Available online at www.sciencedirect.com

Surgical Neurology 71 (2009) 566–572www.surgicalneurology-online.com

Subarachnoid Hemorrhage

Nonaneurysmal nonperimesencephalic subarachnoid hemorrhage:is it a benign entity?

Sunil K. Gupta, MCh a,⁎, Rahul Gupta, MCh a, Virender K. Khosla, MCh a,Sandeep Mohindra, MCh a, Rajesh Chhabra, MCh a, Niranjan Khandelwal, MD b,Vivek Gupta, MD b, Kanchan K. Mukherjee, MCh a, Manoj K. Tewari, MCh a,

Ashish Pathak, MCh a, Suresh N. Mathuriya, MCh a

Departments of aNeurosurgery, and bRadiodiagnosis, Postgraduate Institute of Medical Education and Research, Chandigarh-160012, India

Received 17 January 2008; accepted 11 April 2008

Abstract Background: Although the clinical profile of patients with PMN SAH is well documented, there are

Abbreviations: COcomputed tomographiGOS, Glasgow outcomartery; n-PMN SAHPMN SAH, perimesennoid hemorrhage.

⁎ CorrespondingE-mail address: d

0090-3019/$ – see frodoi:10.1016/j.surneu.2

scarce data available for patients with nonaneurysmal n-PMN SAH. In the present study, the clinicalcharacteristics of patients with n-PMN SAH were analyzed and compared with those of PMN SAHand aneurysmal SAH.Methods: Patients with spontaneous SAH, in whom the initial DSA or 3-dimensional CTA resultwas normal, underwent another investigation (CTA/DSA). If the results of both of these werenegative, a second\ DSAwas done after 4 to 6 weeks. Patients in whom even the second DSA failedto reveal an aneurysm or any other vascular abnormality were labeled as nonaneurysmal SAH.Within this group, 2 different types were identified: PMN SAH and n-PMN SAH.Results: There were 61 patients in whom the results of the first DSA and CTAwere both negative. In2 of these patients, an aneurysm was demonstrated at a second DSA. Seven patients died before asecond DSA could be done. After excluding these, there were 18 patients with PMN SAH and 34with n-PMN SAH. There was no mortality in these patients; and at a mean follow-up of 1.8 years, allpatients with PMN SAH and 94.1% of patients with n-PMN SAH had a good outcome. Associatedcomorbid illnesses were more frequent in patients with PMN SAH and n-PMN SAH as comparedwith the aneurysmal SAH patients.Conclusions: Once an aneurysm is definitely excluded, patients with n-PMN SAH have a goodoutcome, and like PMN SAH, have a benign clinical course. However, a second DSA is mandatoryto avoid missing an aneurysm or any other vascular lesion.© 2009 Elsevier Inc. All rights reserved.

Keywords: Subarachnoid hemorrhage; Negative angiography; Nonperimesencephalic hemorrhage; Perimesencephalic hemorrhage

1. Introduction

In about 15% of patients with spontaneous SAH, novascular abnormality is demonstrated on a DSA[3,9,12,16,21,23]. About 21% to 68% of these patients

PD, chronic obstructive pulmonary disease; CTA,c angiogram; DSA, digital subtraction angiography;e score; H&H, Hunt and Hess; MCA, middle cerebral, nonperimesencephalic subarachnoid hemorrhage;cephalic subarachnoid hemorrhage; SAH, subarach-

author. Fax: +91 172 2744401.rguptasunil@gmail.com (S.K. Gupta).

nt matter © 2009 Elsevier Inc. All rights reserved.008.04.021

are known to have perimesencephalic hemorrhage [11],wherein the blood is confined to the cisterns around themidbrain and the center of bleeding is immediately anteriorto the midbrain [25]. The clinical course of patients withperimesencephalic hemorrhage is much better than that ofpatients with aneurysmal SAH; and therefore, this entityhas been labeled as benign SAH [18,19,25]. There are,however, a significant percentage of patients within thisnonaneurysmal spontaneous SAH group where the patternof blood on CT scan is not perimesencephalic. There isabundant literature available on PMN SAH. However, theincidence, clinical course, and outcome of patients withnonaneurysmal n-PMN SAH are not well described. In the

Fig. 1. Computed tomographic scans of 2 patients with PMN SAH.

