How does spontaneous hemostasis occur in ruptured cerebral aneurysms? Preliminary investigation on 247 clipping surgeries

How does spontaneous hemostasis occur in ruptured cerebral aneurysms? Preliminary investigation on 247 clipping surgeries

Surgical Neurology 66 (2006) 269 – 276 www.surgicalneurology-online.com Aneurysm How does spontaneous hemostasis occur in ruptured cerebral aneurysm...

375KB Sizes 7 Downloads 64 Views

Surgical Neurology 66 (2006) 269 – 276 www.surgicalneurology-online.com

Aneurysm

How does spontaneous hemostasis occur in ruptured cerebral aneurysms? Preliminary investigation on 247 clipping surgeries Tatsuya Ishikawa, MDa,4, Naoki Nakayama, MDb, Tetuyuki Yoshimoto, MDc, Takeshi Aoki, MDd, Shynsuke Terasaka, MDe, Mikio Nomura, MDf, Akihiro Takahashi, MDg, Satoshi Kuroda, MDa, Yoshinobu Iwasaki, MDa a

Department of Neurosurgery, Hokkaido University Graduate School of Medicine, Sapporo, Hokkaido 060-8638, Japan b Department of Neurosurgery, Sapporo Azabu Neurosurgical Hospital, Sapporo, Hokkaido 007-0840, Japan c Department of Neurosurgery, Kashiwaba Neurosurgical Hospital, Sapporo, Hokkaido 062-8513, Japan d Department of Neurosurgery, Hokkaido Neurosurgical Memorial Hospital, Sapporo, Hokkaido 060-0022, Japan e Department of Neurosurgery, Teine Keijinkai Hospital, Sapporo, Hokkaido 006-8555, Japan f Department of Neurosurgery, Keiwakai Ebetsu Hospital, Ebetsu-shi, Hokkaido 069-0817, Japan g Department of Neurosurgery, Sapporo Tokeidai Hospital, Sapporo, Hokkaido 060-0031, Japan Received 26 October 2005; accepted 1 March 2006

Abstract

Background: Rupture of cerebral aneurysms results in subarachnoid hemorrhage. In many cases, bleeding from aneurysms spontaneously arrests. Although bleeding from cerebral aneurysms has been reported to arrest from outside, bleeding from some aneurysms can arrest in different ways. Methods: Between April 2002 and March 2004, we prospectively investigated mechanisms of spontaneous hemostasis in ruptured aneurysms by macroscopic examination when performing craniotomy and clipping surgeries. Results: Hemostatic mechanisms were investigated in 247 patients with ruptured aneurysm (77 men, 170 women; age range, 25-95 years). Hemostatic mechanisms were divided into 3 different patterns. In the most common pattern (79.4%), the surface of the aneurysm rupture point was sealed from the outside by a platelet plug or fibrin net (outside-arrest pattern). In some aneurysms (10.1%), a thrombus or platelet plug was attached to the rupture point from inside the aneurysm (inside-arrest pattern). In a very small number of aneurysms (1.6%), a naked thrombus covered the hole made on the arterial wall or small remnant of the aneurysmal dome (bursting pattern) The mechanism remained unclear in the remaining 8.9% of aneurysms. Multivariate analysis revealed that alert consciousness on admission (WFNS grade I) significantly associated with usual hemostasis (outside-arrest-pattern: OR, 3.8; 95% CI, 1.4-10.0; P = .008). Borderline association with usual hemostasis was found in aneurysms with a size of 5 or smaller than 5 mm (OR, 2.6; 95% CI, 0.99-7.1; P = .052). Conclusions: The present preliminary study revealed that arrest of bleeding from a ruptured cerebral aneurysm does not always occur from outside the aneurysm. Unusual mechanisms of hemostasis are seen in approximately 12% of ruptured aneurysm. The outside-arrest-pattern aneurysm was more common for smaller aneurysms, and these patients tended to be of better grade. Further studies are necessary to explore the mechanism of hemostasis for ruptured cerebral aneurysms. D 2006 Elsevier Inc. All rights reserved.

