Article Text

Download PDFPDF
Original research
Endovascular therapy for middle cerebral artery M2 segment occlusion: subanalyses of RESCUE-Japan Registry 2
  1. Masatomo Miura1,2,3,
  2. Shinichi Yoshimura1,
  3. Nobuyuki Sakai4,
  4. Hiroshi Yamagami5,
  5. Kazutaka Uchida1,6,
  6. Yoichiro Nagao1,3,
  7. Takeshi Morimoto6
  1. 1 Department of Neurosurgery, Hyogo College of Medicine, Nishinomiya, Japan
  2. 2 Department of Neurology, Japanese Red Cross Kumamoto Hospital, Kumamoto, Japan
  3. 3 Department of Neurology, Graduate School of Medical Sciences, Kumamoto University, Kumamoto, Japan
  4. 4 Department of Neurosurgery, Kobe City Medical Center General Hospital, Kobe City, Hyogo, Japan
  5. 5 Division of Stroke Care Unit, NCVC, Suita, Osaka, Japan
  6. 6 Department of Clinical Epidemiology, Hyogo College of Medicine, Nishinomiya, Hyogo, Japan
  1. Correspondence to Professor Shinichi Yoshimura, Department of Neurosurgery, Hyogo College of Medicine, Nishinomiya 663-8131, Japan; hyogoneuro{at}yahoo.co.jp

Abstract

Objective To compare the efficacy of endovascular therapy (EVT) with that of medical treatment in ’real-world ’patients with M2 occlusion.

Methods This was a post hoc analysis of the Recovery by Endovascular Salvage for Cerebral Ultra-acute Embolism Japan Registry 2. Among 2420 patients in the registry, we evaluated patients with isolated M2 occlusion and those with functional independence before the stroke. Multivariable logistic regression analysis was used to evaluate and compare clinical outcomes between EVT and medical treatment. Additional propensity score-matched (PSM) analyses were performed. We performed subgroup analyses of the primary outcome (modified Rankin Scale score 0–2 at 90 days) using forest plots of treatment effects.

Results Overall, 372 patients with M2 occlusion (n=184 EVT; n=188, medical treatment) were evaluated. The EVT group had a higher baseline National Institutes of Health Stroke Scale score (median (IQR), 15 [9–19] vs 10 [5–16]) and earlier onset to hospital door time (110 [50–258] vs 150 [60–343] min) than the medical treatment group. After adjustment, EVT was significantly associated with higher odds of primary outcome (adjusted OR=2.09; 95% CI 1.26 to 3.47) and lower odds of mortality at 90 days (adjusted OR= 0.27; 95% CI 0.08 to 0.93). After PSM analyses (184 patients were 1:1 matched with each group), EVT was effective and safe relative to medical treatment. Effects favoring EVT were present in several subgroups of interest.

Conclusion In patients with M2 occlusion, our registry suggests that EVT is effective and safe.

  • Endovascular therapy
  • middle cerebral artery occlusion
  • stroke
  • thrombectomy

Statistics from Altmetric.com

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Introduction

Several recent randomized clinical trials (RCTs) of patients with acute ischemic stroke due to anterior large vessel occlusion (LVO) have demonstrated the efficacy and safety of endovascular therapy (EVT).1–7 However, they focused on anterior LVO involving the internal carotid artery and proximal middle cerebral artery (MCA). The HERMES collaboration, a meta-analysis of five RCTs (MR CLEAN, ESCAPE, REVASCAT, SWIFT PRIME, and EXTEND-IA) showed that only 94 patients had M2 segment occlusion (51 of these patients received EVT).8 The HERMES collaboration, therefore, failed to show treatment benefits in patients with M2 segment occlusion. The efficacy and safety of EVT for MCA M2 segment occlusion remain debatable.

