Endovascular Therapy for Stroke
2014; Lippincott Williams & Wilkins; Volume: 129; Issue: 10 Linguagem: Inglês
10.1161/circulationaha.113.003703
ISSN1524-4539
AutoresChung‐Huan Sun, Deepak L. Bhatt, Raul G. Nogueira, Rishi Gupta,
Tópico(s)Peripheral Artery Disease Management
ResumoHomeCirculationVol. 129, No. 10Endovascular Therapy for Stroke Free AccessResearch ArticlePDF/EPUBAboutView PDFView EPUBSections ToolsAdd to favoritesDownload citationsTrack citationsPermissionsDownload Articles + Supplements ShareShare onFacebookTwitterLinked InMendeleyReddit Jump toSupplemental MaterialFree AccessResearch ArticlePDF/EPUBEndovascular Therapy for StrokeGetting to the "Heart" of the Matter Chung-Huan J. Sun, BS, Deepak L. Bhatt, MD, MPH, FAHA, Raul G. Nogueira, MD and Rishi Gupta, MD, MBA, FANA Chung-Huan J. SunChung-Huan J. Sun From the Department of Neurology, Emory University School of Medicine, Grady Memorial Hospital, Atlanta, GA (C.-H.J.S., R.G.N.); Brigham and Women's Hospital Heart and Vascular Center and Harvard Medical School, Boston, MA (D.L.B); and Wellstar Neurosurgery, Kennestone Hospital, Marietta, GA (R.G.). , Deepak L. BhattDeepak L. Bhatt From the Department of Neurology, Emory University School of Medicine, Grady Memorial Hospital, Atlanta, GA (C.-H.J.S., R.G.N.); Brigham and Women's Hospital Heart and Vascular Center and Harvard Medical School, Boston, MA (D.L.B); and Wellstar Neurosurgery, Kennestone Hospital, Marietta, GA (R.G.). , Raul G. NogueiraRaul G. Nogueira From the Department of Neurology, Emory University School of Medicine, Grady Memorial Hospital, Atlanta, GA (C.-H.J.S., R.G.N.); Brigham and Women's Hospital Heart and Vascular Center and Harvard Medical School, Boston, MA (D.L.B); and Wellstar Neurosurgery, Kennestone Hospital, Marietta, GA (R.G.). and Rishi GuptaRishi Gupta From the Department of Neurology, Emory University School of Medicine, Grady Memorial Hospital, Atlanta, GA (C.-H.J.S., R.G.N.); Brigham and Women's Hospital Heart and Vascular Center and Harvard Medical School, Boston, MA (D.L.B); and Wellstar Neurosurgery, Kennestone Hospital, Marietta, GA (R.G.). Originally published11 Mar 2014https://doi.org/10.1161/CIRCULATIONAHA.113.003703Circulation. 2014;129:1152–1160IntroductionThose who cannot remember the past are condemned to repeat it.1—George Santayana, 1863 to 1952In the late 1970s, Gruentzig's groundbreaking reports on a series of successfully performed coronary angioplasties created the field of endovascular medicine, setting the precedent for the Food and Drug Administration's 510(k) approval of the procedure in 1981.2 Over the next decade, the number of angioplasties performed in the United States grew exponentially from 1000 procedures in 1980 to >300 000 by 1995.3 Despite robust growth, however, clinical trials failed to demonstrate the benefits of angioplasty over thrombolysis for the treatment of ST-segment–elevation myocardial infarction (STEMI).4–7 The Primary Angioplasty in Myocardial Infarction (PAMI) and ZWOLLE trials in 1993 were the first randomized, controlled studies to demonstrate the efficacy of percutaneous coronary intervention (PCI) over fibrinolytic therapy,8,9 nearly 10 years after reimbursement by the Centers for Medicare & Medicaid (diagnosis-related group 108) was granted.10 Although subsequent studies disputed the merits of PCI,11–13 several meta-analyses provided definitive evidence that PCI was superior for STEMI.14,15 In 1996, the American College of Cardiology/American Heart Association incorporated PCI as a Class I recommendation for STEMI if performed in a "timely fashion…by personnel in high volume center[s]."16 The use of PCI as a facilitated or rescue strategy after fibrinolysis, however, remained controversial, having only recently been shown to be effective.17–19As history repeats itself, many challenges encountered by cardiology are being revisited by acute stroke trialists. Since the publication of the nonpositive results of the Interventional Management of Stroke (IMS) III, The Mechanical Retrieval and Recanalization of Stroke Clots Using Embolectomy (MR RESCUE), and SYNTHESIS Expansion trials,20–22 endovascular procedures for stroke are being met with skepticism, similar to PCI in the early 1990s, with concerns about treatment efficacy and procedural reimbursement.23 Despite differences in pathophysiology and in the impact of disability in stroke, the clinical trial paradigm for proving the benefits of endovascular reperfusion harbors similar features. The interventional cardiologists' belief in promoveo promovi promotum would culminate in the acceptance of primary PCI as the standard of care. This blueprint of clinical trial