Keywords:

Blister-like aneurysm; Intraaneurysmal thrombus; Ruptured cerebral aneurysm; Spontaneous hemostasis

Abbreviations: A-com, anterior communicating artery; CI, confidence interval; CT, computed tomography; GDC, Guglielmi detachable coil; ICA, internal carotid artery; OR, odds ratio; SAH, subarachnoid hemorrhage; WFNS, World Federation of Neurological Surgeons. 4 Corresponding author. Department of Neurosurgery, Saitama Medical Center, Kawagoe-shi, Saitama-ken 350-8550, Japan. Tel.: +81 49 228 3671; fax: +81 49 228 3671. E-mail address: [email protected] (T. Ishikawa). 0090-3019/$ – see front matter D 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.surneu.2006.03.030

1. Introduction The rupture of cerebral aneurysms results in SAH. Recently, mechanisms for aneurysm growing and rupture have been studied by histologic analysis [6]. However, the hemostatic mechanism once after aneurysm rupture has not

270

T. Ishikawa et al. / Surgical Neurology 66 (2006) 269–276

been fully examined. In many cases, bleeding from aneurysms spontaneously arrests. Bleeding from a cerebral aneurysm has been reported as arresting from the outside by a platelet plug that builds up at the surface of the rupture point [8,16]. This platelet plug is reinforced by the deposition of fibrin within [16]. Most of the sealing effect reportedly arises from a mass of fibrin embedded with platelets, lying external to the sac [4,21]. Conversely, ruptured cerebral aneurysms have been described, in which bleeding is arrested by thrombosis within the aneurysm [20]. Another type of aneurysm, the bblister-likeQ aneurysm, displays shape and wall characteristics differing from the usual saccular aneurysms [1,10,14,15,17,18]. Moreover, some reports have noted aneurysms in which the aneurysmal dome has been lost and the point of bleeding is covered solely by thrombus, sometimes associated with the formation of a pseudoaneurysm [9]. Our previous experiences have led us to recognize that a significant number of ruptured aneurysms is associated with unique hemostatic mechanisms [9,10]. However, no previous studies have addressed this issue. The present study therefore involved preliminary examination of patterns of hemostasis in ruptured cerebral aneurysms, with a view to exploring appropriate treatment strategies for ruptured cerebral aneurysms. 2. Materials and methods Between April 2002 and March 2004, mechanisms of hemostasis were prospectively investigated in ruptured aneurysms that were surgically treated at Hokkaido University Hospital and affiliated hospitals. During this period, data on 247 consecutive patients who underwent craniotomy and clipping surgery for aneurysmal subarachnoid hemorrhage was accumulated. On the other hand, 24 aneurysms, mainly situated in the posterior circulation, were treated with Guglielmi detachable coil (GDC) coil embolization. Age, sex, WFNS grade, Fisher CT group, aneurysm location, size,

time interval between the SAH onset and macroscopic examination of the aneurysms, factors which might have influence to the hemostatic mechanism (history of hypertension, coagulopathy, and use of anticoagulant or antiplatelet drugs before SAH onset), and mechanism of hemostasis were identified for these 247 patients who received clipping surgery. Because of the small number of patients with WFNS grade III, patients with WFNS grades II and III were united into a single group. Aneurysms were classified according to the maximum diameter on cerebral angiography as follows: small, 5 mm or smaller; regular, N5 to 15 mm; large, larger than 15 mm. We also investigated their angiographic appearance to determine whether we could identify a thrombus inside the aneurysm. After the aneurysms were secured by neck clipping, the dome of the aneurysm, along with the rupture point, was very carefully dissected from surrounding structures with irrigation system. The rupture point of the aneurysm was then visually inspected with regard to the mechanism of spontaneous hemostasis. Aneurysmal domes were resected for histologic analysis if the resection of domes was possible, keeping the rupture point along with surrounding structures intact. The sections were subjected to the hematoxylin-eosin stain to confirm the findings on visual inspection. Mechanisms of hemostasis were categorized into 3 different patterns based on visual inspection. The most common pattern involved the surface of the rupture point being sealed from the outside by a platelet plug and fibrin net (outside-arrest-pattern). In the outside-arrest-pattern aneurysm, sharp dissection between the aneurysmal dome and the surrounding structures usually results in a deflation of the aneurysmal dome because the remaining blood comes out from the hole (rupture point), which has been sealed from the outside. In some aneurysms, thrombus or platelet plugs had attached to the rupture point from inside the aneurysm (inside-arrest-pattern). In the inside-arrestpattern aneurysm, on the other hand, complete dissection of the aneurysmal dome from the surrounding structures does

