Should the direct aspiration first pass technique be advocated over the stent-retriever technique for acute ischemic stroke? A systematic review and meta-analysis of 7692 patients.

BACKGROUND: The present meta-analysis aimed to synthesize evidence from all published studies with head-to-head data on the outcomes of a direct aspiration first pass technique (ADAPT) and the stent-retriever (SR) in acute ischemic stroke (AIS) patients.
METHODS: We searched PubMed, Scopus, Web of Science, and Cochrane Central Register of Controlled Trials from inception to March 2021 for relevant clinical trials and observational studies. Eligible studies were identified, and all relevant outcomes were pooled in the meta-analysis random-effects model of DerSimonian-Laird.
RESULTS: Thirty studies were included in the meta-analysis with a total of 7868 patients. Compared with the SR, the ADAPT provides slightly higher rates of successful recanalization (RR 1.06, 95% CI [1.02 to 1.10]) and complete recanalization (RR 1.20, 95% CI [1.01 to 1.43]) but with more need for rescue therapy (RR 1.81, 95% CI [1.29 to 2.54]). There were no significant differences between the two techniques in terms of mortality at discharge, mortality at 90 days, change in the National Institutes of Health Stroke Scale score, the favorable outcome (modified Rankin scale (mRS) of 0-2), time to the groin puncture, or frequency of complications as intracerebral hemorrhage (ICH), symptomatic intracranial hemorrhage (sICH), embolus in a new territory (ENT), hemorrhagic infarction, parenchymal hematoma, subarachnoid hemorrhage, or procedural complications (all P > 0.05).
CONCLUSION: Current evidence supports the use of the ADAPT technique to achieve successful and complete recanalization while considering the higher need for rescue therapy in some patients. Copyright:

INTRODUCTION

Acute ischaemic stroke (AIS) is considered to be one of the leading causes of the mortality worldwide.[1] Moreover, stroke is associated with a high rate of disability among the survivors.[8]

There are two major methods to treat AIS; mechanical thrombectomy and standard medical therapy, which can be considered if patients present to a thrombectomy-capable facility in a timely manner. Over the last decade, several randomized controlled trials (RCTs) showed the superiority of mechanical thrombectomy techniques over medical management in thrombectomy-eligible patients.[6,11,34] These RCTs have drawn more attention to mechanical thrombectomy, which is now considered one of the main lines of treatment for AIS.[35]

The reference mechanical thrombectomy technique is the stent-retriever (SR) for selected patients, especially those with AIS due to vascular occlusion in the anterior circulation, according to the results of a recently published meta-analysis by Sivan-Hoffmann et al.,[35] which showed that the SR is a safe method with favorable clinical outcomes. The SR technique is the gold standard technique for mechanical thrombectomy in patients with AIS.

Recently, a direct aspiration first pass technique (ADAPT) was proposed by Turk et al.[39] and started to gain acceptance due to the lower costs of the procedure, increased rates of successful recanalization, and the better clinical outcomes reported by some studies when used either alone or as an adjunctive technique to the SR.[40] However, data from the literature are controversial regarding the comparative outcomes of the ADAPT technique compared with the conventional thrombectomy technique of the SR.

Therefore, the present meta-analysis aimed to synthesize evidence from all published studies with head-to-head data on the outcomes of ADAPT and SR in AIS patients.

MATERIALS AND METHODS

We followed the most recent version of the preferred reporting items for systematic reviews and meta-analysis (PRISMA statement 2020) guidelines during this systematic review and meta-analysis.[23]

Eligibility criteria

Studies satisfying the following criteria were included in this meta-analysis:

Population: studies on patients with AIS undergoing thrombectomy

Intervention: studies where the exposed group received ADAPT

Comparator: studies where the control group received SR

Outcome: studies reporting recanalization outcomes, National Institutes of Health Stroke Scale (NIHSS) score at baseline, 24 h and/or 7 days after the stroke event, mRS at 90 days, complications of the procedure and/or procedure time

Study design: studies with comparative designs, whether RCTs or observational studies comparing the outcomes of ADAPT and SR.

