurysms as reported by Drake et al. (7),the rebleeding rate was 10%. Aneurysm regrowth has beenreported to occur in 3.5 to 15% of patients (6–8). The highrebleeding rate over the long term (10–20 yr) emphasizes theimportance of perfect clipping. However, aneurysms treatedwith perfect clip placement also may rupture or regrow with

FIGURE 3. Digital subtraction an-giograms of a 43-year-old man. A,ruptured aneurysm of the anteriorcommunicating artery and unrup-tured aneurysm of the right middlecerebral artery bifurcation, WorldFederation of Neurosurgical Societ-ies Grade 1. The patient underwentsurgery on Day 2. B, control digitalsubtraction angiogram showing thefundus of the middle cerebral arterybifurcation aneurysm still filling.The aneurysm of the anterior com-municating artery is completely un-clipped. The aneurysm sacs were notresected and coagulated, as is nowour standard method. C, after reop-eration, the middle cerebral arterybifurcation aneurysm is completelyoccluded, but part of the aneurysmsac of the anterior communicatingartery aneurysm remains unclipped(arrow).

KIVISAARI ET AL.

1018 | VOLUME 55 | NUMBER 5 | NOVEMBER 2004 www.neurosurgery-online.com

mass effect (6); this is also our long-term experience inFinland.

Depending on definitions and the use of control angio-grams, the reported incidence of residual neck after surgicalclipping of an aneurysm ranges from 3.8 to 18% (5, 7, 8, 12, 21,23, 24, 29, 31, 32). Despite our extensive experience, our overallresults of 12% incomplete closures and 7.4% unexpected in-complete closures are within this wide range. This might beattributable to nonselection of patients for surgery but also toour strict criteria of the small neck remnants that might beconsidered successful surgical results in some other series. Inour study, even the slightest “dog ear” was considered a neckremnant. Such neck remnants may not be recognized on rou-tine control angiograms or considered as failures in otherseries. Furthermore, postoperative control angiography is notroutinely performed in many institutions, and many surpriseswith partially ligated or filling aneurysms or occluded vesselsremain hidden even in the most experienced neurosurgical

hands. The few institutions that have presented excellent post-operative morphological results may not represent the aver-age results achieved by surgery, or the criteria for aneurysmremnants are different. However, in the series by Drake et al.(7) (Drake and Peerless have a total surgical experience of 5000aneurysm operations) of 1767 vertebrobasilar aneurysms, totalobliteration was achieved in only 82.5% of aneurysms. Thisresult is biased by the extremely difficult aneurysms in rela-tion to site and size that were treated. In addition, in ourstudy, the frequency of unexpected incomplete closure wasextremely high at the location of the anterior communicatingartery (84% unexpected of all incomplete closures), internalcarotid artery (73% unexpected of all incomplete closures),and pericallosal arteries (71% unexpected of all incompleteclosures). Specifically, the high frequency of unexpected clo-sure of the anterior communicating artery was surprising. Thismust depend on its complicated and difficult vascular anat-omy. In clipping aneurysms at these locations, intraoperativeangiography may be required in addition to careful handling.

In a study of timing of operation, Kassell et al. (17) reportedincreased tightness of the brain at early surgery. It was surprisingthat this finding did not result in more difficult dissection of theaneurysm. Accordingly, in our study, the timing of surgery did

TABLE 3. Results of control angiograms

Totalclosure (%)

Neck remnant(%)

Neck and fundusremnant (%)

Major vesselocclusion (%)

Total no. ofaneurysms

Ruptured 424 (86%) 52 (11%) 17 (3%) 30 (6%) 493

Unruptured 287 (91%) 17 (5%) 11 (3%) 14 (4%) 315

Total no. of aneurysms 711 (88%) 69 (9%) 28 (3%) 44 (5%) 808

TABLE 4. Expected and unexpected incomplete closures incontrol angiogramsa

No. (%) of incomplete closures

Expected Unexpected

Aneurysm locationICA 4 (27%) 11 (73%)MCA 14 (58%) 10 (42%)AComA 5 (16%) 26 (84%)Peric 2 (29%) 5 (71%)VBA 13 (65%) 7 (35%)Total 38 (39%) 59 (61%)