Although several studies have reported the efficacy and safety of EVT for patients with M2 segment occlusion, some of those studies were retrospective analyses or systematic reviews, whereas others were post hoc analyses of patient groups that received EVT.9–17 Of these, only a few studies compared the efficacy and safety of EVT with those of medical treatment. Sarraj et al carried out the only study that retrospectively compared the efficacy of EVT and medical treatment in patients with M2 occlusion.9

Our post hoc study evaluated the efficacy of EVT compared with that of medical treatment, including the use or non-use of intravenous tissue plasminogen activator (IV tPA), among real-world patients with MCA M2 segment occlusion.

Materials and methods

Description of the study and patient selection

We included patients from the Recovery by Endovascular Salvage for Cerebral Ultra-acute Embolism Japan Registry 2 (RESCUE-Japan Registry 2). The design of this registry, including patient eligibility and methods, has been described in detail elsewhere.18 The RESCUE-Japan Registry 2 was approved by the ethics committee or institutional review board at each participating center. The need for written informed consent from each patient was waived for the registry because we used clinical information obtained in routine clinical practice. The present post hoc study was approved by a local institutional review board.

In this study, we evaluated patients with isolated M2 segment occlusion, who had functional independence before the stroke, defined as a modified Rankin Scale (mRS) score <1 (possible range: 0 [normal] to 6 [death]). The M2 segment was defined as vertical MCA branches extending from the genu to the top of the Sylvian fissure and circular sulcus.18 We excluded patients with tandem lesions (eg, internal carotid artery and M2 segment occlusion).

Treatment and EVT

Treatment methods (EVT, IV tPA, or medical treatment) were determined according to the Japan Stroke Society guideline. Criteria of candidates for EVT were changed after March 2015 because of the publication of four RCTs of EVT for LVO.1–4 According to the Japan Stroke Society guideline, EVT for M2 occlusion may be reasonable for selected patients with mismatch between the ischemic core and neurological symptoms. Thus, the final decision on the treatment approach was determined by physicians, based on the mismatch between neurological symptoms and neuroimaging, and the affected eloquent area.

We divided patients with M2 occlusion into groups receiving EVT or medical treatment. Both medical treatment and EVT included the use or non-use of IV tPA. All devices used for EVT were approved, such as stent retrievers, aspiration catheters, angioplasty balloons, devices for local intra-arterial fibrinolysis, guidewires and/or microcatheters for piercing, or a combination of these. The stent retrievers used in this study were the Solitaire 2 revascularization device (Covidien, Irvine, California, USA), the Trevo ProVue retriever/Trevo XP ProVue retriever (Stryker, Fremont, California, USA), and the Revive SE (Codman, Raynham, Massachusetts, USA). The aspiration catheter used was the Penumbra system (Penumbra, Alameda, California, USA).

Study outcomes

The primary outcome was a favorable outcome, defined as mRS score 0–2 at 90 days. We defined secondary outcome measures as excellent outcome (mRS score 0–1 at 90 days) or mortality at 90 days. Safety outcomes included any intracranial hemorrhage (ICH) or symptomatic ICH within 72 hours after stroke onset. Symptomatic ICH indicated neurological worsening by more than four points on the National Institutes of Health Stroke Scale (NIHSS).

Data collection and definitions

Clinical information was collected by a review of hospital charts. Follow-up information up to 90 days was mainly collected by a review of hospital charts, and any additional information was collected by making contact with patients, relatives, and referring physicians. In addition to patient characteristics (age, sex, and vascular risk factors), we collected data of the mRS score before the onset of stroke, NIHSS score (range 0–42), time from the onset of symptoms, process times (symptom onset to hospital door, symptom onset to puncture, and symptom onset to reperfusion), treatment with IV tPA, and stroke classification. Acute ischemic stroke was classified as cardioembolic, atherothrombotic, cryptogenic, or other.19

Early ischemic changes were measured by the Alberta Stroke Programme Early CT Score (ASPECTS) using diffusion-weighted imaging or non-contrast-enhanced CT (range 0–10). We classified the location of the occlusion site using magnetic resonance angiography, CT angiography, or digital subtraction angiography.

Statistical analysis

Categorical variables are presented as numbers and percentages, and were compared using the Χ2 test. Continuous variables are described as mean (SD) or median (IQR). Based on distributions, continuous variables were compared using Student’s t-test or Wilcoxon’s rank-sum test.

