INTRODUCTION — Timely restoration of cerebral blood flow using reperfusion therapy is the most effective maneuver for salvaging ischemic brain tissue that is not already infarcted. There is a narrow window during which this can be accomplished since the benefit of reperfusion decreases over time.
This topic will review the use of mechanical thrombectomy for acute ischemic stroke. The approach to reperfusion therapy for acute ischemic stroke, including the use of intravenous thrombolytic therapy (recombinant tissue plasminogen activator or tPA), is reviewed elsewhere. (See "Approach to reperfusion therapy for acute ischemic stroke" and "Intravenous thrombolytic therapy for acute ischemic stroke: Therapeutic use".)
OVERVIEW OF REPERFUSION THERAPY — For eligible patients with acute ischemic stroke, intravenous thrombolytic therapy with alteplase (or tenecteplase) is first-line therapy, provided that treatment is initiated within 4.5 hours since the time the patient was last known to be at their neurologic baseline (eg, time last known well) (table 1). Because the benefit is time dependent, it is critical to treat patients as quickly as possible; eligible patients should receive intravenous thrombolytic therapy without delay even if mechanical thrombectomy is being considered. (See "Intravenous thrombolytic therapy for acute ischemic stroke: Therapeutic use".)
Mechanical thrombectomy is indicated for patients with acute ischemic stroke due to a large artery occlusion in the anterior circulation who can be treated within 24 hours of the time last known to be well (ie, at neurologic baseline), regardless of whether they receive intravenous thrombolytic therapy for the same ischemic stroke event, as discussed in the sections that follow (algorithm 1).
Two issues may limit the widespread clinical use of mechanical thrombectomy. First, only an estimated 10 percent of patients with acute ischemic stroke have a proximal large artery occlusion in the anterior circulation and present early enough to qualify for mechanical thrombectomy within six hours [1-4], while approximately 9 percent of patients presenting in the 6 to 24 hour time window may qualify for mechanical thrombectomy [5]. Second, only a few stroke centers have sufficient resources and expertise to deliver this therapy [6]. However, eligible patients should receive standard treatment with intravenous thrombolysis if they present to hospitals where thrombectomy is not an option, and those with qualifying anterior circulation strokes from large artery occlusion should then be transferred, if at all possible, to tertiary stroke centers in which intra-arterial thrombectomy is available, a strategy called "drip and ship" [7].
PATIENT SELECTION — Patients with ischemic stroke caused by a proximal large artery occlusion in the anterior circulation are candidates for intra-arterial mechanical thrombectomy if they present to or can be transferred expeditiously to, a stroke center with expertise in the use of second-generation stent retrievers for acute ischemic stroke (algorithm 1). Intra-arterial mechanical thrombectomy can be used in addition to treatment with intravenous thrombolysis using recombinant tissue plasminogen activator (tPA) or tenecteplase. Mechanical thrombectomy treatment should be started as quickly as possible and should not be delayed to assess the response to intravenous tPA.
Treatment with intravenous thrombolysis prior to mechanical thrombectomy is also known as bridging therapy. The potential efficacy of bridging therapy compared with mechanical thrombectomy alone is reviewed separately. (See "Approach to reperfusion therapy for acute ischemic stroke", section on 'IVT followed by MT'.)
General criteria — For patients with acute ischemic stroke caused by a proximal large artery occlusion in the anterior circulation who can be treated within 24 hours of the time they were last known to be at their neurologic baseline, we recommend treatment with intra-arterial mechanical thrombectomy (algorithm 1), whether or not the patient received standard treatment with intravenous tPA, if the following general conditions are fulfilled:
●Neuroimaging (eg, computed tomography [CT] without contrast or diffusion-weighted magnetic resonance imaging [MRI]) is consistent with a small infarct core (ie, limited signs of early ischemic change) and excludes hemorrhage.
●Angiography (eg, CT angiography or magnetic resonance [MR] angiography) demonstrates a proximal large artery occlusion in the anterior circulation.
●Thrombectomy is performed at a stroke center with appropriate expertise in the use of stent retrievers.
●The patient has a persistent, potentially disabling neurologic deficit; some guidelines require a National Institutes of Health Stroke Scale (NIHSS) score ≥6 [8].
●Thrombectomy can be started within 24 hours of the time last known to be well.
The specific criteria for patients within 6 hours (see 'Within six hours' below) and for patients within 6 to 24 hours from the last time known to be at neurologic baseline (see '6 to 24 hours' below) are discussed in the sections that follow.
Many patients who are eligible for mechanical thrombectomy will be treated with intravenous thrombolytic therapy using alteplase or tenecteplase prior to mechanical thrombectomy. Patients who are not candidates for intravenous thrombolytic therapy can still be treated with mechanical thrombectomy if otherwise eligible according to the criteria outlined here and below. As an example, patients with infective endocarditis, which is a contraindication to intravenous thrombolysis, may still undergo mechanical thrombectomy if otherwise eligible [9].
Some experts argue that most of the currently used selection criteria for mechanical thrombectomy are merely predictors of outcome rather than true selection criteria, and that strict adherence to these criteria results in over-selection of patients and thereby denies benefit to many more patients who are deprived of a highly effective, safe treatment [10].
