Efficacy and safety of bridging therapy and direct mechanical thrombectomy in large vessel occlusions

To the Editor: Acute large vessel occlusion (LVO) is responsible for most acute ischemic stroke (AIS), a common cause of disability and death worldwide. Randomized controlled clinical trials (RCTs) provided evidence endorsing intravenous thrombolysis (IVT) and endovascular thrombectomy over IVT alone as the current standard treatment for people with LVO in the anterior circulation.^[1] The current American and European guidelines recommend using IVT for all eligible individuals with LVO before direct mechanical thrombectomy (d-MT) (class of recommendation-I), also termed bridging therapy (BT). Recent RCTs suggest that MT was non-inferior to BT in terms of efficacy and safety,^[2,3] which contradict the results from multiple meta-analyses favoring BT over d-MT.^[4] The benefit of routine IVT for eligible individuals before thrombectomy has become controversial. BT is associated with complications, including the risk of vasospasm, distal emboli, or symptomatic intracranial hemorrhage.

Due to the considerable number of studies addressing this issue, we conducted a meta-analysis of observational studies and RCTs published up to October 30, 2021 to compare efficacy and safety outcomes after thrombectomy with and without prior IVT in comparable populations. Searching PubMed, Embase, Web of Science, and Cochrane databases to identify relevant studies, we found 3023 articles. Fifty clinical studies were eventually included (PRISMA flowchart, Supplementary Figure 1, https://links.lww.com/CM9/B300), and the main characteristics of eight studies are summarized in Supplementary Table 1 (https://links.lww.com/CM9/B300). We used the Cochrane Collaboration's tool to assess the risk of RCT bias and the Newcastle–Ottawa quality assessment scale to assess the risk of bias of cohort studies and case-control studies [Supplementary Table 2, https://links.lww.com/CM9/B300].

We collected information from each study on publication year, study design, sample size, countries included, individual characteristics (age, gender, and previous medical history), National Institute of Health Stroke Scale score, Alberta Stroke Program Early CT score, and symptom onset-to-groin timing (OGT). The primary outcome was functional independence at 90 days (modified Rankin Scale score 0–2). The secondary outcomes were (1) successful recanalization (modified Thrombolysis in Cerebral Ischemia score 2b-3), (2) 90-day mortality, (3) symptomatic intracranial hemorrhage, (4) large or malignant middle cerebral artery infarction, (5) any procedural complications, including vessel dissection, contrast extravasation, embolization into new territory, femoral access complications, and transient vasospasms, (6) vasospasm, and (7) emboli into an uninvolved territory.

We used odds ratio (OR) with 95% confidence intervals (CIs) to evaluate dichotomous variables for primary and secondary outcomes. Heterogeneity was assessed using χ² and I² statistics (P value < 0.05 and I² > 50% indicating significant heterogeneity). In this case, a random-effect model was used; otherwise, we used a fixed-effect model.

Subgroup analysis was predefined for (1) study design and (2) OGT based on data after propensity score (PS) matched. For OGT subgroup analysis, the difference between the two groups of no more than 30 min was considered similar.

Pooled OR from unadjusted analyses showed that people in the BT group achieved significantly better clinical outcomes at 90 days (functional independence) than those in the d-MT group (OR = 1.35 [95% CI, 1.21–1.51]) [Table 1]. The benefit of the bridging approach became unclear after adopting PS matched data for adjusted analyses (OR = 1.12 [95% CI, 0.98–1.26]) as well as analyses for RCTs (OR = 0.96 [95% CI, 0.76–1.16]). Subgroup analysis showed that functional independence frequency remained unclear between groups regardless of study design (five prospective, 12 retrospectives), and OGT [Supplemental Figures 2–10, https://links.lww.com/CM9/B300].

Table 1 - Pooled outcomes of bridging therapy and direct mechanical thrombectomy in large vessel occlusions by meta-analysis.

Outcomes	N of studies	Pooled OR	95% CI	I 2 (%)
Functional independence at 90 days
Unadjusted∗	21	1.35	1.21–1.51	40.0
PS matched	17	1.12	0.98–1.26	13.2
RCTs	7	0.96	0.76–1.16	16.5
Successful recanalization
Unadjusted∗	27	1.40	1.23–1.58	48.7
PS matched	15	0.98	0.77–1.18†	55.8
RCTs	6	1.07	0.76–1.39	0
Mortality at 90 days
Unadjusted∗	21	0.68	0.59–0.77	18.6
PS matched∗	17	0.70	0.60–0.80	0
RCTs∗	6	0.62	0.34–0.90†	53.4
Symptomatic intracerebral hemorrhage
Unadjusted	18	1.02	0.79–1.32	0
PS matched	15	0.94	0.72–1.15	39.2
RCTs	5	1.53	0.90–2.15	0

^∗Present results from random-effect model.