Size of aneurysmSmall (�7 mm) 6 (21%) 22 (79%)Medium (7–14 mm) 18 (36%) 32 (64%)Large (15–24 mm) 8 (62%) 5 (38%)Giant (�24 mm) 4 (100%) 0 (0%)NA 2 (100%) 0 (0%)Total 38 (39%) 59 (61%)

a ICA, internal carotid artery; MCA, middle cerebral artery; AComA, ante-rior communicating artery; Peric, pericallosal artery; VBA, vertebrobasilararteries; NA, not available.

TABLE 5. World Federation of Neurosurgical Societies grade,timing of surgery, and complete closures oncontrol angiogramsa

Complete closure (%)

WFNS grade1 178 (86%)2 82 (92%)3 23 (85%)4 68 (81%)5 73 (86%)Total 424 (86%)

Timing of surgery�72 h 347 (87%)3–10 d 54 (86%)�10 d 23 (77%)Total 424 (86%)

a WFNS, World Federation of Neurosurgical Societies (2).

CEREBRAL ANGIOGRAPHY AFTER SURGERY FOR SACCULAR ANEURYSMS

NEUROSURGERY VOLUME 55 | NUMBER 5 | NOVEMBER 2004 | 1019

not affect the number of incomplete closures despite the tightconditions caused by the red and swollen brain tissue at earlysurgery. The most important thing in acute and early aneurysm

surgery is to achieve a slack brain before beginning to clip theaneurysm(s). This can be achieved by anesthesiological meansand especially by opening of the lamina terminals or frontalventriculostomy. These maneuvers markedly reduce the need ofretraction, which results in brain contusion and damage whenuse of brain spatulas is necessary (18).

The difference of occlusion rate between unruptured andruptured aneurysms is significant, and it might be explainedby the bold dissection of unruptured aneurysms and smalleraneurysm size. In previous studies, the incidence of aneurysmneck remnants after surgery is relatively rare in the anteriorcirculation and in small aneurysms and much higher in theposterior circulation and midline, large, or giant aneurysms (5,7, 29). In our study, the incidence of aneurysm neck remnantsin posterior circulation aneurysms hidden deep in small gapswas significantly higher than in middle cerebral artery aneu-rysms, which are the most common aneurysms in Finnishseries. Large aneurysms were more often left with residualneck, often intentionally so as not to occlude any branches.

The reported high rates of unexpected major vessel occlu-sions, which also were observed in our series, harbor a greatrisk for the patient; many occur without symptoms, and theyoccur in the very best hands (7, 12, 16, 24). Those who do notobtain control angiograms do not observe the unexpectedfindings. At surgery, it is often difficult to predict the presenceof residual aneurysm or major vessel occlusion (24). Use ofmini-Doppler ultrasonography has improved the situation,but many vessels are too distal to be visualized. The goodresults of intraoperative angiography in prevention of vesselocclusions and residual aneurysms have been well docu-

TABLE 6. Aneurysm location and size and complete closureson control angiogramsa

Location and size Complete closure (%)

Aneurysm locationICA 147 (91%)MCA 282 (92%)AComA 187 (86%)Peric 38 (84%)VBA 57 (74%)Total 711 (88%)

Size of aneurysmSmall (�7 mm) 405 (94%)Medium (7–14 mm) 266 (84%)Large (15–24 mm) 15 (54%)Giant (�24 mm) 7 (64%)NA 18Total 711 (88%)

a ICA, internal carotid artery; MCA, middle cerebral artery; AComA, ante-rior communicating artery; Peric, pericallosal artery; VBA, vertebrobasilararteries; NA, not available.

FIGURE 4. Flowchart summarizing the included and excluded patients,DSA results, and treatment after DSA.