Patients with severe comorbidities prior to stroke onset (eg, pre-existing severe disability, life expectancy less than six months) are unlikely to benefit from mechanical thrombectomy. However, findings from an observational study suggest that patients with slight or moderate prestroke disability, defined by an mRS score (table 2) of 2 or 3, have a similar likelihood of recovery to their prestroke level of function after treatment with mechanical thrombectomy as do patients who are independent at baseline [11].
Within six hours — For patients who can start treatment (femoral puncture) within six hours of symptom onset, we suggest using the following criteria for mechanical thrombectomy (algorithm 1), which are modified from those used in the MR CLEAN trial [12]:
●A clinical diagnosis of acute stroke
●A deficit on the NIHSS (table 3) of ≥6 points (calculator 1)
●An Alberta Stroke Program Early CT Score (ASPECTS) score ≥6 on noncontrast brain CT or diffusion-weighted MRI (see 'ASPECTS method' below)
●Brain CT or MRI scan ruling out intracranial hemorrhage
●Intracranial arterial occlusion of the distal internal carotid artery (ICA), or the M1 or M2 segments of the middle cerebral artery (MCA), or the A1 or A2 segments of the anterior cerebral artery (ACA), demonstrated with CT angiography, MR angiography, or digital subtraction angiography (see "Neuroimaging of acute stroke")
●Age ≥18 years
Others have used somewhat different, more selective criteria. As examples:
●Guidelines from the American Heart Association/American Stroke Association (AHA/ASA) recommend mechanical thrombectomy for adults with no significant prestroke disability (ie, a modified Rankin Scale [mRS] score of ≤1) and a causative occlusion of the ICA or the M1 segment of the MCA [8]. These guidelines state that benefits are uncertain for patients who have a prestroke mRS score >1, or an NIHSS score <6, or a larger infarct core (ie, ASPECTS score <6).
●Both the ESCAPE and EXTEND-IA trials restricted eligibility to patients who were functioning independently prior to stroke onset [13,14].
●ESCAPE also required evidence of moderate-to-good collateral circulation, defined as the filling of ≥50 percent of the MCA territory pial circulation on CT angiography. The time window in ESCAPE was up to 12 hours from stroke onset, but few patients were enrolled beyond 6 hours [13].
●EXTEND-IA required evidence of salvageable brain tissue and an ischemic core lesion volume of <70 mL on CT perfusion imaging [14].
6 to 24 hours — The DAWN and DEFUSE 3 trials selected patients for treatment beyond 6 hours using imaging-based criteria [15-17]. For patients with ischemic stroke caused by a large artery occlusion in the proximal anterior circulation who are evaluated at stroke centers with automated infarct determination, or who can be transferred to such centers, time permitting, we recommend mechanical thrombectomy when DAWN or DEFUSE criteria are fulfilled (algorithm 1).
Eligibility criteria based upon the DAWN trial for patients who can start treatment (femoral puncture) within 6 to 24 hours of time last known to be at neurologic baseline are as follows [15]:
●Failed or contraindicated for intravenous thrombolytic therapy with alteplase or tenecteplase
●A deficit on the NIHSS (table 3) of ≥10 points (calculator 1)
●No significant prestroke disability: baseline modified Rankin scale (mRS) score ≤1
●Baseline infarct involving less than one third of the territory of the MCA on CT or MRI
●Intracranial arterial occlusion of the ICA or the M1 segment of the MCA
●A clinical-core mismatch according to age:
•Age ≥80 years: NIHSS ≥10 and an infarct volume <21 mL
•Age <80 years: NIHSS 10 to 19 and an infarct volume <31 mL
•Age <80 years: NIHSS ≥20 and an infarct volume <51 mL
Eligibility criteria based upon the DEFUSE 3 trial for patients who can start treatment (femoral puncture) within 6 to 16 hours of time last known to be at neurologic baseline are as follows [16]:
●A deficit on the NIHSS (table 3) of ≥6 points (calculator 1)
●Only slight or no prestroke disability: baseline mRS score ≤2
●Arterial occlusion of the cervical or intracranial ICA (with or without tandem MCA lesions) or the M1 segment of the MCA demonstrated on MR angiography or CT angiography
●A target mismatch profile on CT perfusion or MRI defined as an ischemic core volume <70 ml, a mismatch ratio (the volume of the perfusion lesion divided by the volume of the ischemic core) >1.8, and a mismatch volume (volume of perfusion lesion minus the volume of the ischemic core) >15 mL
●Age 18 to 90 years
For patients with ischemic stroke caused by a large artery occlusion in the proximal anterior circulation who are evaluated at stroke centers that do not use automated infarct volume determination, we suggest mechanical thrombectomy (algorithm 1) if treatment can be started within 6 to 24 hours of the time last known to be well and there is a clinical-ASPECTS mismatch, such as an NIHSS ≥10 and ASPECTS ≥6 (see 'ASPECTS method' below) [18]. However, this approach has not been evaluated in rigorous controlled trials.
Moderate to good collateral flow status on CTA is useful for identifying patients who are likely to benefit from mechanical thrombectomy in cases in which automated infarct volume determination is not available [8,13,19,20]. (See "Neuroimaging of acute stroke".)
