Dynamics of Polymorphonuclear Leukocyte Accumulation in Acute Cerebral Infarction and Their Correlation With Brain Tissue Damage
Background and Purpose This study was performed to study the dynamics of polymorphonuclear leukocyte (PMNL) accumulation in human cerebral infarction and its association with neurological outcome and brain lesion.
Methods A total of 88 patients diagnosed as having hemispheric ischemic stroke were examined. PMNL accumulation was studied using technetium-99m hexamethylpropyleneamine oxime (99mTc HMPAO)−labeled leukocyte brain single-photon emission computed tomography (SPECT). Volume of brain infarction was evaluated by CT scan. The Mathew Scale was used for neurological assessment. Dynamics of PMNL accumulation was studied at 3 to 6, 6 to 12, and 12 to 24 hours and 6 to 9, 28 to 30, and 90 days after stroke onset. In parallel, at admission, at 6 to 9 days, and at 28 to 30 days neurological outcome and infarction volume were evaluated.
Results Generally, PMNL accumulation progressively increased during 6 to 24 hours after stroke, remained at a high level up to 6 to 9 days and then declined. With the use of cluster analysis, all patients were subdivided into three groups: patients with severe PMNL accumulation that dramatically increased within 12 hours after stroke onset and persisted even at 28 to 30 days (group A); those with moderate PMNL accumulation that significantly decreased at 30 days (group B); and those with mild PMNL accumulation that decreased at 6 to 9 days (group C). Baseline neurological deficit and brain tissue damage at admission appear to be at a similar level for all groups of patients. In dynamics, however, in patients with severe PMNL accumulation, neurological outcome was worse and infarction volume larger than in patients with less marked PMNL accumulation.
Conclusions The present clinical study confirms that PMNLs intensively accumulate in the regions of cerebral infarction. The present study revealed that this accumulation correlated with the severity of the brain tissue damage and poor neurological outcome.
A body of evidence suggests that leukocytes, particularly PMNLs, contribute to ischemic damage in different tissues.1 2 In experimental models, it has been repeatedly demonstrated that PMNL accumulation in regions with depressed circulation aggravates ischemic tissue damage in brain and heart, and leukocyte depletion may decrease tissue necrosis.1 2 3 4 In clinical conditions, leukocyte accumulation in the region of cerebral ischemia has also been proved in a few clinical observations with brain scintigraphy or brain SPECT with labeled leukocytes.5 6 In patients with acute ischemic infarction it was demonstrated that leukocyte accumulation occurred in the early stages of stroke and persisted for more than a month. However, it remains unclear whether this phenomenon indeed may be responsible for progression of the brain tissue damage and whether it may alter neurological outcome in human ischemic stroke. Furthermore, dynamics of leukocyte accumulation were studied only during the first week after onset of the symptoms. Meanwhile, the critical period for brain tissue damage appears to be within 12 hours after the onset and only within this period might therapeutic intervention be sufficiently effective.7 8 9 Finally, no attempt has been made to selectively label leukocyte subpopulations for brain imaging, therefore leukocyte accumulation studied to date5 6 represents the accumulation of mixed leukocyte subpopulations, including PMNLs, monocytes, and lymphocytes. However, participation of these subpopulations in tissue damage and the time frames for their accumulation are different. In the early stages, which are critical for infarction, PMNLs are the main leukocyte population participating in the inflammatory reaction of ischemia. After a week, mononuclear leukocyte accumulation occurs and persists for many weeks, indicating completed stages of necrosis.1 2 3 4 10 11
The aim of the present work was to study the dynamics of PMNL accumulation in patients with acute cerebral infarction in the early stages of cerebral ischemia and to identify their association with neurological outcome and brain tissue damage.
Subjects and Methods
We entered 88 patients, 62 men and 26 women. Patient age ranged from 45 to 70 years (mean±SEM, 58±6 years). Patients were eligible for the study if they had acute neurological deficit with weakness against resistance in at least one limb within 12 hours (mean±SEM, 6.2±3.6 hours) prior to examination and were diagnosed as having hemispheric ischemic stroke in the middle cerebral artery territory. Thirty-four patients were admitted within 6 hours of onset, and others were admitted within 6 to 12 hours after onset of stroke. Admission criteria required that the deficit must have been present for more than 24 hours and that the initial CT scan must be free of any evidence of brain hemorrhage or hemorrhagic infarction, brain stem infarction, lacunar infarction, or any serious organic brain disease other than ischemic brain infarction. Follow-up CT scan examinations also did not reveal any signs of subsequently developed hemorrhage. Exclusion criteria required the absence of coma, the need for mechanical ventilation, hypotension (systolic blood pressure <100 mm Hg), and bradycardia (rate <50 beats per minute). Also excluded were patients with severe systemic disease (eg, recent myocardial infarction, cardiogenic shock, or renal or hepatic failure), systemic infection, cancer, and unstable diabetes. Informed consent was obtained from all patients or their responsible relative.