We excluded studies that were not in English language and studies on either ADAPT or SR without direct comparison between the two techniques.

Information sources

We performed a comprehensive search of four electronic databases (PubMed, Scopus, Web of Science, and Cochrane Central Register of Controlled Trials) from inception to March 1, 2021.

Search strategy

We used the following search query ([aspiration OR ADAPT] AND [SR OR Solitaire OR Trevo OR Merci] AND [stroke OR Large vessel occlusion OR LVO]) in the four databases with no filters or limitations.

Selection process

Retrieved records from the literature search were screened in two steps. In the first step, the title and abstracts of all articles were screened for eligibility. Then, the full-text articles of the eligible abstracts were retrieved and further screened for eligibility.

Data collection process and data items

Data were extracted to a uniform data extraction sheet. The extracted data included (1) Characteristics of the included studies, (2) Characteristics of the population of included studies, (3) Risk of bias domains, and (4) Outcome measures.

Study risk of bias assessment

For RCTs, we used the Cochrane Risk of Bias assessment tool (ROB 2.0) while for observational studies; we used the Newcastle Ottawa Scale (NOS scale).

Effect measures

In this meta-analysis, we considered the following outcome measures:

Successful recanalization (%): defined angiographically, according to the modified thrombolysis in cerebral infarction (mTICI) scale, as mTICI2b/3 at the end of the procedure

Complete recanalization (%): defined angiographically, according to the mTICI scale, as mTICI-3 at the end of the procedure

Favorable neurological outcome (%): defined as the number of patients with an mRS score of 0-2 at 90 days after the stroke event

Complications (%): defined as any complication reported in the included studies with a frequency of more than 5%. This includes intracranial hemorrhage (ICH), symptomatic ICH (sICH), Subarachnoid haemorrhage (SAH), parenchymal hematoma, hemorrhagic infarction, and embolization to a new territory (ENT)

Procedural time (%): defined as the time interval from groin puncture to reperfusion time.

Synthesis methods

For outcomes that constitute continuous data, the mean difference (MD) between the two groups from the baseline to the endpoint, with its confidence interval (CI), was pooled in the DerSimonian-Laird random-effect model. In the case of studies reporting data in multiple time points, we considered the last endpoint for the primary analysis. For outcomes that constitute dichotomous data, the frequency of events and the total number of patients in each group were pooled as relative risk between the two groups in the DerSimonian-Laird random-effect model.

Subgroup analysis

We conducted subgroup analysis according to the study design (RCTs only vs. observational studies) and according to the site of stroke (anterior circulation vs. posterior circulation).

Assessing the heterogeneity

Heterogeneity (non-combinability) of the included studies and subgroups was examined by visual inspection of the forest plot and assessed by the Cochrane Q and I-square tests using RevMan version 5.3 for windows. For heterogeneity testing, a P < 0.1 and I-square >50% were considered for significant heterogeneity.

Calculating the missing data

When the MD from baseline to endpoint was not provided, we calculated it from the pre- and post-treatment means (MD = Posttreatment-pretreatment). Then, we calculated the MD between the ADAPT and SR groups as follows: (MD = MD experiemental-MD placebo).

When the standard error (SE) of MD was not provided, we calculated it from the standard deviation [SE=SD⁄(√n)], 95% CI ([upper limit-lower limit]⁄3.92), or 90% CI ([upper limit-lower limit]⁄3.29).

For studies and groups with a sample size of <60 patients, the numbers (3.92 and 3.29) were substituted by a value from the table of t distributions with degrees of freedom equal to the group sample size minus 1.

Reporting bias assessment

To explore the publication bias across studies, we constructed funnel plots to present the relationship between effect size and SE. Two methods assessed evidence of publication bias; (1) Egger’s regression test and (2) Begg and Mazumdar rank correlation test (Kendall’s tau).