Posterior circulation stroke — Although the benefits are uncertain, mechanical thrombectomy may be a reasonable treatment option for patients with acute ischemic stroke caused by occlusion of the basilar artery, vertebral arteries, or posterior cerebral arteries when performed at centers with appropriate expertise [8,21-24]. There is moderate-quality evidence that mechanical thrombectomy is beneficial for patients of Chinese ancestry with basilar artery occlusion who have an NIHSS score ≥10, indicating a moderate to severe stroke; a posterior circulation ASPECTS (pc-ASPECTS) of ≥6, indicating a limited extent of ischemic change on brain imaging; and who can be treated within 24 hours of time last known well. (See 'Basilar artery occlusion' below.)
ASPECTS method — The ASPECTS was developed to provide a simple and reliable method of assessing ischemic changes on head CT scan in order to identify acute stroke patients unlikely to make an independent recovery despite thrombolytic treatment [25]. The ASPECTS method has also been adopted to assess the extent of ischemia on diffusion-weighted MRI (DWI); the ability to detect early ischemic changes by ASPECTS was similar on noncontrast CT and DWI [26].
●Original (MCA territory) ASPECTS – The ASPECTS value is calculated from two standard axial CT cuts; one at the level of the thalamus and basal ganglia, and one just rostral to the basal ganglia (figure 1) [25,27].
•The ASPECTS method divides the middle cerebral artery (MCA) territory into 10 regions of interest.
•Subcortical structures are allotted three points: one each for caudate, lentiform nucleus, and internal capsule.
•MCA cortex is allotted seven points:
-Four of these points come from the axial CT cut at the level of the basal ganglia, with one point for insular cortex and one point each for M1, M2, and M3 regions (anterior, lateral, and posterior MCA cortex).
-Three points come from the CT cut just rostral to the basal ganglia, with one point each for M4, M5, and M6 regions (anterior, lateral, and posterior MCA cortex).
One point is subtracted for an area of early ischemic change, such as focal swelling or parenchymal hypoattenuation, for each of the defined regions.
Therefore, a normal CT scan has an ASPECTS value of 10 points, while diffuse ischemic change throughout the MCA territory gives a value of 0.
●Posterior circulation ASPECTS – The pc-ASPECTS subtracts one point for each ischemic lesion (right or left) of the thalamus, cerebellar hemisphere, or posterior cerebral artery territory, and two points for each lesion in the mesencephalon or pons [28,29]. A normal pc-ASPECTS has a value of 10 points; lower scores indicate greater extent of infarction.
EFFICACY OF MECHANICAL THROMBECTOMY — Early intra-arterial treatment with second-generation mechanical thrombectomy devices is safe and effective for reducing disability and is superior to standard treatment with intravenous thrombolysis alone for the treatment of acute ischemic stroke caused by a documented large artery occlusion in the proximal anterior circulation (figure 2 and figure 3).
Anterior circulation stroke
Benefit of early treatment — Five multicenter, open-label randomized controlled trials (MR CLEAN [12,30], ESCAPE [13], SWIFT PRIME [31], EXTEND-IA [14], and REVASCAT [32]) demonstrated that early intra-arterial treatment with second-generation mechanical thrombectomy devices is safe and effective for reducing disability and is superior to standard treatment with intravenous thrombolysis alone for ischemic stroke caused by a documented large artery occlusion in the proximal anterior circulation [1,33-39]. The number needed to treat (NNT) for one additional person to achieve functional independence in these trials ranged from approximately 3 to 7.5 [12-14,30-32,40].
When the positive results of the MR CLEAN trial were announced in late 2014 [12], the remaining trials (ESCAPE [13], SWIFT PRIME [31], EXTEND-IA [14], and REVASCAT [32]) were stopped early on the basis of positive interim efficacy analyses. All of these trials enrolled overlapping but not identical patient populations and had generally similar results.
In the HERMES meta-analysis of these trials, with pooled patient-level data for 1287 subjects, the rate of functional independence (ie, a 90-day modified Rankin scale [mRS] score of 0 to 2) was significantly greater for the intervention group compared with the control group (46 versus 27 percent, odds ratio [OR] 2.35, 95% CI 1.85-2.98) [33]. Similarly, mechanical thrombectomy led to significantly reduced disability as indicated by an improvement of ≥1 point on the mRS at 90 days (adjusted OR 2.49, 95% CI 1.76-3.53). Mechanical thrombectomy was beneficial across a wide range of patient subgroups, including age ≥80 years, high initial stroke severity, and those not treated with intravenous thrombolytic therapy. There was no significant difference between the mechanical thrombectomy and control groups for rates of symptomatic intracranial hemorrhage or 90-day mortality.
Several additional trials (THERAPY [41], PISTE [42], EASI [43], and RESILIENT [44]) also had point estimates suggesting improved functional outcomes for patients treated with mechanical thrombectomy. The RESILIENT trial, conducted in 12 public hospitals in Brazil, showed that mechanical thrombectomy can be efficacious in a country with limited health care resources [44].
Earlier trials (SYNTHESES Expansion [45], IMS III [46], and MR RESCUE [47]) failed to show benefit for intra-arterial treatment of acute ischemic stroke, in part because they used older-generation stent retriever devices, which were less likely to achieve reperfusion (see 'Devices' below), and because they did not require routine vessel imaging to confirm a large artery occlusion as the cause of the stroke [48].