567S.K. Gupta et al. / Surgical Neurology 71 (2009) 566–572

present study, all patients with spontaneous SAH in whomno vascular abnormality could be identified were studied;and the differences between the PMN SAH and n-PMNSAH were analyzed.

2. Material and methods

A prospective study was done on patients of spontaneousSAH admitted to the department of neurosurgery during a15-month period (March 2005-July 2006). In all patients, aCT scan was done to demonstrate the subarachnoid blood.Patients with a radiological and/or clinical diagnosis of SAHunderwent a DSA and/or a 3-dimensional CTA. Patients in

Fig. 2. Computed tomographic scans showing the p

whom an aneurysm was demonstrated underwent eitherclipping or coiling of the aneurysm. Patients in whom theinitial investigation (DSA/CTA) did not demonstrate ananeurysm were subjected to another investigation (CTA/DSA). If both these failed to demonstrate any aneurysm orany other cause of SAH, these patients were managedconservatively as per standard SAH protocol, whichincluded adequate hydration, anticonvulsants, steroids, andcalcium channel blockers (nimodipine). Coagulation studieswere performed in all patients. The demographic profile andclinical details of all patients including their clinical status atarrival, CT scan findings, neurologic deficits, course in thehospital, and status at discharge were recorded. Patients in

attern of SAH in patients with n-PMN SAH.

able 2unt and Hess grade at arrival to Hospital

&H grade No. of patients

116

I 28131

otal 59

568 S.K. Gupta et al. / Surgical Neurology 71 (2009) 566–572

whom the results of both the investigations were negativeand who were discharged were followed up by a repeatedDSA 6 weeks after the initial investigation. All patients werefollowed up for at least 1 year.

2.1. Inclusion criteria

Patients were classified as having nonaneurysmal SAH ifthey fulfilled the following criteria:

1. Clinical history of spontaneous SAH2. CT scan evidence of SAH3. A negative DSA and CTA result, that is, no evidence

of either an aneurysm or any other vascular lesion4. A negative second DSA result (4-6 weeks after the first

DSA)

Of the patients who met the above criteria, 2 separategroups were identified:

A) PMN SAH: SAH in the perimesencephalic cisternswith the center of hemorrhage located in front of thebrainstem mainly in the interpeduncular cistern with orwithout extension to the ambient, chiasmatic, andhorizontal part of the sylvian cisterns (Fig. 1).

B) n-PMN SAH: SAH primarily in the chiasmatic,sylvian, and interhemisperic cistern. This group alsoincluded those patients who had some perimesence-phalic blood, but in whom the major amount of bloodwas in other cisterns (Fig. 2).

3. Results

During the study period, 292 patients underwent surgeryfor clipping of their intracranial aneurysm(s). Over the sameperiod, there were 61 patients in whom results of both theDSA and the CTA were negative despite a radiologicalevidence of SAH.

Out of these 61, 2 patients were demonstrated to have ananeurysm at a second DSA performed at 6 weeks; so these 2were excluded from the study. Both these patients had anonperimesencephalic pattern of SAH. The male-femaleratio was 2.1: 1. Most of the patients (45, 76.3%) were in thefourth to sixth decades of life (mean age, 48 years). Forty-four (74.6%) reached the hospital within 72 hours of the last

Table 1Clinical presentation

Symptoms/signs No. of patients

Sudden severe headache 58Loss of consciousness 32Vomiting 26Neck rigidity 57Aphasia 16Hemiparesis 12Cranial nerve palsies 6

TH

H

IIIIIIVVT

ictus, 8 (13.1%) between 3 and 7 days, and 7 (11.5%) aftermore than 7 days. Fifty-six (94.9%) were admitted after thefirst episode of SAH, whereas in 3, there was history of aprevious bleed. Fifty-eight (98.3%) had a definite history ofsudden severe headache, whereas loss of consciousness waspresent in 32 (54.2%) (Table 1). Forty-five (76.3%) were inH&H grade I to III at arrival, whereas 14 (23.7%) were inpoor grade (IV-V) (Table 2). Two patients had a history ofdrug abuse, whereas one had a severe ankylosing spondylitiswith a strong family history of this disease. No patient hadany coagulopathy or sickle cell disease.