Certainty assessment

To test the robustness of the evidence, we conducted a certainty assessment through sensitivity analysis (also called, leave-one-out meta-analysis). For every outcome in the meta-analysis, we run sensitivity analysis in multiple scenarios excluding one study in each scenario to make sure the overall effect size was not dependent on any single studies.

RESULTS

Study selection

Our literature search process retrieved 2,832 records. Following titles and abstract screening, 351 articles were eligible for full-text screening. From these 351 studies, 30 studies were included in the meta-analysis.[2-5,7,9-10, 12,14-22,24-28,30,32,33,36-38,40,41] The references of the included studies were manually searched, and no further articles were included. The flow chart of the study selection process is shown in the PRISMA flow diagram in [Figure 1].

Figure 1:

PRISMA flow diagram of the study selection process.

Study characteristics

The population of the studies was homogenous; all studies enrolled 7868 patients with AIS. Two studies were RCTs while 28 studies were observational studies. The characteristics of the included studies are summarized in [Table 1], while summary and baseline characteristics of populations of these studies are shown in [Table 2].

Table 1:

Characteristics of the included studies.

Table 2:

The characteristics of the included studies’ populations.

Risk of bias within studies

The quality of included studies ranged from moderate to high quality according to the Cochrane Risk of Bias assessment tool for RCTs and the NOS for the observational studies.

Improvement in NIHSS score

The overall standardized MD (SMD) of improvement in the NIHSS from baseline did not favor either of the two techniques (SMD 0.01, 95% CI [–0.11 to 0.13]). Subgroup analysis did not show any difference between the two techniques after 24 h (SMD 0.01, 95% CI [–0.14 to 0.15]), after 7 days (SMD –0.04, 95% CI [–0.26 to 0.18]) or at discharge (SMD 0.08, 95% CI [–0.25 to –0.41]), [Figure 2] and [Supplementary File 1]. Subgroup analysis of the improvement in NIHSS according to the study design did not show the superiority of either of the two techniques in the subgroup of the RCTs (SMD –0.01, 95% CI [–0.16 to 0.15]) or in the subgroup of the observational studies (SMD 0.05, 95% CI [–0.16 to 0.25]) [Figure 2] and [Supplementary File 1].

Figure 2:

Forest plot of the pooled SMD of change in National Institutes of Health Stroke Scale score overall and subgroup by different time points, study designs, and location of vascular occlusion; SMD = standardized mean difference of the change from baseline to endpoint between the A direct aspiration first pass technique and Stent-retriever groups.

Similarly, subgroup analysis of the improvement in the NIHSS according to the site of vascular occlusion did not show superiority of either of the two techniques for the 4 studies conducted on patients with AIS in the anterior circulation occlusion (SMD –0.03, 95% CI [–0.16 to 0.09]), the two studies conducted on patients with basilar artery occlusion (SMD –0.11, 95% CI [–0.45 to 0.24]) or the two studies with (unspecified) intracranial arterial occlusion (SMD 0.36, 95% CI [0.04 to 0.68]). The pooled effect estimates on all subgroups were homogenous (Chi-square P > 0.1) [Figure 2] and [Supplementary File 1].

Time to groin puncture

The pooled analysis of the time to groin puncture reported by 18 studies (n = 5729 patients) did not favor either of the two techniques (SMD 0.81, 95% CI [–11.78 to 13.40]). The difference was not statistically significant in the subgroups of the RCTs or the observational studies [Supplementary File 2].

Successful recanalization (TICI 2b-3)

Twenty-nine studies (n = 7560 patients) reported the frequency of achieving successful recanalization (TICI 2b-3) by both techniques. The pooled risk ratio (RR) of successful recanalization (TICI 2b-3) favored aspiration thrombectomy technique (RR 1.06, 95% CI [1.02 to 1.10]) [Figure 3] and [Supplementary File 3].

Figure 3:

Forest plot summarizing the pooled RR of the dichotomous study outcomes between the A direct aspiration first pass technique and stent-retriever groups; outcomes in green are statistically significant. The green colour means a significant difference exists while the red colour means no significant difference exists.