Benefit of later treatment — Mechanical thrombectomy is effective when used from 6 to 24 hours for selected patients who have a clinical deficit that is disproportionally severe compared with the volume of infarction on imaging studies.
●The open-label DAWN trial enrolled 206 adults with acute ischemic stroke who were last known to be normal 6 to 24 hours earlier; all had a stroke caused by occlusion of the intracranial internal carotid artery or the proximal middle cerebral artery and had a clinical mismatch between the severity of the neurologic deficit, as measured by the National Institutes of Health Stroke Scale (NIHSS; median score 17 at baseline), and the infarct volume, as measured by automated software analysis using diffusion-weighted MRI or perfusion CT (median approximately 8 mL) [15]. Approximately 55 percent of the patients in the trial had a "wake-up" stroke (ie, they were last known to be well before going to bed and stroke symptoms were first noted upon awakening). Patients were randomly assigned to thrombectomy plus standard care or to standard care alone (control). The trial was stopped early for efficacy at the first interim analysis. The following observations were noted:
•At 90 days, the rate of functional independence, as defined by a score of 0 to 2 on the mRS, was greater for the thrombectomy group compared with the control group (49 versus 13 percent, adjusted difference 33 percent, 95% CI 24-44). The number needed to treat for one additional patient to achieve functional independence was 3. All other efficacy outcome measures also favored thrombectomy.
•There was no significant difference between the thrombectomy and control groups in the rate of symptomatic intracranial hemorrhage (6 and 3 percent) or 90 mortality (19 and 18 percent).
●The open-label DEFUSE 3 trial enrolled patients with ischemic stroke due to occlusion of the proximal middle cerebral artery or internal carotid artery who were last known to be well 6 to 16 hours earlier [16]. Patients were required to have a target perfusion mismatch characterized by an infarct size of <70 mL and a ratio of ischemic tissue volume to infarct volume of ≥1.8, as measured by automated software processing of diffusion-weighted MRI or CT perfusion imaging. The DEFUSE 3 trial was stopped early for efficacy after randomly assigning 182 patients to thrombectomy plus standard care or to standard care alone. Approximately one-half of the patients in the trial had a "wake-up" stroke. Patients assigned to thrombectomy were treated with stent retrievers or aspiration catheters. At 90 days, the percentage of patients who were functionally independent, defined as an mRS score of 0 to 2, was higher with endovascular therapy compared with medical therapy alone (45 versus 17 percent, difference 28 percent), and therefore the number needed to treat for one additional patient to achieve functional independence was 3.6. There was also a trend to lower mortality with endovascular therapy (14 versus 26 percent). There was no significant difference between groups in the rate of symptomatic intracranial hemorrhage (7 and 4 percent) or serious adverse events (43 and 53 percent).
●The AURORA study analyzed pooled patient-level data from 505 individuals from six randomized controlled trials of mechanical thrombectomy, including DAWN and DEFUSE 3, that included patients enrolled beyond six hours after they were last known well and received treatment with a second-generation stent retriever [49]. At 90 days, mechanical thrombectomy led to higher rates of independence in activities of daily living, defined by an mRS of 0 to 2, than best medical therapy alone (45.9 versus 19.3 percent, adjusted rate ratio 2.19, 95% CI 1.44-3.34, absolute risk reduction approximately 26 percent). The mechanical thrombectomy and best medical treatment groups had similar rates of mortality (16.5 versus 19.3 percent) and symptomatic intracerebral hemorrhage (5.3 versus 3.3 percent).
The AURORA investigators also compared outcomes among three subgroups: first, patients (n = 295) who met criteria for a clinical mismatch profile as used in the DAWN trial; second, patients (n = 359) who met criteria for a target perfusion mismatch profile as used in the DEFUSE 3 trial; and third, patients (n = 132) with an undetermined mismatch profile due to the absence of an adequate CT or MRI perfusion study [50]. At 90 days, mechanical thrombectomy led to reduced disability for the both the clinical mismatch subgroup (odds ratio [OR] 3.57, 95% CI 2.29-5.57) and the target perfusion mismatch subgroup (OR 3.13, 95% CI 2.10-4.66). Importantly, the benefit was significant in both subgroups for the entire 6- to 24-hour time window. There was a trend toward benefit for patients with an undetermined profile that did not reach statistical significance (OR 1.59, 95% CI, 0.82-3.06).
These data support the use of mechanical thrombectomy to treat acute ischemic stroke due to occlusion of the intracranial carotid or proximal middle cerebral artery for patients within 6 to 24 hours of the time last known to be well (ie, at neurologic baseline) who meet either the DAWN trial criteria for a clinical mismatch profile or the DEFUSE 3 trial criteria for a target perfusion mismatch profile [51].
Limitations to these trials include stopping early, which can overestimate treatment effects. However, this drawback is at least partially offset by the relatively large effect size demonstrated in the trials and meta-analysis [15,16,49].
