1156

Grading the Amount of Blood on ComputedTomograms After Subarachnoid Hemorrhage

A. Hijdra, MD, P.J.A.M. Brouwers, MD, M. Vermeulen, MD, and J. van Gijn, MD

According to several studies, the amount of subarachnoid blood on the initial computedtomogram of patients with aneurysmal subarachnoid hemorrhage has predictive value withrespect to infarction and outcome. Of several methods for assessing the amount of subarach-noid blood, none has been subjected to a study of interobserver agreement. We describe our ownmethod, applied in previous studies, in which the amounts of blood in 10 basal cisterns andfissures and in four ventricles are graded separately. In grading single computed tomograms of182 consecutive patients with subarachnoid hemorrhage, the agreement between pairs of threeobservers, studied with K statistics, was relatively good for individual cisterns or fissures (Kbetween 0.35 and 0.65) and ventricles (K between 0.47 and 0.74). The Spearman rankcorrelation coefficients for the sum of the scores for subarachnoid and intraventricular bloodwere very high. Summed scores for extravasated blood are suitable as a baseline variable infollow-up studies of patients with subarachnoid hemorrhage. (Stroke 1990:21:1156-1161)

The amount of aneurysmal subarachnoid bloodon an early computed tomogram (CT scan) isassociated with subsequent vasospasm and

cerebral ischemia and with final outcome.1-13 This isimportant for the management of patients with aneu-rysmal subarachnoid hemorrhage14 and for the evalu-ation of new modes of treatment.15 Most investigatorshave adopted their own method of quantifying theamount of subarachnoid blood on CT scans (Table 1),but none of these studies was primarily concernedwith evaluating the method itself.13-5-7-8-10-14-16-18

Some authors1-2 recorded only the presence orabsence of blood in the subarachnoid space, but since>90% of patients have such blood on a CT scanmade <1 day after onset,19-20 this method seems tooblunt. Quantification of the amount of blood present,however, is difficult. It cannot be assumed thatdensity at a single point in the subarachnoid spacemeasured in Hounsfield units7-8-13 or on a semiquan-titative scale3-11 accurately reflects the total amountof extravasated blood, and measurement of the thick-ness of layers of blood or of clots5-9-11 depends asmuch on the dimensions of the cisterns in question ason the amount of subarachnoid blood. Simple grad-ing scales with a limited number of categories deal

From the University Department of Neurology (A.H.), Acade-misch Medisch Centrum, Amsterdam, the University Departmentof Neurology (P.J.A.M.B., J.V.G.), University Hospital Utrecht,Utrecht, and the University Department of Neurology (M.V.),University Hospital Dijkzigt, Rotterdam, The Netherlands.

Address for reprints: P.J.A.M. Brouwers, Department of Neu-rology, Medisch Spectrum Twente, Postbus 50000, 7500 KA En-schede, The Netherlands.

Received December 1, 1989; accepted April 3, 1990.

inadequately with all possible aneurysm sites.4-81718

Some methods also take into account intracerebraland intraventricular blood,51516 but these types ofhemorrhage can be combined with only a smallamount of subarachnoid blood. Other studies4-58-9

emphasize the extent of subarachnoid blood in theinterhemispheric fissure and the sylvian fissures,which reflects the assumption that vasospasm iselicited by tight clots around major cerebral vessels.However, cerebral infarction after subarachnoidhemorrhage is often a multifocal or diffuse event.21

The initial impact of the hemorrhage could play animportant part, and the total amount of blood couldbe as important for the development of ischemia asits presence in certain cisterns and fissures.