Fig. 1. Illustrative case of an outside-arrest-pattern aneurysm is demonstrated. Intraoperative view of a ruptured middle cerebral artery aneurysm in an 83-yearold housewife. The plug comprised of platelets and fibrin sealing the rupture point of the aneurysm from the outside (left). Photomicrograph showing thrombus and fibrin nets covering the rupture point (arrow) from the outside the aneurysmal wall (hematoxylin-eosin, original magnification 40) (right).

T. Ishikawa et al. / Surgical Neurology 66 (2006) 269–276

not cause a deflation of the aneurysmal dome, and thrombus, to some extent, can be seen through the thin walls of the aneurysm. There is actually some overlapping regarding the presence of thrombus around the rupture point, inside only, outside only, or both inside and outside; the grouping performed based on either side of thrombus has a main sealing effect, that is, whether sharp dissection between the aneurysmal dome and surrounding structures causes a deflation of the aneurysmal dome or not. In a very small number of aneurysms, naked thrombus covered the hole in the arterial wall, formed by a small remnant of aneurysmal dome (bursting pattern). In the bursting-pattern aneurysm, the aneurysmal wall is almost lost, and dissection from the surrounding structures is impossible. After the removal of the attached thrombus, only a small remnant of aneurysmal wall, which was closed by clip, can be observed. Pathological examination investigating hemostatic mechanism was possible in only 17 aneurysms. Pictures and pathologies of representative cases in each pattern are shown in Figs. 1-3. Frequencies of each pattern and differences in associated clinical characteristics were examined. The time interval

271

between the SAH onset and macroscopic examination of the aneurysms was compared using an unpaired t test. The other univariate analyses were performed using v 2 tests. A multivariate logistic regression model was conducted to test the effect on unusual hemostatic pattern (inside-arrest or bursting pattern) of patient sex, age, consciousness disturbance on admission (whether WFNS grade I or II-V), aneurysm location, size (whether smaller or bigger than 5 mm), Fisher CT group, time interval between the SAH onset and macroscopic examination of the aneurysms, and several factors which might have influenced their hemostatic mechanism. Statistical analysis was completed with SPSS for Windows, version 12.0 (SPSS, Inc, Chicago, Ill). 3. Results The mean age of the 247 patients (77 men and 170 women) was 60.1 F 12.5 years (range, 25-95 years). The time interval between SAH onset and macroscopic examination of the aneurysms, factors which might have influenced the hemostatic mechanism, WFNS grade, Fisher CT group, and location and size of aneurysms are listed in Table 1.

Fig. 2. Illustrative case of an inside-arrest-pattern aneurysm is demonstrated. Left internal cerebral artery angiography in a 56-year-old woman shows an A-com aneurysm with a dumbbell shape (upper left). The aneurysm was clipped via an anterior interhemispheric approach, and the dome of the aneurysm was found to be partially thrombosed (upper right). A tiny hole on the aneurysmal wall at the rupture point was sealed mainly by thrombus from inside the aneurysm (lower left). Photomicrograph showing thrombus and fibrin nets exist inside of the aneurysm wall, which is partially organized, covering the rupture point (arrow) (hematoxylin-eosin, original magnification 40) (lower right).