Other approaches
●Large core infarcts – The requirement for a small infarct core as a criterion for mechanical thrombectomy eligibility has been challenged by results from the RESCUE-Japan LIMIT trial, which enrolled 203 patients with acute ischemic stroke due to a proximal middle cerebral artery or internal carotid artery occlusion and a low Alberta Stroke Program Early CT Score (ASPECTS) value of 3 to 5 on CT or diffusion-weighted MRI, consistent with a large infarct core (see 'ASPECTS method' above) [52]. Patients were randomly assigned in a 1:1 ratio to endovascular therapy with medical care or medical care alone; enrolled patients were within six hours after the time last known to be well (n = 145) or within 6 to 24 hours after the time last known well if fluid-attenuated inversion recovery (FLAIR) MRI showed no signal change (n = 58), suggesting very recent infarction. At 90 days, more patients had a "good" outcome, defined by an mRS score of 0 to 3, in the endovascular therapy group compared with the medical care group (31.0 versus 12.7 percent, relative risk [RR] 2.43, 95% CI 1.35-4.37). For the outcome of an mRS of 0 to 2 (ie, functional independence), there was only a trend for benefit with endovascular therapy (14 versus 7.8 percent, RR 1.79, 95% CI 0.78-4.07). The endovascular group had a higher rate of any intracranial hemorrhage (58 versus 31.4 percent, RR 1.84, 95% CI 1.33-2.58) and a nonsignificantly higher rate of symptomatic intracranial hemorrhage (9 versus 4.9 percent, RR 1.84, 95% CI 0.64-5.29).
Limitations of this trial include concerns about generalizability beyond the Japanese population, and relatively small patient numbers in the 6-to-24-hour treatment subgroup [52]. More data are needed to determine if patients with large core infarcts can benefit from mechanical thrombectomy.
●Adjunct intra-arterial thrombolysis – Some patients have poor clinical outcomes after mechanical thrombectomy despite successful reperfusion (ie, a Thrombolysis in Cerebral Infarction [TICI] 2b or 3) of the target large artery; one possible but controversial explanation is persisting impaired reperfusion of the microcirculation (the "no-reflow" phenomenon) [53,54]. The Chemical Optimization of Cerebral Embolectomy (CHOICE) trial investigated the use of adjunct intra-arterial thrombolysis with alteplase to treat hypothesized persistent thrombi in the microcirculation after angiographically successful mechanical thrombectomy [55]. At 90 days, more patients achieved an excellent neurologic outcome (an mRS score of 0 to 1) with intra-arterial alteplase compared with placebo (59 versus 40.4 percent, adjusted absolute risk reduction 18.4 percent, 95% CI 0.3-36.4 percent). There was no increased risk of intracranial hemorrhage or mortality with intra-arterial alteplase.
Limitations of the CHOICE trial include early stopping (due to slow recruitment and inability to obtain placebo), which can lead to overestimation of treatment effects, small patient numbers (and resulting wide confidence intervals with a lower limit of only 0.3 percent absolute risk reduction), and the protocol allowing premature stopping (and therefore potential underdosing) of intravenous alteplase infusion started before the onset of thrombectomy [55,56]. Thus, the benefit of this approach requires confirmation in larger trials.
Basilar artery occlusion — There is moderate-quality evidence that mechanical thrombectomy is beneficial for patients of Chinese ancestry who can be treated within 24 hours of moderate to severe stroke (an NIHSS score ≥10) caused by a basilar artery occlusion if the posterior circulation ASPECTS (pc-ASPECTS) score is consistent with a limited extent of ischemia [57,58].
The ATTENTION trial evaluated patients from China with moderate to severe stroke (with an NIHSS ≥10) due to basilar artery occlusion who were within 12 hours of the estimated time of stroke onset and had a limited degree of early ischemic change, as quantified by the pc-ASPECTS [57]. Patients were randomly assigned in a 2:1 ratio to medical care plus endovascular thrombectomy or medical care alone (control). At baseline, the median NIHSS score was 24 in each group. Approximately one-third of patients in each group received intravenous thrombolysis. At 90 days, the rate of good functional status (ie, an mRS score of 0 to 3) was higher in the thrombectomy group compared with the control group (46 versus 23 percent, adjusted rate ratio [RR] 2.06, 95% CI 1.46-2.91) and the mortality rate was lower in the thrombectomy group (37 versus 55 percent, RR 0.66, 95% CI 0.52-0.82). The rate of functional independence (ie, an mRS score of 0 to 2) was also higher in the thrombectomy group (33 versus 11 percent, RR 3.17, 95% CI 1.84-5.46), and results for most secondary outcomes favored thrombectomy. Symptomatic intracranial hemorrhage occurred in 5 percent of cases in the thrombectomy group versus none in the control group. Mechanical thrombectomy was associated with procedural complications in 14 percent of patients, including one death caused by arterial perforation.
The BAOCHE trial from China evaluated patients within 6 to 24 hours after stroke onset due to basilar artery occlusion [58]. Patients were randomly assigned in a 1:1 ratio to mechanical thrombectomy plus medical care with medical care alone. The trial was stopped early after an interim analysis suggested superiority of thrombectomy. At baseline, the median NIHSS was 20 for the thrombectomy group and 19 for the control group. The rate of intravenous thrombolysis was 14 percent in the thrombectomy group and 21 percent in the control group. At 90 days, the rate of good functional status (ie, an mRS score of 0 to 3) was higher in the thrombectomy group compared with the control group (46 versus 24 percent, RR 1.81, 95% CI 1.26-2.60), and the rate of functional independence (ie, an mRS score of 0 to 2) was also higher in the thrombectomy group (39 versus 14 percent, RR 2.64, 95% CI 1.54-4.50). There was a trend for lower mortality at 90 days favoring the thrombectomy group (31 versus 42 percent, RR 0.75, 95% CI 0.54-1.04). Symptomatic intracranial hemorrhage occurred more often in the thrombectomy group (6 versus 1 percent, RR 5.18, 95% CI 0.64-42.18). Procedural complications occurred in 11 percent of the thrombectomy group.
