Skip main navigation
×

Endovascular Therapy Versus Medical Therapy Alone for Basilar Artery Stroke: A Systematic Review and Meta‐Analysis Through Nested Knowledge

Originally publishedhttps://doi.org/10.1161/SVIN.121.000147Stroke: Vascular and Interventional Neurology. 2022;2:e000147

Abstract

Background

Endovascular thrombectomy (EVT) is an effective treatment for acute ischemic stroke attributable to the anterior circulation large‐vessel occlusion. Randomized trials of patients with posterior circulation large‐vessel occlusion (PC‐LVO) have failed to show a benefit of EVT over medical therapy (MEDT). We performed a systematic review and meta‐analysis to understand better whether EVT is beneficial for PC‐LVO.

Methods

Using the Nested Knowledge AutoLit living review platform, we identified randomized control trials and prospective studies that reported functional outcomes in patients with PC‐LVO treated with EVT versus MEDT. The primary outcome variable was 90‐day modified Rankin scale score of 0 to 3, and secondary outcome variables included 90‐day modified Rankin scale score of 0 to 2, 90‐day mortality, and rate of symptomatic intracranial hemorrhage. A separate random effects model was fit for each outcome measure to calculate pooled odds ratios.

Results

Three studies with 1248 patients, 860 in the EVT arm and 388 in the MEDT arm, were included in the meta‐analysis. The favorable outcome rate (modified Rankin scale score of 0–3) in patients undergoing EVT was 39.9% (95% CI, 30.6%–50.1%) versus 24.5% in patients undergoing MEDT (95% CI, 9.6%–49.8%). Patients undergoing EVT had higher modified Rankin scale score of 0 to 2 rates (31.8% [95% CI, 25.7%–38.5%] versus 19.7% [95% CI, 7.4%–42.7%]) and lower mortality (42.1% [95% CI, 35.9%–48.6%] versus 52.8% [95% CI, 33.3%–71.5%]) compared with patients undergoing MEDT, but neither result was statistically significant. Patients undergoing EVT were more likely to develop symptomatic intracranial hemorrhage (odds ratio, 10.36; 95% CI, 3.92–27.40).

Conclusions

EVT treatment of PC‐LVO trended toward superior functional outcomes and reduced mortality compared with MEDT despite a trend toward increased symptomatic intracranial hemorrhage in patients undergoing EVT. Existing randomized and prospective studies are insufficiently powered to demonstrate a benefit of EVT over MEDT in patients with PC‐LVO.

Nonstandard Abbreviations and Acronyms

AC‐LVO

anterior circulation large‐vessel occlusion

EVT

endovascular thrombectomy

LVO

large Vessel Occlusion

MEDT

medical therapy

MOOSE

Meta‐Analyses Of Observational Studies in Epidemiology

mRS

modified Rankin Scale

PC‐LVO

posterior circulation large‐vessel occlusion

PRISMA

Preferred Reporting Items for Systematic reviews and Meta‐Analyses

sICH

symptomatic intracranial hemorrhage

Clinical Perspective

  • In this novel meta‐analysis, we collated high‐quality studies and sought to evaluate the efficacy of endovascular thrombectomy (EVT) in the treatment of patients presenting with posterior circulation stroke. Our results failed to show a convincing benefit of EVT over medical therapy alone, as long‐term functional and safety outcomes were comparable between groups. However, the directionality and magnitude of our findings suggest that EVT may be a superior treatment option for this patient cohort, which may be definitively established in a study with more power. Notably, our study also introduced a semi‐automated systematic review platform for expeditious and streamlined study collation.

  • While EVT is a standard‐of‐care treatment in anterior circulation stroke, and single‐arm and retrospective studies have shown similar efficacy of EVT in posterior circulation stroke, there is yet to be a high‐quality study concluding the efficacy of EVT in posterior circulation stroke. We sought to do so in this study, and found our analysis to be underpowered. Further randomized control trials and prospective studies to add power towards this research question are warranted, and would aid neurointerventional surgeons in clarifying the clinical decision making for this patient cohort.

Acute ischemic stroke attributable to a large‐vessel occlusion (LVO) can be a debilitating clinical event that results in significant disability and mortality. Patients with acute ischemic stroke are often stratified by the location of their stroke; anterior circulation LVO (AC‐LVO) strokes affect the internal carotid, anterior cerebral, or middle cerebral arteries, and posterior circulation large‐vessel occlusion (PC‐LVO) strokes affect the basilar artery or vertebral arteries. Endovascular thrombectomy (EVT) has been shown to be more effective in the treatment of AC‐LVO stroke compared with medical therapy (MEDT),1, 2, 3, 4, 5, 6, 7 and EVT is the standard of care in appropriately selected patients.8

By contrast, whether EVT improves outcomes in patients with PC‐LVO stroke has not been conclusively determined because of exclusion of these patients from the landmark EVT studies of 2015 and 2018 and the fact that randomized studies of EVT for PC‐LVO treatment to date have been underpowered.9, 10 Reasons for this lack of data may include enrollment bias or lack of pretreatment patient eligibility and selection with advanced imaging. However, prospective studies to determine the efficacy of EVT for PC‐LVO treatment have suggested a benefit for EVT.11 Additional data or conglomeration of existing data is needed to determine whether EVT is an effective treatment for PC‐LVO.