Eighteen patients had nonaneurysmal PMN SAH, whereas41 had nonaneurysmal n-PMN SAH. In the n-PMN SAHgroup, 19 patients had a sylvian fissure bleed, 17 had a diffuseSAH, 4 had an anterior interhemispheric bleed, whereas inone, there was blood in the suprasellar cisterns only.

Seven of the 59 patients died. All the patients who diedhad n-PMN SAH. Five of these 7 were in grade IV/V atarrival, whereas 2 were in grade III. Four had a diffuse SAH,whereas 3 had a sylvian pattern of SAH. There wasassociated intraventicular hemorrhage in 2 and an intracer-ebral bleed in 1. Four patients had angiographic evidence ofvasospasm, and 2 of these 4 had an infarct in the MCA\territory. In these 7 patients, although results of both theinitial DSA and CTA were negative, a repeated DSA couldnot be done because they died before that. Therefore, thepossibility of a missed aneurysm cannot be ruled out.

There was no autopsy in any of these 7 patients. Becausethe possibility of an aneurysm could not be definitelyexcluded in these 7 patients, they were excluded from furtherdiscussion, leaving us with 34 patients of n-PMN SAH.

The mean follow-up period was 1.8 years (range, 1.5-2.5years). At follow-up, all the 18 patients with PMN SAH and32 of the 34 patients with n-PMN SAH had a good outcome(GOS 4 and 5). Two patients were left with severe disability.These 2 patients were also from the n-PMN SAH group(Table 3).

Patient characteristics in these 2 groups were comparedwith those in patients who had their aneurysms surgicallyclipped during the same period (Table 4). Of the 292 patientswith aneurysmal SAH, 24% were in poor H&H grade,whereas none of the patients with PMN SAH and 26.5% ofthe patients with n-PMN SAH were in poor grade.

Comorbid illnesses were more common in both the PMNSAH and n-PMN SAH groups as compared with the

Table 3Glasgow outcome score at 6 months

GOS Good (5) Moderate disability,independent (4)

Severe disability,dependent (3)

Vegetative (2) Dead (1)

PMN SAH (n = 18) 16 2 0 0 0n-PMN SAH (n = 34) 25 7 2 0 0Mean duration of follow-up 1.8 y

569S.K. Gupta et al. / Surgical Neurology 71 (2009) 566–572

aneurysm-positive patients. Statistical analysis of thesedifferences was done by χ 2 test. Associated hypertensionand/or cardiac disease was present in 41.5% of patients withn-PMN SAH, 27.8% of those with PMN SAH, and 10.6%of patients with aneurysmal SAH. These differences werestatistically significant between aneurysmal SAH and boththe negative SAH groups (aneurysmal SAH vs PMN SAH,P value = .044; aneurysmal SAH vs n-PMN SAH, P value =.0001). Similarly, 24.4% of patients with n-PMN SAH,16.7% of those with PMN SAH, and 8.6% of aneurysmpatients had a significant history of smoking and/or COPD(aneurysmal SAH vs PMN SAH, P value = .214;aneurysmal SAH vs n-PMN SAH, P value = .0001). Thedifference was thus significantly different between theaneurysmal SAH and n-PMN SAH groups. A good outcomewas achieved in 100% of patients with PMN SAH, in 94.1%of patients with n-PMN SAH, and in 55% of patients withaneurysm-positive SAH.

