The Impact of Recanalization on Ischemic Stroke Outcome
Background and Purpose— For a biomarker to serve as an auxiliary or surrogate outcome measure, it must be tightly correlated with and causally related to functional clinical outcome. Vessel recanalization is a potential surrogate outcome marker for functional outcome in trials of thrombolytic and mechanical recanalization therapies in acute stroke, but the correlation of recanalization and clinical outcome has not been previously systematically reviewed.
Methods— Through Medline search, we identified and abstracted recanalization and outcome data from all articles published between 1985 and 2002 that assessed vessel recanalization, either spontaneous or therapeutically induced, in acute ischemic stroke.
Results— Fifty-three studies encompassing 2066 patients reported recanalization rates. Recanalization rates categorized according to intervention were: spontaneous (24.1%), intravenous fibrinolytic (46.2%), intra-arterial fibrinolytic (63.2%), combined intravenous–intra-arterial (67.5%), and mechanical (83.6%). Clinical outcome data categorized by success or failure in achieving recanalization was available from 33 articles encompassing 998 patients. Good functional outcomes at 3 months were more frequent in recanalized versus nonrecanalized patients with odds ratio of 4.43 (95% CI, 3.32 to 5.91). Three-month mortality was reduced in recanalized patients (odds ratio, 0.24; 95% CI, 0.16 to 0.35). Rates of symptomatic hemorrhagic transformation did not differ between the 2 groups (odds ratio, 1.11; 95% CI, 0.71 to 1.74).
Conclusions— Formal meta-analysis confirms a strong correlation between recanalization and outcome in acute ischemic stroke. Recanalization is strongly associated with improved functional outcomes and reduced mortality. These findings suggest that recanalization is an appropriate biomarker of therapeutic activity in early phase trials of thrombolytic treatment in acute ischemic stroke.
Thrombolytic stroke therapy is based on the “recanalization hypothesis,” ie, that reopening of occluded vessels improves clinical outcome in acute ischemic stroke through regional reperfusion and salvage of threatened tissues. Though usually taken for granted, the recanalization hypothesis in stroke has sometimes been challenged.1 Several biologic factors may weaken the relationship of recanalization to outcome in acute ischemic stroke patients. Recanalization of large arteries may not yield effective tissue reperfusion because of persisting distal emboli and microcirculatory occlusions (no-reflow phenomenon). Recanalization may occur too late to benefit ischemic tissues. Adequate collateral circulation may have protected tissues even without recanalization. Recanalization may exacerbate tissue injury by promoting reperfusion injury, excessive cerebral edema, and hemorrhagic transformation. Some case series have failed to identify a correlation between recanalization and outcome (see references 44 and 47 of supplemental Table I⇓, available at http://stroke.ahajournals.org).
Confirming the recanalization hypothesis and characterizing the strength of the association of recanalization with clinical outcomes in acute ischemic stroke is particularly important now that clinical trialists and regulatory agencies are increasingly adopting recanalization as a biomarker of treatment activity in human clinical trials. Rates of recanalization are now routinely used as efficacy measures in phase 2 trials of thrombolytic agents, for example, to identify the most promising dose regimens to advance to phase 3 pivotal studies.2 Most recently, the Food and Drug Administration used the efficacy end point of recanalization rate as the basis of labeled approval of the MERCI Retriever mechanical thrombectomy for use in patients experiencing ischemic stroke.3 Under the Food and Drug Administration’s 510K pathway, new mechanical thrombectomy devices can now similarly gain approval on the basis of recanalization, not clinical, efficacy, demonstrating similar or better recanalization rates than the approved MERCI Retriever device.4 For a biomarker to serve as an auxiliary or surrogate outcome measure, it must be not only causally related to but also tightly correlated with functional clinical outcome.5 Consequently, establishing the strength of the association between recanalization and clinical outcomes is critically important.
Because the recanalization hypothesis has not previously been systematically assessed in acute ischemic stroke, we undertook a formal meta-analytic study. The meta-analysis was designed to address several issues: (1) the rates of recanalization that occur with active therapy versus spontaneously, and variations in these rates associated with different recanalization interventions, different target vessels, and different stroke subtypes; (2) the association of recanalization with the clinical outcomes of mortality and functional independence, and influences on these associations of the timing and degree of recanalization; and (3) the association of recanalization with the development of hemorrhagic transformation.
