Cerebrospinal fluid cellular response in uncomplicated acute ischemic stroke

JStroke Cerebrooasc Dis 1992;2:168-172 © 1992 National Stroke Association

Cerebrospinal Fluid Cellular Response in Uncomplicated Acute Ischemic Stroke Manuel Ramirez-Lassepas, M.D., and Barbara K. Patrick, M.D.

Because information regarding cerebrospinal fluid (CSF) cellular response in acute ischemic stroke (AI5) is inconclusive, we conducted the following investigation. Cell counts were performed in C5F obtained within 24 h of onset of symptoms in 118 adult patients 86 years or younger with uncomplicated A15, after hemorrhage and mass effect were ruled out by computed tomography. In 98, C5F examination was repeated on day 7. None of the patients had evidence for systemic infection, inflammatory disease, vasculitis, leukocytosis, anemia or coagulopathy. C5F pleocytosis was defined as white blood cell (WBC) count of five or more mononuclear leukocytes (MNLs) and/or one or more polymorphonuclear leukocytes per cubic milliliter. No cells were found in 78 (66.1%) patients, 25 (21.2%) had one to four MNLs , and 15 (12.7%) had pleocytosis; none had hypoglycorrhachia. Nine patients had traumatic lumbarpunclure (LP); in five of them, pleocytosis persisted after correction for red blood cell counl. The highest WBC count was 15 cells per cubic milliliter. Only two patients had C5F pleocytosis on day 7; both had had traumatic LP and pleocytosis acutely. We found no correlati0D between C5F pleocytosis and cardiac source of embolus or type of ischemic stroke. From our data, we conclude that in uncomplicated AI5, C5F pleocytosis is rare and, when present, is mild; it has no correlation with type of stroke and is of no diagnostic value. Key Words: Acute stroke-Cerebrospinal fluid pleocytosisPrognosis.

Since the advent of computed tomography (CT) of the head, the diagnostic value of cerebrospinal fluid (CSF) examination in acute stroke has been questioned (1,2). It is deemed by some that lumbar puncture (LP) is an unnecessary risk (3,4), since CT is diagnostic for intracerebral hematomas with close to 100% sensitivity and specificity, and, to a somewhat lesser degree of reliability, for subarachnoid hemorrhage (5-7) . However, in ischemic stroke, the diagnostic accuracy of CT is much lower, especially in the acute phase. Early changes suggestive of the diagFrom the Department of Neurology, 51. Paul Ramsey Medical Center, 51. Paul, MN, U.S.A. Address correspondence and reprint requests to Dr. M. Ramirez-Lassepas at Department of Neurology, 51. Paul Ramsey Medical Center, 640 Jackson Street, 51. Paul, MN 55101, U.S.A. 168

/ STROKE CEREBROVASC DIS, VOL. 2, NO.3, 1992

nosis within the first 24 h after onset are seen only in approximately 35% of cases (3,8). At this time, CT is of little help in establishing the possible mechanisms responsible for vascular occlusion (9). The literature on the subject of CSF changes in stroke, for the most part, precedes the generalized use ofCT in the diagnosis of acute stroke (10-21). Studies on cellular response in the CSF are few (1,10-17); the latest was published in 1975 (1). So far the potential value of CSF. examination in cases of acute ischemic stroke (AIS) with atypical presentation, as an aid in determining the possible etiology of vascular occlusion has remained unresolved (22-25) . We had the unique opportunity to revisit this issue when, as part of a trial of a therapeutic agent for the treatment of acute cerebral infarction, the protocol included examination of the patient's CSF within 24 h

CSFPLEOLYfOSIS IN STROKE

of onset of symptoms and again on the seventh day (26). The results of CSF cell counts of patients screened and included in the study protocol are reported in this article.