Detailed neurological examination was performed at admission within 6 to 12 hours after onset, at 6 to 9 days, and at 28 to 30 days. The Mathew Scale12 was used as the main criterion for assessing neurological outcome of the patients. For each patient, the score on the Mathew Scale (the sum of the scores for individual items) was evaluated. Patients were excluded from final analysis if they had on admission a Mathew Scale sum score >80 (ie, mild deficit). Patients were also excluded if they had a preexisting neurological deficit that would interfere with the investigators' ability to determine the outcome of the acute cerebral infarction. The neurological outcome was defined as the relative change of neurological deficit that was calculated using the following equation.13 14 If Yo is Mathew Scale sum score at admission and Yt is Mathew Scale sum score at 6 to 8 or 28 to 30 days, then X=100(Yt–Yo)/(100–Yo) for Yt≥Yo (improvement of neurological status) and X=100(Yt–Yo)/Yo for Yt<Yo (deterioration). For gradual estimation of the improvement of neurological status, three levels of Mathew Scale sum score increase (0% to 20%, 20% to 40%, and >40%) were used.14 Outcome evaluations were performed by neurologists blinded to SPECT data.
Brain SPECT Imaging and CT Scan
The dynamics of in vivo PMNL accumulation in the brain tissue with99mTc HMPAO–labeled cells were studied with SPECT as described in detail by Wang et al.6 15 Autologous leukocytes were isolated from 100 mL of blood. PMNLs were separated by centrifugation on gradient Ficoll-hypaque in accordance with standard procedure16 and labeled with the method of Uno et al.17 The cells were washed twice in phosphate-buffered saline and incubated in 10 to 15 mCi 99mTc HMPAO for 15 minutes at room temperature. The labeled PMNLs were washed and resuspended in 5 mL autologous plasma for reinjection. The in vitro viability study of labeled PMNLs by the trypan blue rejection was greater than 92%. The labeling efficiency varied between 70% and 80% of the dose added in vitro.
SPECT was carried out with a rotating gamma camera (General Electric, 64 projections of 30 seconds each) as described in detail by Laloux et al.18 Data were reconstructed to obtain consecutive axial slices every 12 mm parallel to and above the orbitomeatal line. Twenty symmetrical ROIs were considered over each cerebral hemisphere. In each ROI, differences in total count were expressed as a percentage of the values from the contralateral asymptomatic hemisphere, and AI was calculated using the following equation: AI=(infarcted hemisphere minus contralateral hemisphere)/(infarcted hemisphere plus contralateral hemisphere). Interhemispheric differences of at least 10% were considered to be significant.6 15 For each patient, the degree of PMNL accumulation was characterized by the highest interhemispheric asymmetry. In addition, we determined the size of PMNL accumulation by the number of ROIs with an interhemispheric difference of at least 10%.
First examination was performed 2 hours after the labeled PMNLs were reinjected, such a reinjection was made on the first day of examination for measurement of PMNL accumulation within 3 to 6, 6 to 12, and 12 to 24 hours. Additional injections were made to measure PMNL accumulation at 6 to 9, 28 to 30, and 90 days. The initial brain SPECT imaging with labeled PMNLs was performed within 6 to 12 hours after stroke onset. Follow-up examinations were performed in all patients at 12 to 24 hours, 6 to 9 days, 28 to 30 days, and 90 days. In 34 patients initial examination was performed within the first 3 to 6 hours after stroke onset and then as for other patients: within 6 to 12 hours, 12 to 24 hours, 6 to 9 days, 28 to 30 days, and 90 days after onset. Control examinations for PMNL accumulation were performed in 8 normal healthy volunteers (5 men, 3 women; mean age, 51±4 years).