The consequence is that grading can be onlysemiquantitative and subjective, but subjectivity canbe restricted by choosing a method that gives the bestinterobserver agreement. Some methods have beenfound difficult to apply outside the center in whichthey were developed.16 Based on these consider-ations, we propose that a method of grading theamount of subarachnoid blood on CT scans shouldfulfill the following criteria: 1) the total amount ofsubarachnoid blood must be graded, 2) the distribu-tion and extension of blood among all basal cisternsand fissures must be reflected in the grading system,3) the amount of subarachnoid blood must be gradedindependent of the amounts of intracerebral andintraventricular blood, and 4) such a grading methodmust be tested for interobserver agreement. Wedescribe and evaluate such a method.

by guest on August 26, 2017

http://stroke.ahajournals.org/D

ownloaded from

Hijdra et al Grading Amount of Blood After SAH 1157

TABLE 1. Methods of Grading Amount of Subarachnoid Blood on Computed Tomograms

Author(s)

Takemae et al1

Bell et aPDavis et al4

Fisher et al5

Suzuki et al7

Sano et al8

Taneda14

Allen et al16

Gurusinghe and Richardson10

Mohsen et al11

Pasqualin et al12

Fujita13

Inagawa et al17

Petruk et al18

Yearpublished

19781980198019801980198219821983198419841984198519871988

Anatomicdistribution

--

t-

-----

-

Measurementtechnique

GGG

HHG0M

G, MGHGG

Intraventricularblood graded?

NoNoNo

Yes*NoNoNo

Yes*NoNoNoNoYesYes

- , anatomic distribution subordinate to some focal measurement in assignment of a certain grade; +, definiteinformation on anatomic distribution; ±, some information on anatomic distribution; G, semiquantitative grading; M,anatomic measurement; H, Hounsfield units.

•Intraventricular blood is incompatible with more than small amounts of subarachnoid blood in this grading system.

Subjects and MethodsEach of 10 basal cisterns and fissures was graded

separately on a semiquantitative scale, according tothe amount of extravasated blood: 0, no blood; 1,small amount of blood; 2, moderately filled withblood; or 3, completely filled with blood. The densityof the clot was not considered. Clots that hadexpanded the original size of a cistern or fissure werestill graded as 3. The total amount of subarachnoidblood (sum score) was calculated by adding the 10scores and ranged from 0 to 30. When an occasionalcistern or fissure was considered to be inadequatelyvisualized, we interpolated by assigning to that cis-tern or fissure the average score of the others. Figure

1 shows a CT scan and a corresponding diagram as anexample of the grading system.

The grading scale for the amount of blood in thefour ventricles was constructed in a comparablefashion, as follows: 0, no blood; 1, sedimentation ofblood in the posterior part; 2, partly filled with blood;or 3, completely filled with blood. The total amountof intraventricular blood (sum score) was the total ofthe four scores and ranged from 0 to 12.

Before the study of interobserver agreement, in apilot study three of the authors independently graded10 randomly chosen CT scans and compared theirresults. Discrepancies between scores as well asproblems in the identification of certain cisterns and

FIGURE 1. Computed tomo-

gram of patient 6 hours aftersubarachnoid hemorrhage fromcarotid artery aneurysm on left(L) side. *Top diagram identi-fies 10 basal cisterns and fis-sures: A, frontal interhemi-spheric fissure; B, sylvian fissure,lateral parts; C, sylvian fissure,basal parts; D, suprasellar cis-tern; E, ambient cisterns; F,quadrigeminal cistern. **Bottomdiagram indicates amount ofblood in each cistern and fissure.Sum score is 22 points.

by guest on August 26, 2017

http://stroke.ahajournals.org/D

ownloaded from

1158 Stroke Vol 21, No 8, August 1990

in the differentiation of subarachnoid and intracere-bral blood were discussed. The cisterns, fissures, andventricles, together with the relevant problems iden-tified in this pilot study, follow.