In our primary analysis, we analyzed the entire cohort of patients with M2 occlusion. We compared primary and secondary outcomes and safety outcomes between the EVT and medical treatment groups using univariate and multivariate analyses. We constructed a multivariable logistic regression model by adjusting the following variables: age (≥75 years vs <75 years), ASPECTS (≥6 vs <6), NIHSS score (≥9 vs <9), time from onset of stroke to hospital door (≥285 min vs <285 min), and treatment with IV tPA. The threshold of 9 for the NIHSS score was determined by the cut-off point for predicting poor outcomes of M2 occlusion in a previous study.16 The threshold of 285 min for time from onset of stroke to hospital door was determined by the recommended time window of 6 hours from onset to puncture and 75 min from hospital door to puncture (360–75=285).20 21 The effects of EVT relative to medical treatment are expressed as adjusted odds ratios and 95% CIs.

For the secondary analysis, we developed a propensity score-matched cohort in order to explore the robustness of our analyses. A logistic regression model was used to develop the propensity score for the choice of EVT with six independent variables relevant to that choice. The variables for the propensity score included the following: study period (before February 2015 or after March 2015), age (≥75 years or <75 years), ASPECTS (≥6 or <6), NIHSS score (≥9 or <9), time from onset of stroke to hospital door (≥285 min or <285 min), and treatment with IV tPA. The two study periods were determined by the publications from RCTs of EVT for LVO, which changed the treatment methods used for patients with acute ischemic stroke.1–4 The patients in the medical treatment group were matched to those in the EVT group using a 1:1 matching technique.22 23 We compared primary outcomes, secondary outcomes, and safety outcomes of the propensity score-matched cohort between the EVT and medical treatment groups. The effects of EVT relative to medical treatment are displayed as ORs and 95% CIs. Subgroups for predicting primary outcomes were age (<75 years or ≥75 years), ASPECTS (0–5 or ≥6), NIHSS score (0–8 or ≥9), time from stroke onset to hospital door (≥285 min or <285 min), and treatment with IV tPA. Forest plots were used to determine the treatment effects in the propensity score-matched cohort.

All statistical analyses were conducted in JMP version 13.2.0 software (SAS Institute Inc, Cary, North Carolina, USA). All p-values were two-sided, and we considered p values <0.05 to be statistically significant.

Results

Among the 2420 patients enrolled in the registry, 372 (219 [59%] men; mean (SD) age, 75 [11] years) were enrolled in this study. Of these 372 patients, 184 received EVT and 188 received medical treatment. A flow chart of patients included in this analysis is shown in online supplementary figure 1. Table 1 shows the patients’ baseline characteristics before matching (the entire cohort). Compared with patients in the medical treatment group, those in the EVT group had a higher baseline NIHSS score (median (IQR), 15 [9–19] vs 10 [5–16], p<0.001) and earlier time from onset of stroke to hospital door (median (IQR), 110 [50–258] min vs 150 [60–343], p=0.01). EVT with stent retrievers (with or without aspiration) was used in 102 cases (55%), whereas aspiration alone was used in 53 cases (28.8%). There was no difference among the groups for stroke subtypes and comorbidities, except for diabetes. After the propensity score matching, 184 patients in each group were matched with their counterparts. Baseline characteristics of the two groups after propensity score matching are shown in table 1. No significant difference was noted between the two groups.

Table 1

Baseline characteristics of the patients

Primary clinical outcomes in the entire cohort

Among 372 patients with M2 occlusion, a favorable outcome was observed in 57.0% and 50.5% of EVT and medical treatment patients, respectively (online supplementary figure 2). No significant differences were seen between the EVT and medical treatment groups according to univariate analysis (OR=1.30; 95% CI 0.86 to 1.95; p=0.21) (table 2). After adjustment, EVT was significantly associated with higher odds of a favorable outcome (adjusted OR=2.09; 95% CI 1.26 to 3.47; p=0.004) and lower odds of mortality at 90 days (adjusted OR=0.27; 95% CI 0.08 to 0.93; p=0.03) (table 2). For an excellent outcome, there were no significant differences between the EVT and medical treatment groups in multivariate analyses (adjusted OR=1.53; 95% CI 0.92 to 2.55; p=0.09).

