V

ACOWYDS

Ycs

BI(cicpMWfLavttroRA(i6snspoCADrc

KSc

N2

©3

ascular

ngiographic Assessment oferebral Circulation Time forutcome Prediction in Patientsith Subarachnoid Hemorrhage

uhei Yoshimoto, M.D., Yoshihiro Tanaka, M.D., and Tomoyuki Sanada, M.D.epartment of Neurosurgery, Koshigaya Hospital, Dokkyo University School of Medicine,aitama, Japan

Mhrsiwtaspim

rdgdSipgacfphae

SPOag

oshimoto Y, Tanaka Y, Sanada T. Angiographic assessment oferebral circulation time for outcome prediction in patients withubarachnoid hemorrhage. Surg Neurol 2004;62:115–120.

ACKGROUNDn the acute stage following subarachnoid hemorrhageSAH), prolonged cerebral circulation time reflects in-reased intracranial pressure (ICP), which can lead torreversible brain damage. We evaluated the ability oferebral circulation time to predict outcome in SAHatients.ETHODSe prospectively studied 40 consecutive patients treated

or SAH according to standard intensive care guidelines.ateral views in preoperative carotid digital subtractionrteriograms (DSA) were used to determine arterio-enous transit time (AVTT), defined as interval from ini-ial opacification of the intracranial internal carotid ar-ery to that of the Rolandic vein. We then analyzedelationship of AVTT to other prognostic parameters andutcome.ESULTSVTT depended on initial Glasgow Coma Scale scoreGCS), ranging from 3.4 to 8.0 seconds (mean: 5.4) whennitial GCS was 13 to 15; from 4.5 to 8.7 seconds (mean:.4) when initial GCS was 7 to 12; and from 5.8 to 15.0econds (mean: 9.1 seconds, excluding angiograms withonfilling) when initial GCS was 3 to 6. AVTT correlatedignificantly with prognosis, longer AVTT predictingoorer outcome. No patient with an AVTT above 12 sec-nds recovered to a functioning state.ONCLUSIONSVTT obtained from routinely performed acute-phaseSA in SAH patients reflects ICP and functional prognosis,epresenting a cost-effective, practical, and reliable out-ome predictor. © 2004 Elsevier Inc. All rights reserved.

EY WORDSubarachnoid hemorrhage, cerebral circulation time, intra-ranial pressure, cerebral angiography, transit time.

Address reprint requests to: Yuhei Yoshimoto, M.D., Department ofeurosurgery, Koshigaya Hospital, Dokkyo University School of Medicine,-1-50 Minami-Koshigaya, Koshigaya, Saitama 343–8555, Japan.

aReceived April 17, 2003; accepted August 27, 2003.

2004 Elsevier Inc. All rights reserved.60 Park Avenue South, New York, NY 10010–1710

anagement strategy and outcome predictionfor patients with aneurysmal subarachnoid

emorrhage (SAH) is determined by clinical andadiographic findings as well as intracranial pres-ure (ICP). Several systems for grading SAH accord-ng to clinical status have proven useful and areidely applied for choice of treatment and predic-

ion of outcome [5,9]. Computed tomography (CT)ssessment of extent of subarachnoid clot [6] andtructural brain damage can be used as an outcomeredictor. ICP is regarded as an indicator of revers-

bility and may be a useful indicator for active treat-ent [2,14].Cerebral circulation time in early posthemor-

hagic periods reflects ICP. Absent or extremelyelayed intracranial vascular filling on cerebral an-iography is an indicator or predictor of braineath. In less extreme cases, increased ICP afterAH still impairs cerebral circulation, leading torreversible brain damage [8,13]. In the followingrospective study, we analyzed relationships of an-iographic circulation time to other prognostic vari-bles and outcome. We suspected that cerebralirculation time could be used to select candidatesor active treatment from among poor-grade SAHatients. Although many prognostic parametersave been investigated, no studies have focused onngiographically determined circulation time in thearly stage following SAH.

ubjects and MethodsATIENT POPULATIONur hospital is an academic tertiary referral center;ll patients with SAH are admitted to the neurosur-ical unit except for those with cardiopulmonary

rrest on arrival. Forty consecutive patients re-

0090-3019/03/$–see front matterdoi:10.1016/j.surneu.2003.08.035

fSwuctp(fpc

MAtrde(a(e

cis

mCnc

ADRwatwltvvrwa

TOtctmwpqoi

(tinRt

wlofl

cirCaam

G

A

G

A

D

dbb

116 Surg Neurol Yoshimoto et al2004;62:115–120

erred to the neurosurgical unit within 3 days afterAH onset between January and December 2000ere studied prospectively. Patients who did notndergo cerebral angiography because of gravelinical condition were excluded from study. Pa-ient characteristics are shown in Table 1. In allatients SAH was verified by computed tomographyCT) on admission. Angiographic studies per-ormed during the same period in 10 age-matchedatients with unruptured aneurysms were used asontrols.