Materials and Methods
Medline was searched to identify all published reports addressing recanalization and clinical outcome in acute ischemic stroke published through the end of 2002. Search key words included recanalization, reperfusion, reopening, thrombolysis, or thrombolytic therapy, combined with ischemic stroke, cerebral ischemia, or cerebrovascular disorder. Published conference proceedings (Thrombolytic Therapy in Acute Ischemic Stroke I & II, 1991 and 1998, respectively) were also searched. The inclusion criterion for the analysis of recanalization rates by treatment modality was that a study reported recanalization rates separately for each treatment modality. The inclusion criterion for the analysis of the association of recanalization with clinical outcome was that a study reported clinical outcome separately in patients experiencing and not experiencing recanalization. Exclusion criteria were: (1) single case report or small case series with <5 patients; (2) study reports recanalized cases only or nonrecanalized cases only; and (3) assessment of recanalization performed >24 hours after symptom onset. When multiple articles drew on the same datasets, data were abstracted only once from the most comprehensive available report. Case series, case-controlled studies, and randomized controlled trial were all included. After one of the authors (J.-H.R.) searched and selected the relevant articles, the other author (J.L.S.) reviewed and confirmed the inclusion of the study. When further data were needed, personal communication to the corresponding author was attempted.
For timing analyses, timing of assessment of recanalization was dichotomized into early (≤6 hours after symptom onset) and late (6 to 24 hours). Sometimes the actual recanalization time was not equal to the assessment time of the recanalization (such as intravenous [IV] thrombolysis and later assessment of recanalization). In many intra-arterial (IA) fibrinolysis series, only the time to start of fibrinolytic therapy was presented, with no data presented regarding the exact time of recanalization. In these cases, if recanalization was achieved, it was presumed that recanalization occurred within 2 hours after procedure start. To rate degree of recanalization, the Thrombolysis in Myocardial Ischemia scale was used for the angiographic studies,6 and the Thrombolysis in Brain Ischemia scale for transcranial Doppler studies.7 Recanalization was considered achieved when Thrombolysis in Myocardial Ischemia scale grade was 2 (partial) or 3 (complete), and when Thrombolysis in Brain Ischemia scale grade was 2 to 3 (partial) or 4 to 5 (complete). For vessel location analyses, vessel sites were classified into 3 groups: the middle cerebral artery and anterior cerebral artery, the internal carotid artery, and the vertebral and basilar arteries.
Good outcome was defined as functional independence, demonstrated by modified Rankin score ≤2. When no modified Rankin score assessment was recorded in an article, the most comprehensive measure of global outcome reported in the study was used instead. When the authors of the study had their own definition of functional independence, that definition was just followed (in 14 studies). When the National Institute of Health Stroke Scale was the most comprehensive measure (in 3 studies), a score ≤4 at subacute stage defined good outcome. When available, outcome status 3 months poststroke was used. If assessments at 3 months were not performed, the reported final assessment closest in time to 3 months was used, ranging from status at discharge to status 6 months after onset.
For symptomatic intracerebral hemorrhage, cases rated as symptomatic intracerebral hemorrhage, according to whatever definition the authors used, were considered to have symptomatic intracerebral hemorrhage. If no rating of symptomatic intracerebral hemorrhage was advanced, but radiologic classification of hemorrhages was performed using published criteria,8 parenchymal hematoma was considered to be equivalent to symptomatic hemorrhagic transformation, whereas hemorrhagic infarction was considered to be asymptomatic.
For the comparison of recanalization and hemorrhage rate, Pearson χ2 test was used. For the meta-analysis of good functional outcome and mortality at 3 months, forest plots and statistical analysis including Peto odds ratios with 95% CIs based on the fixed model were performed using Review Manager 4.1. Statistical tests of heterogeneity were performed by Mantel-Haenszel method and revealed no definite heterogeneity. The presence of publication bias was interrogated by funnel plot and showed no evidence of bias.
Initial Medline search identified 72 articles published between 1985 and 2002 reporting some information about recanalization in acute cerebral ischemia. Among these, 19 were excluded for the following reasons: data overlapped with data in other articles (5), recanalization assessment took place >24 hours after onset (5), data were reported on successfully recanalized cases only (5), and insufficient data or case number (4). The remaining 53 studied 2066 patients and reported recanalization information on 1774 patients. Study features are detailed in the Table⇓ (and full citations for the studies are available in supplemental Table I). Forty-four of the studies were case series, 3 were case-control studies, and 6 were randomized control trials. Recanalization was assessed by catheter angiography in 46 studies. The other 7 studies used transcranial Doppler, magnetic resonance angiography, and single proton emission computed tomography, alone or in combination. Recanalization interventions studied were: intra-arterial fibrinolysis in 30 reports, intravenous fibrinolysis in 14, spontaneous recanalization in 8, endovascular mechanical therapies in 7, and combined IV/IA fibrinolysis in 7. In 13 studies, >1 treatment method was used.