TABLE 7. Unexpected major vessel occlusions oncontrol angiogramsa

No. (%) of major vesselocclusions

Expected Unexpected

Aneurysm locationICA 3 (25%) 9 (75%)MCA 3 (21%) 11 (79%)AComA 2 (40%) 3 (60%)Peric 1 (17%) 5 (83%)VBA 3 (43%) 4 (57%)Total 12 (27%) 32 (73%)

Size of aneurysmSmall (�7 mm) 3 (17%) 15 (83%)Medium (7–14 mm) 3 (16%) 16 (84%)Large (15–24 mm) 2 (67%) 1 (33%)Giant (�24 mm) 4 (100%) 0 (0%)Total 12 (27%) 32 (73%)

a ICA, internal carotid artery; MCA, middle cerebral artery; AComA, ante-rior communicating artery; Peric, pericallosal artery; VBA, vertebrobasilararteries.

KIVISAARI ET AL.

1020 | VOLUME 55 | NUMBER 5 | NOVEMBER 2004 www.neurosurgery-online.com

mented in three recent studies (1, 19, 30). Because vesselocclusion is diagnosed early, an immediate repositioning ofthe clip probably prevents ischemic damage to the brain. Allefforts should be made during surgery to save the vessels andreplace the clip; this is the golden moment for the patient andthe surgeon.

Small residual necks detected on a postoperative angiogramare left alone because of very difficult conditions at surgery(large and giant aneurysms), and reoperation probably wouldnot achieve a better result. This leaves the responsibility to theexperienced surgeon, who must make a judgment regardingeach patient and aneurysm individually. In 7% of all patients,findings lead to reoperation or coiling. Considering the risk ofboth residual aneurysms and ischemic deficits caused by ma-jor vessel occlusion, virtually all patients should undergo in-traoperative angiography. We have tested the value of intra-operative angiography in complex (large, giant, fusiform, andvertebrobasilar artery) aneurysms. If a simpler method weremade available, intraoperative angiography should be usedeven in patients with uncomplicated aneurysms. Control an-giography and postoperative computed tomography are rec-ommended even when surgeons are highly experienced, assurprising findings can be observed despite the good condi-tion of the patient. Experience slowly improves the results.

The prevalence of middle cerebral artery aneurysms is typ-ically high in the Finnish population. In addition, the female-to-male ratio usually has been 50:50. In this series, the mostcommon ruptured aneurysm is the anterior communicatingartery aneurysm, and the number of female patients is some-what higher than that of males. The reason for this is un-known. The study group has undergone 3 months offollow-up at our clinic. This is a short recovery period foraneurysmal SAH. For complete recovery, up to 2 years isneeded, and a longer follow-up period probably would havedemonstrated better results.

The conclusion to be drawn from the Eastern Finland study(21, 31) and the International Subarachnoid Aneurysm Trial

(25) is that only competent aneurysm surgeons should con-tinue to perform open aneurysm surgery. In the presence ofinexperience, the aneurysm should be coiled, even if the ex-pected results are inferior to those that might be achieved viasurgical clipping (20, 33). Competent aneurysm and endovas-cular surgeons should form neurovascular teams to discussand tailor an individual treatment plan for each patient andensure satisfactory results. Furthermore, today and in thefuture, a great amount of research effort should involve iden-tification and treatment of aneurysms before their rupture,which will improve management results far more than anytechnical or medical advance.

CONCLUSION

Most intracranial aneurysms, ruptured or nonruptured, aretreated perfectly and permanently. Even when surgeons arecompetent and have extensive experience with aneurysm sur-gery, patients should undergo postoperative DSA, as surpris-ing findings of incomplete occlusion of the aneurysm andunplanned vessel occlusion are observed in one-sixth of pa-tients. Because aneurysms located at the posterior circulationand large- or giant-sized aneurysms either in anterior or pos-terior circulation are more prone to inadequate clipping, theirtreatment should involve intra- and postoperative angiogra-phy. If a simpler method becomes available, intraoperativeangiography should be used even in uncomplicated aneurysmcases. This series can be used as a basis to compare the resultsof other series reporting surgical or endovascular treatmentand intraoperative angiography.