The initial brain CT scan was performed in all patients within 6 to 12 hours after onset and was repeated at 6 to 9 days and 28 to 30 days. In the final analysis, patients included were only those in whom the initial or at least the second CT scan revealed focal hypodensities frequently associated with some mass effects and contrast enhancements. Contiguous axial slices (matrix of 512 pixels) were obtained parallel to the canthomeatal line. Slice thickness was 8 mm. The infarction volume (in cm3) was calculated by outlining areas of low density on axial images and multiplying them by the interimage distance.19 Lesion size was also expressed as a percentage of brain volume.19
Statistical analysis software (SPSS PC+, SPSS Inc) was used to analyze the data. All results are expressed as mean±SD. Differences between means were assessed using the Kruskal-Wallis test (nonparametric ANOVA). To allow us to attribute individual profiles of PMNL accumulation to homogeneous groups (classes), hierarchical cluster analysis was used.20 Each patient was included in this analysis as an independent object characterized by the set of parameters that represents data of PMNL accumulation at 6 to 12 hours, 12 to 24 hours, 6 to 9 days, 28 to 30 days, and 90 days. Data for PMNL accumulation at 3 to 6 hours were not used for clustering because they were not available for all patients. After clustering, the mean values (centroids) on each parameter characterizing patients separated into different groups (classes) were calculated and levels of significance comparing the differences between the group means were evaluated.20
In normal subjects from the control group, brain SPECT images revealed very little but symmetrical hemispheric uptakes of labeled PMNLs with negligible AIs for symmetrical ROIs. However, in all patients with acute cerebral infarction well-defined brain asymmetries showing greater uptake of labeled leukocytes were observed in the infarcted hemisphere. As shown in Fig 1A⇓, AIs were minimal at 3 to 6 hours after onset, but at 6 to 12 hours AIs increased. The increase progressed for 12 to 24 hours. The AI remained at high levels for 6 to 9 days but decreased at 28 to 30 days and virtually returned to control levels at 90 days. The size of PMNL accumulation showed similar dynamics (Fig 2A⇓).
The dynamics of PMNL accumulation varied significantly among different patients. Cluster analysis classified all profiles of AI and PMNL accumulation sizes into three (A, B, and C) homogeneous groups (classes). Fig 3⇓ shows the dendrogram of the cluster analysis. Dynamics of AI and size of PMNL accumulation for patients classified into these groups are summarized in Figs 1B and 2B⇑⇑. Generally, among patients from group A PMNL accumulation was markedly increased within 12 hours and remained significantly elevated for up to 6 to 9 days. In patients from group C, changes in AIs and sizes of PMNL accumulation within 12 hours remained trivial, PMNL accumulation did increase at 12 to 24 hours, but at 6 to 9 days PMNL accumulation declined. In group B, intermediate variants of PMNL accumulation profiles were included.
On the basis of the above ABC classification, we correlated clinical characteristics, neurological outcome, and lesion size for the patients identified as groups A, B, and C. There were no statistically significant differences between the three groups for age, sex, localization of infarction, or risk factors (Table 1⇓). The Mathew Scale sum score on admission (within 12 hours) for all three groups was also similar (Table 1⇓). However, we found marked differences for the relative changes in neurological deficit at 6 to 9 days and particularly at 28 to 30 days (Table 2⇓). The numbers of patients with good improvements of their neurological deficits were higher in group C than in groups B and, especially, A. At 28 to 30 days, improvement >40% was observed in 62.5% of patients in group C and in only 3.7% of patients from group A. Vice versa, in group A 33.3% of patients showed deterioration of neurological deficit, whereas in group C such a deterioration was observed in only one patient (4.2%).
The same differences were found in the dynamics of infarct volume. The volume at <12 hours from stroke onset did not differ between patients of the three groups (Table 3⇓). On admission, only 55.7% of patients showed positive CT scans for new cerebral infarctions. Numbers of patients with positive CT scans on admission were similar for all groups. At 6 to 9 days, the mean volumes of the infarctions, as well as lesion sizes, expressed as a percentage of brain volume, were significantly higher in group A than in group C (Table 3⇓). At 28 to 30 days these differences also remained statistically significant (Table 3⇓).