To determine whether existing literature supports EVT for PC‐LVO stroke treatment, we performed a systematic review and meta‐analysis of prospective and randomized studies that compared EVT and MEDT in these patients. We performed our study using a novel software platform (AutoLit; Nested Knowledge, St. Paul, MN) that allows for the rapid identification, collation, synthesis, and analysis of data.

Methods

The data that support the findings of this study are available from the corresponding author on reasonable request.

Nested Knowledge Systemic Review Platform

A Preferred Reporting Items for Systematic reviews and Meta‐Analyses (PRISMA) and Meta‐analyses Of Observational Studies in Epidemiology (MOOSE)‐compliant systematic review of the literature was undertaken on the PubMed database through the Nested Knowledge platform. Before study selection and screening, 2 authors (G.A. and J.J.H.) established the framework for the study by writing up a protocol for the systematic review that included acceptable study designs, intervention arms, patient characteristics to collect as baseline variables, and outcome variables. This was presented in both written format and “nest” format, as depicted in Figure 1B.

FIGURE 1

FIGURE 1 Basilar stroke study nest on Nested Knowledge. A, A simple flow diagram of the data gathering process for this nest (full instructions for use of the Nested Knowledge software available at https://wiki.nested‐knowledge.com/doku.php?id=wiki:autolit). B, The qualitative synthesis of our nest, with the patient characteristics, interventions, and outcomes stratified by category (the interactive version is available at https://nested‐knowledge.com/nest/qualitative/390).

Literature Search and Study Selection

Initial search was completed on the PubMed database using the Application Program Interface in the AutoLit platform; all study metadata and abstracts from the search results were also obtained via Application Program Interface. We sought to include randomized controlled trials (RCTs) and prospective cohort studies that reported EVT treatment of patients presenting with acute ischemic stroke attributable to basilar artery occlusion, and we captured these using the search strings:

  1. basilar AND “ischemic stroke” AND (RCT OR “randomized controlled trial”)

  2. basilar AND (stent‐triever OR aspiration OR thrombectomy) AND (IVT OR IV‐tPA OR thrombolysis) AND stroke

  3. “basilar artery occlusion” AND “randomized controlled trial”

Studies identified in the initial literature search were screened automatically by the AutoLit platform for 2 preconfigured automated exclusion criteria: (1) published before January 1, 2014; and (2) published in a non‐English language. Two independent raters (G.A. and K.H.) then used the AutoLit dual screening module to vet the remaining studies and included studies if they were: (1) RCTs and prospective cohort studies with treatment arms of patients treated with EVT and patients treated using medical therapy alone and (2) reported 90‐day modified Rankin scale (mRS) scores of 0 to 3 and 0 to 2, 90‐day mortality, and symptomatic intracranial hemorrhage (sICH) among the outcome variables. Studies that were not RCTs or prospective cohort studies by design were excluded. Other manual exclusion criteria included:

  • Includes pediatric patients

  • Does not relate to basilar acute ischemic stroke

  • Does not report an EVT arm

  • Only reports an EVT treatment arm, without an MEDT arm

  • Does not report patient outcomes

  • MERCI thrombectomy device used in EVT arm

After manual screening was completed, discordant study ratings were independently adjudicated by another rater (J.J.H.). Detailed results of our study search, screening, and data extraction process are hosted on the Nested Knowledge website (www.nested‐knowledge.com).

Extraction of Patient Characteristics and Outcome Variables

Extraction of the data from each selected study was completed by 1 author (G.A.) and confirmed for accuracy independently by 2 other authors (H.K. and K.H.), as displayed in Figure 1A. Patient characteristics collected included the following: age, sex, baseline National Institutes of Health Stroke Scale score, onset‐to‐needle time, and onset‐to‐puncture time. Our primary outcome was a favorable functional outcome, which was defined as mRS score of 0 to 3 at 90 days. Secondary outcomes included functional independence (mRS score of 0–2 at 90 days) after treatment, sICH, 90‐day mortality, and successful recanalization (thrombolysis in cerebral infarction≥2b). Because of heterogeneity in outcome reporting and limited data, only comparisons of 90‐day mRS score of 0 to 3, 90‐day mRS score of 0 to 2, 90‐day mortality, and sICH were eventually included for quantitative synthesis.