success may be replicated in acute ischemic stroke (AIS) by considering the salient overlaps with respect to efficient treatment times, enhanced patient selection, and composite outcome measures (Figure 1).Download figureDownload PowerPointFigure 1. Factors influencing clinical trial success for interventional therapies in both ST-segment–elevation myocardial infarction (STEMI) and acute ischemic stroke (AIS). CSC indicates comprehensive stroke center; ED, emergency department; IAT, intra-arterial therapy; IC, interventional cardiology; mRS, modified Rankin Scale; NCCT, noncontrast computed tomography; OSH, outside hospital; TIMI, Thrombolysis in Myocardial Infarction; and UA/NSTEMI, unstable angina/non–ST-segment–elevation myocardial infarction."Time is Myocardium"…and BrainThe advantages of primary PCI over fibrinolysis have not always been universally accepted in cardiology. In 1989, DeWood4 showed no benefit of direct PCI versus intravenous duteplase in a sample of 90 patients. Subsequent studies by Ribeiro et al5 and Gibbons et al6 arrived at a similar conclusion: PCI had neither a physiological or clinical benefit over medical management. Although these studies lacked proper statistical power, delays in door-to-balloon (D2B) times of 2 hours affected the efficacy of PCI. Registry studies further highlighted the absence of a mortality benefit at D2B times approaching 2 hours.24 In contrast, both the PAMI and Zwolle studies had either a randomization-to-angiography or D2B time of 1 hour, respectively, which coincided with the trials being the first studies to demonstrate the efficacy of PCI over thrombolysis.8,9Cannon et al25 demonstrated that delays to PCI with D2B times beyond 2 hours resulted in a 40% to 60% higher odds of mortality. A separate study revealed that every minute of difference between time to balloon in PCI and time to needle in fibrinolytic therapy correlated with a 1% absolute reduction in mortality benefit of PCI. Despite higher recanalization rates with PCI, the amount of myocardial salvage decays with time, reiterating the importance of timely delivery of care.26 The tipping point when primary PCI is no longer superior to immediate administration of fibrinolysis depends on the ischemic and bleeding risks of the patient.19,27 For example, an older patient at higher risk of intracranial hemorrhage from fibrinolysis having a large anterior myocardial infarction would benefit from slightly delayed primary PCI compared with prompt fibrinolysis.Analogous to the early PCI trials, the variable of time may be a critical denominator confounding the outcomes from recent AIS trials. In SYNTHESIS Expansion, the time from symptom onset to the initiation of intra-arterial therapy (IAT) was an hour longer than intravenous thrombolytics (225 versus 165 minutes), reflecting significant delays in the delivery of endovascular care.21 Similarly, in IMS III, the delay to IAT after intravenous tissue-type plasminogen activator (t-PA) delivery was nearly 90 minutes and may have contributed to the negative trial result.20 Analogous to the myocardium, the impact of time delays despite higher recanalization rates with IAT precludes salvage of ischemic tissue.28The intrinsic relationship between the time to IAT and patient recovery has been shown in the IMS I/II trial, in which faster reperfusion correlated with improved clinical outcomes.29 Recent studies have supported this relationship, demonstrating that every half-hour delay to groin puncture results in an 18% reduction in favorable outcomes.30 Indeed, the importance of time cannot be understated in the setting of clinical trials. In IMS III, patients who received IAT within 2 hours of intravenous t-PA initiation had outcomes favoring the IAT arm. These benefits were accretive among the IAT patients who received t-PA within 2 hours of symptom onset. Efforts to reduce the time to IAT may prove to be critical for achieving a positive clinical trial. As in PAMI, clinical trials may need to be selective of sites that demonstrate the ability to achieve timely reperfusion.Time Metrics for Ischemic Stroke ManagementAnother criticism of the recent AIS trials has been the heterogeneity among the sites in terms of patient volume, operator experience, and efficient delivery of IAT.31 Without performance metrics, the heterogeneity of care among stroke centers continues to challenge trialists. This problem may be addressed through the development of standard time metrics.As shown by interventional cardiologists, the American Heart Association first proposed D2B times 8 hours from last known normal state in whom salvageable tissue can be preserved among those with superior