The present study confirms that in humans leukocytes infiltrate cerebral ischemic infarction. The dynamics of leukocyte accumulation show that PMNLs accumulate at 6 to 12 hours after stroke onset, remains high however with decreasing levels at 6 to 9 days, and then returns toward normal. Present data differ from those of Wang et al6 who revealed high levels of leukocyte accumulation up to 5 weeks after stroke. However, these authors studied mixed leukocyte accumulations; whereas only PMNL subfractions were used in the present study for labeling. Consequently, we assume that prolonged leukocyte accumulation observed by Wang et al6 was due to replacement of PMNLs by mononuclear leukocytes, as confirmed in experimental models.10 11 Our data emphasize the participation of PMNLs in the early stages of inflammatory reactions of cerebral ischemia, which may be important for tissue damage.1 3 11
The fact that PMNLs accumulate in areas of infarction does not, by itself, clarify its pathophysiological importance. For the first time, in the present work, PMNL accumulations have been correlated with neurological outcomes and signs of brain tissue lesions. Such analyses made it possible to separate patients into several groups characterized by their different but homogeneous types of dynamics of PMNL accumulation. The patients were objectively separated into groups: those with severe PMNL accumulations dramatically increasing within 12 hours after stroke onset and persisting for at least 28 to 30 days (group A); those with moderate PMNL accumulation that decreased significantly at 30 days (group B); and those with mild PMNL accumulation that decreased at 6 to 9 days (group C). It should be noted that on admission there were no statistically significant differences in clinical characteristics of the patients, neurological scores, or initial volumes and sizes of brain infarcts. Thus, baseline neurological deficit and brain tissue damage at admission appeared to be similar for all groups of patients. Accordingly, it may be speculated that in these three groups differences in the processes of PMNL accumulation are not determined by any initial heterogeneity of the patients.
The major finding of the present study is that among patients with acute cerebral infarction, the neurological outcome and lesion sizes differ depending on the various types of PMNL accumulation dynamics. Clear evidence for poor neurological outcome and large infarction volumes in patients with severe PMNL accumulations points to the importance of PMNL involvement in brain tissue damage after cerebral ischemia. At the moment, it is not clear what component of PMNL accumulation predominates in the infarcted tissue. PMNLs may accumulate within the vascular compartment, presumably due to adhesion on the vessel wall, and/or infiltrate into the brain tissue. Both these processes may produce the tissue damage.21 22 23 24 Intravascular accumulation of PMNLs can aggravate microthrombosis, thereby augmenting microcirculatory disorders and initiating vasospastic reactions and endothelium damage.1 2 3 4 22 24 Accumulated PMNLs may also augment tissue necrosis by producing free oxygen radicals1 4 21 and may release lysosomal proteolytic enzymes and cytokines.3 4 22 23 Finally, after transvascular migration and infiltration into ischemic tissue, PMNLs may directly induce cytotoxic effects.11 From a pathophysiological viewpoint, all these factors may increase hypoperfusion and necrosis, particularly in the ischemic penumbral zone. In patients in group C, with minimal PMNL accumulation, the infarction volume was minimal and neurological outcome was better compared with the other two groups.
This evidence for different PMNL accumulations in different patients with initially similar neurological status is not easy to explain. Mechanisms responsible for PMNL accumulation in ischemia are far from being fully understood. It is likely that when cerebral ischemia occurs, an array of different factors initiate PMNL recruitment from the circulatory bed. Factors include complement activation, expression of adhesion molecules, and release of chemotactic factors, such as platelet-activating factor.1 2 3 4 11 21 22 23 24 25 It seems likely that expression of all these factors as well as PMNL reactivity vary from patient to patient, thereby determining variation in PMNL accumulation.
It was not the intent of this work to study the influences of the degree of PMNL accumulation on the efficiency of therapeutic intervention. However, we hypothesize that the intensity of PMNL accumulation might alter the results of treatment of cerebral infarction, particularly since they may increase and accelerate irreversible tissue damage. There is evidence that cerebral ischemia, especially within the ischemic penumbra, is susceptible to therapeutic interventions when the latter are begun no later than 3 to 6 hours after stroke onset. When treatment is started later than 12 hours after stroke, it appears generally to be ineffective.8 9 Our data show for the first time that PMNL accumulation beginning 3 to 6 hours after the onset of ischemic stroke increases during the next 6 hours and reaches maximum within 12 to 24 hours. This is in concert with some experimental observations demonstrating that several hours are necessary to produce intensive leukocyte infiltration in ischemia.23 25 26 Presumably, this delayed process might reduce the treatment efficiency 6 hours after stroke onset because PMNLs accumulate mainly in the ischemic penumbra, which is the region most susceptible to therapeutic intervention, since it is not irreversibly damaged.27 Generally, PMNLs are likely to complete the formation of infarcted brain tissue, thereby decreasing effectiveness of any pharmacological intervention. Indeed, clinical experience shows that therapeutic nihilism has characterized the clinical approach to therapy of the “completed stroke.”7
In conclusion, results of the present clinical study confirm experimental data showing that PMNLs intensively accumulate in the regions of cerebral infarction. The present study revealed that this accumulation correlated with the severity of the brain tissue damage and with poor neurological outcome. This is consistent with the hypothesis that PMNL accumulation is a critical factor that aggravates cerebral ischemia and contributes to tissue necrosis, as has been demonstrated in experimental models of cerebral ischemia.1 24 Certainly, the clinical significance of these findings remains debatable and needs further corroboration. Nevertheless, these observations give hope that further investigations of the involvement of PMNLs in cerebral ischemia may clarify new diagnostic and therapeutic approaches.