Intervention-Specific Recanalization Rates
Overall, recanalization (partial or complete) was achieved in 977 of 1774 patients (55.1%) in whom an active revascularization treatment method was specified. In contrast, the overall spontaneous recanalization rate within 24 hours was 24.1% (61/253).
The recanalization rate with IV fibrinolysis was 46.2% (219/474) overall. One group reported a much higher rate (66.2%, 43/65; P<0.01) when IV fibrinolysis was continuously monitored by transcranial Doppler (reference 37 of supplemental Table I). IA fibrinolysis showed a recanalization rate of 63.2% (549/868), which was substantially higher than the IV fibrinolysis rate (P<0.01). The recanalization rate of combined IV/IA thrombolysis did not significantly differ from IA alone (67.5%, 56/83; P=0.45). The highest recanalization rate was observed with mechanical thrombolysis, 83.6% (92/110; P<0.01, compared with IA fibrinolysis) (Figure 1).
Information regarding recanalization rates in specific target vessels was available in 37 studies and 1054 patients. The recanalization rate across all vessels for this subset of patients was 58.9% (621/1054). For middle cerebral artery/anterior cerebral artery vessels, the recanalization rate was 61% (spontaneous 12/54, IV 99/181, IA 259/389, combined IV/IA 24/36, mechanical 40/51, total 434/711). For vertebral and basilar arteries vessels, it was 66.2% (IV 4/5, IA 73/115, combined IV/IA 2/3, mechanical 7/7, total 86/130). The internal carotid artery showed the lowest recanalization rate, at 49.8% (IV 5/36, IA 59/121, combined IV/IA 23/38, mechanical 14/18, total 101/213; P<0.01, versus all others).
Limited data were available for the recanalization according to stroke etiologic subtype. Only for cardioembolic stroke mechanism was even modest data available, and the reported recanalization rate was 61.8% (55/89), but a comparison could not be made because other studies did not specify the stroke subtype.
Clinical outcome data according to occurrence or nonoccurrence of recanalization was available in 33 articles for 998 patients (Figure 2). Good outcome was achieved in 58.1% of recanalized patients versus 24.8% of nonrecanalized patients (odds ratio, 4.43; 95% CI, 3.32 to 5.91). In the subgroup of patients in whom occurrence of recanalization was assessed within 6 hours of onset, good outcome was achieved in 50.9% of recanalized patients versus 11.1% of nonrecanalized patients, (odds ratio, 6.36; 95% CI, 3.32 to 12.17). The frequency of fatal outcomes was 14.4% in recanalized patients versus 41.6% in nonrecanalized patients (odds ratio, 0.24; 95% CI, 0.16 to 0.35; Figure 3). In the subgroup of recanalization assessment within 6 hours, fatal outcomes occurred in 12.1% of recanalized patients versus 41.1% of nonrecanalized patients (odds ratio, 0.22; 95% CI, 0.10 to 0.51).
Information regarding symptomatic hemorrhage linked to recanalization was available in 24 studies for 678 patients. The hemorrhagic transformation rate was 13.7% (49/358) for recanalized versus 12.5% (40/320) for nonrecanalized patients, showing no statistically significant difference (P=0.65).
The goal of recanalization therapy in acute ischemic stroke is to improve clinical outcome by restoring anterograde perfusion and salvaging ischemic brain. This systematic review of the literature confirms the wisdom of this strategy, demonstrating that occurrence of early vessel recanalization is linked powerfully to final clinical outcome in acute ischemic stroke. The occurrence of recanalization is associated with a 4- to 5-fold increase in the odds of good final functional outcome and a 4- to 5-fold reduction in the odds of death. The recanalization hypothesis is confirmed. Despite the theoretical limitations and risks of recanalization, rapid restoration of anterograde perfusion and salvage of ischemic brain generally confers substantial benefit.