The ATTENTION and BAOCHE trial results are not generalizable to all patients with basilar artery stroke. The Chinese population has higher rates of large artery intracranial atherosclerotic disease relative to other populations, and many patients in the thrombectomy groups of both trials were also treated with angioplasty and/or stenting of the basilar artery. The low rates of treatment with intravenous thrombolysis in both trials may have reduced the rates of good outcomes particularly affecting the control groups and biased the results in favor of thrombectomy.
Earlier trials were also limited by methodologic issues. A randomized trial (BEST) comparing endovascular treatment (mechanical thrombectomy) with standard medical care for patients with acute vertebrobasilar occlusion who could be treated within eight hours was stopped early for slow recruitment and high crossover rate after enrolling 131 patients [21]. Compared with standard medical care, patients assigned to endovascular therapy had similar rates of favorable outcome and 90-day mortality by intention-to-treat analysis. The BASICS trial of 300 patients with acute ischemic stroke attributed to basilar artery occlusion found no statistically significant difference in outcomes for endovascular therapy compared with medical therapy [23]. However, there was a nonsignificant trend of benefit with endovascular treatment in both trials [21,23].
Larger randomized controlled trials in more diverse populations are needed to assess the efficacy of endovascular therapy for posterior circulation stroke due to large artery occlusion.
PROCEDURE
Overview — General anesthesia or conscious sedation may be used for the procedure, depending upon local preference and experience (see 'Anesthesia' below). (See 'Devices' below.)
Catheterization is commonly performed with femoral artery puncture. The catheter is guided to the internal carotid artery and beyond to the site of the intracranial large artery occlusion. The stent retriever is then inserted through the catheter to reach the clot. The stent retriever is deployed and grabs the clot, which is removed as the device is pulled back. The initial goal is to achieve reperfusion, defined by a modified Thrombolysis in Cerebral Infarction (mTICI) perfusion grade 2b (anterograde reperfusion of more than half in the downstream target arterial territory) or grade 3 (complete anterograde reperfusion of the downstream target arterial territory) (table 4), as early as possible [8,59]. In a meta-analysis of five trials that evaluated treatment within six hours of symptom onset, over 500 patients received mechanical thrombectomy and substantial reperfusion (mTICI score of 2b or 3) was achieved in 71 percent of this group [39].
Following the procedure, most centers monitor patients in an intensive care unit setting until stable.
Devices — Both second-generation stent retrievers and catheter aspiration devices can be used for mechanical thrombectomy. The choice between them depends mainly upon local expertise and availability [60]. In some cases, treatment using stent retrievers and aspiration techniques in combination may be appropriate.
●Stent retrievers – A number of mechanical thrombectomy devices are approved in the United States and Europe for clot removal in patients with acute ischemic stroke due to large artery occlusion. These include the first-generation Merci Retriever and Penumbra System devices, the second-generation Solitaire Flow Restoration Device and Trevo Retriever, and the third-generation Tigertriever. The first-generation Merci and Penumbra devices may increase recanalization rates in carefully selected patients, but their clinical utility for improving outcomes after stroke is unproven [61-63]. When compared directly with the Merci retriever in small randomized trials, the second-generation Solitaire and Trevo neurothrombectomy devices achieved significantly higher reperfusion rates and better patient outcomes [64,65]. In a single-arm study, the Tigertriever device achieved higher reperfusion rates, improved patient outcomes, and had similar safety outcomes compared with historical controls from studies of the Solitaire and Trevo devices [66].
In light of these data and the positive thrombectomy trials discussed above [12-14,31,32], which preferentially used the second-generation devices, only the second-generation or later devices should be used to treat patients with acute ischemic stroke.
●Catheter aspiration devices – Catheter aspiration devices are another option for mechanical thrombectomy. This method employs a catheter to aspirate the thrombus as the first approach to performing thrombectomy; if aspiration alone does not achieve reperfusion after one or more passes, a stent retriever can be inserted through the catheter to complete the thrombectomy.
Mounting evidence suggests that catheter aspiration devices can attain rates of revascularization [41,67] and good functional outcome [68,69] that are similar to the rates achieved with second-generation stent retrievers. The open-label, multicenter COMPASS trial randomly assigned 270 patients within six hours of symptom onset to mechanical thrombectomy with either catheter aspiration as first pass treatment or stent retriever first-line [68]. At 90 days, a good functional outcome (mRS score of 0 to 2) was achieved by a similar number of patients in each treatment group (52 versus 50 percent for aspiration first-pass and stent retriever first-line, respectively), indicating that aspiration first-pass was noninferior to stent retriever first-line treatment. In addition, secondary efficacy and angiographic outcome measures did not differ between treatment groups, and there were no significant differences in mortality, symptomatic intracranial hemorrhage, or other safety outcomes.