272

T. Ishikawa et al. / Surgical Neurology 66 (2006) 269–276

Fig. 3. Illustrative case of a bursting-pattern aneurysm is demonstrated. Left cerebral artery angiography in an 84-year-old woman taken on admission shows no aneurysm (upper left). Repeated angiography on day 4 after onset demonstrates aneurysmal protrusion at the A-com (upper right). Clipping surgery was performed via bifrontal craniotomy and an anterior interhemispheric approach. Intraoperatively, thrombus was seen attached to the aneurysm (lower left). The aneurysm wall was almost completely lost, and thrombus just covered the remnant dome (lower right).

A pattern of hemostasis was apparent intraoperatively in 225 of the 247 ruptured aneurysms. Of these, 196 aneurysms (79.4%) were outside-arrest-pattern, 25 (10.1%) were insidearrest-pattern, and 4 (1.6%) were bursting pattern. In the remaining 22 aneurysms, inspection of the rupture point was impossible for several reasons, predominantly premature rupture before dissection of the dome. The result of visual inspection completely corresponded with histologic findings in both inside- and outside-arrest-pattern aneurysms, which were performed in 17 aneurysmal dome specimens. There were 13 outside-arrest-pattern aneurysms and 4 insidearrest-pattern aneurysms. The specimens for the burstingpattern aneurysms were not harvested. The time interval between the SAH onset and macroscopic examination of the aneurysms ranged from 3 to 504 (median, 9) hours. The median time interval between the SAH onset and macroscopic examination of the aneurysm was 9 hours in the outside-arrest pattern and 8.5 hours in the inside-arrest pattern, whereas it was 33 hours in the bursting pattern. Although there were no statistically

significant differences among the patterns of hemostasis, the extended time interval in the bursting pattern is due to 2 of the 4 bursting-pattern aneurysms found on repeated angiography and then surgically treated. There were only 2 patients with coagulopathy because of advanced liver cirrhosis in both. There were no patients receiving anticoagulants, but antiplatelet drugs were given to 4. There was no significant difference in the incidence of factors such as history of hypertension, coagulopathy, and antiplatelet or anticoagulant drugs among each pattern. Patients harboring aneurysms with the usual mechanisms of hemostasis (outside pattern) displayed significantly better WFNS grade ( P b .05, v 2 test) than patients with insidearrest or bursting pattern aneurysms. On multivariate analysis, alert consciousness on admission (WFNS grade I) was significantly associated with usual hemostasis (OR, 3.8; 95% CI, 1,4-10.0; P = .008). There were no significant differences in the Fisher CT group among each pattern of hemostasis. Findings for cerebral aneurysms, including location and size, were retrospectively analyzed. No differences in

T. Ishikawa et al. / Surgical Neurology 66 (2006) 269–276

angiography and appeared as small but distinct aneurysms on repeated angiography.

Table 1 Patient and aneurysm characteristics and pattern of hemostatis Outside-arrest Inside-arrest Bursting pattern pattern pattern

273

Unknown

Number 196 (79.4%) 25 (10.1%) 4 (1.6%) 22 (8.9%) Age (y) 25-89 50-86 42-65 23-95 Mean + SD 59.0 + 11.5 65.4 + 11.9 50.5 + 10.2 65.7 + 18.1 Sex 62:134 10:15 2:2 3:19 (male:female) Interval between the SAH onset and macroscopic examination of the aneurysms (h) Median 9 8.5 33 8.25 Minimum4-384 3-504 6-120 5-432 maximum Factors which might have influenced the hemostatic mechanism Hypertension 62 (32%) 8 (32%) 2 (50%) 6 (27%) Coagulopathy 2 (1%) 0 (0%) 0 (0%) 0 (0%) Antiplatelet 4 (2%) 0 (24%) 0 (0%) 0 (0%) drugs WFNS grade I 105 8 0 8 II/III 36 5 1 6 IV 29 5 2 4 V 25 8 1 4 Aneurysm location ICA 55 3 0 11 A-com 53 9 3 2 MCA 64 9 0 4 ACA 15 3 0 3 VA-BA 9 1 1 2 Aneurysm size Small ( b 5 mm) 90 4 4 8 Regular 95 16 0 12 (5 b 15 mm) Large or giant 11 5 0 2 ( N 15 mm) Fisher CT group 1 12 2 0 1 2 60 2 0 4 3 89 14 2 15 4 35 6 2 4 MCA, middle cerebral artery; ACA, anterior cerebral artery; VA, vertebral artery; BA, basilar artery.

