Endovascular treatment of acute ischemic stroke

Introduction
Endovacular treatment for all eligible ischemic stroke patients arriving at the hopsital within 14.5 hours of last known well is a Neurocritical Care Society Clinical Performance Measure. This can be accomplished by performing the treatment at the institution itself, or transferring the patient for mechanical thrombectomy.

Tenecteplase
The Tenecteplase versus Alteplase before Endovascular Therapy for Ischemic Stroke (EXTEND-IA TNK, 2018) trial randomized 202 patients with LVO to either tenecteplase (0.25 mg/kg) or alteplase (0.5 mg/kg) within 4.5 hours of stroke onset, followed by thrombectomy. The patients given tenecteplase had a higher rate of >50% reperfusion, and showed better 90-day functional outcomes as measured by median mRS compared with alteplase (common OR 1.7, 95% CI 1.0-2.8), with similar rates of symptomatic intracerebral hemorrhage. A larger dose of tenecteplase (0.40 mg/kg) did not confer any benefit over the 0.25 mg/kg dose prior to planned thrombectomy in the EXTEND-IA TNK Part 2 trial.

Intra-arterial thrombolytics
In patients in whom TICI 3 recanalization is not achieved, administering intra-arterial urokinase appears to be safe. In one series of 53 patients with partial or near-complete reperfusion treated with IA urokinase, 60% had early reperfusion improvement, and 34% had an improvement in TICI grade. Accordingly, patients treated with IA urokinase also had higher rates of functional independence (aOR 1.93, 95% CI 1.11-3.37). This is not a proven therapy, but seems reasonable in select cases.

Pediatric patients
Although data is limited, there is no good reason to withhold endovascular intervention in children compared to adults as it pathophysiologically makes sense, appears like it likely benefits them, and appears to be safe. The Save ChildS Study, a retrospective multicenter cohort study, evaluated 73 patients from 27 stroke centers, with median age of 11.3 years (IQR 7-15 years). Neurologic outcome improved from a median PedNIHSS of 14.0 (IQR 9.2-20.0) at admission to 4.0 (IQR 2.0-7.3) at day 7. Median mRS score was 1.0 at 6 and 12 months. The proportion of patients with hemorrhagic events was similar to adult patients. One patient (1%) had postinterventional bleeding and 4 patients (5%) had transient peri-interventional vasospasm.

Vertebrobasilar occlusion
A large cohort study suggests benefit of thrombectomy in basilar artery occlusion. The BEST trial showed no difference but was woefully underpowered, and did show a trend towards improved outcomes with thrombectomy. Thus the overall evidence favors performing thrombectomy in these cases. The Basilar Artery Occlusion Endovascular Intervention versus Standard Medical Treatment (BEST) trial set out to randomize patients with vertebrobasilar artery occlusion to endovascular treatment vs best medical management. They enrolled only 131 patients of a planned 344 patients due to poor recruitment and loss of equipoise over the course of the trial. The primary outcome of mRS 0-3 at 90 days occurred in 42% of the intervention group vs 32% in the control group, with aOR 1.74 (0.81-3.74). However, there was significant crossover (14 of 65 patients in control group received endovascular treatment and 3 of 66 patients in the intervention group did not receive intervention). In both per-protocol and as-treated analysis, the intervention group had improved outcomes for mRS 0-3 at 90 days (as treated aOR 3.02, 95% CI 1.31-7.00), mRS 0-2 at 90 days (as treated aOR 2.90, 95% CI 1.20-7.04), and ordinal mRS at 90 days (as treated aOR 2.36, 95% CI 1.20-4.62). In the as-treated analysis, there was a trend towards lower mortality with intervention (29.9% vs 44.4%, OR 0.53 (0.26-1.10), despite a higher rate of symptomatic intracranial hemorrhage (6.5% vs 0%). However, this study was very underpowered to detect a difference. In a nonrandomized cohort study called BASILAR, they enrolled 829 consecutive patients at 47 centers in China over five years. Patients were divided into groups receiving medical treatment alone (n=182) or medical treatment plus endovascular intervention (n=182).  90 day functional outcome (measured by shift on mRS) was improved in the patients who received endovascular treatment, (aOR 3.08, 95% CI 2.09-4.55).  Endovascular treatment was also associated with higher rates of 90-day mRS 0-3 (aOR 4.70, 95% CI 2.53-8.75), 90-day mRS 0-2 (aOR 4.90, 95% CI 2.71-8.83), and 90-day mRS 0-1 (aOR 4.49, 95% CI 2.31-8.76). Lastly, thrombectomy was associated with a lower rate of 90-day mortality (aOR 0.34, 95% CI 0.24-0.49) despite a higher likelihood of hemorrhage (7.1% vs 0.5%, p <0.001).