(RBC) counts in cases of traumatic LP in which cell counts were corrected by using a RBC: white blood cell (WBC) count ratio of 700:1, (17,29). Chemistry and enzyme determinations were performed on the first three tubes, and the fourth tube of CSF was deep frozen for later date determination of 5-hydroxytryptamine and hydroxyindolacetic acid (26). Pleocytocis in the CSF was defined as WBC counts of five or more mononuclear leukocytes (MNL) and/or one or more polymorphonuclear leukocytes (PML). In addition to CT and LP, all patients had carotid Doppler examination and transthoracic cardiac echocardiography, as well as electrocardiography (12lead). Forty-seven patients had cerebral angiography. Nine patients suffered complications during their hospitalization; four died; autopsy was performed in three and cerebral infarction was corroborated; only one patient was judged to have died as a consequence of repeated cerebral embolization; the other three suffered cardiac deaths.

Methodology Over a 3-year period, 194 adult patients suffering from AIS syndromes, diagnosed by previously established criteria (27), were judged to be candidates for the aforementioned trial by meeting the following characteristics: younger than 86 years with moderate to moderately severe neurologic deficit (28) who did not have evidence of systemic infection or inflammatory disease; leukocytosis (WBC > 10,000); anemia or coagulopathy; and intracerebral or subarachnoid hemorrhage, as well as mass effect due to cerebral edema, ruled out by CT. Of these patients, 118 underwent LP within 24 h of onset of symptoms after signing informed consent, and CSF examination was repeated on day 7 in 98. LP was performed under local anesthesia following standard techniques; No. 22 French needles were used to obtain CSF in aliquots of 2 ml (tubes 1, 2, and 3) and 5 ml (tube 4), after pressure was measured and recorded. Cell counts in CSF were performed using an "improved" Neubauer 0.1 mm-deep hemocytometer by a trained hematology technician at the section of hematology of the Department of Laboratory Medicine and Clinical Pathology of the Medical Center. Counts were performed in tubes 1, 2, and 3; results from tube 1 were used as "true values." Those of tubes 2 and 3 were used for corroboration and to determine reduction in the number of red blood cell

Table 1.

Results Table 1 shows the cell counts in CSF obtained within 24 h of onset of AIS, as well as the results of RBC counts in those patients with traumatic LP and pleocytosis and the results of CSF examination at 48 h after traumatic LP and on day 7 after AIS. Pleocytosis was found in 15 (12.7%) patients; the highest cell count was 15 WBC/cu ml (10 MNLand 5 PML). In 78 (66.1%) patients, no cells were found, whereas 25 (21.2%) had one to four MNL. Five of nine patients who had traumatic LPs had CSF pleocytosis, and these were the same two patients who

Results ofcell counts periormed ill CSFobtained unthin 24 h of onset ofacute ischemic stroke" No.PML

No.MNL

0

0 1 2 3 4 5

78 11 10 3 111 0 0

6 5+ 0 0 0 0 0

103

11

10

Total patients

1

No. patients

NoRBC traumatic"

48·h LP RBC

MNU

PMU

38611 354+ 350§ 4133,513t 1,259'

NoLP No cells No cells No cells 468 812

4 (5) 12 (6)

2 (0) 2 (1)

l'

85 16 11 3 1 1 1

1

118

2

3

4

5

1§ 0 0 0 0 10

0 0 0 0 0 0 0

0 0 It 0 0 0 0

0 0 0 0 0 0

2

0

1

'PML, polymorphonuclear leukocytes; MNL, mononuclear leukocytes; RBC, red blood cells. bThe values in this column were obtained in the patients with the corresponding symbols in the white blood cell count section to the left. 'Results of CSF examination on day 7 are given in parentheses. / STROKE CEREBROVASC DIS, VOL. 2, NO.3, 1992

169

M. RAMIREZ-LASSEPAS AND B. K. PATRICK

Table 2. Relationship of CSFcellular response to type ofstroke Cellular response in CSF Type of stroke Card ioembolic Large vessel thrombosis Arterioembolic Small vessel occlusion Undetermined cause Totals

WBC (%)" 6 4 2 1 2

(15.8) (15.4) (14.3) (5.9) (8.7)

15 (12.7)