Across the series studied in this analysis, rates of spontaneous recanalization by 24 hours were substantial, occurring in one-quarter of patients. Other studies have shown that after 24 hours, recanalization continues to occur, appearing in 52.7% (33.3% to 69.6%) of cases up to 1 week after stroke.9–11 Late spontaneous recanalization is the most common natural outcome of acute cerebral vascular occlusions. However, only early recanalization while substantial penumbral tissues still persist to be salvaged has a substantial impact on outcome. Most studies we reviewed assessed recanalization at only one time point, often somewhat variable, limiting meta-analytic exploration of the critical time window for recanalization. However, recanalization documented within 6 hours of onset tended to be even more strongly associated with good clinical outcomes than recanalization documented within 24 hours of onset. This finding accords with studies performed with continuous transcranial Doppler monitoring, which suggest that recanalization up to 6 hours after onset will confer benefit in most patients and studies performed with multimodal magnetic resonance or computed tomography imaging indicating that select patients will still harbor substantial residual penumbra beyond 6 hours and benefit from reperfusion.12
The recanalization rate with IV fibrinolytic therapy was approximately twice the spontaneous rate, but still less than half. This recanalization rate correlates well with the observation that IV fibrinolysis confers clinical benefit in approximately one-third of treated patients.13 Techniques to enhance this modest recanalization rate are urgently needed to improve intravenous fibrinolytic therapy. Enhanced recanalization has been observed with sonothrombolysis (reference 37 of supplemental Table I),14 and multiple strategies of combination pharmacological therapy (eg, fibrinolytic plus a glycoprotein IIb/IIIa antagonist) are undergoing investigation.15 A substantially higher rate of recanalization occurred in IA fibrinolysis series, with reperfusion achieved in nearly two-thirds of cases. The rate of recanalization with combined IV/IA fibrinolysis was similar to that of IA alone, suggesting that it is the IA component of combined therapy that determines the final recanalization rate, although the IV component may increase the speed at which recanalization is achieved in cases re-opening in response to initial intravenous agents.
The predominant mechanical therapy used in the series we analyzed was intracranial angioplasty in Asian populations in whom intracranial atherosclerosis with supervening in situ thrombosis is a common stroke mechanism. The high recanalization rate achieved, with vessel reopening in 4 of 5 cases, suggests that primary angioplasty is a very promising strategy for intracranial atherothrombosis.
It has often been suggested that recanalization might have a deleterious influence on the hemorrhagic conversion of the infarct, but the pooled data failed to demonstrate an increased rate of hemorrhagic transformation in patients experiencing recanalization. This finding suggests that the severity of the ischemic injury to the blood–brain barrier is an equal or greater determinant of hemorrhagic transformation than whether reperfusion occurs under high orthograde pressure through a recanalized vessel or under less retrograde pressure through collateral vessels.16,17 If this is true, then recanalization might prevent hemorrhagic complications in patients when vessel re-opening occurs very early, before advanced end-organ injury, while predisposing to hemorrhagic complications when vessel re-opening occurs late, after loss of vascular integrity.
Recanalization rates differed by vessel site, with more frequent vessel re-opening in middle cerebral artery–anterior cerebral artery and vertebral and basilar arteries than in the intracranial internal carotid artery. Clot burden is likely a major determinant of this site-specific difference in vessel recanalization rates, with the volume of thrombus to be digested by intrinsic or therapeutically administered fibrinolytic agents much larger in the intracranial carotid artery than in the middle cerebral artery–anterior cerebral artery and posterior circulation vessels.
This study has several limitations. We sought data from abstracts and from direct communication with investigators, as well as from full-length published articles; nonetheless, the results are subject to publication bias. Investigators may have reported recanalization and clinical outcome associations when they accorded with expectations and failed to report unexpected results. Because this investigation is a meta-analysis of study-level data, rather than a pooled analysis of individual patient-level data, we could not adjust for baseline imbalances in characteristics between recanalized and nonrecanalized patients. However, the benefits of recanalization therapy in randomized trials suggest that recanalization itself produces good clinical outcomes, rather than simply being a marker of patients bound to have good clinical outcomes. Some studies did not use the modified Rankin scale or other standard measure of disability, so we had to use the most comprehensive global measure of outcome available in lieu of the modified Rankin score.