Subgroup analysis of the rates of successful recanalization (TICI 2b-3) according to the study design showed that the observational studies (1.06, 95% CI [1.02 to 1.11]) but not the RCTs (1.05, 95% CI [0.97 to 1.15]) had statistically significant pooled RR in favor of the ADAPT group.

Complete recanalization (TICI 3)

Seventeen studies (n=3824 patients) reported the frequency of achieving complete perfusion (mTICI 3) by both techniques, the pooled RR of complete perfusion (mTICI 3) favored the ADAPT technique (RR 1.20, 95% CI [1.01 to 1.43]) [Figure 3] and [Supplementary File 4].

Subgroup analysis of the complete perfusion (mTICI 3) according to the study design showed that the pooled RR did not favor any of both techniques either in the RCTs (RR 1.10, 95% CI [0.83 to 1.46]) or in the observational studies (RR 1.22, 95% CI [1.0 to 1.49]).

Mortality

Five studies (n = 729 patients) reported the in-hospital mortality while ten studies (n = 2901 patients) reported the 90-day mortality. Neither the pooled RR of the in-hospital mortality nor the 90-day mortality favored either of the two groups (RR 0.89 and RR 0.92; both P > 0.05) [Figure 3] and [Supplementary Files 5 and 6].

Favorable outcome (mRS of 0-2)

The frequency of patients with favorable outcome according to the mRS score (ranging from 0 to 2) was reported by 22 studies (n = 6244 patients), the pooled RR of favorable outcome (mRS 0-2) did not favor either of the two techniques (RR 0.99, 95% CI [0.93 to 1.05]) [Figure 3] and [Supplementary File 7].

Rescue therapy

The frequency of the patients who required rescue therapy was reported by 15 studies (n = 3079 patients), the pooled RR of rescue therapy showed that more patients in the ADAPT group required rescue therapy compared with the SR group (RR 1.81, 95% CI [1.29 to 2.54]) [Figure 3] and [Supplementary File 8].

Prior thrombolysis

The frequency of patients with prior thrombolysis was reported by 23 studies (n = 6763 patients), the pooled RR of prior thrombolysis did not favor either of the two techniques (RR 1.00, 95% CI [0.93 to 1.07]) [Figure 3] and [Supplementary File 9].

General anesthesia

The frequency of patients who underwent general anesthesia was reported by 7 studies (n = 3561 patients), the pooled RR of the frequency of general anesthesia did not favor either of the two techniques (RR 0.81, 95% CI [0.48 to 1.36]) [Figure 3] and [Supplementary File 10].

Intracerebral haemorrhage

The frequency of patients with ICH was reported by 8 studies (n = 2063 patients), the pooled RR of intracranial haemorrhage did not favor either of the two techniques (RR 1.22, 95% CI [0.90 to 1.66]) [Figure 3] and [Supplementary File 11].

Occurrence of embolus in a new territory

The frequency of patients with ENT was reported by 11 studies (n = 1876 patients), the pooled RR of occurrence of an embolus in a new territory did not favor either of the two techniques (RR 1.13, 95% CI [0.73 to 1.73]) [Figure 3] and [Supplementary File 12].

Symptomatic intracerebral haemorrhage

The frequency of patients with sICH was reported by 14 studies (n = 4504 patients), the pooled RR of symptomatic intracranial haemorrhage did not favor either of the two techniques (RR 0.91, 95% CI [0.54 to 1.54]) [Figure 3] and [Supplementary File 13].

Haemorrhagic infarction

The frequency of the patients with haemorrhagic infarction was reported by 3 studies (n = 626 patients), the pooled RR of haemorrhagic infarction did not favor either of the two techniques (RR 1.04, 95% CI [0.63 to 1.72]) [Figure 3] and [Supplementary File 14].

Parenchymal hematoma

The frequency of patients with parenchymal hematoma was reported by 9 studies (n = 1389 patients), the pooled RR of parenchymal hematoma did not favor either of the two techniques (RR 0.83, 95% CI [0.58 to 1.19]) [Figure 3] and [Supplementary File 15].