characteristic locations were identified between outsideand inside-arrest-pattern aneurysms. On univariate analysis, the frequency of small aneurysms was significantly less for inside-arrest-pattern aneurysms than for outside-arrestpattern aneurysms ( P b .01, v 2 tests). On multivariate analysis, the smaller aneurysm size (5 or b 5 mm in diameter) was a borderline predictor of usual hemostasis (outside-arrest-pattern; OR, 2.6; 95% CI, 0.99-7.1; P = .052). However, we could not detect the presence of inside thrombus on a cerebral angiogram. Bursting-pattern aneurysms were quite rare (n = 4; 1.6%) and were predominantly seen at the A-com artery (3 of 4; 75%). The remaining bursting-pattern aneurysm was identified at the branching of the posterior inferior cerebellar artery from the vertebral artery. In the bursting pattern, all aneurysms were small. As an angiographic characteristic, 2 of the 4 bursting-pattern aneurysms were missed on initial

4. Discussion In the present study, the mechanism of the hemostasis was grouped based on mainly visual inspection during the clipping surgery. One would criticize that visual inspection only means nothing without the pathologic confirmation. However, in a clinical setting, neurosurgeons know that structures covering rupture points were usually lost with surgical procedures. Therefore, harvesting the aneurysm wall, along with the rupture point as well as the structures covering it, for the purpose of investigating hemostatic mechanism, is nearly impossible in many cases. Indeed, in 17 specimens we have harvested, the pathologic finding was concurrent with grouping based on visual inspection. In this sense, visual inspection provides us irreplaceable and practical information in this kind of study, although it must be followed by further sophisticated investigations. Another criticism would be that the time interval between SAH onset and visual inspection would relate to thrombus formation and melting. However, time interval was not statistically different among the patterns of hemostasis. Moreover, if we limited the analysis on the patients who were treated within 24 hours from the onset, a unique bursting-pattern aneurysm might not be identified. The hemostatic mechanism immediately after the rupture of aneurysms has not been fully examined. However, as previously reported, the sealing effect at the rupture point results from outside the aneurysm by platelet plugs and fibrin nets in many cases [4,8]. In some patients, the hemostatic process can take place from inside the aneurysm. Actually, bleeding from a ruptured cerebral aneurysm can be arrested by thrombosis within the aneurysm. In an old report by Shunk [20], 26 of 110 aneurysms manifested intraaneurysm thrombosis to various extents, although their study included aneurysms of various sizes, locations, and both ruptured and unruptured status. In the present study, about 10.1% of ruptured cerebral aneurysms displayed hemostatic processes occurring from inside the aneurysms. The arrest of bleeding by complete thrombosis within the aneurysm is not common, and temporary thrombosis is considered one reason for the initial angiography missing the aneurysm. Actually, about 15% of patients with subarachnoid hemorrhage are angiographically occult [12]. Among such patients, repeated angiography detects the presence of a ruptured aneurysm after recanalization [2,5,11,14,19]. Recent pathologic studies have reported the disappearance of normal intima and smooth muscle cells in the wall of a ruptured aneurysm, associated with macrophage infiltration into the wall [3,13]. Those studies speculated that the appearance of macrophages might cause aneurysm rupture, but not result from aneurysm rupture. Ruptured