Risk‐of‐Bias Assessment

The risk of bias and levels of evidence of each study were scored using the Scottish Intercollegiate Guidelines Network checklists for controlled clinical trials and cohort studies.12 Within their separate checklists, non‐RCTs were rated no higher than 1+. The risk‐of‐bias assessment was completed independently by 2 authors (K.H. and H.K.). Any disagreements were discussed and resolved by a third author (J.M.P.).

Statistical Analysis

Data were extracted within the Nested Knowledge interface, exported as a .csv file, and imported to RStudio (Version 1.3.959; RStudio, PBC, Boston, MA) running on R‐4.0.2 for analysis. The “meta” (Version 4.18‐0) and “metafor” (Version 2.4‐0) packages were used to perform meta‐analyses.13, 14

Effect sizes from each study were computed as logarithmically transformed odds ratios (ORs) with random‐effects, Mantel‐Haenszel weighting. Logarithmic transformations were used to correct for skewed marginal distributions and to shrink the influence of high leverage outliers. To aid in interpretation, logarithmically transformed pooled effect sizes were back‐transformed to their original scale. The between‐study variance component of random‐effects models were estimated using restricted effects maximum likelihood with 95% CIs computed using the Q‐profile method.15 Because of the small number of patients included in the meta‐analysis, 95% CIs around pooled effect sizes were calculated using Hartung‐Knapp adjustment to provide a more conservative estimate of the true intervention effect and to reduce the risk of false positives.16 For each aggregated result, Higgin I2 statistics were used to measure the percentage of the total variability in effect estimates that can be attributed to heterogeneity rather than sampling error.17 I2 values of <25%, 25% to 75%, and >75% were considered low, moderate, and high between‐study variability in effect estimates, respectively. The absolute value of the true variance in effect sizes is indicated by τ2 values in forest plots, which were estimated using restricted effects maximum likelihood.

Finally, the data from these studies were reported by randomization status, as “intention‐to‐treat” results. An “as‐treated” analysis would require patient‐level data from the respective trials and is beyond the scope of this study.

Results

Literature Search Results and Risk‐of‐Bias Assessment

Our initial search identified 237 studies, with 2 additional records identified through expert recommendation (Supplementary Figure S1). After removing duplicates, a total of 226 articles were screened for inclusion. A total of 223 articles were excluded after screening based on title and abstract. A total of 3 full‐text articles, the Basilar Artery Occlusion Endovascular Intervention versus Standard Medical Treatment (BEST), Basilar Artery International Cooperation Study (BASICS), and Endovascular Treatment for Acute Basilar Artery Occlusion (BASILAR) studies, were assessed for eligibility and included in the final quantitative meta‐analysis.9, 10, 11

Our risk‐of‐bias assessment identified 1 RCT of high quality (1++),10 1 RCT of moderate quality (1+),9 and 1 nonrandomized study of high quality (2++).11 The results of our quality appraisal are summarized in Supplementary Table S1.

Study‐Specific Patient Demographics and Outcomes

From the 3 studies identified in our systematic review, a total of 1248 patients were available for analysis. A total of 388 (31.1%) patients were treated with MEDT alone and 860 (68.9%) patients were treated with EVT. Study‐specific patient characteristics and outcomes are provided in the Table 1.

John Wiley & Sons, Ltd.

TABLE 1 Patient Characteristics and Function, Angiographic, and Safety Outcomes of Patients in the 3 Studies Included in the Meta‐Analysis

VariableLangezaal et al (BASICS)Liu et al (BEST)Zi et al (BASILAR)
ControlEVTControlEVTControlEVT
No. of patients1461546059182647
Age, mean±SD, y67.2±11.966.8±13.1N/AN/AN/AN/A
Age, median (quartile 1–quartile 3), yN/AN/A68 (57–74)62 (50–74)67 (59–76)64 (56–73)
Sex, male/female ratio96:50100:5452:848:11129:53483:164
Baseline NIHSS score, median (quartile 1–quartile 3)222126 (13–37)32 (18–38)26.5 (16–33)27 (17–33)
mRS score of 0–3, n/total (%)55/146 (37.7)68/154 (44.2)21/60 (35.0)28/59 (47.5)17/182 (9.3)207/647 (32.0)
mRS score of 0–2, n/total (%)44/146 (30.1)54/154 (35.1)18/60 (30.0)22/59 (37.3)13/182 (7.1)177/647 (27.4)
TICI≥2b, n/total (%)N/A63/88 (71.6)9/14 (64.3)45/59 (76.3)11/182 (6.0)522/647 (80.7)
Mortality, n/total (%)63/146 (43.2)59/154 (38.3)25/60 (41.7)22/59 (37.3)130/182 (71.4)299/647 (46.2)
sICH, n /total (%)1/146 (0.7)7/154 (4.5)0/60 (0)5/59 (8.5)1/182 (0.5)45/636 (7.1)