^†Statistically significant. CI: Confidence interval; OR: Odds ratio; PS: Propensity score; RCTs: Randomized controlled trials.

Pooled OR showed that individuals from the BT group have higher rates of successful recanalization in unadjusted analysis (OR = 1.40 [95% CI, 1.23–1.58]) [Table 1]. There was no significant difference for adjusted analysis (OR = 0.98 [95% CI, 0.77–1.18]) and RCTs between the two groups (OR = 1.07 [95% CI, 0.76–1.39]). The results remained similar after adopting subgroup analysis [Supplemental Figures 2–10, https://links.lww.com/CM9/B300].

Pooled OR from the adjusted, unadjusted, and RCTs analysis showed that people in the BT group had significantly lower mortality at 90 days than those in the d-MT group [Table 1]. Subgroup analysis indicated that the 90-day mortality was significantly lower in the BT group regardless of study design or OGT, except for one subgroup (when OGT of intervention group is less than that of control group; OR = 1.35 [95% CI, 0.66–2.03]) [Supplemental Figures 2–10, https://links.lww.com/CM9/B300].

The occurrence of symptomatic intracerebral hemorrhage showed no significant difference between the two groups in unadjusted analysis (OR = 1.02 [95% CI, 0.79–1.32]) and adjusted analysis (OR = 0.94 [95% CI, 0.72–1.15]) [Table 1]. The results remained similar after in the subset of RCTs. No significant difference was found in the subgroup analysis by study design or OGT [Supplemental Figures 2–10, https://links.lww.com/CM9/B300].

Pooled OR showed no significant difference in the occurrence of large or malignant middle cerebral artery infarction, complications, vasospasm, and emboli between the two groups [Supplemental Figures 2–10, https://links.lww.com/CM9/B300]. We have not done subgroup analysis for these outcomes.

Our study is an extensive meta-analysis assessing the comparative safety and efficacy outcomes of BT and d-MT, including data from the most recent RCTs. It included 50 clinical studies, with 12,166 individuals in the BT group and 12,497 in the d-MT group. In the unadjusted analysis, we found that the bridging approach had better functional outcomes and lowered mortality at 90 days than the d-MT group without increasing the risk of symptomatic hemorrhagic complications or other complications. The BT group also had a higher rate of successful recanalization than the d-MT group. The BT group's better functional outcomes and higher recanalization rate became unclear after adopting PS matched data and in the subset of RCT studies. These results added to the growing evidence supporting BT and d-MT as an effective treatment for AIS with LVOs.

Previous studies addressing the same issue reported inconsistent conclusions. Studies advocating BT claimed that, when preceding IVT, a bridging approach could accelerate the endovascular process and achieve successful recanalization by partially cracking and fragmenting the clot, not directly accessible by thrombectomy devices.^[5] Other studies recommending d-MT noted several problems in such studies, including bias in individual selection and exclusion, cost issues, and delays in treatment time from symptom onset due to the nature of BT in terms of mixing and infusion.^[2] We considered all these confounding factors and pooled our results with adjusted analysis, showing that BT and d-MT have similar therapeutic outcomes.

Most previous meta-analyses concluded that the people in the BT group obtained significantly better functional independence and higher recanalization rate at 90 days than those in the d-MT group;^[6] our findings did not support this. RCTs results and adjusted analysis suggested that d-MT is non-inferior to BT. This discrepancy in our findings could be explained by adjusting to eliminate bias and confounding factors.

Conversely, we observed no increase in the risk of post-MT complications of BT compared to d-MT with more included studies, participants, and outcome measurements, than a previous study with consistent conclusions.^[6]

Compared with multiple previously published meta-analysis, we have a few strengths. Analyses before 2018 did not systematically include sufficient efficacy and safety outcomes as we did. A few primary studies included in the meta-analysis did not use the latest MT devices, which limited the interpretability and generalizability of their results.^[7] We reduced the possibility of potential bias by avoiding them. A recent meta-analysis up to 2021, the latest meta-analysis of three RCTs, for instance, included fewer studies than we did.^[8] As the number of included studies increased, we could provide more reliable conclusions.

Our study has limitations. First, we have included several studies to increase the effective size and statistical power, but biases could still be introduced into the estimates due to missing data in some covariates and clinical outcomes. Second, participants who dropped out after some clinical improvements after IVT pretreatment were excluded from all of the studies, which may lead to an overestimation of BT benefit in our analysis. As included in studies published before October 30, 2021, people in the d-MT group undergoing different devices for thrombectomy would result in different clinical outcomes. It could contribute to the instability in our results of the d-MT group.

As ongoing trials (DIRECT-SAFE in Australia, SWIFT DIRECT in Switzerland) may provide further evidence for determining the best treatment strategy between BT and d-MT, future meta-analysis with detailed substratification is needed to enable individualized treatment for people with AIS.

Conflicts of interest

None.