RBCb

2! 2 1 0 1 6

<4 MNL (%) 8 6 3 2 4

(21.0) (23.1) (21.4) (11.8) (17.4)

25 (21.2)

None (%) 24 16 9 14 17

(63.1) (61.5) (64.3) (82.3) (63.9)

78 (66.1)

Total (%) 38 26 14 17 23

(32.2) (22.0) (11.9) (14.4) (19.5)

118 (100)

HTCT(%)< 6 4 2 1 3

(16.6) (16.0) (15.4) (6.2) (13.6)

16 (14.3)

"WBC, pleocytosis. bRBC, all considered traumatic LP, all had WBC. 'HTCT, hemorrhagic transformation by CT; (%), percent of 112. dpatients had persistent pleocytosis on day 7.

had pleocytosis on day 7. All patients with CSF pleocytosis had repeated LP on day 7. Table 2 shows the type of stroke in relation to the presence of CSF pleocytosis and presumed traumatic LP. The types of stroke were defined as follows: cardioembolic, when there was a cardiac source of embolism (CSOE) demonstrated, such as cardiac arrhythmia's (atrial fibrillation, sick sinus syndrome), or cardiac echography showed a cardiac clot, ventricular wall akinesia, or aneurysm, or major left valvular abnormality associated with arrhythmia (30). Largevessel thrombosis was diagnosed when carotid Doppler and/or cerebral angiography demonstrated carotid and/or middle, anterior, or posterior main branch occlusion in the absence of cardiac source of embolus. In eight patients, angiography demonstrated distal occlusion associated to proximal stenosis, with or without ulceration and no CSOE; these patients were considered to have had arterioembolic stroke. Six patients in whom CT showed relatively small cortical infarcts, and carotid Doppler or angiography disclosed carotid artery stenosis of >50%, and no CSOE was demonstrated, were also considered to have had arterioembolic stroke. Patients presenting with one of the lacunar stroke syndromes (31) and/or those in whom CT demonstrated a small white matter lesion were diagnosed as having small vessel occlusion. When none of the above criteria was found, patients were classified as having had a stroke of undetermined etiology. None of the patients had indication of hemorrhagic infarction by CT acutely; 16 of 112 (14.3%) had changes on the second CT study (days 5-8) suggestive of either hemorrhagic transformation of pale infarction or "Iuxury perfusion" (32); these changes did not correlate with CSF pleocytosis, nor with the type of stroke diagnosed (Table 2). 170

I STROKE CEREBROVASC DIS,

VOL 2, NO.3, 1992

Discussion The presence of pleocytosis in the CSF following AIS was reported as far back as 1912 by Babinski and Gendron, who reported on two patients with presumably embolic AIS due to mitral valve disease, who on the second day after stroke had CSF pleocytosis of 190 and 450 WBC, 90% of which were PMLs (10). In 1935, Aring and Merritt reported that, of 43 patients with cerebral thrombosis, 7 (16.3%) had CSF pleocytosis and one had more than 51 WBC/cu ml of CSF. Of 13 patients with cerebral embolism, 9 (69.2%) had CSF pleocytosis and 5 had more than 51 WBC/cu ml, one patient having 1,900 and another 3,600. The timing of LP in relation to the onset of symptoms and the presence of associated illness was not clearly indicated (12). Merritt and FremontSmith later reported that 7 of 51 patients (13.7%) with autopsy-proven cerebral thrombosis had CSF pleocytosis (13). Townsend et al. reported on selected cases of cerebral "softening" and intracerebral hemorrhage (ICH), concluding that significant CSF pleocytosis suggesting aseptic meningitis may be seen in both these conditions (14). By far, the most comprehensive study on the subject is that of Somas et aI. (16), who studied 108 patients with serial LP, the first of which was performed within 72 h of onset of symptoms. Sixteen of these patients had intracerebral hematomas and 92 cerebral infarction (CI), which was classified into three groups according to the results of cerebral angiography. There were 26 patients with normal angiography classified as "infarcts of the internal carotid artery area without any collateral circulation" (Group 1). Of these, three (11.5%) had CSF pleocytosis; one had 7 MLN, and the other two had 20 and 80 WBC, with 60% and 80% PML, respectively. Eleven (42.3%) had