4. Discussion

4.1. Role of repeated angiography

In a significant percentage of patients with spontaneousSAH, the initial angiographic studies do not reveal ananeurysm or any other vascular abnormality [7,8]. In some ofthese patients, the angiograms can yield false-negative resultsfor a variety of reasons: vasospasm, vascular thrombosis,obliteration of the aneurysm by the effects of an adjacenthematoma, or simply a technically inadequate examination[13,22]. Patients with spontaneous SAH and with negativeresults in initial diagnostic studies are subjected to subsequentstudies to find out the cause of SAH. In recent years, spiralCTA has proven to be a valuable modality for detection of

Table 4Comparison of clinical features between aneurysmal SAH, PMN SAH, and n-PM

Clinical feature Aneurysmal SAH (n = 292)

Mean age (y) 44.8Male-female ratio 1.1:1Hypertension/cardiac disease 10.6%COPD/smoking 8.6%Diabetes 1.0%Alcoholism 2%Poor H&H grade (4 and 5) 24%Good outcome (GOS 4 and 5) 55%Mortality 29.8%

aneurysms. In the present study, at the time of initialpresentation, all patients had both a DSA and a CTA; and ifthe results of both were negative, a second DSAwas done at 4to 6 weeks after the first angiography. The rate of detection ofan underlying cause of SAH with a repeated DSA rangesfrom 2% to 21% [12,19], the positivity rate being higher inpatients with a non-PMN pattern of SAH on CT scan.Topcuoglu et al [22] assessed the diagnostic yield of imagingtests in 86 patients in whom the initial catheter angiographyfailed to reveal the cause of SAH. Forty-one of these had n-PMN SAH. In their study, 3 aneurysms were picked up onsecond angiography and one on third angiography, all inpatients with n-PMN SAH. Repeated imaging studies werenot fruitful in diagnosis in patients with PMN SAH. In thepresent study, out of a total of 61 patients, an aneurysm wasdetected in 2 (3.28%) only, at a second DSA; both of thesehad n-PMN SAH. Some have advocated even a third DSA[22], although others have considered it superfluous [10]. Inour opinion, if technically adequate DSA and CTA areperformed and read by experienced neuroradiologists andneurosurgeons, and a second DSA is also done, a third DSA isprobably not fruitful.

4.2. Clinical profile

In patients in whom the initial DSA and CTA and thesecond DSA also do not reveal any vascular abnormality,2 patterns of subarachnoid bleed could be identified: PMNSAH and n-PMN SAH. The n-PMN SAH group includespatients with anterior interhemispheric, sylvian, suprasellar,and diffuse SAH. There was no significant difference in themean age of patients in the aneurysmal SAH group as well asthe PMN SAH and n-PMN SAH groups. Flaherty et al [6]reported that patients with PMN SAH were younger and lesslikely to be female than other patients with SAH. In another

N SAH

PMN SAH (n = 18) n-PMN SAH (n = 34)

48.3 42.83.5: 1 1.6: 127.8% 41.5%16.7% 24.4%16.7% 4.8%11% 7.3%0% 26.5%

100% 94.1%0% 0%

570 S.K. Gupta et al. / Surgical Neurology 71 (2009) 566–572

report, these patients were noted to be younger andpredominantly male [3]. In a literature review of 132 patientswith PMN SAH, 52%were male [11,21]. However Ildan et al[11] noted a female predominance (75.8%). In the presentstudy, there was no significant sex difference in theaneurysmal SAH group. However, there was a strong malepredominance in the PMN SAH patients (3.5:1); and in then-PMN SAH group, the ratio was 1.6: 1. Thus, in the presentstudy also, the male predominance in the nonaneurysmalSAH patients has been observed, especially in the PMN SAHpatients. There was a significant difference in the neurologicstatus at admission between patients with PMN SAH andn-PMN SAH. No patient with PMN SAH had a poor H&Hgrade at presentation, whereas about 26.5% of patients withn-PMN SAH were in poor grade.