ANAGEMENTll patients were treated according to standard in-

ensive care guidelines. Patients showed a wideange of SAH severity. Of the 40 patients, 8 (20%)id not undergo treatment for ruptured aneurysm,ither because the neurologic condition was poor6 cases) or repeated angiography did not discloseneurysm (2 cases). The remaining 32 patients80%) underwent surgical clipping for ruptured an-urysms. The day of SAH was defined as Day 0.Postoperative management for all patients in-

luded mild hypervolemic therapy using an IV dripnfusion of low molecular weight dextran as neces-

1 Patient Characteristics

CHARACTERISTIC

SAH GROUPNO. (%)*(N � 40)

CONTROL GROUPNO. (%)*(N � 10)

endermale 18 (45) 5 (50)female 22 (55) 5 (50)

ge, yearsrange 38–80 43–69mean � SD 58 � 10 56 � 8

CS on admission13–15 15 (38)7–12 7 (18)3–6 18 (44)

neurysm locationICA 11 (28)MCA 11 (28)ACA 13 (31)VBA 4 (10)unknown 1 (3)

ay of surgery0–1 24 (60)2–3 6 (15)4– 2 (5)no surgery 8 (20)

SAH, subarachnoid hemorrhage; GCS, Glasgow Coma Scale; SD, stan-ard deviation; ICA, internal carotid artery territory; MCA, middle cere-ral artery territory; ACA, anterior cerebral artery territory; VBA, verte-robasilar artery territory.

*Except for age.

ary. Dobutamine and nicardipine were given to c

ost patients during the risk period for vasospasm.ontinuous cisternal, ventricular, or lumbar exter-al drainage was established when necessary toontrol ICP.

NGIOGRAPHIC STUDIES ANDATA ANALYSES

outine preoperative angiography was performedithin 3 days after SAH. When the patients showedcute hydrocephalus on admission, which seemedo be the main cause of poor neurologic conditions,e initially performed ventricular drainage fol-

owed by cerebral angiography. A 5-French selec-ive catheter (Medikit, Tokyo, Japan) was insertedia the femoral artery into the common carotid andertebral arteries, and conventional anteroposte-ior and lateral views were obtained. Contrast agentas injected into the common carotid artery withn autoinjector (5 mL/sec; total, 7 mL).A digital subtraction unit (DFP-2000A, Toshiba,

okyo, Japan) was used for angiographic studies.ther equipment used included a 1000 mA X-ray

ube and a trimode cesium iodide image intensifieroupled to a television camera with a 1000:1 signal-o-noise ratio. Image quality was maximized by re-asking, reregistration of the mask, and a pixel shiftith reference to the contrast agent to correct foratient motion. The conventional exposure se-uence used for cerebral angiography was 5 sec-nds � 3 seconds followed by 2 seconds until the

ntracranial dural sinus was fully opacified.Lateral views of right and left carotid angiograms

CAG) were used for assessment of arteriovenousransit time (AVTT), which was defined as timenterval between initial opacification of the intracra-ial internal carotid artery (ICA) and that of theolandic vein (Figure 1). Mean AVTT (mAVTT) was

he mean of AVTTs from the right and left CAG:

AVTT � TV � TA

mAVTT � (AVTTR-CAG � AVTTL-CAG) / 2here AVTTR-CAG (or L-CAG) is AVTT in the right (or

eft) CAG, TA is time from injection until initialpacification of the intracranial ICA, and TV is time

rom injection until initial opacification of the Ro-andic vein.

The neurologic condition of all patients was re-orded throughout the hospital stay by the attend-ng neurosurgeons. The following information wasecorded for each patient: gender, age, Glasgowoma Scale score (GCS) on admission, SAH severityccording to admission CT (Fisher group) [6], andneurysm location. Outcome was assessed at 3onths after onset according to the Glasgow Out-

ome Scale (GOS), representing good recovery

((

tyew3adtpctu1

RNpnmqh

culwtiew

aorS3i(aaG

pt

117Circulation Time in SAH Surg Neurol2004;62:115–120

GR), moderate disability (MD), severe disabilitySD), vegetative state (VS), or death (D) [10].

