Reperfusion and Metabolic Recovery of Brain Tissue and Clinical Outcome After Ischemic Stroke and Thrombolytic Therapy
Background and Purpose—It is unclear from recent clinical trials whether thrombolytic agents are capable of facilitating reperfusion and metabolic recovery over time or whether a beneficial effect is counteracted by an increase in the risk of brain hemorrhage. We studied the effect of thrombolytic treatment on metabolic recovery after reperfusion and clinical outcome.
Methods—Patients were prospectively studied with 99mTc-ethyl cysteinate dimer single photon emission computed tomography (99mTc-ECD-SPECT) before treatment with recombinant tissue plasminogen activator (rTPA; 0.9 mg/kg IV; n=26) or placebo (n=26) 6 to 8 hours after treatment and at 7±1 days. Activity deficits were graded, compared between the treatment groups, and correlated with clinical outcome and the incidence of brain hemorrhage. Metabolic recovery of ischemic brain tissue was defined as a 25% decrease on the SPECT graded scale.
Results—Patients with metabolic recovery (n=28) had a better chance of being functionally unimpaired 3 months after stroke than patients without recovery (n=24) (OR 4.5, 95% CI 1.09 to 18.89) and had smaller infarcts on follow-up CT (36±38 versus 167±162 mL), regardless of whether metabolic recovery was observed within 6 to 8 hours of treatment or at 7 days. None of the 28 patients with metabolic recovery had a fatal parenchymal hemorrhage versus 5 of 24 patients without recovery (P=0.016). Treatment did not affect the incidence of brain tissue metabolic recovery.
Conclusions—Brain tissue metabolic recovery after ischemic stroke was associated with a beneficial effect on clinical outcome and was not facilitated by treatment with 0.9 mg of intravenous rTPA.
Controlled clinical trials often aim to discern whether a new drug affects predefined clinical end points, but they are seldom designed to study the rationale for the new treatment. The rationale of thrombolytic therapy for ischemic stroke is the recanalization of occluded arteries to increase cerebral perfusion and thus to reestablish normal brain metabolism and function by saving viable brain tissue before brain damage becomes irreversible. A successful thrombolytic recanalization may completely prevent ischemic tissue damage or reduce the size of infarction and ultimately improve the patient’s long-term clinical outcome. So far, only 1 study1 clearly showed a beneficial effect of recombinant tissue plasminogen activator (rTPA) treatment on clinical outcome. This study, like other clinical trials on thrombolysis,2 3 4 5 did not monitor arterial occlusion, recanalization, reperfusion, or metabolic recovery. Moreover, a beneficial effect of rTPA and streptokinase on clinical outcome in other controlled clinical trials was less impressive.2 3 4 5 6 It is therefore unclear whether the intravenous application of 0.9 mg/kg rTPA as done in clinical trials can reestablish brain tissue perfusion and metabolism after focal ischemia in time to be effective and is not counteracted by an increased risk of brain hemorrhage.
To date, the effect of thrombolysis on brain tissue perfusion has been studied with 99mTc-hexamethylpropyleneamine single photon emission computed tomography (99mTc-HMPAO-SPECT) in 5 small studies with inconsistent results.7 8 9 10 11 In contrast to 99mTc-HMPAO-SPECT, uptake of the tracer 99mTc-ethyl cysteinate dimer (99mTc-ECD) by the brain tissue depends on the capacity of cellular metabolism in addition to perfusion and can be regarded as a marker of nutritional blood flow.12 13 We performed 99mTc-ECD-SPECT scans to prospectively study the initial metabolic impairment of brain tissue and its recovery 6 to 8 hours after initiation of treatment and at day 7 after ischemic hemispheric stroke in a subgroup of patients randomized to rTPA or placebo from a single site in the Second European Cooperative Acute Stroke Study (ECASS II).6 We wanted to know whether and when treatment with 0.9 mg/kg rTPA IV facilitates recovery of ischemic brain tissue and whether this recovery is associated with an improved clinical course and outcome.
Subjects and Methods
Patients and Treatment
For a substudy of ECASS II,6 we consecutively included 52 patients (35 men, 17 women) who provided informed consent and who met the inclusion and exclusion criteria in this double-blind, randomized, placebo-controlled, multicenter, multinational phase III study between October 1996 and December 1997. Randomization was stratified by center. Treatment was initiated within 6 hours of symptom onset with 0.9 mg of rTPA/kg (maximum total dose given intravenously, 90 mg) or placebo. We treated all patients on the critical care unit of our department of neurology. The local ethics committee approved the study.