In the frontal interhemispheric fissure (A in Figure1) the falx can be mistaken for blood. When later CTscans are available, densities of the structure inquestion can be compared. If the densities do notdiffer over time, the structure is the falx. Distinguish-ing subarachnoid from intracerebral blood may bedifficult; a hematoma can often be identified becauseit deviates from the midline at some point, andsometimes cerebrospinal fluid can be identified alongthe edge of the hematoma. The lateral parts of thesylvian fissure (B in Figure 1), an insular cistern, aremostly in the sagittal plane and perpendicular to thebasal parts of the sylvian fissure (C in Figure 1),which are in the frontal plane. The basal parts of thesylvian fissure lie anterosuperior to the tip of thetemporal lobe and run from the suprasellar cistern(D in Figure 1) laterally to the sphenoid angle. Theinterpeduncular fossa is considered part of the supra-sellar cistern. Left and right parts of the suprasellarcistern are graded separately since blood is oftendistributed asymmetrically. The shape of the supra-sellar cistern depends on the angle of the CT slice.Blood in the ambient cisterns (E in Figure 1) can beconfused with blood on the tentorium. In the lattercase, cerebrospinal fluid is visible between the mid-brain and the tentorial edge, and the line of blooddoes not curve inward into the quadrigeminal cistern(F in Figure 1) but continues in a posterolateraldirection. Asymmetrical distribution of blood in thequadrigeminal cistern always depends on unevendistribution between the two ambient cisterns, andtherefore the left and right parts are not gradedseparately. The lateral ventricles often contain asmall amount of sedimented blood in the posteriorhorns; some CT scans show a narrow layer of bloodnear the foramina of Monro. One point was assignedto these findings.

We studied interobserver agreement in the gradingof admission CT scans of 182 consecutive patientswith subarachnoid hemorrhage who were investi-gated <72 hours after the initial hemorrhage. Ananeurysmal origin of the hemorrhage was proven byangiography or autopsy in 130 patients and wasstrongly suggested by CT scan evidence22 in 44; in theremaining eight patients the cause of the hemorrhagewas nonaneurysmal (perimesencephalic hemor-rhage23 or arteriovenous malformation). All CT scanswere included, and all showed at least some blood inthe subarachnoid space or the ventricles. Routinestandard series without contrast enhancement wereobtained. Two different machines were used, an EMICT 1010 (Hayes, Middlesex, England) and a PhilipsTomoscan 310 (Eindhoven, the Netherlands), whichdisplayed the pictures on 160x160 and 256x256matrices, respectively. Transparencies with a diminu-tion factor of approximately Vi were used for assess-

ment. Three observers separately judged the 182 CTscans.

Agreement between pairs of observers for eachcistern, fissure, and ventricle was calculated with Kstatistics.24 The observed agreement was correctedfor chance agreement, with K values ranging from 0(total disagreement) to 1 (perfect agreement). If alldegrees of disagreement are of equal importance, K isexpressed as (p0-pc)/(100-pc), where p0 is the per-centage observed agreement and pc is the percentageagreement expected by chance. Weighted K is usedwhen the degree of disagreement is taken intoaccount.24 In our calculations, a linear weight factorof Vs was given to a difference of one grade in thefour-category grading scale. We also calculated KvaluesPfor two contracted scales: K3 for the threecategories 0, (1+2), and 3 and K2 for the two catego-ries 0 and (1+2+3). K values were adjusted formissed observations.25 Interobserver agreement forthe sum scores was studied with Spearman rankcorrelation coefficients.

ResultsOf the 1,820 subarachnoid cisterns and cerebral

fissures, 83 (5%) were considered inadequately visu-alized for grading by one or more observers; fourcisterns and fissures (0.2%) could not be graded byany observer. Only one of the 182 CT scans (0.5%)had more than two inadequately visualized cisternsor fissures. Of the 728 ventricles, 15 (2%) wereconsidered inadequately visualized by one or moreobservers, and three ventricles (0.4%) could not begraded by any observer. There were no CT scans withmore than one inadequately visualized ventricle.