CSFPLEOLYfOSIS IN STROKE

no cells in the CSF, and 12 (46fl %) had less than 5 waC/cu ml of CSF. Of 25 patients with abnormal cerebral angiograms classified as "infarcts in the internal carotid artery area with arterial occlusion and collateral circulation" (Group 2), eight (32%) had CSF pleocytosis that ranged from 6 to 60 WBC/cu ml with PML ranging from 10 to 90%. Only 1 (4%) patient in this group had no cells in the CSF, and the remaining 16 (64%) had less than 5 WBC/cu ml; all had from 10 to 60% PML This group was supposed to represent hemorrhagic cerebral infarction (HCI). There were 41 patients with CSOE classified as "embolic infarcts"; 12 (29%) had CSF pleocytosis, 1 had only 6 MNL, and of the others, 8 had 7-20 WBC with 10-90% PMLand 3 had 150, 500, and 1,100 with 70%, 85%, and 95% PMLs, respectively. Only 4 patients (9.7%) had no cells in the CSF, and 9 (22%) had less than 4 WBC/cu ml. In 18 of their patients, (19.5%) pathologic confirmation of CI was obtained; seven died within 7 days of onset and four had pale and three hemorrhagic infarctions. The authors found a correlation between number of PMLs and hemorrhagic infarction. The actual RBC counts are not given. The authors used Soma's own formula to calculate pleocytosis in CSF (33) and the method of spectroscopic oxyhemoglobin absorption and the presence of erythro- and siderophagocytes as an indicator of whether the CSF was hemorrhagic or not. Oxyhemoglobin absorption was negative in the patients in group 1, and no erythro- or sideroblasts were found in their CSF, whereas patients in groups 2 and 3 had 17 and 21 %, respectively, and those with ICH, 38%. In the patients with hemorrhage, 13 (81.2%) had CSF pleocytosis,S had from 9 to 50 WBC, 6 from 110 to 1,000; 1 had 5,000, and another 11,000 WBC. Patients with pleocytosis had from 50% to 100% PML The authors concluded that CSF pleocytosis, mainly the presence ofPML, is determined by contamination of the subarachnoid space with RBC from the infarcted area; therefore, in HCI, there would be CSF pleocytosis, whereas in ischemic (pale) cerebral infarction (ICI) there would not (16). In their retrospective study, Lee et al. (1) looked at the results of CSF examination in 45 patients with autopsy, verified CI and 16 with ICH. Their analysis focused on the diagnostic value of CSF examination in differentiating CI from ICH, which was often the purpose of the studies published prior to the widespread use of CT (10-16). Twenty-seven patients had HCI, 10 of them (37%) had more than 10 RBC/cu ml of CSF, and three had more than 3,000. Six (33%) of the 18 patients with ICI also had more than 10 RBC/cu ml; the highest count was 500 RBC. CSF pleocytosis was found in 7 of 21 (33.3%) patients with HCI and 5