One trial found a trend to higher rates of near-total or total reperfusion for combined stent retriever plus aspiration compared with stent retriever alone, but the difference did not achieve statistical significance (64.5 versus 57.9 percent, risk difference, 6.6 percent, 95% CI -3.0 to 16.2) [70].
Anesthesia — Either monitored anesthesia care (also called conscious sedation) or general anesthesia may be used for procedural sedation during mechanical thrombectomy. The anesthetic technique should be chosen based upon individual patient risk factors, preferences, and institutional experience [8]. (See "Anesthesia for endovascular therapy for acute ischemic stroke in adults", section on 'Choice of anesthetic technique: General anesthesia versus monitored anesthesia care'.)
The type of anesthesia used for mechanical thrombectomy in patients with ischemic stroke may have some impact on short- and long-term outcomes, but data are inconsistent. This issue is reviewed in detail separately. (See "Anesthesia for endovascular therapy for acute ischemic stroke in adults", section on 'Literature comparing general anesthesia with monitored anesthesia care or conscious sedation'.)
Risk of periprocedural antithrombotics — There is no indication for the routine use of periprocedural antithrombotic agents. Based upon the results of the MR CLEAN-MED trial, the use of periprocedural aspirin or unfractionated heparin in patients undergoing endovascular therapy for acute ischemic stroke increases the risk of symptomatic hemorrhagic transformation and may increase the risk of worse outcomes [71]. However, antithrombotic agents may be indicated in specific instances (eg, if a stent gets deployed, or if there is distal embolism).
Blood pressure management — We suggest keeping systolic blood pressure (SBP) between 150 to 180 mmHg prior to reperfusion; SBP ≥150 mmHg may be useful for maintaining adequate collateral blood flow during the time the large artery remains occluded [8,13]. Some experts suggest no use of antihypertensives prior to reperfusion unless SBP exceeds 200 mmHg for patients not being treated with intravenous thrombolysis, or unless SBP exceeds 185 mmHg for patients who are candidates for intravenous thrombolysis [72].
Once reperfusion is achieved with mechanical thrombectomy, we suggest targeting SBP to <140 mmHg [15,73]. More intensive blood pressure lowering (eg, targeting SBP <120 mmHg) may be harmful, particularly in Asian populations where the prevalence of large artery atherosclerosis is high [74,75]. However, the optimal blood pressure range with mechanical thrombectomy is not well defined, and there are few data to guide periprocedural management [8,76,77]. Others suggest keeping SBP <160 to 170 mmHg for patients with successful reperfusion (ie, TICI 2b or 3) and targeting SBP of 170 mmHg for patients with less successful reperfusion (ie, TICI 0 to 2a) [72].
Many patients undergoing mechanical thrombectomy will have been treated with intravenous thrombolytic therapy (recombinant tissue plasminogen activator or tPA) in the first hours after stroke symptom onset and should be managed accordingly, with systolic/diastolic blood pressure maintained at ≤180/105 mmHg during and for 24 hours following alteplase infusion or tenecteplase injection; a higher blood pressure may increase the risk of hemorrhage in ischemic brain regions even when thrombolytic agents are not used. (See "Intravenous thrombolytic therapy for acute ischemic stroke: Therapeutic use", section on 'Management of blood pressure'.)
Adverse effects — In the MR CLEAN trial, clinical signs of a new ischemic stroke in a different vascular territory within 90 days of treatment were more common in the intra-arterial group compared with no endovascular therapy (5.6 versus 0.4 percent) [12]. Device-related serious adverse events are uncommon but include access site hematoma and pseudoaneurysm, arterial perforation, and arterial dissection [13,14,31,32]. Transient intraprocedural vasospasm is also uncommon but is sometimes treated.
Mechanical thrombectomy is not associated with increased rates of symptomatic intracranial hemorrhage or mortality. In meta-analysis of five trials, with pooled patient-level data for 1287 subjects, there was no significant difference between the intervention population and control population for 90-day symptomatic intracranial hemorrhage (4.4 versus 4.3 percent) or mortality (15 versus 19 percent) [33].
Approach to tandem lesions — Fifteen to 30 percent of patients eligible for mechanical thrombectomy present with tandem lesions characterized by extracranial carotid artery stenosis or occlusion and a downstream, ipsilateral intracranial large vessel occlusion [12,13,32,78]. Mechanical thrombectomy is directed at revascularization of the intracranial occlusion, but the best approach to management of the extracranial carotid lesion is uncertain [78,79]. Options include acute treatment of the extracranial carotid lesion with stent placement (anterograde or retrograde), angioplasty alone, or thrombo-aspiration alone, versus deferred or no revascularization of the extracranial carotid artery lesion (figure 4) [80]. Deferred revascularization options include eventual carotid endarterectomy or carotid artery stenting. (See "Management of symptomatic carotid atherosclerotic disease".)
Available data from observational studies suggest that acute carotid stenting for patients with tandem lesions who are undergoing mechanical thrombectomy is associated with a higher rate of favorable outcomes at 90 days compared with no stenting [81,82]. A subgroup analysis from one study further suggests that stenting is associated with improved outcomes in patients with carotid lesions caused by atherosclerosis but not in patients with carotid lesions caused by dissection [82].