Table 2

Clinical outcomes: entire cohort

For the safety outcome in univariate or multivariate analyses, symptomatic ICH occurred in seven patients (3.8%) in the EVT group, showing no significant difference between the EVT and medical treatment groups (table 2). Therefore, any ICH at 72 hours did not differ significantly between the groups.

Clinical outcomes in the propensity score-matched cohort

In the propensity score-matched patients with M2 occlusion, a favorable outcome was observed in 57.1% and 32.6% of EVT and medical treatment patients, respectively (online supplementary figure 3). EVT was significantly associated with higher odds of a favorable outcome (OR=2.74; 95% CI 1.79 to 4.19; p<0.0001) and excellent outcome (OR=2.01; 95% CI 1.26 to 3.19; p=0.003) (table 3).

Table 3

Clinical outcomes: propensity score-matched cohort

For the safety outcome, no significant differences were seen in symptomatic ICH (OR=0.72; 95% CI 0.26 to 2.00; p=0.61) and any ICH at 72 hours (OR=0.83; 95% CI 0.53 to 1.31; p=0.48) between the groups.

Subgroup analyses of a favorable outcome suggested that EVT was generally effective in all subgroups, although the ORs for treatment were not significant for the patients with a low ASPECTS (0–5) or NIHSS score (0–8), and in those with more than 285 min from stroke onset to hospital door (figure 1). Effects favoring EVT were present in the patients regardless of age and treatment with IV tPA. No interaction p-values were statistically significant.

Figure 1

Forest plot showing treatment effects for favorable outcome (mRS score 0–2 at 90 days) in propensity score-matched patients with M2 occlusion. ASPECTS, Alberta Stroke Programme Early CT Score; EVT, endovascular therapy; M2, M2 segment of the middle cerebral artery; mRS, modified Rankin Scale; NIHSS, National Institutes of Health Stroke Scale.

Discussion

The main findings of this study were that (1) in real-world patients with M2 occlusion, EVT resulted in a favorable clinical outcome without increasing ICH; (2) results of the propensity score-matching analysis were robust and indicated that EVT for M2 occlusion was effective and safe relative to medical treatment; and (3) EVT was generally effective in all subgroups, and effects favoring EVT were present in several subgroups of interest, including age ≥75 years, NIHSS score ≥9, and ineligible for treatment with IV tPA.

Although the efficacy of EVT in patients with M2 occlusion is still unclear, efficacy involving the M2 segment has occasionally been reported in previous retrospective studies and systematic reviews.9 12 15 17 24 Of these, only a few studies compared the efficacy and safety of EVT with those of medical treatment. The study by Sarraj et al was the only large retrospective cohort study that compared the efficacy of EVT and medical treatment in patients with M2 occlusion.9 Our study showed the efficacy of EVT for a favorable outcome relative to medical treatment, as in study by Sarraj et al.9 The percentage of favorable functional outcomes in the EVT group was 57.1%, and this was not much lower than results in previous meta-analysis reports (62% according to Chen et al, 62% according to Phan et al, and 59% according to Saber et al).12 17 25

On the other hand, in the medical treatment group, the percentage of favorable functional outcomes was 50.5%, which was higher than that reported by Sarraj et al (35.4%).9 This was possibly owing to the difference in patient characteristics. Specifically, M2 occlusion included many patients with M2 proximal occlusion in the report by Sarraj et al, and the medical treatment group included a large percentage of patients with a high NIHSS score at onset. Previous reports on the natural history of M2 occlusion, which mean medical treatment alone, showed that the percentages of an mRS score 0–2 at 90 days were 53% according to Hernández-Pérez et al and 54% according to Lima et al.26 27 These percentages of favorable functional outcomes were similar to the results from our study. As a study on M2 occlusion, this study may more accurately reflect clinical practice. Even with such a patient background, the efficacy of EVT for M2 occlusion was demonstrated. This was one of the most significant and useful findings from this study.