4.3. PMN SAH

It has been well documented in the literature that patientswith a negative spontaneous SAH and a PMN SAH have agood prognosis, and these patients have been reported tohave a good outcome in a number of series [1-6,9,11,14-21,23-25]. This entity has thus been labeled as benign SAH. Ifthe result of the initial investigation (either a DSA or aCTA) is negative, there may not be a need for any furtherinvestigation in these patients. The etiology of SAH in thisgroup is possibly of venous origin [24]. All 18 patientswith PMN SAH and a negative angiogram result did well.There was no mortality; and at follow-up, all were in goodGOS. There is no specific CT pattern or clinicalpresentation that can differentiate a PMN SAH due to aposterior circulation aneurysm from a benign SAH [11].Therefore, all these patients need an initial DSA or CTA.However, a second DSA may not be necessary.

4.4. n-PMN SAH

In contrast to patients with PMN SAH, there is relativepaucity of data on the course of patients with negative resultin initial angiography and with n-PMN SAH. These patientspossibly have a better prognosis than those with aneurysmalSAH, but not as good as with the PMN SAH group [1]. Thepattern of SAH in this group of patients is no different fromthe hemorrhage seen in patients with aneurysmal SAH.The main risk in this group is the possibility of missingan aneurysm by initial studies. In the present study, allthe 7 patients who died belonged to this group. In these7 patients, results of both the DSA and CTA were negative;but they died from the effects of their initial bleed before asecond DSA could be done. Therefore, the possibility of amissed aneurysm remains. Most of these 7 patients were inpoor grade at presentation and had either vasospasm or anintracranial infarct. As mentioned, it is possible that thepresence of vasospasm may have masked the presence of ananeurysm. Apart from these 7 patients, most (94.1%) patientswith n-PMN SAH and a negative second DSA result had agood outcome; and there was no mortality. One can presume

that if an aneurysm is definitely excluded, most patients withn-PMN SAH also do well, although the outcome is not asgood as that of the PMN SAH group. The natural course ofthese patients is therefore significantly better than that ofaneurysmal SAH and more like the classic “benign” PMNSAH. The cause of bleed in these patients remainsspeculative, and there are not much data available on thisissue. As in the PMN SAH group, it is possible that the bloodcould be of venous origin. On the other hand, Canhao et al [2]postulated that the cause of bleed in patients with PMN SAHcould be a minimal arterial leak, probably located in smallperforating arteries in their subarachnoid course beforeentering the brainstem. The same hypothesis may hold truefor n-PMN SAH as well.

As there are hardly any studies on n-PMN SAH, there areno defined guidelines for management of these patients.Because the pattern of SAH on CT scans is similar to that ofpatients with aneurysmal SAH, one has to be extra careful soas not to miss an aneurysm or any other vascular lesion.Therefore, even when the results of the first DSA and CTAare negative, a second DSA is mandatory before labelingthese patients as nonaneurysmal n-PMN SAH.

The technical quality of the DSA and CTA also has to beconsidered, as a suboptimal examination again increases thechances of missing an aneurysm. Once an aneurysm hasdefinitely been excluded, this group of patients can also beexpected to behave like benign SAH, with an outcomealmost as good as those with PMN SAH.

4.5. Associated comorbidities

In recent years, the role of vascular risk factors inspontaneous SAH has been assessed. Canhao et al [2]reported that hypertension is an independent risk factor forPMN SAH. In their study, they compared the incidence withcontrols who did not have SAH. However, hypertension,smoking, and alcohol abuse have been shown to besignificant risk factors for aneurysmal SAH as well.Therefore, in the present study, we studied the presence ofthese factors in the aneurysmal group and in the PMN SAHand n-PMN SAH groups. The incidence of hypertension/cardiac disease was significantly more in both the PMN SAHgroup and the n-PMN SAH group as compared with theaneurysm-positive SAH patients, especially so in the n-PMNSAH patients . Similarly, COPD was more commonly seenin both the PMN SAH and n-PMN SAH groups; but thisdifference was statistically significant only between theaneurysmal SAH and n-PMN SAH groups. The incidence ofdiabetes and alcohol intake was also higher in patients withnegative angiogram results as compared with aneurysm-positive patients. It is possible that these comorbidities playan etiologic role in the causation of SAH in patients withspontaneous SAH and negative angiogram results. Thesefindings are at variance with some other reports. Klienpeterand Lehr [15] as well as Flaherty et al [6] reported thatpatients with PMN SAH were less likely to be hypertensivethan other patients with SAH.