Rescue therapy for failed MT — Approximately 8 to 30 percent of patients fail to achieve substantial reperfusion with MT, with failure defined by mTICI scores (table 4) of 2a or less [33,83-85]. In such cases, urgent rescue therapy with intracranial angioplasty/stenting, intravenous glycoprotein IIb/IIIa inhibitors, or intravenous P2Y12 receptor inhibitors is sometimes attempted [83]. Limited observational data suggest that these interventions are safe [83,86], but prospective studies are lacking, and the optimal approach to is uncertain.
Intracranial stenting is the best-studied option [87-89]. The retrospective, multicenter SAINT study of patients who failed MT compared those who received acute rescue stenting (n = 107) with propensity-score matched patients who did not receive rescue stenting (n = 107) [87]. At 90 days, rescue stenting was associated with a shift to lower rates of disability in the overall modified Rankin Scale score distribution (adjusted odds ratio [OR] 3.74, 95% CI 2.16–6.57), increased functional independence (34.6 versus 6.5 percent, OR 10.91, 95% CI 4.11–28.92), decreased mortality (29.9 versus 43.0 percent, OR 0.49, 95% CI 0.25–0.94), and comparable rates of symptomatic intracranial hemorrhage (7.5 versus 11.2 percent, OR, 0.87, 95% CI 0.31–2.42]. These results are limited by retrospective design and wide confidence intervals, and further study is needed to determine the benefit of this intervention for failed MT.
SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Stroke in adults".)
SUMMARY AND RECOMMENDATIONS
●Efficacy of mechanical thrombectomy – Early intra-arterial treatment with mechanical thrombectomy is safe and effective for reducing disability and is superior to standard treatment with intravenous thrombolysis alone for ischemic stroke caused by a documented large artery occlusion in the proximal anterior circulation. (See 'Efficacy of mechanical thrombectomy' above.)
●Selection of patients within six hours – For patients with ischemic stroke caused by a large artery occlusion in the proximal anterior circulation who can start treatment (femoral puncture) within six hours of stroke symptom onset, we recommend treatment with intra-arterial mechanical thrombectomy using a second-generation stent retriever device (algorithm 1), whether or not the patient received treatment with intravenous thrombolytic therapy (recombinant tissue plasminogen activator or tPA), if the following conditions are fulfilled (Grade 1A) (see 'General criteria' above and 'Within six hours' above):
•Neuroimaging (eg, CT without contrast or diffusion-weighted MRI) is consistent with a small infarct core (eg, limited signs of early ischemic change; an Alberta Stroke Program Early CT Score [ASPECTS] score ≥6) and excludes hemorrhage
•The patient has a persistent, potentially disabling neurologic deficit, defined by a National Institutes of Health Stroke Scale (NIHSS) score ≥6
•Thrombectomy is performed at a stroke center with expertise in the use of stent retrievers
●Selection of patients from 6 to 24 hours – Mechanical thrombectomy can reduce disability when started 6 to 24 hours from the time last seen well for patients who have a clinical deficit that is disproportionally severe compared with the volume of infarction on imaging studies (see 'Benefit of later treatment' above). The volume of infarction can be measured directly at stroke centers with automated software, or can be estimated from noncontrast CT or diffusion MRI using the ASPECTS method at stroke centers without automated infarct volume determination. (See '6 to 24 hours' above.)
•For patients with ischemic stroke caused by a large artery occlusion in the proximal anterior circulation who are evaluated at stroke centers with automated infarct determination (algorithm 1), we recommend mechanical thrombectomy if treatment can be started within 6 to 24 hours of the time last known to be well, and there is either a clinical-core mismatch as defined by the DAWN trial or an imaging-target mismatch as defined by the DEFUSE 3 trial (Grade 1B). (See 'Benefit of later treatment' above.)
•For patients with ischemic stroke caused by a large artery occlusion in the proximal anterior circulation who are evaluated at stroke centers that do not use automated infarct volume determination (algorithm 1), we suggest mechanical thrombectomy if treatment can be started within 6 to 24 hours of the time last known to be well and there is a clinical-ASPECTS mismatch (eg, National Institutes of Health Stroke Scale [NIHSS] ≥10 and ASPECTS ≥6) (Grade 2C). (See '6 to 24 hours' above and 'ASPECTS method' above.)
●Use in posterior circulation stroke – Although the benefits are uncertain, mechanical thrombectomy within 24 hours of the time last known to be well may be a reasonable treatment option for patients with acute ischemic stroke caused by occlusion of the basilar artery, vertebral arteries, or posterior cerebral arteries when performed at centers with appropriate expertise. Moderate-quality evidence supports the benefit of mechanical thrombectomy for patients of Chinese ancestry with basilar artery occlusion who have an NIHSS score ≥10, indicating a moderate to severe stroke; a posterior circulation ASPECTS (pc-ASPECTS) of ≥6, indicating a limited extent of ischemic change on brain imaging; and who can be treated within 24 hours of time last known well. (See 'Posterior circulation stroke' above.)
●Procedure – Second-generation stent retriever devices or catheter aspiration devices should be used for mechanical thrombectomy. Other aspects of the mechanical thrombectomy procedure, including anesthesia, blood pressure management, and adverse events, are discussed above. (See 'Procedure' above.)