All end points were assessed blinded to treatment allocation. The clinical end point was functional outcome after 3 months (Barthel Index, Modified Rankin Scale) according to the ECASS protocol. The primary hemodynamic end point for this substudy was the reperfusion rate at 6 to 8 hours after treatment, defined as an increase in tracer uptake by ≥25% compared with the baseline image.
Clinical Examinations and Follow-Up
The Scandinavian Stroke Score and the National Institute of Health Stroke Score (NIHSS), including the score for distal motor function, were determined at the time of randomization, 24±2 hours after the start of treatment, after 7±1 days, after 30±2 days, and after 90±14 days. We defined clinical improvement within the first 24 hours as a decrease in the NIHSS by ≥4 points between the assessments at baseline and at 24±2 hours. After 90±14 days, the Modified Rankin Scale and the Barthel Index were determined.
Each patient underwent 3 CT examinations: before randomization, 22 to 36 hours after the start of treatment, and on day 7±1 after stroke onset. The CT scans were evaluated for the extent of parenchymal hypoattenuation due to acute ischemic edema, well-demarcated ischemic lesions, and cerebral hemorrhage by an investigator who was blinded to treatment allocation and the results of SPECT. We categorized the extent of parenchymal hypoattenuation due to acute ischemic edema on baseline CT as normal, hypoattenuation ≤33% of the middle cerebral artery (MCA) territory, and hypoattenuation >33% of the MCA territory. We measured the ischemic lesion with and without hemorrhagic transformation on follow-up CT using the formula for irregular volumes. The ECASS-I classification was used to distinguish between hemorrhagic infarction and parenchymal hematoma (PH).2 14
The 99mTc-ECD-SPECT studies were performed with a brain-dedicated SPECT camera (Ceraspect, DSI) with 3 rotating parallel hole collimators. Patients received 99mTc-ECD (400 MBq) before the start of therapy and were scanned after treatment, so that treatment was not delayed. The exact SPECT technique and SPECT analysis used were described in a previous study.15 A first follow-up SPECT study with 600 MBq of 99mTc-ECD was performed 6 to 8 hours after therapy, and a third SPECT examination (600 MBq of 99mTc-ECD) followed after 7±1 days.
For the semiquantitative region-of-interest (ROI) analysis of all 3 SPECT examinations, 5 transversal and 3 coronal slices were selected in each patient at predefined distances from the commissura anterior–commissura posterior (CA-CP) line (transversal slices: Talairach coordinates −20 mm, +1 mm, +8 mm, +21 mm, and +34 mm) and from the line perpendicular to the CA-CP line intersecting the commissura anterior (coronal slices: Talairach coordinates +5 mm, −16 mm, and −37 mm), respectively. In these 8 slices, 88 ROIs were generated with a commercial program (Ceraspect, DSI) and were assigned to anatomic structures by use of the stereotactic atlas.16 Count densities of ROIs of the symptomatic hemisphere were related to those of the corresponding contralateral regions and classified as abnormal if a deficit exceeded 10% (ratio ≤0.90), in agreement with widely accepted standards.17 We used the SPECT graded scale, a measure of the intensity and spatial extent of activity deficits.18 Each ROI was given a score of 0 to 9, where 0 indicated a ratio of ≥0.91, 1 indicated a ratio of 0.81 to 0.9 (corresponding to 81% to 90% activity compared with the contralateral side), 2 indicated a ratio of 0.71 to 0.8, etc. The scores for all individual ROIs were summed to produce the SPECT graded scale, extending from 0 to 792. Early (within 6 to 8 hours of treatment) and late (within 7±1 days of treatment) metabolic recovery was prospectively defined as a decrease in the SPECT graded scale >25% between baseline and the first or second control (Figure 1⇓).9 11 We did not study metabolic recovery in patients with a baseline SPECT graded scale <15.
We analyzed the SPECT scans blinded to treatment allocation. For 3 patients who died between the second and third SPECT studies, the SPECT graded scale of the second SPECT study was carried forward for the data analysis.