Selected Abbreviations and Acronyms
|ROIs||=||regions of interest|
|SPECT||=||single-photon emission computed tomography|
|99mTc HMPAO||=||technetium-99m hexamethylpropyleneamine oxime|
- Received December 15, 1995.
- Revision received June 24, 1996.
- Accepted June 24, 1996.
- Copyright © 1996 by American Heart Association
Kochanek PM, Hallenbeck JM. Polymorphonuclear leukocytes and monocytes/macrophages in the pathogenesis of cerebral ischemia and stroke. Stroke. 1992;23:1367-1379.
Granger DN, Kubes P. The microcirculation and inflammation: modulation of leukocyte-endothelial cell adhesion. J Leukoc Biol. 1994;55:662-675.
Pozzilli C, Lenzi GL, Argentino C, Carolei A, Rasura M, Signore A, Bozzano L, Pozzilli P. Imaging of leukocytic accumulation in human cerebral infarcts. Stroke. 1985;16:251-255.
Wang P-Y, Kao C-H, Mui M-Y, Wang S-J. Leukocyte accumulation in acute hemispheric ischemic stroke. Stroke. 1993;24:236-240.
Wong MCW, Haley EC. Calcium antagonists: stroke therapy coming of age. Stroke. 1990;21:494-501.
Akopov SE. Calcium antagonists in cerebral ischemia: experimental basis and clinical experience. Drugs Today. 1994;30:171-210.
Brierly JB, Graham DI. Hypoxia and vascular disorders of the central nervous system. In: Adams J, Corsellis JAN, Duchen LW, eds. Greenfield's Neuropathology. Baltimore, Md: Edward Arnold Ltd; 1984:126-207.
Gelmers HJ, Hennerici M. Effect of nimodipine on acute ischemic stroke. Stroke. 1990;21(suppl IV):IV-81-IV-84.
Martinez-Vila E, Guillen F, Villanueva J, Matias-Guiu J, Bigorra J, Gil P, Carbonell A, Martinez-Lage J. Placebo-controlled trial of nimodipine in the treatment of acute ischemic cerebral infarction. Stroke. 1990;21:1023-1028.
Uno K, Imazeki K, Yoshikawa K, Arimizu N. Clinical use of technetium-99 m HMPAO labeled leukocyte for inflammatory imaging. J Nucl Med. 1987;28:648.
Laloux P, Richelle F, Jamart J, De Coster P, Laterre C. Comparative correlations of HMPAO SPECT indices, neurological score, and stroke subtypes with clinical outcome in acute carotid infarcts. Stroke. 1995;26:816-821.
Brott T, Marler JR, Olinger CP, Adams HP, Tomsick T, Barsan WG, Biller J, Eberle R, Hertzberg V, Walker M. Measurement of acute cerebral infarction: lesion size by computed tomography. Stroke. 1989;20:871-975.
Hair JF, Anderson RE, Tatham RL, Black WC. Multivariate Data Analysis. New York, NY: Macmillan Publishing Co; 1992.
Lucchesi BR. Complement activation, neutrophils, and oxygen radicals in reperfusion injury. Stroke. 1993;24(suppl I):I-41-I-49.
Lefer AM, Ma XL. Cytokines and growth factors in endothelial dysfunction. Crit Care Med. 1993;21(suppl 2):S9-S14.
Akopov S, Sercombe R, Seylaz J. Cerebrovascular reactivity: role of endothelium/platelet/leukocyte interactions. Cerebrovasc Brain Metab Reviews. 1996;8:11-94.
Carlos TM, Harlan JM. Leukocyte-endothelial adhesion molecules. Blood. 1994;84:2068-2101.
Strangaard S, Paulson OB. Pathophysiology of stroke. J Cardiovasc Pharmacol. 1990;15(suppl 1):S38-S42.