For statistical analyses, mAVTT (more vs. lesshan 7 seconds), age (older vs. younger than 60ears), and Fisher group (I or II vs. III or IV) werexpressed dichotomously, and GCS on admissionas expressed trichotomously (13–15 vs. 7–12 vs.–6). Relationship between mAVTT and other vari-bles were analyzed using the chi-squared test. Ad-itionally, mAVTT was tested for association withrichotomized GOS (GR or MD vs. SD vs. VS or D). Arobability less than 0.05 was considered to indi-ate statistical significance. All statistical calcula-ions were performed with a personal computersing a statistical software package (SPSS, version0.0; SPSS Japan, Tokyo).

esultso patient suffered from major surgical com-lication. Six patients showed delayed ischemiceurologic deterioration (temporary in 5 and per-anent in 1), and nine patients subsequently re-

uired shunt procedure for normal pressureydrocephalus.Correspondence between AVTT in right and left

arotid angiograms in each patient is shown in Fig-re 2. The difference in AVTT between sides was

ess than 2 seconds in most patients; the 2 patientshose right-left differences exceeded 2 seconds had

emporal lobe intracerebral hematomas with signif-cant shift of midline structures, and AVTT ipsilat-ral to the hematoma was longer. These 2 patientsere excluded from statistical analyses.

1 AVTT was defined as the interval between the first apartery and the first opacification of the Rolandic ve

Figure 3 shows the relationship between AVTT s

nd outcome in SAH patient groups defined by GCSn admission. AVTT in patients in the control groupanged from 3.2 to 5.1 seconds (mean: 4.4). AVTT inAH groups depended on initial GCS, ranging from.4 to 8.0 seconds (mean: 5.4) in patients with annitial GCS of 13 to 15; from 4.5 to 8.7 secondsmean: 6.4) in those with an initial GCS of 7 to 12;nd from 5.8 to 15.0 seconds (mean: 9.1, excludingngiograms with nonfilling) in those with an initialCS of 3 to 6. During the study period, although

rance of contrast agent in the intracranial internal carotid

2 Scatter plot showing the differences in AVTT be-tween right and left carotid angiograms in individual

atients. Differences in AVTT between sides were lesshan 2 seconds, except for 2 patients in whom CT showed

peain.

ignificant midline shift from intracerebral hematoma.

sahtco

moarmpcap

afsa

opwm

T

G

A

G

F

O

Gv

r

118 Surg Neurol Yoshimoto et al2004;62:115–120

ome patients with GCS 13 to 15 showed moder-tely long AVTT (6–8 seconds), all these patientsad favorable outcomes. In patients with a GCS of 7o 12 and 3 to 6, AVTT clearly was related to out-ome, with longer AVTT corresponding to poorerutcome.Table 2 shows relationships between dichoto-ized mAVTT (more vs. less than 7 seconds) and

ther prognostic variables or outcome. Initial GCSnd Fisher (CT) groups on admission were closelyelated to mAVTT. Older patients were somewhatore likely to have longer mAVTT than youngeratients, although this tendency was not signifi-ant. We found no relationship between mAVTTnd gender. Longer mAVTT was strongly related tooor outcome.The main purpose of this study was to determinethreshold value that might serve as an indication

or active treatment. Figure 4 shows the relation-hip between arbitrarily chosen upper thresholdsnd occurrence of any functional recovery (GR, MD,

3 Scatter plot showing the relationship between AVTTand outcome in patients in control and SAH groups.

he latter groups are defined by GCS on admission.

w

r SD) in patients with AVTTs exceeding the cutoffoint. Likelihood of functional recovery in patientshose AVTT exceeded the threshold decreased al-ost linearly as the cutoff point was increased. No

2 Relationship of mAVTT to Other Parametersand Outcome

MAVTT INPREOPERATIVE ANGIOGRAM

<7 SEC >7 SEC P VALUES

CS on admission �0.00113–15 14 17–12 4 33–6 3 13

ge 0.39�60 yrs 14 9� 60 yrs 7 8

ender 0.79Male 9 8Female 12 9

isher group 0.01I, II 7 0III, IV 14 17utcome �0.001GR, MD 20 3SD 1 6VS, D 0 8

mAVTT, mean arteriovenous transit time; GCS, Glasgow Coma Scale;R, good recovery; MD, moderate disability; SD, severe disability; VS,egetative state; D, death.