SAH

The frequency of SAH in the two groups was reported by 9 studies (n = 1289 patients), the pooled RR of SAH did not favor either of the two techniques (RR 0.78, 95% CI [0.45 to 1.37]) [Figure 3] and [Supplementary File 16].

Procedural complications

The frequency of patients with procedural complications was reported by 9 studies (n = 3916 patients), the pooled RR of procedural complications did not favor either of the two techniques (RR 0.93, 95% CI [0.80 to 1.08]) [Figure 3] and [Supplementary File 17].

We conducted subgroup analysis for the main outcomes according to the study design (RCTs only vs. observational studies vs. all studies). Data showed consistent results in both RCTs and observational studies except for the outcome of successful recanalization (TICI 2b-3) where RCTs showed no difference between the ADAPT, and the SR while observational studies reported significantly higher successful recanalization rates in ADAPT compared with the SR [Table 3].

Table 3:

Summary of the subgroup analysis results, data are stratified according to the study design into RCTs only, observational studies only, and all studies.

Further, we conducted subgroup analysis for the main outcomes according to the site of stroke (anterior circulation vs. posterior circulation). Data were consistent in the subgroups of anterior circulation and posterior circulation except that the risks of emboli in a new terrorist was significantly lower in the ADAPT compared with the SR in the subgroup of posterior circulation [Table 4].

Table 4:

Summary of the subgroup analysis results, data are stratified according to the site of stroke into anterior circulation, posterior circulation, and all studies.

DISCUSSION

The development of mechanical thrombectomy technology has revolutionized the treatment of patients with AIS. The SR technique is the current gold-standard mechanical thrombectomy technique used in patients with AIS. However, despite the results of RCTs showing favorable outcomes of this technique, the rate of successful revascularization is still considered suboptimal. Therefore, using large-bore aspiration catheters during routine clinical practice has been widely debated in medical literature.

The ADAPT method was developed by Turk et al.[13] to achieve a higher recanalization rate in a shorter period of time. The method was initially described as fast, safe, simple, and effective.[39] However, the major limitation of the ADAPT method was the unavailability of the catheter technology needed to perform such a procedure. This technique started to gain acceptance recently after solving the catheter availability problem by developing the latest generation of tractable large-bore aspiration catheter, which provides sufficient aspiration force and easy manoeuvrability to navigate through the cerebral vasculature. However, it is still debatable whether the ADAPT technique should be preferred over the conventional SR technique for AIS owing to the inconsistent data reported in the literature. Therefore, we conducted this meta-analysis to synthesize evidence from published studies on ADAPT outcomes compared with SR using data from head-to-head comparative studies.

Our meta-analysis provides evidence that the ADAPT technique achieves slightly higher rates of successful recanalization and complete recanalization than the conventional SR technique. However, the subgroup analysis showed that this significant effect size in successful recanalization was mainly driven by observational studies but not RCTs. On the contrary, the ADAPT method was associated with higher need for rescue therapy (defined as the use of another endovascular strategy after failure of the initially used technique [mTICI 0-2a]) as compared with the SR. There were no significant differences between the two techniques in terms of mortality at discharge, mortality after 90 days, change in NIHSS score, the favorable outcome (mRS of 0-2), time to the groin puncture, or frequency of complications as ICH, sICH, the occurrence of an embolus in a new territory, hemorrhagic infarction, parenchymal hematoma, SAH, and procedural complications.

The main finding of our meta-analysis that ADAPT provides higher successful recanalization rates compared with the conventional SR is concordant with the findings of Phan et al.[31] and Ye et al.[42] The meta-analysis of Phan et al.[31] provided an indirect comparison between the two techniques, pooling single-arm data into two subgroups. Our findings are consistent with Phan et al.[31] that ADAPT provides a higher successful recanalization rate, but we found a superiority of ADAPT in the rate of complete perfusion, which was not significant in their meta-analysis. A limitation of their meta-analysis methods was the indirect comparison between the two arms, limiting the generalizability of their findings. Our meta-analysis provides more robust evidence by including comparative data from a head-to-head comparison between the two techniques.