Numerical variables are presented as means with standard deviations. Clinical data, SPECT, and CT findings between groups were compared with the U test developed by Whitney and Mann for nonparametric variables, the Student’s t test (parametric variables) for unpaired data, and the χ2 test or Fisher’s exact test (proportions). We accepted 0.05 as a level of significance.
Findings at Baseline
Fifty-two patients (16 women, 36 men, mean age 61±12 years) were included in the study. Twenty-six patients each were allocated to rTPA or placebo and treated 115 to 355 minutes after the onset of symptoms. We did not detect significant differences between treatment groups regarding vascular risk factors, the time period between the onset of symptoms and the start of treatment, severity of neurological deficit at the time of randomization, or findings on baseline CT (Table 1⇓). One patient with parenchymal hypoattenuation of >33% of the MCA territory on CT was allocated in the rTPA group. According to SPECT, all patients had a perfusion deficit that varied between 4 and 238 on the SPECT graded scale (extremes, 0 to 792). The perfusion deficit was relatively mild in 8 patients with a SPECT score <15. The SPECT score correlated positively with stroke severity as assessed with NIHSS (r=0.33, P=0.0158, Figure 2⇓) at baseline. Patients in the rTPA group tended to have a more severe initial perfusion deficit. The mean SPECT score, however, was not statistically different between the placebo- and rTPA-treated patients (P=0.375, Mann-Whitney U test). The mean SPECT score was 47±53 in patients with normal CT, 83±52 in patients with hypoattenuation ≤33% of the MCA territory, and 148 in 1 patient with hypoattenuation >33% of the MCA territory (P=0.007, Kruskal-Wallis test). Compared with the entire ECASS II population, this study subgroup was somewhat younger, showed similarly severe neurological deficits, and less frequently had a normal CT. More patients were treated within 3 hours after the onset of symptoms.
Follow-Up and Outcome
We did not detect any statistically significant differences between the 2 treatment groups with regard to follow-up SPECT, metabolic recovery after reperfusion, follow-up CT, clinical improvement within 24 hours, or clinical outcome at 90 days (Table 2⇓). The relative mean improvement on the NIHSS between baseline and day 90 was 0.5±15 in the placebo group and 3.2±17 in the rTPA group. SPECT showed a mean decrease by 8±29 points in placebo-treated patients and by 8±28 points in rTPA-treated patients (P=0.942) 6 to 8 hours after treatment initiation. The decrease was 22±30 and 18±38 points, respectively, at 7 days after stroke onset. Five patients had a PH within 24 hours and died within the first week of stroke. These 5 patients had high SPECT scores (between 80 and 238) at baseline (Figure 2⇑). The mean baseline SPECT score of patients who developed PH and died (166±59) was higher than the mean SPECT score of patients with hemorrhagic infarction (86±61, P=0.006) and higher than the mean SPECT score of patients without secondary intracranial hemorrhage (58±41, P<0.0001).
The follow-up CT after 7±1 days did not detect an infarct in 6 patients (12%); 15 patients (29%) had an infarct ≤33% of the MCA territory, 24 (46%) had an infarct >33% of the MCA territory, 4 (8%) had an isolated infarct of the anterior or posterior territory, and 3 (6%) had a brain stem infarction. Patients with brain stem infarctions had SPECT scores at baseline and follow-up that did not exceed 12. The SPECT scores of patients without visible infarctions on CT varied between 7 and 29 (median 10.5) on day 7. Twenty-nine patients (56%) had a hemorrhagic transformation, among them 5 patients with fatal PH. Four of these 5 patients were treated with rTPA (Table 2⇑).
Comparison of Patients With and Without Metabolic Recovery After Reperfusion
According to our definition of metabolic recovery after reperfusion as assessed by 99mTc-ECD-SPECT, the patients were grouped in 3 categories: 18 patients (35%) with early metabolic recovery within 6 to 8 hours of treatment, 10 (19%) with late recovery not detected within 6 to 8 hours of treatment initiation but at 7±1 days after stroke, and 16 (31%) without recovery within the observation period of 7±1 days (Figure 1⇑). For this analysis, we excluded all 8 patients with SPECT scores <15 at baseline, among them the 3 patients with brain stem infarcts. We compared these 3 groups regarding their baseline characteristics, CT findings, clinical improvement, and long-term outcome in a post hoc analysis.