Table 2 shows the interobserver agreement onscores for the 10 individual cisterns and fissures,represented by unweighted (K4) and weighted (KW4)values for each pair of observers. Table 3 containssimilar results for the four ventricles, /c, values forobserver pairs A-C and B-C are generally lower thanthose of pair A-B. Within observer pairs there wereno striking differences for individual cisterns, fis-sures, or ventricles except for relatively low agree-ment for the basal parts of the sylvian fissure (allpairs: *c 0.37-0.52), the interhemispheric fissure(pairs A-C and B-C: /c, 0.38 and 0.35, respectively),and the left and right part of the suprasellar cistern(pairs A-C and B-C: K, 0.35-0.43). /c, values for theother cisterns and fissures and for the ventricles wererelatively good, that is, between 0.45 and 0.74.

In general, the K values increased considerablywhen we contracted the scale (data not shown). Themean increases in K3 and K2 values for the 10 cisternsand fissures was 0.09 and 0.17, respectively. For thefour ventricles the increases were 0.03 and 0.06,respectively.

K values adjusted for inadequately visualized cis-terns, fissures, and ventricles (data not shown) wereonly slightly lower than the values shown in Tables 2and 3 (differences 0.00-0.05 for pair A-B and 0.00-0.06 for pairs A-C and B-C).

by guest on August 26, 2017

http://stroke.ahajournals.org/D

ownloaded from

Hijdraetal Grading Amount of Blood After SAH 1159

TABLE 2. K, and KW, Values for 10 Basal Cisterns and Fissures for Three Observer Pairs

Cistern/fissure

Interhemispheric fissureSylvian fissure

Lateral partLeftRight

Basal partLeftRight

Suprasellar cisternLeftRight

Ambient cisternLeftRight

Quadrigeminal cisternStandard error

MeanRange

A-B

K4

0.60

0.650.63

0.370.45

0.560.58

0.650.640.60

0.048

KW4

0.70

0.750.75

0.500.53

0.710.74

0.780.770.71

0.0390.045-0.055 0.032-0.056

Observer pair

A-C

0.38

0.600.53

0.480.43

0.430.39

0.540.510.51

0.0500.046-0.054

KW4

0.54

0.730.69

0.580.55

0.590.57

0.690.640.66

0.0420.036-0.051

B-C

K4

0.35

0.500.52

0.420.52

0.400.35

0.470.590.45

0.050

KW4

0.49

0.650.67

0.560.62

0.550.55

0.630.690.61

0.0430.047-0.053 0.038-0.047

id, unweighted K for four-category grading scale; KWJ, weighted K (weight factor Vn) for four-category grading scale.

For each CT scan we calculated the sum score forsubarachnoid blood and for intraventricular blood(data not shown). The Spearman rank correlationcoefficients were very high (pair A-B: 0.92 and 0.87,pair A-C: 0.91 and 0.75, and pair B-C: 0.89 and 0.74,respectively). To study the possibility that oneobserver had scored consistently higher than theother two, we calculated the mean differencesbetween the sum scores of each observer pair (Table4). Ideally, this mean difference would be 0. For eachobserver pair the mean difference between the sumscores was roughly one grade or less, which was verygood, considering the ranges of possible sum scores.

DiscussionThere is no consensus on what constitutes an

acceptable K value,26-28 but Landis and Koch29 haveproposed that values between 0.40 and 0.80 can be

considered to indicate moderate to substantial agree-ment. In our study, the rvalues for most cerebrospinalspaces were within this range. On the other hand, oneshould be cautious in comparing K values from dif-ferent studies because the values depend not only onthe actual agreement between observers but also onthe number of categories used in their computation.