of 15 (33.3%) with ICI. Eleven of the 12 patients with pleocytosis had only 1-5 PML, and the other patient, with HCI, had 6 MNLIcu ml of CSF; in addition 3 HCI and 2 ICI patients had 1-4 MNLIcu ml of CSF. The authors concluded that CSF examination was of limited value in differentiating ICI from HCI. Frankly hemorrhagic CSF was diagnostic of ICH; however, 25% of the patients had clear CSF (1). In our patient population, severe neurologic illness and associated conditions that could influence CSF pleocytosis were excluded so that the patient population is not comparable to the above-reviewed studies (1,11-16); however, the low numbers ofWBC found by us compare with those reported by Lee et al. (1). We could not find a correlation between supposedly hemorrhagic embolic infarction and CSF pleocytosis; this is contrary to the findings reported by Somas et al. (16). In a recent exchange of opinions, Powers (22,23) and Hart (24,25) discussed the value of CSF examination in atypical cases of AIS, especially in cases in which AIS could be an embolic complication of infective endocarditis (IE) in which heparin anticoagulation would be contraindicated (34). The literature on the subject (35-38) agrees that CSF pleocytosis is a common finding in as many as 50% of patients suffering from AIS due to IE embolization. From our findings, one could argue that CSF pleocytosis in AIS is an indicator of stroke complicated by systemic illness such as IE. On the other hand, as Hart points out (24,25), even though as many as 20% of patients with IE have complicating cerebral embolization, only 2% of those AIS present as an isolated or first symptom; AIS due to IE complication represents less than 1% of all strokes (25). Our results show that, in uncomplicated AIS, CSF pleocytosis is rare; when present, it is mild, selflimited, and of no real diagnostic value. Therefore, in agreement with the opinions of others (2-4,24,25), we conclude that LP, under these circumstances, is an unnecessary risk and should not be performed.

References 1. Lee MC, Heaney LM, Jacobson RL, Klassen AC. Cerebrospinal fluid in ce rebral hemorrhage and infarction. Stroke 1975 ;6:638-41. 2. Ruff RL, Dougherty JH Jr. Evaluation of acute ischemia for anticoagulant therapy. Computed tomography or lumbar puncture? Nellrology 1981;31:736-40. 3. Ramirez-Lassepas M, Quinones MR Heparin therapy for stroke: hemorrhagic complications and risk factors for intracerebral hemorrhage. Nellrology 1984;34:1147. / STROKE CEREBROVASC DIS, VOL 2, NO.3, 1992

171

M. RAMIREZ-LASSEPASA ND B. K. PATRICK

4. Walsh MN, Fisher GG, Anderson D, Mastri A. Fatal intracranial extension of spinal hemorrhage after lumbar puncture. Arch Neuro11984, 41:987-9. 5. Barraquer-Bordas L, IlIa I, Escartin A, Marti-Vilalta JL Thalamic hemorrhage. A study of 23 patients with diagnosis by computed tomography. Stroke 1981;12: 524-7. 6. Heros RC, Zervas NT. Subarachnoid hemorrhage. Alllw Rev Med 1983;34:367-75 . 7. Houser OW, Campbell JK, Baker HL [r, Sundt TS Jr. Radiologic evaluation of ischemic cerebrovascular syndromes with emphasis on computed tomography. Radiol Clin NOr/It Am 1982;20:123-42. 8. Hakim AM, Ryder-Cooke A, Melanson D. Sequential computerized tomographic appearance of strokes. Stroke 1983;14:893-7. 9. Mohr JP, Barnett HJM. Classification of ischemic strokes. In: Barnett HJM, Mohr JP, Stein BM,Yatsu FM, eds. Stroke: pathophysiology, diagnosis and management. New York: Churchill Livingstone, 1986:281-91. 10. Babinski J, Gendron A. _Leucocytose du Iiquide cephalorachidien au cours du ramollissement du l'ecore cerebraIe. Bull Soc Med Hop Paris 1912;33:370-4. 11. Cone WV, Barrera SE. The brain and the cerebrospinal fluid in acute aseptic cerebral embolism. Arch Neurol Psychiatr 1931;25:523-47. 12. Aring CD, Merritt HH. Differential diagnosis between cerebral hemorrhage and cerebral thrombosis. Arch Intern Med 1935;56:435-6. 13. Merritt HH, Fremont-Smith H. The cerebrospinalfluid. Philadelphia: Saunders, 1937:197-203. 14. Townsend SR, Craig RL, Braunstein AL Neutrophilic leukocytosis in spinal fluid associated with cerebral vascular accidents. Arcl: Intern Med 1939;63:848-57. 15. Schaasfma S. On the differential diagnosis between cerebral hemorrhage and infarction. / Neurol Sci 1968; 7:83-95. 16. Somas R, Ostlund H, Muller R. Cerebrospinal fluid cytology after stroke. Arch NeuroI1972;26:489-501. 17. Westlake PT, Markovits CT, Stellar S. Cytologic evaluation of cerebrospinal fluid with clinical and histologic correlation. Acta CytoI1972;16:224-39. 18. Meyer JS, Stoica E, Pasku I, Shimazu K, Hartmann A. Cathecholamine concentrations in CSF and plasma of patients with cerebral infarction and hemorrhage. Brain 1973;96:277-88. 19. Meyer JS, Welch KMA, Okamoto 5, Shimazu K Disordered neurotransmitter function. Demonstration by measurement of norepinephrine and 5-hydroxytryptamine in CSF of patients with recent cerebral infarction. Brain 1974;97:655-65. 20. Allori L, Cioli V, Silvestrini B. Experimental and clinical data indicating a potential use of trazodone in acute stroke. Curr TI,er Res 1975;18:410-6 . 21. Welch KMA, Meyer [S, Neurochemical alterations in