4 Correlation between arbitrarily chosen upperthresholds of AVTT for active treatment and occur-

ence of any functional recovery by 3 months in patients

ith AVTTs exceeding these cutoff points.

pt

DInsdtctfcSdTtfeRtwto

ticttmiteetl

bpdAhsabsS

psa(

wCuv

bisbootsldibSgeisvhrpcppsdtowcPpiIip

CAgrapctAw

119Circulation Time in SAH Surg Neurol2004;62:115–120

atient with an AVTT longer than 12 seconds at-ained functional recovery.

iscussionntracranial circulation time can be measured usingondiffusible substances. Various methods for mea-urement of transit time have been described usingynamic CT [1,4,12] or DSA [11,15]. Mean transitime (MTT) has been assessed as an indicator oferebral circulation time; the time-density curve inhe region of interest was fitted to a gamma variateunction to calculate MTT. We initially attempted toalculate MTT, but motion artifact in acute-stageAH patients caused significant distortion of time-ensity curves, precluding reliable assessment.herefore, we used direct inspection of DSA moni-

oring to calculate circulation time as an intervalrom when contrast material filled the intracranialntry vessel (ICA) to when contrast entered to theolandic vein to exit to the dural sinus. We believe

hat our method proved to be a useful and practicalay to evaluate cerebral circulation time. Little

ime was needed for analysis, and no equipmentther than a conventional DSA unit was required.In occlusive cerebrovascular disease, prolonga-

ion of cerebral circulation time reflects an increasen cerebral blood volume (CBV) resulting from de-reased perfusion pressure and compensatory dila-ion of the vascular bed. In acute-stage SAH pa-ients, however, elevated ICP instead would be theain factor delaying cerebral circulation to result

n decreased cerebral blood flow (CBF). However,he acute phase response has not been fully delin-ated, and acute vasoconstriction after SAH appar-ntly occurs independently of changes in ICP [3];his may be associated with delayed cerebral circu-ation and decreased CBF [16].

In our study, AVTT did not differ significantlyetween right and left carotid angiograms in mostatients. Two exceptions involving a large ICH andeviation of midline structures showed a longerVTT (difference, �2 seconds) ipsilateral to theematoma. This implies that an intracranial pres-ure gradient existed in these patients [17]. Gener-lly, though, we could determine AVTT on one ar-itrarily chosen side as an indicator of ICP andeverity of SAH, provided that CT showed typicalAH findings without a large space-occupying mass.In this prospective study, AVTT during the acute

hase following SAH proved to be associated witheverity of SAH indicated by clinical grade (GCS ondmission) and extent of clot demonstrated by CT

Fisher group). Longer AVTT also was associated f

ith poorer short-term outcome (at 3 months).linical usefulness of AVTT appeared to be partic-larly promising in patients with moderate to se-ere SAH (GCS 3–12).Our study has several limitations. First, applica-

ility of our results is limited by the study design,nvolving a relatively small number of patients at aingle institution; as a result, a potential exists forias that might render predictions inapplicable tother populations. However, we believe that studyf prospectively collected consecutive SAH pa-ients during a relatively short period can minimizeelection bias. Second, cerebral circulation time,ike clinical grading scores [5,9,18], is time-depen-ent; we did not control for the effect of the time

nterval from onset of SAH to angiography. ICP haseen reported to dramatically increase just afterAH to approach arterial blood pressure, and thenradually decrease [7]. Outcome might be influ-nced by the duration and extent of the associatednitial intracranial circulatory arrest [7]. In ourtudy, angiography was performed at various inter-als from onset, usually after at least several hoursad passed to allow stabilization of patients’ neu-ologic condition, but always during the acutehase (within 3 days). Therefore, no conclusionsan be drawn from our data regarding the temporalrofile of AVTT or the effect of timing of angiogra-hy on the applicability of AVTT to treatment deci-ions. Finally, we did not analyze circulation timeuring any period of vasospasm, an event followinghe acute phase that contributes significantly toverall mortality and morbidity. Evaluation of MTTith dynamic DSA during treatment of symptomaticerebral vasospasm was reported to be useful [15].resumably, prolonged circulation time in this laterhase would be mainly caused by arterial narrow-

ng and cerebral ischemia rather than increasedCP. Further studies are needed to elucidate themportance of data characterizing the ischemichase in addition to the initial phase.

onclusionsVTT calculated from acute-phase routine angio-rams in SAH patients reflects increased ICP and iselated to prognosis. AVTT assessment performeds part of the standard admission protocol shouldrovide clinicians with a cost-effective, useful out-ome predictor that is especially applicable to pa-ients with poor-grade SAH. Extremely delayedVTT would indicate irreversible brain damage;hen AVTT exceeds 12 seconds, our results (zero

unctional recovery rate) would not support inter-

vc

R

1

1

1

1

1

1

1

1

1

CTmd(tasnpt3i1vogaabpu

120 Surg Neurol Yoshimoto et al2004;62:115–120

entions such as surgical clipping or intravascularoil embolization.