At baseline, patients without subsequent metabolic recovery had a significantly higher mean SPECT score than patients with early or late recovery and tended to have a higher incidence of ischemic edema on baseline CT and a worse neurological score. These 3 groups did not differ with regard to the interval between symptom onset and treatment, sex, or age (Table 3⇓).
Within 6 to 8 hours of treatment (up to 14 hours after stroke onset), the SPECT score decreased significantly in patients with early recovery compared with patients with late or no recovery (P=0.0002). At 7 days, the SPECT score was unchanged in patients with no recovery in contrast to patients with early or late recovery (Figure 3⇓).
Patients with metabolic recovery after reperfusion according to SPECT had a remarkably better clinical course and outcome at 3 months after stroke regardless of whether recovery was observed within 6 to 8 hours of treatment or even later before day 7 (Figure 4⇓). They showed a continuous improvement of their mean stroke scores at all control examinations in contrast to patients without metabolic recovery, who continuously deteriorated so that the stroke severity was significantly different at all follow-up examinations (Table 4⇓). Mortality was 50% (8/16) without metabolic recovery; 5 patients died of intracerebral hemorrhage within 1 week of stroke onset. Among patients with metabolic recovery, only 1 died (15 days after stroke, of a noncerebral cause).
The mean volume of ischemic lesions was significantly smaller in patients with early or delayed metabolic recovery on CT at 22 to 36 hours and 7±1 days after treatment.
We studied a subpopulation of ECASS II with repeated 99mTc-ECD-SPECT measurements before and after treatment with rTPA or placebo. Although we did not find significant differences in the baseline characteristics between this subpopulation and the entire ECASS II population, our study population may not represent the ECASS population because it is too small (6.5% of the entire ECASS population) and was selected from a single site. When we designed the study simultaneously with ECASS II, we did not foresee that the effect of 0.9 mg/kg rTPA IV on the proportion of patients with very good clinical outcome (Rankin 0 and 1) would be too small to be proven in 800 patients. Our substudy clearly was not powered to detect minor effects of rTPA on clinical outcome. It would be of great interest, however, to determine whether an effect of rTPA on nutritional blood flow with metabolic recovery was too small, too late, or counteracted by secondary hemorrhage to become relevant for functional recovery and clinical outcome. Because of the time constraints of the study design, we could not directly assess arterial obstruction and recanalization. It is more meaningful, however, to determine the extent of hypoperfusion and metabolic impairment, which may vary even among patients with identical arterial occlusion sites because of different collateral blood flow capacities and perfusion pressures.
We preferred 99mTc-ECD to 99mTc-HMPAO for studying whether rTPA can enhance nutritional blood flow and improve brain metabolism. It is assumed that the lipophilic complex of ECD is intracellularly oxidized to a polar monoacid-monoester complex in normal brain tissue.19 These polar metabolites cannot diffuse back through the blood-brain barrier, and >70% of the metabolite remains in the cytosolic fraction if the metabolism is undisturbed.20 Thus far, neither the corresponding enzyme nor the follow-up products have been identified.21 A reduction of 99mTc-ECD uptake by ischemic or infarcted brain areas could therefore be caused by various mechanisms: impaired cerebral blood flow, diminished tracer extraction from the arterial blood, and a reduced brain metabolism.22 Low local radioactivity on 99mTc-ECD-SPECT can thus be interpreted as diminished cerebral blood flow with (still) intact brain metabolism, a disturbance of both perfusion and metabolism, or metabolic disturbance with normal blood flow or even hyperperfusion, eg, postischemic luxury perfusion.23 An increase of a former diminished tracer uptake requires both improvement of cerebral blood flow and metabolic recovery.
Our observations support the notion that 99mTc-ECD uptake is in fact an indicator for nutritional cerebral blood flow. The SPECT graded scale reflected cerebral pathophysiology: it correlated significantly with neurological impairment during the clinical course and with the extent of ischemic edema at baseline, although it did not differentiate between small areas of severe metabolic impairment and more widespread areas of mild impairment.