172

/ STROKE CEREBROVASC DIS, VOL. 2, NO.3, 1992

22. 23. 24. 25. 26.

27. 28.

29. 30. 31.

32.

33. 34. 35. 36.

37.

38.

cerebrospinal fluid in stroke. In: Wood JH, ed. Neurobiology of cerebrospinalfluid. New York: Plenum Press, 1980:325-44. Powers WJ. Should lumbar puncture be part of the routine evaluation of patients with cerebral ischemia (letter). Stroke 1985;16:737 . Powers WJ. Letter. Stroke 1986;17:332-3. Hart RG. Response to Powers WJ (letter). Stroke 1985; 16:737. Hart RG. Letter . Stroke 1986;17:333. Ramirez-Lassepas M, Patrick BK, Snyder BD, Lakatua DJ. Failure of central nervous system serotonin blockage to influence outcome in acute cerebral infarction. A double blind randomized trail. Stroke 1986;17: 953-6. Classification of cerebrovascular diseases. III. Special report from the National Institute of Neurological Disorders and Stroke. Stroke 1990;21:637-76. Strand K, Asplund K, Ericksson 5, Hagg E, Lithner F, Wester PO. A randomized controlled trial of hemodilution therapy in acute ischemic stroke. Stroke 1984; 15:980-9. McMenemy WHo The significance of subarachnoid bleeding. Proc R Soc Med 1954;47:701-4. Rarnirez-Lassepas M, Cipolle RJ, Bjork RJ, et al. Can embolic stroke be diagnosed on the basis of neurologic clinical criteria? Arch NeuroI1987;44:87-9. MohrJP. Lacunes. In: BarnettHJM,MohrJP,SteinBM, Yatsu FM, eds. Stroke: pathophysiology, diagnosis and management. New York: Churchill Livingstone, 1986: 475-96. Hornig CR, Busse 0, Buettner T, Dorndorf W, Agnoli A, Akengin Z. CT contrast enhancement on brain scans and blood-CSF barrier disturbances in cerebral ischemic infarction. Stroke 1985;16:268-73 . Somas R. A new method for the cytological examination of the cerebrospinal fluid. / Neurol Neurosurg Psychiatry 1967;30:568-77. Cerebral Embolism Study Group. Immediate anticoagulation of embolic stroke: a randomized trail. Stroke 1983;14:668-76. Harrison MJG, Hampton JR. Neurological presentation of bacterial endocarditis. Br Med / 1967;1:14851. Pelletier LL, Petersdorf RG. Infective endocarditis: a review of 125 cases from the University of Washington Hospitals, 1963-1972. Medicine (Baltimore) 1977;56: 287-313. Garvey GJ, Neu He. Infective endocarditis-an evolving disease. A review of endocarditis at the ColumbiaPresbyterian Medical Center, 1968-1973. Medicille (Baltimore) 1987;57:105-27. Pruitt AA, Rubin RH, Karchrner AW, Duncan GW. Neurologic complications of bacterial endocarditis. Medicine (Baltimore) 1978;57:329-43 .