REFERENCES

1. Chiang VL, Gailloud P, Murphy KJ, Rigamonti D, Tamargo RJ: Routineintraoperative angiography during aneurysm surgery. J Neurosurg 96:988–992, 2002.

TABLE 8. Outcome according to World Federation of Neurosurgical Societies grade at 3 monthsa

WFNS gradeGOS score

TotalGR (%) MD (%) SD (%) VS (%) Dead (%)

1 154 (73%) 36 (17%) 12 (6%) 2 (1%) 4 (2%) 208

2 43 (48%) 19 (21%) 20 (22%) 2 (2%) 5 (6%) 89

3 7 (26%) 7 (26%) 8 (30%) 0 (0%) 5 (19%) 27

4 19 (21%) 26 (30%) 26 (31%) 1 (1%) 12 (14%) 84

5 6 (7%) 12 (14%) 28 (33%) 15 (18%) 24 (28%) 85

Total 229 (46%) 100 (20%) 94 (19%) 20 (4%) 50 (10%) 493

a WFNS, World Federation of Neurosurgical Societies (2); GOS, Glasgow Outcome Scale (13); GR, good recovery; MD, moderate disability; SD, severe disability;VS, vegetative state.

CEREBRAL ANGIOGRAPHY AFTER SURGERY FOR SACCULAR ANEURYSMS

NEUROSURGERY VOLUME 55 | NUMBER 5 | NOVEMBER 2004 | 1021

2. Committee of the World Federation of Neurological Surgeons: Report ofWorld Federation of Neurological Surgeons Committee on a UniversalSubarachnoid Hemorrhage Grading Scale. J Neurosurg 68:985–986, 1988.

3. Dandy W: Intracranial aneurysms of the internal carotid artery: Cured byoperation. Ann Surg 107:654–659, 1938.

4. David CA, Vishteh AG, Spetzler RF, Lemole M, Lawton MT, Partovi S: Lateangiographic follow-up review of surgically treated aneurysms.J Neurosurg 91:396–401, 1999.

5. Drake CG, Allcock JM: Postoperative angiography and the “slipped” clip.J Neurosurg 39:683–689, 1973.

6. Drake CG, Friedman AH, Peerless SJ: Failed aneurysm surgery: Reoperationin 115 cases. J Neurosurg 61:848–856, 1984.

7. Drake CG, Peerless SJ, Hernesniemi JA: Surgery of Vertebrobasilar Aneurysms:London, Ontario Experience on 1,767 Patients. New York, Springer-Verlag,1996.

8. Feuerberg I, Lindquist C, Lindqvist M, Steiner L: Natural history of post-operative aneurysm rests. J Neurosurg 66:30–34, 1987.

9. Fogelholm R, Hernesniemi JA, Vapalahti MP: Impact of early surgery onoutcome after aneurysmal subarachnoid hemorrhage: A population-basedstudy. Stroke 24:1649–1654, 1993.

10. Henkes H, Fischer S, Weber W, Miloslavski E, Felber S, Brew S, Kuehne D:Endovascular coil occlusion of 1811 intracranial aneurysms: Earlyangiographic and clinical results. Neurosurgery 54:268–285, 2004.

11. Hernesniemi JA: Distal PICA aneurysms. J Neurosurg 98:1144, 2003 (com-ment).

12. Hernesniemi JA, Vapalahti MP, Niskanen M, Tapaninaho A, Kari A,Luukkonen M, Puranen M, Saari T, Rajpar M: One-year outcome in earlyaneurysm surgery: A 14 years experience. Acta Neurochir (Wien) 122:1–10,1993.

13. Jennett B, Bond M: Assessment of outcome after severe brain damage.Lancet 1:480–484, 1975.

14. Juvela S, Porras M, Poussa K: Natural history of unruptured intracranialaneurysms: Probability of and risk factors for aneurysm rupture.J Neurosurg 93:379–387, 2000.

15. Juvela S, Poussa K, Porras M: Factors affecting formation and growth ofintracranial aneurysms: A long-term follow-up study. Stroke 32:485–491,2001.

16. Karhunen PJ: Neurosurgical vascular complications associated with aneu-rysm clips evaluated by postmortem angiography. Forensic Sci Int 51:13–22, 1991.