collateral systems.45 For all other patients, image acquisition and interpretation should be streamlined to eliminate the delays associated with poorer patient outcomes. Noncontrast head computed tomography can be evaluated both for core infarct with the Alberta Stroke Programme Early CT Score and for the presence of a hyperdense thrombus with the use of thin-cut reconstruction.46,47 Although such strategies may not be as precise as advanced imaging, there does not appear to be sufficient evidence supporting the clinical benefits of advanced imaging (Figure 3). There is evidence, however, that time to reperfusion matters in both the coronary and cerebral vascular beds.26,29,30,34 Clinical examination and noncontrast computed tomography may be sufficient to select patients for IAT, allowing systems to streamline care and thus reduce times to IAT.Download figureDownload PowerPointFigure 3. Summary of studies involving patient selection for intra-arterial treatment (IAT) based on imaging modality. Only IAT arms are included. Good outcome was defined as 90-day modified Rankin Scale score of 0 to 2. Successful reperfusion was based on the definition used by the individual study. Statistics were calculated with 1-way ANOVA. All included studies are listed in the online-only Data Supplement. CT indicates computed tomography; GP, groin puncture; MRI, magnetic resonance imaging; n, number of prospective or retrospective studies; NIHSS, National Institutes of Health Stroke Scale; and Sx, symptom.Optimizing Patient SelectionStroke SeverityPatient selection is another area that plays a critical role in determining IAT efficacy. In a recent study, patients with large-vessel occlusions treated with IAT had significantly smaller final infarct volumes than those receiving intravenous t-PA alone, which was reflective of the reperfusion status. In particular, the maximal benefit of IAT over intravenous t-PA was observed in patients with more proximal occlusions (ie, internal carotid artery terminus) and higher National Institutes of Health Stroke Scale (NIHSS) scores.48Intravenous t-PA may be as effective as IAT in patients with distal clot locations and lower NIHSS scores, but its efficacy is limited with more proximal occlusions and higher NIHSS scores.49 As suggested by Rangaraju et al,48 this cutoff may be an NIHSS score of 14. Earlier t-PA trials, including National Institute of Neurological Disorders and Stroke (NINDS) and European Cooperative Acute Stroke Study (ECASS) III, have suggested an NIHSS threshold of 20 as the tipping point for treatment efficacy with intravenous t-PA. Patients with an NIHSS score >20 had a lower likelihood of good outcomes at 90 days and higher rates of intracranial hemorrhages after t-PA administration compared with those with an NIHSS score ≤20.50,51Results from IMS III similarly revealed that outcomes tended to be more favorable with IAT when patients presented with an NIHSS score >20, albeit not statistically significantly.20 In a post hoc analysis, endovascularly treated patients with internal carotid artery terminus occlusions or tandem M1/internal carotid artery terminus occlusions also had greater recanalization rates and better outcomes than the intravenous t-PA cohort.52 These findings suggest that a more homogeneous patient cohort with a higher NIHSS score and proximal clot locations may provide the basis for investigating the efficacy of IAT.Strategies exploring the stratification of patients based on ischemic location and severity have been explored in cardiology. The magnitude of benefit of rapid PCI within the first hours from symptom onset appears to be greatest among STEMI patients compared with patients with unstable angina (UA) or non-STEMI (NSTEMI). Unlike STEMI, the magnitude of benefit of PCI over medical therapy appears to be smaller with UA/NSTEMI; the number needed to treat to prevent 1 case of short-term mortality is 50 for STEMI patients15 and nearly 500 for UA/NSTEMI patients.53 This number, however, underestimates the benefit of PCI in UA/NSTEMI with respect to the substantial reductions in reinfarction rates.Guidelines have suggested using a risk score to determine the patients who may maximally benefit from acute intervention for UA/NSTEMI. A history of recurrent angina, elevated cardiac biomarkers, new or evolving ST-segment depression, and high Thrombolysis in Myocardial Infarction (TIMI) risk scores are all indications that a patient with UA or NSTEMI may preferentially derive benefit from an invasive strategy.54One can argue that the role of IAT in AIS may benefit from a risk score much like that used in patients with UA/NSTEMI. Two scoring systems have already been