274

T. Ishikawa et al. / Surgical Neurology 66 (2006) 269–276

aneurysms displayed loss of intima and a covering of fibrin and blood cells. In conditions where the intima is lost, hemostatic processes may develop to form an intraaneurysmal thrombus. Actually, the percentage of small aneurysms was significantly lower for inside- than for outside-arrest-pattern aneurysms. On multivariate analysis, a small-size aneurysm (5 or b5 mm in diameter) was a borderline predictor for usual hemostasis (outside-arrestpattern). Aneurysms may need to reach a certain size threshold to form an intraaneurysmal thrombus or platelet plug. It is reasonable that larger aneurysms may produce stronger turbulent flow within the aneurysms and then, easily, thrombi formation. By means of our investigation, however, the intraaneurysmal thrombi that existed cannot always be differentiated from those formed after rupture. Large aneurysms tend to be partially thrombosed, therefore, inside-arrest-pattern aneurysms may be larger than outside-arrest-pattern aneurysms. There is actually some overlapping regarding the presence of thrombus around the rupture point (inside only, outside only, or both inside and outside), by means of intraoperative visual inspection. However, we could easily judge which side has a main sealing effect for the rupture point in all aneurysms. As described before, the grouping was performed based on which side of the thrombus had a main sealing effect, that is, whether sharp dissection between the aneurysmal dome and the surrounding structures causes a deflation of aneurysmal dome or not. A rare and unique hemostatic pattern was identified in this study—the bursting pattern. In such aneurysms, naked thrombus covers the hole in the arterial wall or small remnant of the aneurysmal dome. These aneurysms resemble the blister-like aneurysms frequently seen on the dorsal wall of the ICA although, in our present series, did not include any dorsal ICA aneurysms. Dorsal ICA aneurysms are known to carry a higher risk of intra- and postoperative massive hemorrhage [1,10,14,15,17]. The pathogenesis of blister-like ICA dorsal aneurysms remains unclear [17]. Although Ohkuma et al [18] recently recategorized these aneurysms as dissecting aneurysms, our experience leads us to consider that the pathology is attributable, at least in part, to laceration of the arterial wall plugged by a naked thrombus [10]. The mechanism of hemostasis for ICA dorsal aneurysms may display characteristics in common with the bursting-pattern aneurysms described herein. In our present series, bursting-pattern aneurysm was rare but predominantly seen in the A-com. Another characteristic is that bursting-pattern aneurysms appeared to be relatively small. The other most notable characteristic is that these aneurysms are easily missed on initial angiography and appear as small aneurysms on repeated angiography. As described earlier, the appearance of initially absent aneurysms on repeated angiography is seen with totally thrombosed aneurysms [2,5,11,14,19]. Such angiographic findings are highly suggestive of aneurysms with an unusual hemostatic mechanism (inside-arrest or bursting pattern).

It has been explained that at the moment an aneurysm ruptures, intracranial pressure rises to equal systemic arterial blood pressure then normalizes [7,16]. During the small interval when intracranial and arterial blood pressures are equivalent, hemostatic processes occur at the rupture point from either inside or outside the aneurysm. If intracranial pressure rises to the blood pressure, then the patient would become unconscious within a few seconds to minutes. However, not all patients with SAH in WFNS grade I become unconscious. In patients with outside-arrest-pattern aneurysms, the blood-stream from the rupture point contacts with the arachnoid membrane and/or brain tissue, and thus, thrombus is formed from outside the aneurysm leading to an arrest of bleeding within a very short period before intracranial pressure rises. Indeed, the alert consciousness on admission (viz, WFNS grade I) was an independent predictor of usual hemostasis (outside-arrest-pattern). Impact from the first aneurysmal bleed was much greater with unusual patterns of hemostasis. Aneurysms of the bursting pattern might reasonably be expected to take much longer to achieve hemostasis because the hole in the aneurysm can be substantially larger than in other patterns. In the insidearrest-pattern, a longer time is required for thrombus formation within the aneurysm where turbulent flow is present, and the impact of rupture on the brain is stronger as a result. Coil embolization does not appear theoretically suitable for bursting-pattern aneurysms, in which the aneurysm sac has been lost. Coil embolization in these aneurysms may result in massive rupture. Fortunately, bursting-pattern aneurysms are angiographically small and do not usually seem to be selected for coil embolization. Moreover, for bursting-pattern aneurysms, clipping surgery theoretically carries a higher risk for intraoperative rupture as described above. We can also speculate that, when inside-arrestpattern aneurysms were treated with coil embolization, the aneurysm reopening may occur along with the regression of thrombus inside. 5. Conclusions The arrest of bleeding from ruptured cerebral aneurysms does not always occur from outside the aneurysms. Unusual mechanisms of hemostasis are seen in approximately 12% of ruptured aneurysms. The outside-arrest-pattern aneurysm was more common for smaller aneurysms and these patients tended to be of better grade. Further studies are necessary to explore the mechanism of hemostasis for ruptured cerebral aneurysms. Acknowledgments The authors thank Drs H Saitoh and H Yasuda of Sapporo Azabu Neurosurgical Hospital, Dr T Aida of Hokkaido Neurosurgical Memorial Hospital, Drs T Kashiwaba and S Kaneko of Kashiwaba Neurosurgical Hospital,