EVT indicates endovascular thrombectomy; mRS, modified Rankin scale; N/A, not available; NIHSS, National Institutes of Health Stroke Scale; sICH, symptomatic intracranial hemorrhage; TICI, thrombolysis in cerebral infarction; BEST, Basilar Artery Occlusion Endovascular Intervention versus Standard Medical Treatment; BASICS, Basilar Artery International Cooperation Study; and BASILAR, Endovascular Treatment for Acute Basilar Artery Occlusion Study.

Qualitative Appraisal of Data

Qualitatively, both RCTs9, 10 showed that EVT compared with MEDT alone failed to demonstrate a significant difference in sICH, 90‐day functional outcome, or mortality at 90 days. Conversely, the findings from Zi et al11 suggested that EVT was associated with better functional outcomes and reduced mortality rate at 90 days; however, sICH was significantly higher with EVT compared with MEDT alone. Because of the considerable differences in findings between the RCTs and the nonrandomized study, we performed additional sensitivity analyses restricting to only the higher‐quality RCTs9, 10 for each separate random‐effects model, which are presented in turn below.

Primary Outcome: 90‐Day mRS Score of 0 to 3

The overall rate of 90‐day mRS score of 0 to 3 for the EVT group was 39.9% (95% CI, 30.6%–50.1%), and for the MEDT group was 24.5% (95% CI, 9.6%–49.8%). There was no statistically significant difference in the odds of mRS score 0 to 3 between the EVT and the MEDT group (OR, 2.17; 95% CI, 0.41–11.52; P=0.185; Figure 2A). The estimated between‐study variability in effect estimates unrelated to sampling error ranged from moderate to high (I2=84.3% [95% CI, 53.2%–94.8%]).

FIGURE 2

FIGURE 2 Forest plot of comparisons of modified Rankin scale score of 0 to 3 at 90 days. A, Results from the full meta‐analysis. B, Results from a meta‐analysis of only randomized controlled trials. Pooled results were computed via the Mantel‐Haenszel method, using restricted effects maximum likelihood for estimation of the between‐study variance component and Hartung‐Knapp adjustment to calculate a 95% CI around the pooled effect size. A 95% prediction interval (PI) for the pooled results was also computed for the full meta‐analysis. EVT indicates endovascular thrombectomy; and OR, odds ratio.

When the analysis was restricted to RCTs only, the overall mRS score of 0 to 3 rate for the EVT group was 45.1% (95% CI, 38.5%–51.8%) versus 36.9% (95% CI, 30.6%–43.7%) in the MEDT group. There was no statistically significant difference in the odds of mRS score of 0 to 3 at 90 days between the EVT group and the control group (OR, 1.40; 95% CI, 0.34–5.81; P=0.203; Figure 2B). With removal of the prospective, but nonrandomized, study by Zi et al from the meta‐analysis, the estimated between‐study variability unrelated to sampling error was low (I2=0.0%). However, given the limited number of studies included in the analysis, a CI around the I2 point estimate could not be produced, and the true variability unrelated to sampling error is likely underestimated by the I2 statistic. This point is true for all subsequent analyses.

Secondary Outcome: 90‐Day Functional Independence (mRS Score of 0–2)

The overall rate of 90‐day functional independence (mRS score of 0–2) for the EVT group was 31.8% (95% CI, 25.7%–38.5%) and 19.7% (95% CI, 7.4%–42.7%) in the MEDT group. There was no statistically significant difference in the odds of mRS score of 0 to 2 between the EVT and the MEDT groups (OR, 2.05; 95% CI, 0.31–13.67; P=0.246; Figure 3A). The estimated between‐study variability in effect estimates unrelated to sampling error ranged from moderate to high (I2=85.0%; 95% CI, 55.6%–94.9%).

FIGURE 3

FIGURE 3 Forest plot of comparisons of modified Rankin scale score of 0 to 2 at 90 days. A, Results from the full meta‐analysis. B, Results from a meta‐analysis of only randomized controlled trials. Pooled results were computed via the Mantel‐Haenszel method, using restricted effects maximum likelihood for estimation of the between‐study variance component and Hartung‐Knapp adjustment to calculate a 95% CI around the pooled effect size. A 95% prediction interval (PI) for the pooled results was also computed for the full meta‐analysis. EVT indicates endovascular thrombectomy; and OR, odds ratio.