17. Kassell NF, Torner JC, Jane JA, Haley EC Jr, Adams HP: The InternationalCooperative Study on the Timing of Aneurysm Surgery: Part 2—Surgicalresults. J Neurosurg 73:37–47, 1990.

18. Kivisaari RP, Salonen O, Ohman J: Basal brain injury in aneurysm surgery.Neurosurgery 46:1070–1076, 2000.

19. Klopfenstein JD, Spetzler RF, Kim LJ, Feiz-Erfan I, Han PP, Zabramski JM,Porter RW, Albuquerque FC, McDougall CG, Fiorella DJ: Comparison ofroutine and selective use of intraoperative angiography during aneurysmsurgery: A prospective assessment. J Neurosurg 100:230–235, 2004.

20. Koivisto T: Prospective outcome study of aneurysmal subarachnoid hemor-rhage: Endovascular versus surgical therapy (thesis). Available at: http://www.uku.fi/tutkimus/vaitokset/2002/isbn951-781-884-X.pdf.

21. Koivisto T, Vanninen R, Hurskainen H, Saari T, Hernesniemi JA, VapalahtiMP: Outcomes of early endovascular versus surgical treatment of rupturedcerebral aneurysms: A prospective randomized study. Stroke 31:2369–2377,2000.

22. Le Roux PD, Elliott JP, Eskridge JM, Cohen W, Winn HR: Risks and benefitsof diagnostic angiography after aneurysm surgery: A retrospective analysisof 597 studies. Neurosurgery 42:1248–1255, 1998.

23. Lin T, Fox AJ, Drake CG: Regrowth of aneurysm sacs from residual neckfollowing aneurysm clipping. J Neurosurg 70:556–560, 1989.

24. Macdonald RL, Wallace MC, Kestle JR: Role of angiography followinganeurysm surgery. J Neurosurg 79:826–832, 1993.

25. Molyneux A, Kerr R, Stratton I, Sandercock P, Clarke M, Shrimpton J,Holman R: International Subarachnoid Aneurysm Trial (ISAT) of neurosur-gical clipping versus endovascular coiling in 2143 patients with rupturedintracranial aneurysms: A randomised trial. Lancet 360:1267–1274, 2002.

26. Ohman J, Heiskanen O: Timing of operation for ruptured supratentorialaneurysms: A prospective randomized study. J Neurosurg 70:55–60, 1989.

27. Rinne JK, Hernesniemi JA: De novo aneurysms: Special multiple intracranialaneurysms. Neurosurgery 33:981–985, 1993.

28. Ronkainen A, Hernesniemi JA, Puranen M, Niemitukia L, Vanninen R,Ryynanen M, Kuivaniemi H, Tromp G: Familial intracranial aneurysms.Lancet 349:380–384, 1997.

29. Sindou M, Acevedo JC, Turjman F: Aneurysmal remnants after microsurgi-cal clipping: Classification and results from a prospective angiographicstudy (in a consecutive series of 305 operated intracranial aneurysms). ActaNeurochir (Wien) 140:1153–1159, 1998.

30. Tang G, Cawley CM, Dion JE, Barrow DL: Intraoperative angiographyduring aneurysm surgery: A prospective evaluation of efficacy. J Neurosurg96:993–999, 2002.

31. Vanninen R, Koivisto T, Saari T, Hernesniemi JA, Vapalahti MP: Rupturedintracranial aneurysms: Acute endovascular treatment with electrolyticallydetachable coils—A prospective randomized study. Radiology 211:325–336,1999.

32. Yasargil MG: Microneurosurgery: Microsurgical Anatomy of the Basal Cisternsand Vessels of the Brain. Stuttgart, Georg Thieme, 1984, vol I.

33. Yasargil MG: Reflections on the thesis “Prospective outcome study of an-eurysmal subarachnoid hemorrhage” of Dr. Timo Koivisto. Available at:http://www.uku.fi/tutkimus/vaitokset/2002/isbn951-780-338-9.pdf.