By applying a criterion for metabolic recovery after reperfusion and by excluding 8 patients with baseline scores <15 (among them, 3 patients with brain stem infarctions), we identified 3 groups with a different pattern of metabolic recovery: 18 patients with “early” recovery within 6 to 8 hours of treatment, 10 patients with “late” recovery not detected at 6 to 8 hours after treatment initiation but at 7 days, and 16 patients without recovery within our observation period. The 28 patients with metabolic recovery developed smaller ischemic lesions, improved clinically, and had a beneficial clinical course regardless of whether metabolic recovery was observed early or late, in contrast to patients without metabolic recovery. It appeared as though patients with higher SPECT scores and larger ischemic edema on CT at baseline had smaller chances for metabolic recovery. Patients who later developed fatal PH had high SPECT scores at baseline. In agreement with previous observations,24 25 26 27 28 this observation suggests that a severe perfusion deficit is associated with poor prognosis and may cause secondary hemorrhage.
All fatal brain hemorrhages and, with 1 exception, all deaths occurred in patients without metabolic recovery. Imaging with 99mTc-ECD-SPECT, however, did not allow us to determine whether persistent hypoperfusion had caused hemorrhage or edema and prevented metabolic recovery or whether reperfusion itself was detrimental. Studies of reperfusion with angiography, positron emission tomography, and 99mTc-HMPAO-SPECT have shown that even delayed reperfusion is associated with smaller infarctions and better clinical outcome and do not support the concept of reperfusion injury.10 29 30 31 32
In contrast to observations with 99mTc-HMPAO-SPECT,10 treatment with rTPA did not affect either the SPECT graded scale on follow-up scans or the clinical course. Using 99mTc-ECD-SPECT, we may have missed nonnutritional reperfusion into areas with irreversible brain tissue damage and impairment of tracer uptake. Four of 5 patients with fatal brain hemorrhage received rTPA. We therefore cannot exclude the possibility that rTPA caused PH by facilitating reperfusion into irreversibly damaged brain tissue. Given the insignificant effect of rTPA on clinical outcome (Rankin 0 and 1) in the entire study,6 it is not surprising that we did not find a beneficial effect of rTPA in this small population. The rTPA-treated patients of our substudy had a slightly more severe perfusion deficit at baseline. This and a lack of effect on nutritional blood flow may be responsible for the missed beneficial clinical effect of rTPA in this subgroup and probably also in ECASS 2.
In summary, we observed that metabolic recovery of brain tissue after reperfusion, even if observed later than 8 hours but within 7 days after symptom onset, is associated with significantly better long-term clinical outcome. Treatment with rTPA did not facilitate metabolic recovery in this small ECASS-2 subpopulation. The effect of 0.9 mg/kg rTPA IV was either too small to be detected or was counteracted by brain hemorrhage that prevented metabolic recovery after reperfusion.
This study was supported by Boehringer Ingelheim, Germany.
- Received February 7, 2000.
- Revision received April 14, 2000.
- Accepted April 21, 2000.
- Copyright © 2000 by American Heart Association
Hacke W, Kaste M, Fieschi C, Toni D, Lesaffre E, von Kummer R, Boysen G, Bluhmki E, Höxter G, Mahagne M, Hennerici M. Intravenous thrombolysis with recombinant tissue plasminogen activator for acute hemispheric stroke: the European Cooperative Acute Stroke Study (ECASS). JAMA. 1995;274:1017–1025.
Hacke W, Kaste M, Fieschi C, von Kummer R, Davalos A, Meier D, Larrue V, Bluhmki E, Davis S, Donnan G, Schneider D, Diez-Tejedor E, Trouillas P. Randomised double-blind placebo-controlled trial of thrombolytic therapy with intravenous alteplase in acute ischaemic stroke (ECASS II). Lancet. 1998;352:1245–1251.
Overgaard K, Sperling B, Boysen G, Pedersen H, Gam J, Ellemann K, Karle A, Arlien-Soborg P, Olsen T, Videbaeck C, Knudsen J. Thrombolytic therapy in acute ischemic stroke: a Danish pilot study. Stroke. 1993;24:1439–1446.
Baird A, Donnan G, Austin M, Fitt G, Davis S, McKay W. Reperfusion after thrombolytic therapy in ischemic stroke measured by single-photon emission computed tomography. Stroke. 1994;25:79–85.
Grotta J, Alexandrov A. tPA-associated reperfusion after acute stroke demonstrated by SPECT. Stroke. 1998;29:429–432.
Yasaka M, O’Keefe G, Chambers B, Davis S, Infeld B, O’Malley H, Baird A, Hirano T, Donnan G. Streptokinase in acute stroke: effect on reperfusion and recanalization: Australian Streptokinase Trial Study Group. Neurology. 1998;50:626–632.