Agreement was better for the ambient and quad-rigeminal cisterns, the lateral parts of the sylvianfissure, and the ventricles than for the basal parts ofthe sylvian fissure, the interhemispheric fissure, andthe suprasellar cistern. The difference suggests thatlow K values cannot be attributed solely to problemsin applying the semiquantitative four-category grad-ing system. The lower interobserver agreement ingrading the basal parts of the sylvian fissure, thesuprasellar cistern, and the interhemispheric fissureshould probably be attributed to anatomic and tech-

TABLE 3. K4 and KW4 Values for Four Ventricles for Three Observer Pairs

Ventricle

Lateral ventricleLeftRight

Third ventricleFourth ventricleStandard error

MeanRange

K4

0.740.740.690.59

0.0530.046-0.065

A-B

KW4

0.800.790.810.73

0.0420.037-0.052

Observer

A-C

K4

0.630.570.580.47

0.0590.053-0.068

pair

KW4

0.670.650.720.65

0.0510.047-0.057

K4

0.540.560.470.54

0.0600.056-0.065

B-C

KW4

0.590.620.650.71

0.0520.051-0.053

K>, unweighted K for four-category grading scale; «w4, weighted « (weight factor VS) for four-category grading scale.

by guest on August 26, 2017

http://stroke.ahajournals.org/D

ownloaded from

1160 Stroke Vol 21, No 8, August 1990

TABLE 4. Differences Between Subarachnoid and Intraventricu-lar Sum Scores for Three Observer Pairs

Observerpair

Subarachnoid bloodA-BA-CB-C

Intraventricular bloodA-BA-CB-C

Mean

0.86-0.24-1.10

0.06-0.27-0.33

Difference

95% Confidenceinterval

0.45 to 1.26-0.69 to 0.22-1.59 to-0.60

-0.07 to 0.19-0.46 to -0.09-0.52 to -0.15

nical factors. Because the first two structures lie justabove the base of the skull, it is difficult to identifysmall amounts of blood and on some CT slices thesecisterns are inadequately visualized. Grading of theinterhemispheric fissure can also be complicated by acalcified falx or by the presence of blood in thefrontal lobe(s).

The number of inadequately visualized cisterns, fis-sures, and ventricles was relatively low and hardlyaffected the K values. As expected, KW4 values and K3and K2 values indicated better interobserver agreement.The error introduced into the sum score when aninterpolated grade is assigned to an inadequately visu-alized cistern or fissure is small and therefore accept-able as long as only one or two of the 10 structures areinadequately visualized. The range of sum scores cal-culated from many more than four categories (31 forsubarachnoid blood and 13 for intraventricular blood)is too large to calculate K statistics. However, theSpearman rank correlation coefficients for the sumscores were very high and, together with the low meandifferences between the sum scores of observer pairs,represent good interobserver agreement.

In summary, our method has some drawbacks for thestudy of the exact anatomic distribution of subarach-noid blood, but it can be used without restriction as ameasure for the total amount of subarachnoid andintraventricular blood. In multicenter studies withmany observers, the use of three or two categories percistern or fissure will probably give satisfactory interob-server agreement. We obtained experience with thefour-category grading system in some of our earlierstudies,30-32 in which the sum scores proved to beimportant prognostic factors. The use of the entirerange of possible sum scores as prognostic categorieswill be impracticable. Broader categories can be cho-sen, the number of which depends on the kind of studyfor which the grading system is used.

AcknowledgmentsWe thank Mrs. I. van der Tweel and Dr. H.

Frericks for their help with statistical calculations.

References1. Takemae T, Mizukami M, Kin H, Kawase T, Araki G:

Computed tomography of ruptured intracranial aneurysms inacute stage: Relationship between vasospasm and high densityon CT scan. Brain Nerve 1978;8:861-866

2. Mizukami M, Takemae T, Tazawa T, Kawase T, Matsuzaki T:Value of computed tomography in the prediction of cerebralvasospasm after aneurysm rupture. Neurosurgery 1980;7:583-586

3. Bell BA, Kendall BE, Symon L: Computed tomography inaneurysmal subarachnoid haemorrhage. / Neurol NeurosurgPsychiatry 1980;43:522-524

4. Davis JM, Davis KR, Crowell RM: Subarachnoid hemorrhagesecondary to ruptured intracranial aneurysm: Prognostic sig-nificance of cranial CT. AJNR 1980;l:17-21