AcknowledgmentsThis study was supported by the Maire Taponen Foundation. We thank Avula

Chakrawarthi, M.D., for the preparation of the DSA studies for analysis.

COMMENTS

The neurosurgery group at Helsinki University CentralHospital presents a large and well-managed series of sur-

gically treated cerebral aneurysms. More than 800 intracranialaneurysms were surgically clipped, followed by postoperativeangiography. In this series, 12% of the aneurysms that wereclipped were incompletely occluded. Sixty-one percent ofthese incomplete closures were totally unexpected on postop-erative angiograms. Many of these patients were subjected torepeat operation for clip repositioning. Some patients weresent for endovascular treatment. Similarly, 44 major vessel orbranch occlusions were noted on postoperative angiograms,and 32 of these occlusions were totally unexpected. Althoughlarge, giant, and complex aneurysms had significantly higherrates of residuals and vessel occlusions, it is interesting to notethat many of the relatively simple anterior-circulation aneu-rysms also had unexpected findings on postoperative angiog-raphy. These data make a strong and convincing argument forthe use of intraoperative angiography.

It is likely that intraoperative angiography should be usedon all aneurysm patients, because problems with clipping aresometimes unexpected, even on relatively simple aneurysms.It is often too late to prevent a stroke if there is a vesselocclusion that is picked up after surgery rather than withinminutes after the clip has been placed. Similarly, the risks oftaking the patient back to the operating room after a routinepostoperative angiogram can be largely eliminated with theuse of intraoperative angiography. For the past few years, Ihave used intraoperative angiography routinely on all aneu-rysm patients. Previously, I operated without the use of intra-operative angiography, but I believe that this technique is

KIVISAARI ET AL.

1022 | VOLUME 55 | NUMBER 5 | NOVEMBER 2004 www.neurosurgery-online.com

worth the small amount of inconvenience and the slight in-crease in time necessary to perform intraoperative studies. Inretrospect, I realize that there were several intraoperative ca-tastrophes that might have been avoided had intraoperativeangiography been used. Since using intraoperative angiogra-phy, I have discovered several situations in which faulty clipplacement was easily rectified.

Robert A. SolomonNew York, New York

Kivisaari et al. provide a good contribution to the literaturesupporting the benefits of routine cerebral angiography

after surgery for saccular aneurysm. They report a retrospec-tive series of 622 patients with 955 aneurysms, of which 808were surgically clipped and underwent postoperative angiog-raphy. They showed that even in very experienced hands, thefrequency of unexpected incomplete closure was very high.This proves that it is worthwhile to perform a routine cerebralangiogram after clipping saccular aneurysms.

They also showed, as a confirmation of what has been wellestablished in the literature, that the size of the aneurysm hada significant impact on the closure rate. Large, complex, andgiant aneurysms had a rather relevant smaller closure rate. Insuch cases, it is preferable to perform an intraoperativeangiography.

The occurrence of unexpected major vessel occlusions wasalso high in their series. Such occurrences can ideally be dealtwith when an intraoperative angiogram is performed. How-ever, the authors were able to reoperate on 14 patients withsuch unexpected major artery occlusions, in 2 by performingbypass surgery and in 12 by repositioning the clip. In total, 45patients of the 808 were reoperated on or treated by theendovascular route because of the postoperative controlangiogram.

Another interesting finding of this article is that, brieflystated, experience accounts for better final results, as shownby the percentage of fundus remnant in the series of the seniorauthor compared with those of the other 11 surgeons. After allthat has been said, we must simply agree with the authors’conclusion: postoperative cerebral angiography should be per-formed in all patients, and in the case of complex and giantaneurysms, the intraoperative angiography is mandatory.