With prior oral anticoagulation
A registry analysis of patients showed that those on DOACs had better functional outcomes and less mortality than those on VKAs at the time of thrombectomy.

With low NIHSS
In one retrospective review from Switzerland, endovascular treatment was not clearly better (and possibly associated with slightly higher mortality) compared with IV thrombolysis alone. A multicenter study and meta-analysis of prior multicenter studies showed no difference in outcomes between endovascular treatment vs medical management in patients with NIHSS < 6, for favorable outcome, functional independence, mortality, or functional improvement.

With large area of core
With one trial showing clear evidence of functional benefit to mRS 0-3 despite an increase in overal ICH (but no definitive increase in symptomatic ICH and a trend toward decreased mortality), thrombectomy despite large core of ASPECTS 3-5 is very reasonable. In one prospective cohort study of 105 patients with poor ASPECTS (0-5) or core ischemia on CT perfusion of >50 mL, with treatment up to 24 hours patients still had improvement with thrombectomy. However, this signal was mostly driven by baseline disability as the results disappeared when adjusted for age and stroke severity, so evidence is limited.

The RESCUE-Japan LIMIT trial including 202 patients with ASPECTS 3-5 and LVO in the ICA or M1, randomized to endovascular treatment of medical care only. The endovascular treatment group was more likely to achieve mRS 0-3 at 90 days (31% vs 12.7%, RR 2.43 with 95% CI 1.35-4.37, p=0.002). There was more ICH within 48 hours with endovascula treatment (58% vs 31.4%, p<0.001) but only a trend towards more symptomatic ICH (9% vs 4.9%, p=0.25). There was a trend towards lower 90 day mortality in the endovascular group (18% vs 23.5%).

Blood pressure
Blood pressure during thrombectomy should probably be maintained with MAP in the 70-90 mmHg range, with an err on higher pressures rather than lower, as MAP <70 mmHg for only 10 minutes total is associated with worsened outcomes. In a retrospective cohort study of 365 patients in the SIESTA, ANSTROKE, and GOLIATH trials undergoing thrombectomy for anterior circulation ischemic stroke, they achieved TICI 2b-3 recanalization in 75-85% of patients, and found that duration with MAP <70 mmHg for 10 minutes or more was associated with worsened mRS at 90 days (aOR 1.51, 95% CI 1.02-2.22), and with MAP <70 mmHg for 20 minutes or more also (aOR 2.30, 95% CI 1.11-4.75) corresponding to number needed to harm (NNH) of 10 and 4, respectively. Also, for MAP >90 mmHg for 45 minutes or more was associated with worsened outcome (aOR 1.49, 95% CI 1.11-2.02), and for 115 minutes or more as well (aOR 1.89, 95% CI 1.01-3.54) for NNH of 10 and 6, respectively.

General anesthesia vs. conscious sedation vs. local anesthesia only
There may be benefits to avoiding general anesthesia. Without general anesthesia, local anesthesia only may be best, as conscious sedation is associated with a higher risk of pneumonia and even worsened outcomes compared with general anesthesia in one study of the MR CLEAN registry. If general anesthesia is used, a strict SBP > 140 mmHg should be targeted during anesthesia induction and throughout the procedure until reperfusion is obtained.

Sedation vs. Intubation for Endovascular Stroke Treatment (SIESTA, 2016)
This trial randomized 150 patients with anterior circulation AIS to general anesthesia or conscious sedation, and found no difference in early neurological improvement. They targeted maintaining SBP 140-160 mmHg in both groups. There were higher rates of complications in the general anesthesia group, with hypothermia or hyperthermia, delayed extubation, and ventilation-associated complications, but no differences in mortality. Functional independence at 90 days (mRS 0-2) was noted in 37% of the general anesthesia group and 18.2% of the conscious sedation group (p=0.01) but post-hoc analyses showed no difference for mRS 0-1 or 0-3 outcomes, suggesting this may not have been a true difference.