5. Fisher CM, Kistler JP, Davis JM: Relation of cerebral vaso-spasm to subarachnoid hemorrhage visualized by computer-ized tomographic scanning. Neurosurgery 1980;6:l-9

6. Fraser J, Johnson S, Ray M, Robertson JT: Prediction ofcerebral vasospasm with subarachnoid hemorrhage due toruptured intracranial aneurysm by computed axial tomography(abstract). Neurosurgery 1980;6:686-687

7. Suzuki J, Komatsu S, Sato T, Sakurai Y: Correlation betweenCT findings and subsequent development of cerebral infarc-tion due to vasospasm in subarachnoid haemorrhage. AdaNeurochir 1980;55:63-70

8. Sano H, Kanno T, Shinomiya Y, Katada K, Katoh Y, Naka-gawa T, Adachi K: Prospection of chronic vasospasm by CTfindings. Ada Neurochir 1982;63:23-30

9. Kistler JP, Crowell RM, Davis KR, Heros R, Ojemann RG,Fisher CM: The relation of cerebral vasospasm to the extentand location of subarachnoid blood visualized by CT scan: Aprospective study. Neurology 1983;33:424-436

10. Gurusinghe NT, Richardson AE: The value of CT in aneurys-mal subarachnoid hemorrhage: The concept of a CT score. /Neurosurg 1984;60:763-770

11. Mohsen F, Pomonis S, Illingworth R: Prediction of delayedcerebral ischemia after subarachnoid haemorrhage by computedtomography. / Neurol Neurosurg Psychiatry 1984;47:1197-1202

12. Pasqualin A, Rosta L, Da Pian R, Cavazzani P, Scienza R: Roleof computed tomography in the management of vasospasm aftersubarachnoid hemorrhage. Neurosurgery 1984;15:344-353

13. Fujita S: Computed tomographic grading with Hounsfieldnumber related to delayed vasospasm in cases of rupturedcerebral aneurysm. Neurosurgery 1985;17:609-612

14. Taneda M: Effect of early operation for ruptured aneurysmson prevention of delayed ischemic symptoms. / Neurosurg1982;57:622-628

15. Pickard JD, Murray GD, Illingworth R, Shaw MDM, TeasdaleGM, Foy PM, Humphrey PRD, Lang DA, Nelson R, Richards P,Sinar J, Bailey S, Skene A: Effect of oral nimodipine on cerebralinfarction and outcome after subarachnoid haemorrhage: BritishAneurysm Nimodipine Trial. Br Med J 1989;298:636-642

16. Allen GS, Ann HS, Preziosi TJ, Battye R, Boone SC, ChouSN, Kelly DL, Weir BK, Crabbe RA, Lavik PJ, RosenbloomSB, Dorsey FC, Ingram CR, Mellits DE, Bertsch LA, BoisvertDPJ, Hundley MB, Johnson RK, Strom JA, Transou CR:Cerebral arterial spasm: A controlled trial of nimodipine inpatients with subarachnoid hemorrhage. TV Engl J Med 1983;308:619-624

17. Inagawa T, Kamiya K, Ogasawara H, Yano T: Rebleeding ofruptured intracranial aneurysms in the acute stage. SurgNeurol 1987;28:93-99

18. Petruk CK, West M, Mohr G, Weir BKA, Benoit BG, GentiliF, Disney LB, Khan MI, Grace M, Holness RO, Karwon MS,Ford RM, Cameron GS, Tucker WS, Parves GB, Miller JDR,Hunter KM, Richard MT, Durity FA Chan R, Clein LJ,Maroun FB, Godon A: Nimodipine treatment in poor gradeaneurysm patients. / Neurosurg 1988;68:505-517

19. Van Gijn J, Van Dongen KJ: The time course of aneurysmalhaemorrhage on computed tomograms. Neuroradiology 1982;23:153-156