When the analysis was restricted to RCTs only, the overall mRS score of 0 to 2 rate for the EVT group was 35.7% (95% CI, 29.5%–42.4%) compared with 30.1% (95% CI, 24.2%–36.7%) in the MEDT group. There was no statistically significant difference in the odds of mRS score of 0 to 2 at 90 days between the EVT group and the MEDT group (OR, 1.29; 95% CI, 0.71–2.33; P=0.115; Figure 3B).

Safety Outcome: sICH

Two studies reported sICH based on the Heidelberg Bleeding Classification criteria, and 1 study reported sICH defined as evidence of intracranial hemorrhage on imaging and an increase of ≥4 points on National Institutes of Health Stroke Scale within 24 hours after randomization.18 The overall rate of sICH for the EVT group was 6.8% (95% CI, 5.3%–8.7%) versus 0.7% (95% CI, 0.2%–2.2%) in the MEDT group. The EVT group had significantly higher odds of sICH in comparison to the MEDT group (OR, 10.36; 95% CI, 3.92–27.40; P=0.009; Figure 4A). The between‐study variability in effect estimates unrelated to sampling error was low (I2=0.0%; 95% CI, 0.0%–11.0%).

FIGURE 4

FIGURE 4 Forest plot of comparisons of symptomatic intracranial hemorrhage. A, Results from the full meta‐analysis. B, Results from the meta‐analysis of only randomized controlled trials. Pooled results were computed via the Mantel‐Haenszel method, using restricted effects maximum likelihood for estimation of the between‐study variance component and Hartung‐Knapp adjustment to calculate a 95% CI around the pooled effect size. A 95% prediction interval (PI) for the pooled results was also computed for the full meta‐analysis. EVT indicates endovascular thrombectomy; and OR, odds ratio.

When the analysis was restricted to RCTs only, the overall rate of sICH for the EVT group was 5.9% (95% CI, 3.2%–10.7%) versus 0.7% (95% CI, 0.2%–3.5%) in the MEDT group. There was no statistically significant difference in the odds of sICH between the EVT and the MEDT groups (OR, 8.40; 95% CI, 0.27–261.45; P=0.081; Figure 4B). However, as reflected by the wide CI around the pooled effect, the results of this meta‐analysis may not exclude a substantial increase in odds of sICH when using EVT, as was indicated in the global meta‐analysis of sICH rate.

Safety Outcome: Mortality

The overall mortality rate for the EVT group was 42.1% (95% CI, 35.9%–48.6%), and for the MEDT group was 52.8% (95% CI, 33.3%–71.5%). There was no statistically significant difference in the odds of mortality between the EVT and the MEDT groups (OR, 0.59; 95% CI, 0.16–2.15; P=0.222; Figure 5A). The estimated between‐study variability in effect estimates unrelated to sampling error ranged from moderate to high (I2=80.8%; 95% CI, 40.0%–93.9%).

FIGURE 5

FIGURE 5 Forest plot of comparisons of mortality at 90 days. A, Results from the full meta‐analysis. B, Results from a meta‐analysis of only randomized controlled trials. Pooled results were computed via the Mantel‐Haenszel method, using restricted effects maximum likelihood for estimation of the between‐study variance component and Hartung‐Knapp adjustment to calculate a 95% CI around the pooled effect size. A 95% prediction interval (PI) for the pooled results was also computed for the full meta‐analysis. EVT indicates endovascular thrombectomy; and OR, odds ratio.

When the analysis was restricted to RCTs only, the overall mortality rate for the EVT group was 38.0% (95% CI, 31.8%–44.7%) versus 42.7% (95% CI, 36.1%–49.6%) in the MEDT group. There was no statistically significant difference in the odds of mortality at 90 days between the EVT group and the MEDT group (OR, 0.82; 95% CI, 0.75–0.91; P=0.025; Figure 5B).

Discussion

In this study of the BEST, BASICS, and BASILAR studies, there was no statistically significant evidence to demonstrate superiority of EVT over MEDT in the treatment of PC‐LVO stroke, both in terms of functional outcomes and safety outcomes. However, the directionality and magnitude of difference between the outcomes suggest that EVT may be a better treatment option for patients with PC‐LVO; the MEDT group performed better only in terms of rates of sICH.

In recent years, the success of EVT for the treatment of AC‐LVO stroke has led to better functional outcomes for patients, especially those presenting outside of the traditional 3‐ to 4.5‐hour window for intravenous thrombolysis.1, 2, 3, 4, 5, 6, 7, 8 As neurointerventional technique and device development continues, focus must be placed on determining the applicability of successful procedures to patient cohorts with inconclusive treatment decision algorithms.