Our finding that the ADAPT provides a higher successful recanalization rate and a higher complete recanalization rate (mTICI 3) is reasonable and is supported by the previous literature. Phan et al.[31] reported from an indirect meta-analysis that ADAPT patients tended to have more excellent neurologic outcomes (P = 0.11), although the difference was not statistically significant. There is now a general understanding that patients with mTICI 3 are more likely to have the excellent neurologic outcome (mRS 0-1) after a stroke event. Therefore, our findings that ADAPT provides a higher complete recanalization rate corroborates the trend analysis of excellent neurologic outcome (mRS 0-1) reported by Phan et al.[31] previously. While this difference in complete recanalization rate was not statistically significant in previous individual studies, our meta-analysis provides a larger sample size and higher statistical power to allow small effect estimates to be detectable. However, the clinical significance of this slight difference remains questionable.

In terms of the need for rescue therapy, Gory et al.,[10] Lapergue et al.,[21] and Nishi et al.[29] reported that more patients in the ADAPT group required rescue therapy which is in line with our findings. Ye et al.[42] did not find any differences between the ADAPT and the SR groups in terms of the need for rescue therapy. However, our meta-analysis showed that patients who underwent the ADAPT required more rescue therapy than those who underwent the SR technique. This discrepancy could be justified by our meta-analysis pooling data from a larger set of studies (30 studies) and a larger sample size (7868 patients), which provides the high statistical power to detect small differences between the thrombectomy techniques.

In [Table 5], we provide a summary of findings from 3 previous meta-analyses that directly and indirectly provided partial evidence on the comparison between ADAPT and SR for AIS. However, it is noteworthy that these studies have major limitations, including (1) dropping significant portions of the literature at the screening process, or (2) the authors selected a particular type of stroke patients to study; Ye et al.[42] compared the ADAPT and SR for acute basilar artery occlusion, or (3) providing indirect rather than a direct head-to-head comparison between the two arms, which is unreliable method to establish the superiority of a technique as long as direct evidence exists in the literature.

Table 5:

A summary of the findings of previous meta-analyses.

We conducted subgroup analysis according to the study design and site of stroke. These results were consistent across the strata except that the difference in the successful recanalization was significant in observational studies but not in the RCTs suggesting that this difference was mainly driven by observational studies which have less internal validity and higher risk of confounders compared with well-designed RCTs. Besides, in the subgroup of posterior circulation occlusion, the occurrence of ENT was significantly lower in the ADAPT group compared with the SR. These results are consistent with the findings of Ye et al.[42] who meta-analyzed data from 5 studies to compare both ADAPT and SR in basilar artery occlusion. The authors explained this difference by the fact that ADAPT does not require passing through the thrombus and therefore carries a lower risk of thrombus fragmentation, in addition to the lower risks of endothelial injury with ADAPT as reported in studies on experimental animal models. However, this difference in emboli occurrence was not observed in the case of anterior circulation. This variation in the risk of ENT in anterior and posterior circulations remains open for discussion.

The strengths of our meta-analysis are the following: (1) we ran an extensive search on multiple medical electronic databases; (2) we included all observational studies and clinical trials comparing the two techniques; (3) we followed the guidelines of the Cochrane Handbook for Systematic Reviews of Interventions when conducting this systematic review and we reported this manuscript according to the PRISMA statement. The major limitation of our meta-analysis is that most of the included studies are observational; it is known that observational studies might suffer from confounding bias and are not reliable in establishing a causal relationship between the intervention and the clinical outcome. Only two studies were described as well-designed RCTs, and 28 studies were observational studies, which invites future research to compare both techniques in RCT design to confirm and update our findings.

CONCLUSION

Current evidence supports the use of the ADAPT technique to achieve successful and complete recanalization while considering the higher need for rescue therapy in some patients.