571S.K. Gupta et al. / Surgical Neurology 71 (2009) 566–572

5. Conclusions

Patients with nonaneurysmal n-PMN SAH generally havea good outcome. These patients could therefore be labeled ashaving a benign SAH similar to patients with PMN SAH.However, one has to be careful not to miss an aneurysm orother vascular malformation; therefore, a repeated DSA isnecessary. Associated comorbidities like hypertension,smoking, and diabetes were more present in this group ascompared with the aneurysmal SAH patients and may have arole to play in the causation of SAH.

References

[1] Alen JF, Lagares A, Lobato RD, Gomez PA, Rivas JJ, Ramos A.Comparison between perimesencephalic nonaneurysmal subarachnoidhemorrhage and subarachnoid hemorrhage caused by posteriorcirculation aneurysms. J Neurosurg 2003;98(3):529-35.

[2] Canhao P, Falcao F, Pinho e Melo T, Ferro H, Ferro J. Vascular riskfactors for perimesencephalic nonaneurysmal subarachnoid hemor-rhage. J Neurol 1999;246(6):492-6.

[3] Cioffi F, Pasqualin A, Cavazzani P, Da Pian R. Subarachnoidhemorrhage of unknown origin: clinical and tomographical aspects.Acta Neurochir (Wien) 1989;97(1-2):31-9.

[4] Eskesen V, Sorensen EB, Rosenorn J, Schmidt K. The prognosis insubarachnoid hemorrhage of unknown etiology. J Neurosurg 1984;61(6):1029-31.

[5] Farres MT, Ferraz-Leite H, Schindler E, Muhlbauer M. Spontaneoussubarachnoid hemorrhage with negative angiography: CT findings.J Comput Assist Tomogr 1992;16(4):534-7.

[6] Flaherty ML, Haverbusch M, Kissela B, Kleindorfer D, Schneider A,Sekar P, Moomaw CJ, Sauerbeck L, Broderick JP, Woo D.Perimesencephalic subarachnoid hemorrhage: incidence, risk factors,and outcome. J Stroke Cerebrovasc Dis 2005;14(6):267-71.

[7] Forster DM, Steiner L, Hakanson S, Bergvall U. The value of repeatpan-angiography in cases of unexplained subarachnoid hemorrhage.J Neurosurg 1978;48(5):712-6.

[8] Gilbert JW, Lee C, Young B. Repeat cerebral pan-angiography insubarachnoid hemorrhage of unknown etiology. Surg Neurol 1990;33(1):19-21.

[9] Gomez PA, Lobato RD, Rivas JJ, Cabrera A, Sarabia R, Castro S,Castaneda M, Canizal JM. Subarachnoid hemorrhage of unknownaetiology. Acta Neurochir (Wien) 1989;101(1-2):35-41.

[10] Hashimoto H, Iida J, Hironaka Y, Okada M, Sakaki T. Use of spiralcomputerized tomography angiography in patients with subarachnoidhemorrhage in whom subtraction angiography did not reveal cerebralaneurysms. J Neurosurg 2000;92(2):278-83.

[11] Ildan F, Tuna M, Erman T, Gocer AI, Cetinalp E. Prognosis andprognostic factors in nonaneurysmal perimesencephalic hemorrhage: afollow-up study in 29 patients. Surg Neurol 2002;57(3):160-5[discussion 165-6].

[12] Iwanaga H, Wakai S, Ochiai C, Narita J, Inoh S, Nagai M. Rupturedcerebral aneurysms missed by initial angiographic study. Neurosurgery1990;27(1):45-51.

[13] Kaim A, Proske M, Kirsch E, von Weymarn A, Radu EW, SteinbrichW. Value of repeat-angiography in cases of unexplained subarachnoidhemorrhage (SAH). Acta Neurol Scand 1996;93(5):366-73.

