Article Text
Abstract
Background Thrombectomy for anterior large vessel occlusion less than 24 hours since last known well is now standard of care. Certain aspects of clinical trials may limit generalizability to ’real-world' practice.
Objective To compare revascularization rates and outcomes for direct aspiration (ADAPT) and stent retriever thrombectomy following anterior acute ischemic stroke (AIS) in a real-life setting.
Methods Data from the most recent 20 consecutive patients with AIS treated with mechanical thrombectomy between 2015 and 2016 were collected from 15 high-volume stroke centers across North America for a total of 300 cases. Patients with proximal anterior large vessel occlusions were dichotomized by primary treatment technique. Ordinal logistic regression assessed the effects of clinical variables on patient disability using 90-day modified Rankin Scale (mRS) scores.
Results Adequate revascularization (Thrombolysis in Cerebral Infarction ≥2b) was ultimately achieved in 91.2% of first-line direct aspiration (ADAPT) cases with an average of 1.9±1.9 passes and in 87.5% of stent retriever cases with an average of 1.7±1.0 passes. Time from groin puncture to revascularization was shorter for ADAPT cases. The mean 90-day mRS score for both groups was 3.0±2.4. Number of passes using primary technique, and postintervention intracranial hemorrhage, were significant predictors of 90-day mRS scores after ADAPT, while age and preprocedure mRS score were predictive of outcomes following first-line stent retriever.
Conclusions Our data show similar adequate revascularization rates and 90-day functional outcomes for first-line direct aspiration and stent retrievers for anterior large vessel occlusion in a real-world setting. These results support the findings of other prospective trials evaluating the two techniques.
- ischemic stroke
- mechanical thrombectomy
- stent-retriever
- direct aspiration
- revascularization
Statistics from Altmetric.com
Introduction
Five landmark, prospective randomized clinical trials (RCTs) conducted in 2015 established mechanical thrombectomy as an efficacious treatment for acute ischemic stroke (AIS) due to large vessel occlusions in the anterior circulation.1–5 A subsequent meta-analysis of these RCTs, the Highly Effective Reperfusion evaluated in Multiple Endovascular Stroke trials (HERMES) study, further demonstrated the therapeutic benefit of mechanical thrombectomy in particular subpopulations as well. These included patients aged >80, patients receiving treatment more than 6 hours after symptom onset, and patients ineligible for intravenous alteplase.6 The DAWN7 and DEFUSE-38 trials have since confirmed the benefit of mechanical thrombectomy in patients treated up to 24 hours from last known well using perfusion CT selection.
Patients enrolled in the aforementioned studies were treated with stent retriever technology. The comparative efficacy of direct aspiration as a first-line approach (ADAPT) has been less extensively evaluated. Recently, the ASTER trial, a prospective, randomized, open-label, blinded endpoint study concluded that first-line direct aspiration was non-inferior to first-line stent retriever for revascularization following AIS in the anterior circulation.9 10
Occasionally, certain aspects of the study design of RCTs can limit the generalizability of the findings to clinical practice in the ‘real-world’. These can include the patient inclusion and exclusion criteria, the selection of treatment centers that participate in the study, and the experience of the device operators in the study. To our knowledge, no recent studies have examined the periprocedural clinical metrics and outcomes for patients undergoing mechanical thrombectomy for large vessel anterior AIS as a means of providing a real-life comparison of direct aspiration and stent retriever as first-line techniques. The goal of this study was to compare revascularization rates and patient outcomes for first-line direct aspiration and first-line stent retriever thrombectomy following anterior AIS in a real-life setting and to describe predictors of clinical outcomes for these techniques.
Methods
Data collection
Data from the 20 most recent, consecutive patients with AIS treated with mechanical thrombectomy in 2015–2016 were collected from 15 high-volume stroke centers across the United States. These data were obtained as part of the selection process for an individual center’s eligibility for participation in the COMPASS trial, a multicenter prospective RCT comparing ADAPT versus stent retrievers in anterior circulation emergent large vessel occlusion.