The frequency of any ICH in the EVT group was 27.7%, and this was higher than the complication rates reported by a previous meta-analysis—11.2% for any ICH, according to Phan et al.25 However, the frequency of symptomatic ICH (sICH) in the EVT group was 3.8%, and it was equivalent to or slightly lower than the previous value reported in meta-analyses—sICH frequency of 5–10%.17 25 At the point of using devices, the rate of using stent retrievers (with or without aspiration) was higher than the rate of using aspiration alone (55% vs 29%) in our study. In a previous report using stent retrievers for distal MCA, the rate of sICH was 4.6%, which was similar to our result.28 Because there were no significant differences of ICH between stent retrievers and aspiration for M2 occlusion in a previous report, device selection did not affect the frequency of ICH.14 Moreover, the percentage of any ICH in the medical treatment group in this study was 21.8%, indicating no significant difference between the EVT and medical treatment groups. Percentages of hemorrhagic complications were similar to results from the overall analysis of the RESCUE-Japan Registry 2, signifying that hemorrhagic complications do not occur more frequently in EVT of the M2 than in EVT of other intracranial vessels.18 This result is similar to that of the report by Salahuddin et al, who compared M1 and M2 occlusions (sICH; M1 occlusion vs M2 occlusion, 3.3% vs 3.4%, p=1.0)15 and the report by Sarraj et al (ICH; EVT vs medication, 10.8% vs 9.4%).9 These findings, including our results, suggest that use of EVT is safe in M2 occlusion.

We used a propensity score-matched cohort to confirm the primary clinical outcomes, because treatment methods were determined by physicians; thus, patients’ characteristics were very different between the EVT group and medical treatment group. As a result, in propensity score-matching analysis, EVT for M2 occlusion was significantly associated with higher odds of a favorable outcome and also with an excellent outcome. There was also no difference in safety outcomes. These results were robust, and they indicate that EVT for M2 occlusion was effective and safe relative to medical treatment.

Our study showed that EVT was generally effective in all subgroups, although the ORs for treatment were not significant for the patients with a low ASPECTS (0–5) or NIHSS score (0–8), and in those with more than 285 min from stroke onset to hospital door. A previous study also showed that younger age and higher ASPECTS were independently associated with a favorable clinical outcome in patients with M2 occlusion who received EVT.9

Effects favoring EVT were present in patients with an NIHSS score ≥9 (OR=3.44; 95% CI 2.05 to 5.77). The cut-off point for predicting outcomes of M2 occlusion was NIHSS ≥9, according to Rai et al.16 This cut-off point is higher than that in recent guidelines recommending EVT for LVO (NIHSS score ≥6).20 This indicates that a higher level of neurological symptoms may be required for M2 occlusion relative to LVO. RCTs of EVT for LVO have suggested that patients with a mismatch between the ischemic core and penumbra could benefit from reperfusion.8 However, the size of the mismatch between the ischemic core and ischemic penumbra of distal intracranial occlusions was theoretically less than that of LVO. This indicates that depending on M2 occlusion, patients develop symptoms in the eloquent area (language and motor areas), and the NIHSS score could be an important factor in comparison with the mismatch between the volumes of the ischemic core and penumbra. In other words, neurological symptoms, rather than a mismatch between the volumes of the ischemic core and penumbra, must be taken into account when considering the effects of EVT in M2 occlusions. In that respect, our study showed that an NIHSS score ≥9 at onset could become a criterion for EVT eligibility in patients with M2 occlusion.

Effects favoring EVT were present in patients regardless of treatment with IV tPA. The rate of patients who received IV tPA before EVT in this study was 50%, which was only slightly lower than that in previous meta-analyses of EVT for M2 occlusion (58–59%).17 25 This study suggested that EVT is efficacious in patients with M2 occlusion who are ineligible to receive treatment with IV tPA. However, this finding does not recommend that IV tPA pretreatment should be withheld in patients who are eligible for such treatment. Rather, we confirmed that EVT can be considered in patients with M2 occlusion, even if they are not eligible for treatment with IV tPA.