Atos Alves de SousaMinas Gerdis, Brazil

The authors present the largest series yet published on theefficacy of routine postoperative angiography for the eval-

uation of aneurysmal remnants. The authors carefully andmethodically analyzed a variety of factors that might contrib-ute to incomplete clipping. As noninvasive imaging methodsadvance and the use of intraoperative angiography becomesmore common, the routine use of angiography is controver-sial. Our institution recently completed a prospective evalua-tion of intraoperative angiography as well as an evaluation ofthe surgeon’s perceived need for angiography. We now be-

lieve that the benefits of intraoperative angiography outweighthe risks, and we have eliminated the use of routine postop-erative studies. The findings from the study, designed toevaluate the surgeon’s perceived need for angiography, sug-gested that the rate of unanticipated incomplete clipping ishigher than reported. This aspect is very difficult to evaluateretrospectively.

Ruth BristolRobert F. SpetzlerPhoenix, Arizona

The authors present a retrospective analysis of 622 patientswith cerebral aneurysms who underwent surgical clipping

and postoperative angiography. Incomplete closures of aneu-rysms were detected in 97 patients by postoperative angiog-raphy. Thirty-eight of 97 incomplete closures were intentional,but 59 of 97 incomplete closures were unexpected. Conversely,44 cases of major vessel occlusions were revealed by postop-erative angiography, and 32 of 44 cases were unexpected. It issurprising that such a lot of unexpected findings were re-vealed by postoperative angiography, although experiencedneurosurgeons performed clipping. It is believed that the com-plete closure of an aneurysm is crucial for the prevention of itssubsequent growth and rupture. However, postoperative an-giography recently tends to be avoided in many neurosurgicalinstitutes. Hence, this article is considered to be importantbecause it raises a warning against this tendency.

In this article, the authors have analyzed the findings ofpostoperative angiography, but not intraoperative angiogra-phy, in aneurysm surgery. However, they conclude and rec-ommend that intraoperative angiography should be used inmore basic aneurysm cases, if a simpler method is available.Intraoperative angiography is considered to be very useful inaneurysm surgery, because we can detect unexpected residualaneurysms or vessel occlusions during operations. However,it has been reported that the false-negative intraoperativeangiography rate is 1.1 to 8.3% (1–3). Therefore, intraoperativeangiography should be conducted under ideal conditions andinterpreted by experienced neurosurgeons and neuroradiolo-gists. Hence, the continued necessity for postoperative angiog-raphy should be emphasized.

Masato HojoNobuo HashimotoKyoto, Japan

1. Chiang VL, Gailloud P, Murphy KJ, Rigamonti D, Tamargo RJ: Routineintraoperative angiography during aneurysm surgery. J Neurosurg 96:988–992, 2002.

2. Tang G, Cawley CM, Dion JE, Barrow D: Intraoperative angiography duringaneurysm surgery: A prospective evaluation of efficacy. J Neurosurg 96:993–999, 2002.

3. Vitaz TW, Gaskill-Shipley M, Tomsick T, Tew JM Jr: Utility, safety, andaccuracy of intraoperative angiography in the surgical treatment of aneu-rysms and arteriovenous malformations. AJNR Am J Neuroradiol 20:1457–1461, 1999.

CEREBRAL ANGIOGRAPHY AFTER SURGERY FOR SACCULAR ANEURYSMS

NEUROSURGERY VOLUME 55 | NUMBER 5 | NOVEMBER 2004 | 1023

The authors’ important conclusion is that intraoperative an-giography should be performed in virtually all patients. Sup-

port for their argument can be found in the data for both rup-tured and unruptured aneurysms when the majority of remnantswere unexpected. Of 44 vessel occlusions, 32 were unexpected.Obviously, the demand for intraoperative angiography must betaken in the context of some associated risk to the patient asso-ciated with this procedure. In high-volume centers, however,

intraoperative studies should be achievable with minimal riskand minimal extension of the operative procedure. There is littledoubt that this practice should be applied to aneurysms of theanterior communicating artery complex, because 84% of residualaneurysms were unexpected in this series.

H. Hunt BatjerChicago, Illinois

Thirty Systematic Color Series (oil on linen, 1950–1955) by Richard Paul Lohse (courtesy of the Richard Paul Lohse Foundation, on loan to Kunsthaus, Zürich).

KIVISAARI ET AL.

1024 | VOLUME 55 | NUMBER 5 | NOVEMBER 2004 www.neurosurgery-online.com