EFERENCES1. Axel L. Cerebral blood flow determination by rapid-

sequence computed tomography: a theoretical anal-ysis. Radiology 1980;137:679–86.

2. Bailes J, Spetzler RF, Hadley MN, Baldwin ME. Man-agement morbidity and mortality of poor grade aneu-rysm patients. J Neurosurg 1990;72:559–66.

3. Bederson JB, Levy AL, Ding WH, et al. Acute vasocon-striction after subarachnoid hemorrhage. Neurosur-gery 1998;42:352–60.

4. Beringer WH, Axel L, Norman D, Napel S, RedingtonRW. Functional imaging of the brain using computedtomography. Radiology 1981;138:711–6.

5. Drake CG. Report of World Federation of NeurologicalSurgeons Committee on a universal subarachnoidhemorrhage grading scale. J Neurosurg 1988;68:985–6.

6. Fisher CM, Kistler JP, Davis JM. Relation of cerebralvasospasm to subarachnoid hemorrhage visualizedby computerized tomographic scanning. Neurosur-gery 1980;6:1–9.

7. Grote E, Hassler W. The critical first minutes aftersubarachnoid hemorrhage. Neurosurgery 1988;22:654–61.

8. Hase U, Reulen HJ, Fenske Schumann K. Intracranialpressure and pressure volume relation in patientswith subarachnoid haemorrhage (SAH). Acta Neuro-chir 1978;44:69–80.

9. Hunt WE, Hess RM. Surgical risk as related to time ofintervention in the repair of intracranial aneurysms.J Neurosurg 1968;28:14–20.

0. Jennett B, Bond M. Assessment of outcome after se-vere brain damage. A practical scale. Lancet 1975;1:480–4.

1. Kwan ESK, Hall Enzmann DR. Quantitative analysis ofintracranial circulation using rapid-sequence DSA.Am J Roentgenol 1986;146:1239–45.

2. Norman D, Axel L, Beringer WH, et al. Dynamic com-puted tomography of the brain: technique, data anal-ysis, and applications. AJNR 1981;2:1–2.

3. Nornes H. The role of intracranial pressure in thearrest of haemorrhage in patients with ruptured in-tracranial aneurysm. J Neurosurg 1973;39:226–34.

4. Nowak G, Schwachenwald R, Arnold H. Early manage-

ment in poor grade aneurysm patients. Acta Neuro-chir 1994;126:33–7.

5. Touho H. Hemodynamic evaluation with dynamicDSA during the treatment of cerebral vasospasm.Surg Neurol 1995;44:63–74.

6. Umansky F, Kaspi T, Shalit MN. Regional cerebralblood flow in the acute stage of experimentally in-duced subarachnoid hemorrhage. J Neurosurg 1983;58:210–6.

7. Wolfla CE, Luerssen TG, Bowman RM, Putty TK. Braintissue pressure gradients created by expanding fron-tal epidural mass lesion. J Neurosurg 1996;84:642–7.

8. Yoshimoto Y, Wakai S, Ochiai C, Nagai M. Significanceof papillary reactivity in poor-grade aneurysm pa-tients as prognostic factor and an indication for ac-tive treatment. Br J Neurosurg 1997;11:25–31.

OMMENTARYhe authors described the angiographical assess-ent of cerebral circulation time for outcome pre-

iction in patients with subarachnoid hemorrhageSAH). They concluded that arteriovenous transitime (AVTT) calculated from acute-phase of routinengiograms reflected increased intracranial pres-ure (ICP) and was related to prognosis. As mosteurosurgeons have experienced, the patients withoor grade recovered dramatically within a shortime after admission or did not recover throughout

days after the onset. In Figure 3, the authorsndicated that the patients who showed GCS of 3 to2 on admission and AVTT of 6 to 12 seconds hadarious outcomes. Therefore, further investigationn relationship among AVTT, patients’ neurologicalrade and the interval from onset will be needed,llowing us to evaluate the relation between AVTTnd outcome more precisely. However, the AVTTased on the routine angiogram that can be easilyerformed in every institution will provide clinicalsefulness for outcome prediction on SAH patients.

Namio Kodama, M.D.Department of Neurosurgery

Fukushima Medical University

Fukushima, Japan

Medicare and Medicaid recipients are dependent on programsdoomed to bankruptcy. Thus, the natural state is of Universal

Uninsurance. If one demands the unachievable endpoint, thechimera of Universal Insurance, then mandatory governmentcoverage must be the ultimate agenda.

—AAPS News 2004;60:1