Limitations

This study has several limitations. First, and most critically, this study was a prospective registry, and the selection of EVT, including the selection of the device and methods, depended on the practicing physicians. Thus, selection biases would be present in comparisons between EVT and medical treatment. By taking into account this bias, we used a propensity score-matched cohort to confirm the robustness of our analyses. Second, although we described a unified anatomical definition of intracranial vessels in the registry protocol, we did not conduct imaging analysis in a central core laboratory. Thus misclassification of the clot location is possible. In addition, because RESCUE-Japan Registry 2 was not a randomized study, the potential for unknown confounding remains. Third, RESCUE-Japan Registry 2 was a prospective registry of consecutive patients with acute LVO and did not focus on patients with M2 occlusion. Hence, the sample size was still not sufficiently large to fully evaluate the effects of EVT for M2 occlusion, although the number of patients in this study was almost the same as in previous comparisons of EVT with medical treatment in patients with M2 occlusion.9 Finally, because symptoms in the eloquent area differ depending on the site of distal MCA occlusion, bilateral differences in the occlusion site or differences in the occluded branch of the M2 segment might have affected the results. However, we did not perform bilateral assessments or evaluations of branches in the M2 segment in this study.

Conclusions

Our registry suggests that EVT for patients with M2 segment occlusion may be effective and safe compared with medical treatment. EVT was generally effective in all subgroups, and effects favoring EVT were present in several subgroups of interest, such as those with NIHSS score ≥9, and regardless of age and treatment or not with IV tPA. These findings can help in the design of future randomized prospective studies and in guiding patient selection for treatment of M2 occlusion.

Acknowledgments

The authors would like to thank all the RESCUE-Japan Registry 2 investigators.

References

Footnotes

  • Contributors MM performed the statistical analysis and wrote the first draft of the article. This article was supervised by KU, SY, NS, HY, and TM. Those authors made substantial contributions to the conception and design of the RESCUE-Japan Registry 2. KU and YN made contributions to the data analysis and interpretation of data. KU, SY, NS, HY, YN, and TM also contributed to drafting the article and its critical revision for important intellectual content. All authors have approved the final manuscript for submission.

  • Funding The Recovery by Endovascular Salvage for Cerebral Ultra-acute Embolism Japan Registry 2 was supported in part by the Japan Agency for Medical Research and Development; the Japanese Society for Neuroendovascular Therapy; the Ministry of Health, Labour, and Welfare of Japan; Medtronic; Stryker; and Medicos Hirata. The funding sources did not participate in any part of the study, from study conception to manuscript preparation.

  • Competing interests MM reports no disclosures. SY discloses research grants from Medtronic, Medicos Hirata, Termo, Bristal-Myers Squibb, and Otsuka and lecturer’s fees from Boehringer-Ingelheim, Daiichi Sankyo, Otsuka, Bayer, Sanofi, Phizer, Bristal-Myers Squibb, Stryker, Medtronic, and Mitsubishi Tanabe. NS reports research a grant from Termo; lecturer’s fees from Jimro, Otsuka, Johnson & Johnson, Medtronic, Stryker, and Medicos Hirata; and membership of the advisory boards for Medtronic and Jimro. HY discloses research grants from Bristol-Myers Squibb; lecturer’s fees from Bayer, Daiichi-Sankyo, Bristol-Myers Squibb, Boehringer Ingelheim, and Striker; and membership of the advisory boards for Daiichi-Sankyo, Bayer, and Boehringer Ingelheim. KU reports lecturer’s fees from Nihon Medi-Physics. YN reports no disclosures. TM reports a research grant from Nexis; lecturer’s fees from AbbVie, AstraZeneca, Daiichi Sankyo, Kyorin, Mitsubishi Tanabe, and Phizer; manuscript fee from Pfizer; and membership of the advisory boards for Asahi Kasei, Boston Scientific, and Bristal-Myers Squibb.

  • Ethics approval Hyogo College of Medicine.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Data sharing statement There are no additional unpublished data available.