T. Ishikawa et al. / Surgical Neurology 66 (2006) 269–276

275

Dr K Kazumata of Teine Keijinkai Hospital, and Dr S Takikawa of Chitose Municipal Hospital for supporting the prospective study. We also thanks to Dr T Abumiya of Hokkaido University Hospital and Dr H Kamiyama of Asahikawa Red-Cross Hospital for helpful discussions.

[21] Suzuki J, Ohara H. Clinicopathological study of cerebral aneurysms. Origin, rupture, repair, and growth. J Neurosurg 1978;48:505 - 14.

References

The authors present novel hypotheses regarding the mechanisms responsible for spontaneous hemostasis in acute intracranial aneurysm rupture. They describe 3 hemostatic mechanisms based upon the appearance of the aneurysms at surgery—intraluminal thrombus/platelet plug, extraluminal platelet plug/fibrin net, and naked thrombus in the aneurysm fundus. In a small number of cases, the authors confirm intraoperative findings with histologic analysis of the rupture site. This report sheds some much-needed light on the mechanisms responsible for the cessation of bleeding in intracranial aneurysm rupture. Their findings may explain why many patients survive aneurysmal subarachnoid hemorrhage without loss of consciousness, whereas others survive with extensive local brain injury. A logical next step in this work would be examination of animal models of subarachnoid hemorrhage to see whether similar hemostatic mechanisms can be identified and studied. From a practical clinical perspective, the importance of platelet plugs in the hemostatic mechanism may argue against the routine use of antiplatelet drugs for coronary prophylaxis in patients with both coronary disease and unruptured aneurysms.