20. Adams HP, Kassell NF, Tomer JC, Sahs AL: CT and clinicalcorrelations in recent aneurysmal subarachnoid hemorrhage:

by guest on August 26, 2017

http://stroke.ahajournals.org/D

ownloaded from

Hijdra et al Grading Amount of Blood After SAH 1161

A preliminary report of cooperative aneurysmal study. Neu-rology 1983;33:981-988

21. Hijdra A, Van Gijn J, Stefanko S, Van Dongen KJ, VermeulenM, Van Crevel H: Delayed cerebral ischemia after aneurysmalsubarachnoid hemorrhage: Clinico-anatomical correlations.Neurology 1986;36:329-333

22. Van Gijn J, Van Dongen KJ: Computerized tomography insubarachnoid hemorrhage: Difference between patients with andwithout an aneurysm on angiography. Neurology 1980;30:538-539

23. Van Gijn J, Van Dongen KJ, Vermeulen M, Hijdra A:Perimesencephalic hemorrhage: A nonaneurysmal and benignform of subarachnoid hemorrhage. Neurology 1985;35:493-497

24. Fleiss JL: Statistical Methods for Rates and Proportions, ed 2.New York, Wiley-Interscience, 1981, pp 218-225

25. Schouten HJA: Nominal scale agreement among observers.Psychometrika 1986;51:453-466

26. Koran LM: The reliability of clinical methods, data andjudgments. N EnglJMed 1975;293:642-646, 695-701

27. Teasdale G, Knill-Jones R, Van der Sande J: Observervariability in assessing impaired consciousness and coma. JNeurol Neurosurg Psychiatry 1978;41:603-610

28. Maas AIR, Braakman R, Schouten HJA, Minderhoud JM,Van Zomeren AH: Agreement between physicians on assess-ment of outcome following severe head injury. / Neurosurg1983;58:321-325

29. Landis JR, Koch GG: The measurement of observer agree-ment for categorical data. Biometrics 1977;33:159-174

30. Van Gijn J, Hijdra A, Wijdicks EFM, Vermeulen M, VanCrevel H: Acute hydrocephalus after aneurysmal subarach-noid hemorrhage. / Neurosurg 1985;63:355-362

31. Hijdra A, Van Gijn J, Nagelkerke NJD, Vermeulen M, VanCrevel H: Prediction of delayed cerebral ischemia, rebleeding,and outcome after aneurysmal subarachnoid hemorrhage.Stroke 1988;19:1250-1256

32. Brouwers PJAM, Wijdicks EFM, Hasan D, Vermeulen M,Wever EFD, Frericks H, Van Gijn J: Serial electrocardio-graphic recording in aneurysmal subarachnoid hemorrhage.Stroke 1989;20:1162-1167

KEY WORDS • interobserver agreement • tomography, x-raycomputed • subarachnoid hemorrhage

by guest on August 26, 2017

http://stroke.ahajournals.org/D

ownloaded from

A Hijdra, P J Brouwers, M Vermeulen and J van GijnGrading the amount of blood on computed tomograms after subarachnoid hemorrhage.

Print ISSN: 0039-2499. Online ISSN: 1524-4628 Copyright © 1990 American Heart Association, Inc. All rights reserved.

is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231Stroke doi: 10.1161/01.STR.21.8.1156

1990;21:1156-1161Stroke. 

http://stroke.ahajournals.org/content/21/8/1156World Wide Web at:

The online version of this article, along with updated information and services, is located on the

  http://stroke.ahajournals.org//subscriptions/

is online at: Stroke Information about subscribing to Subscriptions: 

http://www.lww.com/reprints Information about reprints can be found online at: Reprints:

  document. Permissions and Rights Question and Answer available in the

Permissions in the middle column of the Web page under Services. Further information about this process isOnce the online version of the published article for which permission is being requested is located, click Request

can be obtained via RightsLink, a service of the Copyright Clearance Center, not the Editorial Office.Stroke Requests for permissions to reproduce figures, tables, or portions of articles originally published inPermissions:

by guest on August 26, 2017

http://stroke.ahajournals.org/D

ownloaded from