Numerous retrospective and single‐arm studies have found EVT to be effective for PC‐LVO stroke treatment, but a lack of evidence from published prospective trials and RCTs remains to demonstrate EVT efficacy in this population.19, 20, 21, 22, 23 We evaluated and pooled 3 studies that allow this comparison, with a separate sensitivity analysis to study the 2 RCTs alone given the discordant interpretation of the prospective cohort study data.9, 10, 11 Our results failed to show a convincing benefit for EVT compared with MEDT for PC‐LVO stroke treatment, but there are several explanations for our unexpected findings. In both the global meta‐analyses and the RCT‐restricted analyses, the differences in rates of 90‐day mRS score of 0 to 3 or 90‐day functional independence between the EVT and the MEDT group were not significant. However, the magnitude and directionality of these differences must be noted. In the RCT‐restricted analysis, the EVT group had a 90‐day mRS score of 0 to 3 rate of 45.1% compared with 36.9% for the MEDT group. Similarly, the 90‐day functional independence (mRS score of 0–2) rate in the EVT group was 31.8% compared with 19.7% for the MEDT group. Future RCTs on this topic may add sufficient power to this analysis to produce significant and clinically relevant findings for patients presenting with PC stroke.

Safety outcomes were also evaluated in our meta‐analysis, which produced similarly inconclusive results. Mortality was not significantly different between groups irrespective of global meta‐analysis or RCT‐restricted analysis. This is consistent with existing literature on EVT treatment in AC‐LVO stroke.1, 2, 3, 4, 5, 6, 7 With respect to sICH, discordant results were found between our global meta‐analysis, which showed a significant increase in rate of sICH with EVT, and our RCT‐restricted analysis, which found no significant differences between the 2 groups. As with functional outcomes, the magnitude and directionality of this trend must be noted. If future RCTs similarly report a greater risk of sICH with EVT treatment in PC‐LVO stroke, it would be valuable to identify the physiological and mechanical factors behind the difference in these rates compared with EVT in AC‐LVO stroke, which has repeatedly been shown to have similar sICH rates to medical therapy alone.1, 2, 3, 4, 5, 6, 7, 8 It is possible that an increased frequency of underlying intracranial atherosclerotic disease in patients with PC‐LVO may be an important contributor to high sICH rates.24

The selection of patients for inclusion in the 3 studies in our meta‐analysis may be overly broad, which may limit our analysis in the detection of significant outcome differences between patients undergoing EVT and MEDT. The patients included in these studies required evidence of a basilar artery occlusion on computed tomography angiography, but there was no qualitative or quantitative estimate of ischemic injury on the preenrollment noncontrast head computed tomography or standardized computed tomography perfusion estimation of ischemic injury. These limitations are similar to prior criticisms of the International Management of Stroke III (IMS‐III) and the Intra‐arterial Versus Systemic Thrombolysis for Acute Ischemic Stroke (SYNTHESIS EXPANSION) studies, which failed to show a difference between EVT and MEDT for patients with AC‐LVO ischemic strokes.25, 26 It is possible that more refined imaging selection of patients before EVT treatment of PC‐LVO may lead to a greater treatment impact and demonstration of EVT efficacy. For example, we see a benefit in future studies using new advanced imaging techniques, such as computed tomography perfusion or magnetic resonance imaging, to delineate the ischemic core before treatment in this population.

Several limitations of our meta‐analysis must be noted. A major limitation is the number of studies and patients with available data, which has likely resulted in an underpowered study. Another limitation is that Zi et al is a nonrandomized study and had a substantially different interpretation of findings in comparison to the RCTs included in the meta‐analysis, which required us to perform RCT‐restricted sensitivity analyses. We were also unable to make definitive conclusions or stratify results based on the specific type of endovascular or nonendovascular therapy. Although most landmark AC‐LVO stroke studies defined the endovascular arm as “mechanical thrombectomy,” the studies included in this article were inclusive of other types of endovascular treatment, such as intra‐arterial thrombolysis. Because of our inability to access individual patient‐level data, we were unable to rigorously control for differences in therapies and various other patient and study characteristics. Furthermore, the lack of individual patient‐level data prevents us from being able to definitively remove patient crossover between groups as a limiting factor. Despite these limitations, we were able to attenuate the risk of violating the assumption of exchangeability of studies by performing sensitivity analyses on RCTs. Results from this meta‐analysis show that current pooled sample sizes from existing RCTs are insufficient to confirm the superiority or noninferiority of EVT in comparison to medical therapy for treatment of acute basilar artery ischemic stroke. Future RCTs on this topic should aim for a larger sample size to achieve a confirmatory result.