In general in clinical trials, surrogate end points should be simple to ascertain, easily quantifiable, inexpensive, linked to the biological mechanism of the therapy, and tightly correlated with clinical outcome.5 Vessel recanalization meets these criteria for surrogacy. For endovascular intervention trials in which catheterization are already a part of the procedure, recanalization is indeed simple to ascertain, easily quantifiable, and inexpensive. With transcranial Doppler, computed tomographic angiography, and magnetic resonance angiography advances, recanalization can also increasingly easily be determined in trials of intravenous therapy. The findings of this meta-analysis now demonstrate the requisite close correlation between early recanalization and late clinical outcomes of direct relevance to the patient, with recanalization within 6 hours of onset increasing the odds of a nondisabled outcome 6-fold. These results lend strong support to the use of restoration of vessel patency as a surrogate end point in phase 2 trials of pharmacological recanalization agents and in trials comparing novel to existing predicate recanalization devices in acute ischemic stroke.
The authors thank Dr Andrei Alexandrov and Dr Osama Zaidat for sharing unpublished data for this analysis.
Sources of Funding
This study was supported in part by NIH-NINDS Award P50 NS044378.
J.H.R. is a practicing vascular neurologist who administers recanalization therapies. J.L.S. has been an unfunded site subinvestigator in multicenter and single-center NIH trials in which recanalization devices or drugs were provided at no cost by Concentric Medical, Ekos, Genentech, and Talacris; is Director of an NIH Program Project which included one recanalization subproject for which devices were provided at no cost by Concentric Medical; has been an unfunded site subinvestigator in multicenter recanalization trials sponsored by Eli Lilly and Neurobiological Technologies, Inc; has served on Scientific advisory boards for Boehringer Ingelheim (secondary prevention) and Nuvelo; has served on a Speaker’s Bureau for Boehringer Ingelheim (secondary prevention); serves as Director of the UCLA Stroke Center, which provides recanalization therapies; is a practicing vascular neurologist who administers recanalization therapies.
- Received September 26, 2006.
- Accepted October 20, 2006.
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.
Hacke W, Albers G, Al-Rawi Y, Bogousslavsky J, Davalos A, Eliasziw M, Fischer M, Furlan A, Kaste M, Lees KR, Soehngen M, Warach S; The DIAS Study Group. The Desmoteplase in Acute Ischemic Stroke Trial (DIAS): a phase II MRI-based 9-hour window acute stroke thrombolysis trial with intravenous desmoteplase. Stroke. 2005; 36: 66–73.
Smith WS, Sung G, Starkman S, Saver JL, Kidwell CS, Gobin YP, Lutsep HL, Nesbit GM, Grobelny T, Rymer MM, Silverman IE, Higashida RT, Budzik RF, Marks MP; The MERCI Trial Investigators. Safety and efficacy of mechanical embolectomy in acute ischemic stroke: results of the MERCI trial. Stroke. 2005; 36: 1432–1438.
Felten RP, Ogden NR, Pena C, Provost MC, Scholsser MJ, Witten CM. The Food and Drug Administration medical device review process: clearance of a clot retriever for use in ischemic stroke. Stroke. 2005; 36: 404–406.
Demchuk AM, Burgin WS, Christou I, Felberg RA, Barber PA, Hill MD, Alexandrov AV. Thrombolysis in brain ischemia (TIBI) transcranial doppler flow grades predict clinical severity, early recovery, and mortality in patients treated with intravenous tissue plasminogen activator. Stroke. 2001; 32: 89–93.
Larrue V, von Kummer R, del Zoppo G, Bluhmki E. Hemorrhagic transformation in acute ischemic stroke: potential contributing factors in the European Cooperative Acute Stroke Study. Stroke. 1997; 28: 957–960.
Kaps M, Damian MS, Teschendorf U, Dorndorf W. Transcranial Doppler ultrasound findings in middle cerebral artery occlusion. Stroke. 1990; 21: 532–537.
Zanette EM, Roberti C, Mancini G, Pozzilli C, Bragoni M, Toni D. Spontaneous middle cerebral artery reperfusion in ischemic stroke: a follow-up study with transcranial Doppler. Stroke. 1995; 26: 430–433.
Kidwell CS, Alger JR, Saver JL. Evolving paradigms in neuroimaging of the ischemic penumbra. Stroke. 2004; 35: 2662–2665.
Molina CA, Saver JL. Extending reperfusion therapy for acute ischemic stroke: emerging pharmacological, mechanical, and imaging strategies. Stroke. 2005; 36: 2311–2320.
Kidwell CS, Saver JL, Carneado J, Sayre J, Starkman S, Duckwiler G, Gobin YP, Jahan R, Vespa P, Villablanca JP, Liebeskind DS, Vinuela F. Predictors of hemorrhagic transformation in patients receiving intra-arterial thrombolysis. Stroke. 2002; 33: 717–724.