shown to predict poor outcomes after IAT but have not been used in clinical trial design.55,56Role of Imaging in IAT SelectionPatient selection for IAT may also benefit from the integration of imaging parameters to strengthen predictive outcomes after acute intervention. At present, clinical variables such as age, baseline glucose, reperfusion, and NIHSS scores have been implicated in the treatment efficacy of IAT,57 whereas imaging parameters have been less conclusive. Although advanced imaging modalities identify patients with favorable penumbral patterns, delays to treatment potentially offset the benefits of improved selection.22,44 The Alberta Stroke Programme Early CT Score system is a feasible alternative, with treatment scores ≤7 being associated with poorer functional outcomes and increased risk of symptomatic hemorrhage.46The Houston IAT 2 score is a predictive model that accounts for both the imaging (ie, Alberta Stroke Programme Early CT Score) and clinical parameters of ischemic stroke patients before IAT. Patients with Houston IAT 2 scores ≥5 had 6 times the probability of a poor outcome, both at discharge and at the 90-day follow-up, compared with those with scores of 8 years.Download figureDownload PowerPointFigure 4. A, Absolute mortality difference of patients at 30 days for all trials involving aggressive percutaneous coronary intervention (PCI; primary or immediate) vs conservative management (fibrinolysis alone or delayed PCI). B, Absolute difference in the composite outcome of mortality, reinfarction, or stroke for patients at 30 days in all trials involving aggressive PCI (primary or immediate) vs conservative management (fibrinolysis alone or delayed PCI). All trials included in the figure are listed in the online-only Data Supplement. AHA indicates American Heart Association.A classic example stems from the Global Use of Strategies to Open Occluded Coronary Arteries in Acute Coronary Syndromes (GUSTO IIb) trial, in which patients assigned to primary angioplasty did not have better results compared with patients receiving thrombolytic therapy with respect to mortality (5.7% versus 7.0%; P=0.37), reinfarction (4.5% versus 6.5%; P=0.13), or disabling stroke (0.2% versus 0.9%; P=0.11). Yet, the composite end point revealed that patients treated with PCI had a reduction in short-term event rates (9.6% versus 13.7%; P=0.033).12 The meta-analysis by Keeley et al15 of 23 primary PCI trials from 1986 to 2003 reinforced the advantages of primary PCI over thrombolysis, revealing both short-term (6-weeks) and long-term (1-year) benefits in the composite end point and the individual variables of mortality and nonfatal reinfarction.Whether redefining outcomes can play a similar role in ischemic stroke trials remains to be seen. The 90-day modified Rankin scale (mRS) has been the gold standard for the evaluation of functional stroke recovery and has been used in all successful stroke trials to date60,61 (Figure 5). Nevertheless, the optimal mode of analysis for this 7-point ordinal scale has been under scrutiny.62 The majority of the endovascular AIS trials use a dichotomized interpretation of the mRS, defining functional independence or good outcomes as an mRS score of 0 to 2 and poor outcomes as an mRS score of 3 to 6. This cut point was selected on behalf of patients but has not been validated with patient preferences. Dichotomous end points may be easier to use to design trials but may not fully realize the impact of treatment effect, particularly in selecting the dichotomization threshold (ie, mRS score of 0–1 or 0–2). For instance, a patient who improves from an mRS score of 4 to 3 would still be classified as having a poor outcome.Download figureDownload PowerPointFigure 5. Evolution of primary outcome measures for interventional trials between ST-segment–elevation myocardial infarction (STEMI) and acute ischemic stroke (AIS). Each primary outcome is listed on the basis of the first time it was used in a randomized, controlled trial. Subsequent trials using the same primary outcome measure were not recorded. All secondary outcomes were excluded from the analysis. CHF indicates congestive heart failure; IAT, intra-arterial therapy; PCI, percutaneous coronary intervention; TVR, target vessel revascularization; and VA, major ventricular arrhythmia. The adoption of a shift analysis has been suggested as a strategy to overcome this problem, allowing treatment and control groups to be compared across the entire range of the ordinal scale.63 This approach may be advantageous among treatments that confer mild and uniform benefits to patients over a wide range of stroke severities.64 The Abciximab Emergent