Conclusions

In this meta‐analysis of the BEST, BASICS, and BASILAR studies, EVT was not shown to be superior to MEDT alone in improving the long‐term functional outcomes of patients presenting with PC‐LVO stroke. The magnitude and directionality of our results suggest, however, that more robust RCTs may show EVT to be a clinically superior treatment and further clarify the clinical decision‐making for this patient cohort.

Sources of funding

None.

Conflict of interest

J. M. Pederson works for and holds equity in Nested Knowledge, Inc, and Superior Medical Experts, Inc. K. Hutchison works for Nested Knowledge, Inc. K. M. Kallmes works for and holds equity in Nested Knowledge, Inc, works for Conway Medical LLC, and holds equity in Superior Medical Experts, Inc. N. Hardy works for and holds equity in Nested Knowledge, Inc.

Acknowledgements

None.

Footnotes

*Correspondence to: Jeremy Heit, MD, PhD, Center for Academic Medicine, Department of Radiology, Stanford University, MC 5659, 453 Quarry Rd, Palo Alto, CA 94304.
E‐mail:

References

  • 1 Berkhemer OA, Fransen PSS, Beumer D, Van Den Berg LA, Lingsma HF, Yoo AJ, Schonewille WJ, Vos JA, Nederkoorn PJD, Wermer MJH, et al. A randomized trial of intraarterial treatment for acute ischemic stroke. N Engl J Med. 2015; 372:11–20.CrossrefMedlineGoogle Scholar
  • 2 Goyal M, Demchuk AM, Menon BK, Eesa M, Rempel JL, Thornton J, Roy D, Jovin TG, Willinsky RA, Sapkota BL, et al. Randomized assessment of rapid endovascular treatment of ischemic stroke. N Engl J Med. 2015; 372:1019–1030.CrossrefMedlineGoogle Scholar
  • 3 Jovin TG, Chamorro A, Cobo E, De Miquel MA, Molina CA, Rovira A, San Román L, Serena J, Abilleira S, Ribó M, et al. Thrombectomy within 8 hours after symptom onset in ischemic stroke. N Engl J Med. 2015; 372:2296–2306.CrossrefMedlineGoogle Scholar
  • 4 Saver JL, Goyal M, Bonafe A, Diener H‐C, Levy EI, Pereira VM, Albers GW, Cognard C, Cohen DJ, Hacke W, et al. Stent‐retriever thrombectomy after intravenous t‐pa vs. T‐pa alone in stroke. N Engl J Med. 2015; 372:2285–2295.CrossrefMedlineGoogle Scholar
  • 5 Campbell BCV, Mitchell PJ, Kleinig TJ, Dewey HM, Churilov L, Yassi N, Yan B, Dowling RJ, Parsons MW, Oxley TJ, et al. Endovascular therapy for ischemic stroke with perfusion‐imaging selection. N Engl J Med. 2015; 372:1009–1018.CrossrefMedlineGoogle Scholar
  • 6 Nogueira RG, Jadhav AP, Haussen DC, Bonafe A, Budzik RF, Bhuva P, Yavagal DR, Ribo M, Cognard C, Hanel RA, et al. Thrombectomy 6 to 24 hours after stroke with a mismatch between deficit and infarct. N Engl J Med. 2018; 378:11–21.CrossrefMedlineGoogle Scholar
  • 7 Albers GW, Marks MP, Kemp S, Christensen S, Tsai JP, Ortega‐Gutierrez S, Mctaggart RA, Torbey MT, Kim‐Tenser M, Leslie‐Mazwi T, et al. Thrombectomy for stroke at 6 to 16 hours with selection by perfusion imaging. N Engl J Med. 2018; 378:708–718.CrossrefMedlineGoogle Scholar
  • 8 Powers WJ, Rabinstein AA, Ackerson T, Adeoye OM, Bambakidis NC, Becker K, Biller J, Brown M, Demaerschalk BM, Hoh B, et al. 2018 Guidelines for the early management of patients with acute ischemic stroke: a guideline for healthcare professionals from the American Heart Association/American Stroke Association. Stroke. 2018; 50(12):e344–e418.Google Scholar
  • 9 Liu X, Dai Q, Ye R, Zi W, Liu Y, Wang H, Zhu W, Ma M, Yin Q, Li M, et al. Endovascular treatment versus standard medical treatment for vertebrobasilar artery occlusion (BEST): an open‐label, randomised controlled trial. Lancet Neurol. 2020; 19:115–122.CrossrefMedlineGoogle Scholar
  • 10 Langezaal LCM, Van Der Hoeven EJRJ, Mont'alverne FJA, De Carvalho JJF, Lima FO, Dippel DWJ, Van Der Lugt A, Lo RTH, Boiten J, Lycklama Nijeholt GJ, et al. Endovascular therapy for stroke due to basilar‐artery occlusion. N Engl J Med. 2021; 384:1910–1920.CrossrefMedlineGoogle Scholar