Ishizu K, Yonekura Y, Magata Y, Okazawa H, Fukuyama H, Tanaka F, Hattori N, Kitano H, Fujita T, Tamaki N, Konishi J. Extraction and retention of technetium-99m-ECD in human brain: dynamic SPECT and oxygen-15-water PET studies. J Nucl Med. 1996;37:1600–1604.
Berrouschot J, Barthel H, von Kummer R, Knapp W, Hesse S, Schneider D. 99mTechnetium-ethyl-cysteinate-dimer single-photon emission CT can predict fatal ischemic brain edema. Stroke. 1998;29:2556–2562.
Talairach J, Tournoux P. Co-Planar Stereotactic Atlas of the Human Brain. New York, NY: Thieme Verlag; 1988.
Podreka I, Suess E, Goldenberg G, Steiner M, Brucke T, Muller C, Lang W, Neirinckx R, Deecke L. Initial experience with technetium-99m HMPAO brain SPECT. J Nucl Med. 1987;28:1657–1666.
Hanson SK, Grotta JC, Rhoades H, Tran HD, Lamki LM, Barron BJ, Taylor WJ. Value of single-photon emission-computed tomography in acute stroke therapeutic trials. Stroke. 1993;24:1322–1329.
Léveillé J, Demonceau G, De Roo M, Rigo P, Taillefer R, Morgan R, Kupranick D, Walovitch R. Characterization of technetium-99m-L, L-ECD for brain perfusion imaging, part 2: biodistribution and brain imaging in humans. J Nucl Med. 1989;30:1902–1910.
Walovitch R, Hill T, Garrity S, Cheesman E, Burgess B, O’Leary D, Watson A, Ganey M, Morgan R, Williams S. Characterization of technetium-99m-L, L-ECD for brain perfusion imaging, part 1: pharmacology of technetium-99m ECD in nonhuman primates. J Nucl Med. 1989;30:1892–1901.
Léveillé J, Demonceau G, Walovitch R. Intrasubject comparison between technetium-99m-ECD and technetium-99m-HMPAO in healthy human subjects. J Nucl Med. 1992;33:480–484.
Lassen N, Sperling B. 99mTc-bicisate reliably images CBF in chronic brain diseases but fails to show reflow hyperemia in subacute stroke: report of a multicenter trial of 105 cases comparing 133Xe and 99mTc-bicisate (ECD, neurolite) measured by SPECT on same day. J Cereb Blood Flow Metab. 1994;14(suppl 1):S44–S48.
Giubilei F, Lenzi G, Di Piero V, Poizzilli C, Pantano P, Bastianello S, Argentino C, Fieschi C. Predictive value of brain perfusion single-photon emission computed tomography in acute ischemic stroke. Stroke. 1990;21:895–900.
Bozzao L, Angeloni U, Bastianello S, Fantozzi L, Pierallini A, Fieschi C. Early angiographic and CT findings in patients with hemorrhagic infarction in the distribution of the middle cerebral artery. AJNR Am J Neuroradiol. 1992;12:1115–1121.
del Zoppo G, von Kummer R, Hamann G. Ischemic damage of brain microvessels: inherent risks for thrombolytic treatment in stroke. J Neurol Neurochir Psychiatr. 1998;65:1–9.
Ogata J, Yutani C, Imakita M, Ishibashi-Ueda H, Saku Y, Minematsu K, Sawada T, Yamaguchi T. Hemorrhagic infarct of the brain without a reopening of the occluded arteries in cardioembolic stroke. Stroke. 1989;20:876–883.
von Kummer R, Holle R, Rosin L, Forsting M, Hacke W. Does arterial recanalization improve outcome in carotid territory stroke? Stroke. 1995;26:581–587.
Barber P, Davis S, Infeld B, Baird A, Donnan G, Jolley D, Lichtenstein M. Spontaneous reperfusion after ischemic stroke is associated with improved outcome. Stroke. 1998;29:2522–2528.
Marchal G, Furlan M, Beaudouin V, Rioux P, Hauttement J, Serrati C, de la Sayette V, Le Doze F, Viader F, Derlon J, Baron J. Early spontaneous hyperperfusion after stroke: a marker of favourable tissue outcome? Brain. 1996;119:409–419.