Stroke Treatment Trial (AbESST) II trial used an alternative strategy with a responder analysis that adjusted the primary outcome end point to the severity of the index NIHSS score.65Another limitation with the mRS scoring system is the interobserver variability inherent in ascribing scores, which is contingent on the interpretation of functional independence and disability.66 The mRS serves to capture a global picture of patient disability but fails to discern non–stroke-related features of impairment (ie, hip fractures), as well as the emotional and socioeconomic factors associated with recovery.67 In contrast to the more objective measures of mortality, reinfarction, and stroke used in the PCI trials, the nonspecific nature of Rankin scores limits the interpretation of outcome results.Infarct Volume as a Surrogate for OutcomePostprocedural infarct volume has been suggested as a surrogate biomarker for outcomes in acute stroke trials.68 At a threshold of 40 cubic cm, infarct volume can be a stronger predictor of patient outcome than other variables such as NIHSS score, Alberta Stroke Programme Early CT Score, and even recanalization rates.The use of a postprocedural imaging metric as a surrogate for clinical outcomes has been explored in the cardiac literature, in which left ventricular ejection fraction (LVEF) was proposed as a surrogate for long-term mortality.69 Early PCI trials routinely investigated the impact of aggressive management on LVEF70 and considered the composite end point of short-term mortality or severe heart failure (New York Heart Association class III–IV) as an outcome measure for rescue angioplasty.71 Criticisms, however, concerning the use of LVEF as an end point hinged on the following reasons: 5% to 20% of patients do not get echocardiograms; 10% to 20% of echocardiograms are technically inadequate for interpretation; LVEF does not solely account for patient survival72; and an open infarct-related vessel is a stronger predictor of patient survival than LVEF.73 More recently, cardiac magnetic resonance imaging has been used to quantify infarct size, and this tool has been shown to be useful in early clinical trials.74An advantage of using infarct volumes in stroke is that it shows the absolute treatment effect of IAT without confounders in the ensuing 89 days that may affect 90-day mRS. Nevertheless, there are still concerns about missing data in patients who die before magnetic resonance image acquisition or those unable to undergo magnetic resonance imaging because of pacemakers or other implants.Device IterationsAlthough not the focus of the present review, the role of device technology in recent acute stroke trials cannot be understated. Patients in the endovascular arm of IMS III were treated predominantly with the Merci Retriever, a first-generation device with lower recanalization rates and worse neurological outcomes than its stent-retriever counterparts.37,38 Adoption of newer technologies in clinical trials may push endovascular treatment toward greater efficacy and significance over intravenous thrombolytics, especially with an anticipated 26% to 37% improvement in successful recanalization. One criticism, however, stems from the fact that clinical outcomes are not concordant with improved reperfusion rates. This is likely attributable to a combination of device technology, patient selection, and systems of care.30,31Historically, improvements in reperfusion rates in acute coronary ischemia have followed a stepwise trajectory of intravenous thrombolytics to intra-arterial thrombolytics to balloon angioplasty, stents, manual thrombus aspiration catheters, and drug-eluting stents.75–77 During this period of growth from the early 1980s to the present, complete reperfusion rates increased at a rate of 3.0%/y,78 which coincided with a significant reduction in the risk-standardized mortality rates of patients with acute myocardial infarction in the United States.79 Over the last decade, complete recanalization rates have been >90%, yet D2B times have continued to improve from 96 minutes in 2006 to 63 minutes in 2010.80,81Acute stroke interventions have followed a similar evolution in the use of intravenous thrombolytics to mechanical thrombectomies, yet the progression of improved recanalization rates and faster treatment times has not been at the same pace. From 1992 to 2009, complete reperfusion rates for treatment of ischemic strokes increased at 0.9%/y,78 with subsequent symptom onset-to-treatment times hovering at 4 to 5 hours in all endovascular trials to date.20–22,36–38 Indeed, improved reperfusion rates must come hand in hand with improvements in system proc
Referência(s)