  • 11 Writing Group for the BG , Zi W, Qiu Z, Li F, Liu H, Liu W, Huang W, Shi Z, Bai Y, Liu Z, Wang L, et al. Assessment of endovascular treatment for acute basilar artery occlusion via a nationwide prospective registry. JAMA Neurol. 2020.Google Scholar
  • 12 Scottish Intercollegiate Guidelines Network SIGN 50: A Guideline Developer's Handbook. 2nd ed. Scottish Intercollegiate Guidelines Network; 2011.Google Scholar
  • 13 Balduzzi S, Rucker G, Schwarzer G. How to perform a meta‐analysis with R: a practical tutorial. Evid Based Ment Health. 2019; 22:153–160.CrossrefMedlineGoogle Scholar
  • 14 Viechtbauer W. Conducting meta‐analyses in R with the Metafor package. J Stat Softw. 2010; 36:1–48.CrossrefGoogle Scholar
  • 15 Viechtbauer W. Confidence intervals for the amount of heterogeneity in meta‐analysis. Stat Med. 2007; 26:37–52.CrossrefMedlineGoogle Scholar
  • 16 Hartung J, Knapp G. A refined method for the meta‐analysis of controlled clinical trials with binary outcome. Stat Med. 2001; 20:3875–3889.CrossrefMedlineGoogle Scholar
  • 17 Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta‐analyses. BMJ. 2003; 327:557–560.CrossrefMedlineGoogle Scholar
  • 18 Von Kummer R, Broderick JP, Campbell BCV, Demchuk A, Goyal M, Hill MD, Treurniet KM, Majoie CBLM, Marquering HA, Mazya MV, et al. The Heidelberg bleeding classification. Stroke. 2015; 46:2981–2986.AbstractGoogle Scholar
  • 19 Gory B, Mazighi M, Blanc R, Labreuche J, Piotin M, Turjman F, Lapergue B. Mechanical thrombectomy in basilar artery occlusion: influence of reperfusion on clinical outcome and impact of the first‐line strategy (ADAPT vs stent retriever). J Neurosurg. 2018; 129:1511–1521.Google Scholar
  • 20 Kaneko J, Ota T, Tagami T, Unemoto K, Shigeta K, Amano T, Ueda M, Matsumaru Y, Shiokawa Y, Hirano T. Endovascular treatment of acute basilar artery occlusion: Tama‐REgistry of Acute Thrombectomy (TREAT) study. J Neurol Sci. 2019; 401:29–33.Google Scholar
  • 21 Shu L, Salehi Ravesh M, Jansen O, Jensen‐Kondering U. Stent retriever thrombectomy potentially increases the recanalization rate, improves clinical outcome, and decreases mortality in acute basilar occlusion: a systematic review and meta‐analysis. Cerebrovasc Dis Extra. 2019; 9:46–56.Google Scholar
  • 22 Hu SY, Yi HJ, Lee DH, Hong JT, Sung JH, Lee SW. Effectiveness and safety of mechanical thrombectomy with stent retrievers in basilar artery occlusion: comparison with anterior circulation occlusions. J Korean Neurosurg Soc. 2017; 60:635–643.Google Scholar
  • 23 Kang DH, Jung C, Yoon W, Kim SK, Baek BH, Kim JT, Park MS, Kim YW, Hwang YH, Kim YS, et al. Endovascular thrombectomy for acute basilar artery occlusion: a multicenter retrospective observational study. J Am Heart Assoc. 2018; 7:e009419.LinkGoogle Scholar
  • 24 Kim JS, Nah HW, Park SM, Kim SK, Cho KH, Lee J, Lee YS, Kim J, Ha SW, Kim EG, et al. Risk factors and stroke mechanisms in atherosclerotic stroke: intracranial compared with extracranial and anterior compared with posterior circulation disease. Stroke. 2012; 43:3313–3318.AbstractGoogle Scholar
  • 25 Broderick JP, Palesch YY, Demchuk AM, Yeatts SD, Khatri P, Hill MD, Jauch EC, Jovin TG, Yan B, Silver FL, et al. Interventional management of stroke (IMS) III investigators. Endovascular therapy after intravenous t‐PA versus t‐PA alone for stroke. N Engl J Med. 2013; 368:893–903.CrossrefMedlineGoogle Scholar
  • 26 Ciccone A, Valvassori L, Nichelatti M, Sgoifo A, Ponzio M, Sterzi R, Boccardi E. SYNTHESIS expansion investigators. Endovascular treatment for acute ischemic stroke. N Engl J Med. 2013; 368:904–913.CrossrefMedlineGoogle Scholar