[14] Kitahara T, Ohwada T, Tokiwa K, Kurata A, Miyasaka Y, Yada K, KanS. Clinical study in patients with perimesencephalic subarachnoidhemorrhage of unknown etiology. No Shinkei Geka 1993;21:903-8[Japanese].

[15] Kleinpeter G, Lehr S. Characterization of risk factor differences inperimesencephalic subarachnoid hemorrhage. Minim Invasive Neuro-surg 2003;46(3):142-8.

[16] Nishioka H, Torner JC, Graf CJ, et al. Cooperative study of intracranialaneurysms and subarachnoid hemorrhage: a long term prognosticstudy. III. Subarachnoid hemorrhage of undetermined etiology. ArchNeurol 1984;41:1147-51.

[17] Rinkel GJ, Wijdicks EF, Vermeulen M, Ramos LM, Tanghe HL, HasanD, Meiners LC, van Gijn J. Nonaneurysmal perimesencephalicsubarachnoid hemorrhage: CT and MR patterns that differ fromaneurysmal rupture. Am J Neuroradiol 1991;12(5):829-34.

[18] Rinkel GJ, Wijdicks EF, Vermeulen M, Hasan D, Brouwers PJ, vanGijn J. The clinical course of perimesencephalic nonaneurysmalsubarachnoid hemorrhage. Ann Neurol 1991;29(5):463-8.

[19] Ruelle A, Lasio G, Boccardo M, Gottlieb A, Severi P. Long-termprognosis of subarachnoid hemorrhages of unknown etiology. J Neurol1985;232(5):277-9.

[20] Ruigrok YM, Rinkel GJ, Buskens E, Velthuis BK, van Gijn J.Perimesencephalic hemorrhage and CT angiography: a decisionanalysis. Stroke 2000;31(12):2976-83.

[21] Schwartz TH, Solomon RA. Perimesencephalic nonaneurysmalsubarachnoid hemorrhage: review of the literature. Neurosurgery1996;39(3):433-40 [Review].

[22] Topcuoglu MA, Ogilvy CS, Carter BS, Buonanno FS, Koroshetz WJ,Singhal AB. Subarachnoid hemorrhage without evident cause on initialangiography studies: diagnostic yield of subsequent angiography andother neuroimaging tests. J Neurosurgery 2003;98:1235-40.

[23] Van Calenbergh F, Plets C, Goffin J, Velghe L. Nonaneurysmalsubarachnoid hemorrhage: prevalence of perimesencephalic hemor-rhage in a consecutive series. Surg Neurol 1993;39(4):320-3.

[24] van der Schaaf IC, Velthuis BK, Gouw A, Rinkel GJ. Venous drainagein perimesencephalic hemorrhage. Stroke 2004;35(7):1614-8.

[25] van Gijn J, Rinkel GJ. Subarachnoid hemorrhage: diagnosis, causesand management. Brain 2001;124(Pt 2):249-78 [Review].

Commentary

Based on their experiences of 34 cases, the authorsconcluded that the patients with the nonaneurysmal nonper-imesencephalic subarachnoid hemorrhage (non-PMN SAH)have good outcomes. Although a good prognosis of SAHwithunknown etiology is well known, the impact of this study isthat they elucidated outcomes of non-PMNSAH in addition tononaneurysmal PMNSAH. There are still a couple of points toreadwith caution in their conclusion. First, 7 patients, who diedbefore a second imaging study, were excluded from the non-PMN SAH group. If they had survived until the secondimaging study and if they had been an all-negative study, theoutcome of their results would have been a different one.Second, they operated on all aneurysms, including grade Vpatients, with either clipping or coiling. The policy mayinfluence the outcome of aneurysmal SAH group. Even withthese concerns, the good outcome of non-PMN SAH afterrepeated\ imaging study is likely to be correct.

Nobuhito Saito, MD, PhDDepartment of NeurosurgeryUniversity of Tokyo Hospital

Tokyo 113-8655, Japan

Patients with subarachnoid hemorrhage (SAH) ofunknown origin have been considered to have a good