Anesthesia During Stroke trial (AnStroke, 2017)
This trial randomized 90 patients to general anesthesia or conscious sedation. They targeted maintaining SBP 140-180 mmHg and showed no difference in functional outcome, successful recanalization, or mortality at 3 months.

Brinjikji et al. Meta-Analysis (2017)
Although not a randomized trial, this study included 4,716 patients and showed that general anesthesia was associated with a lower odds of achieving mRS 0-2 (OR 0.59, 95% CI 0.29-0.94). However, this result was not significant when only considering studies performed in the era of stent-retrievers and aspiration techniques.

General or Local Anesthesia in Intra Arterial Therapy (GOLIATH, 2018)
This trial randomized 128 patients with anterior circulation AIS to general anesthesia or conscious sedation, and showed no significant difference in infarct growth, the primary outcome. However, it showed a trend towards lower infarct growth in the general anesthesia group (8.2 vs 19.4 mL, p=0.10). It also showed that successful reperfusion was higher in the general anesthesia arm (76.9% vs 60.3%, p=0.04) and that there were higher odds in the general anesthesia arm to shift to a lower mRS score (OR 1.91, 95% CI 1.03-3.56). Of note, this trial specifically attempted to maintain SBP > 140 mmHg and MAP >70 mmHg, but failed to do this in 35% of the general anesthesia patients and 16% of the conscious sedation patients. However, the median time with low BP was only 2 minutes in the general anesthesia group and 6.5 minutes in the conscious sedation group.

HERMES Anesthesia Meta-analysis (2018)
This was not a trial but an analysis of data from the 7 trials included in the HERMES collaboration, analyzing outcomes by type of anesthesia received, and included 1,690 patients. After adjustment for baseline prognostic variables, outcomes were significantly better for patients who did not receive general anesthesia compared with those who did. For achieving independence (mRS 0-2), this occurred in 31% of medical management only group, 40% for thrombectomy with general anesthesia, and 50% for thrombectomy without general anesthesia (aOR 1.65 for no GA vs GA, 95% CI 1.14-2.38). For achieving excellent functional outcome (mRS 0-1), this occurred in 17% of the medical management only group, 23% for thrombectomy with general anesthesia, and 32% for thrombectomy without general anesthesia (aOR 1.68 for no GA vs GA, 95% CI 1.12-2.52). The authors suggested that this difference was seen because there was no specific protocol for the general anesthesia group regarding blood pressure control, which is different than in the SIESTA, AnStroke, and GOLIATH trials where they set specific blood pressure goals (SBP > 140 mmHg).

Other data
Several studies have analyzed anesthetic data from the MR CLEAN Trial and the MR CLEAN registry. In an initial analysis of the trial from 2016, only treatment without general anesthesia was associated with a significant treatment benefit compared to medical management. At Erasmus Medical Center, some patients in the trial underwent local anesthesia only as the procedure was started before the anesthesia team was available. In 146 patients, there was a shift towards worse mRS scores in the conscious sedation group (OR 0.4, 95% CI 0.2-0.7) compared with local anesthesia. Conscious sedation did not reduce duration of the procedure and was not associated with an increase in procedure-related complications.

An analysis of the MR CLEAN registry included 1,376 patients, of whom 821 underwent local anesthesia only, 381 underwent general anesthesia, and 174 underwent conscious sedation. In these patients, the local anesthesia group did the best, followed by general anesthesia, and the conscious sedation group did the worst for achieving functional outcomes and mortality. There were higher rates of pneumonia in the conscious sedation group (20%) compared with the local anesthesia (11%) and general anesthesia (10%) groups. However, one important caveat to this study is that ASPECTS scores were lower in the general anesthesia and conscious sedation groups, so they may have been less likely to improve from that alone, which seriously limits the results of this study.

Blood pressure
In a prospective cohort of 65 patients undergoing thrombectomy for ischemic stroke, the ideal range of autoregulation was calculated using cerebral oximetry (NIRS), and the percentage of time exceeding values outside of ideal range correlated with risk of hemorrhagic transformation and worsened neurological outcomes.

In a retrospective study of 212 patients, higher SBP on days 1-3 after thrombectomy were associated with a lower likelihood of functional independence and higher mortality. Specifically on day 3, SBP <140 mmHg was associated with higher rates of functional independence (OR 4.3, 95% CI 2.0-9.9).

Prognosis
In a review of data from the THRACE clinical trial, Increasing numbers of white matter hyperintensities correlate with a greater likelihood of mRS 3-6 after thrombectomy.