[1] Abe M, Tabuchi K, Yokoyama H, Uchino A. Blood blisterlike aneurysms of the internal carotid artery. J Neurosurg 1988;89:419 - 24. [2] Bohmfalk GL, Story JL. Intermittent appearance of a ruptured cerebral aneurysm on sequential angiograms. Case report. J Neurosurg 1980; 52:263 - 5. [3] Chyatte D, Bruno G, Desai S, Tordor R. Inflammation and intracranial aneurysms. Neurosurgery 1999;45:1137 - 47. [4] Fisher CM, Ojemann RG. Basal rupture of saccular aneurysm. A pathological case report. J Neurosurg 1978;48:642 - 4. [5] Forster DMC, Steiner L, Hakanson S, Bergvall U. The value of repeat pan-angiography in cases of unexplained subarachnoid hemorrhage. J Neurosurg 1978;48:712 - 6. [6] Frfsen J, Pippo A, Paetau A, Kangasniemi M, Niemel7 M, Hernesniemi J, et al. Remodeling of saccular cerebral artery aneurysm wall is associated with rupture. Histological analysis of 24 ruptured and 42 ruptured cases. Stroke 2004;35:2287 - 93. [7] Grote E, Hassler W. The critical first minutes after subarachnoid hemorrhage. Neurosurgery 1988;22:654 - 61. [8] Houkin K, Kuroda S, Takahashi A, Takikawa S, Ishikawa T, Yoshimoto T, et al. Intra-operative premature rupture of the cerebral aneurysms. Analysis of the causes and management. Acta Neurochir (Wien) 1999;141:1255 - 63. [9] Ishikawa T, Kamiyama H, Kazumata K, Takizawa K. How to prevent or manage laceration of aneurysms at their neck. Surg Cereb Stroke (Jpn) 2002;30:153 - 8. [10] Ishikawa T, Nakamura N, Houkin K, Nomura M. Pathological consideration of a bblister-likeQ aneurysm at the superior wall of the internal carotid artery: case report. Neurosurgery 1997;40:403 - 6. [11] Iwanaga H, Wakai S, Ochiai C, Narita J, Inoh S, Nagai M. Ruptured cerebral aneurysms missed by initial angiographic study. Neurosurgery 1990;27:45 - 51. [12] Jafar JJ, Weiner HL. Surgery for angiographically occult cerebral aneurysms. J Neurosurg 1993;79:674 - 9. [13] Kataoka K, Taneda M, Asai T, Kinoshita A, Ito M, Kuroda R. Structural fragility and inflammatory response of ruptured cerebral aneurysms. A comparative study between ruptured and unruptured cerebral aneurysms. Stroke 1999;30:1396 - 401. [14] Lorenzo ND, Guidetti G. Anterior communicating aneurysm missed at angiography: report of two cases treated surgically. Neurosurgery 1988;23:494 - 9. [15] Nakagawa F, Kobayashi S, Takemae T, Sugita K. Aneurysms protruding from the dorsal wall of the internal carotid artery. J Neurosurg 1986;65:303 - 8. [16] Nornes H. The role of intracranial pressure in the arrest of hemorrhage in patients with ruptured intracranial aneurysm. J Neurosurg 1973; 39:226 - 34. [17] Ogawa A, Suzuki M, Ogasawara K. Aneurysms at nonbranching sites in the supraclinoid portion of the internal carotid artery: internal carotid artery trunk aneurysms. Neurosurgery 2000;47:578 - 86. [18] Ohkuma H, Nakao T, Manabe H, Suzuki S. Subarachnoid hemorrhage caused by a dissecting aneurysm of the internal carotid artery. J Neurosurg 2002;97:576 - 83. [19] Spetzler RF, Winestock D, Newton HT, Boldrey EB. Disappearance and reappearance of cerebral aneurysm in serial arterograms. J Neurosurg 1974;41:508 - 10. [20] Shunk H. Spontaneous thrombosis of intracranial aneurysms. Am J Roentgenol Radium Ther Nucl Med 1964;92:1327 - 38.

Commentary

Phillip S. Dickey, MD New Haven, CT 06510, USA

The authors studied the gross features contributing to the arrest of aneurysmal rupture in 247 ruptured aneurysms. They noted 2 main patterns of thrombus formation. In most cases (about 80%), there was mainly an external thrombus plugging the rupture point. In the remaining 20%, half had thrombus sealing the aneurysm from inside. Two percent showed most of the aneurysm destroyed by bleeding. Smaller aneurysms tended to be more likely to have an outside-arrest-pattern and to be in better grade. A multivariate analysis was not done to determine the relation of various factors to the hemorrhage arrest pattern. This would be difficult to do because of the small number of cases and the large number of possible factors. One might need to address the role of blood pressure and intracranial pressure (was the patient straining?) at the time of rupture, whether the patient had no antiplatelet agents or anticoagulants and the like. The suggestion that smaller aneurysms were associated with better clinical grade is not what would be predicted from other studies. Russell et al [1] reported that smaller aneurysms were associated with larger volume subarachnoid hemorrhage. There is a rough correlation between more subarachnoid hemorrhage and worse clinical