Clinical outcomes
Age, gender, site of vessel occlusion, imaging modality used, intra-arterial tissue plasminogen activator (tPA) administration, preprocedure modified Rankin Scale (mRS) and National Institutes of Health Stroke Scale (NIHSS) scores, time points from last known well to revascularization, and number of devices and passes needed to achieve revascularization were among the clinical variables considered. The final Thrombolysis in Cerebral Infarction (TICI) scores, the postprocedural mRS and NIHSS scores, 90-day mRS and NIHSS scores, and postintervention intracranial hemorrhage (ICH) were used as outcome metrics. Postprocedure intracranial hemorrhage was defined as symptomatic parenchymal or subarachnoid hemorrhage using the European Cooperative Acute Stroke Study (ECASS) classification.11 All data were recorded in an Excel spreadsheet (Microsoft, Inc, Redmond, Washington, USA).
Statistical analysis
Patients were dichotomized according to whether direct aspiration or a stent retriever was used as the first-line technique for revascularization. A two-sided t test was used to assess significant differences in quantitative variables, while a Chi-Squared test was used for categorical variables. Fisher’s Exact test was used for contingency tables with expected counts less than one under the null hypothesis of independence. A significance threshold of p=0.05 was implemented across all analyses. Univariate and multivariate ordinal logistic regression was performed to evaluate how well certain clinical variables predict 90-day mRS scores in both cohorts. All statistical analysis was performed using R 3.4.1 and all figures were created using the ggplot package in R (The R Foundation for Statistical Computing, 2016).
Results
Cohort demographics and clinical characteristics
After excluding patients with AIS in the posterior circulation, we identified 274 patients in our dataset who experienced an anterior AIS, with 107 first treated with direct aspiration and 121 with a stent retriever. Complete revascularization data were available for 257 patients, including 102 direct aspiration cases and 120 stent retriever cases. Table 1 reports the baseline clinical information of each treatment cohort. Both cohorts had similar age distributions, numbers for each sex, sites of vessel occlusion, and side of occlusion. Both groups also shared similar preprocedure NIHSS scores, but the mean preprocedure mRS scores were significantly higher in the direct aspiration cohort (p=0.002).
Procedural characteristics for each treatment cohort
Table 2 reports the characteristics of the endovascular procedures in both treatment cohorts. The patients in the direct aspiration cohort were more likely to have had received a CT scan on hospital arrival (p=0.0003) than patients in the stent retriever cohort. On the other hand, patients in the stent retriever cohort were significantly more likely to have had CT angiography (p=0.0009) and CT perfusion study (p=0.0021) than patients in the direct aspiration cohort. Just over half of patients in both cohorts were administered IA tPA on hospital arrival. As expected, choice of secondary technique differed significantly between the treatment cohorts (p=0.014). Of note, the median time from stroke onset to hospital arrival for the direct aspiration cohort was nearly twice that of the stent retriever cohort (p=0.010). The stent retriever cohort, however, had significantly greater median times from hospital arrival to groin puncture (p=0.039) as well as from groin puncture to revascularization (p=0.006). Ultimately, median time from stroke onset to revascularization did not differ significantly between the two cohorts (p=0.112).
Procedural outcomes between patient cohorts
Table 3 shows details of the clinical outcome metrics for the two treatment cohorts. Rates of adequate revascularization, defined as a final TICI score ≥2b, with the first-line treatment technique were not significantly different between the two cohorts and both first-line techniques took a similar number of passes to achieve revascularization, on average. The proportion of patients in each cohort who were adequately revascularized by the end of the intervention also did not differ significantly between the two mechanical thrombectomy techniques used (p=0.805). Final, adequate revascularization (TICI ≥2b) was achieved in 91.2% of the primary ADAPT group and 87.5% of the primary stent retriever group.
Postoperatively, 19 (17.8%) patients in the ADAPT cohort experienced an ICH, while 5 (4.7%) experienced a procedure-related adverse event. Similar rates were seen in the stent retriever cohort, in which 14 (11.6%) experienced an ICH (p=0.178) and 4 (3.3%) experienced a procedure-related adverse event (p=0.587). In the ADAPT cohort, the procedure-related adverse events included a large vessel vasospasm, three subarachnoid hemorrhages, and a groin hematoma. In the stent retriever cohort, the adverse events included two vasospasms in large vessels and two emboli to new territories. Both cohorts shared similar mean mRS and NIHSS scores immediately after the intervention. After 90 days, the mean mRS scores for both cohorts were 3.0±2.4 (p=0.939), while their NIHSS scores were slightly, but not significantly, different (direct aspiration: 4.5±8.7; stent retriever: 9.6±15.1; p=0.150).
Cohort analysis for outcome predictors by primary technique
Table 4 reports the results from applying univariate and multivariate ordinal logistic regression modeling to identify patient and periprocedural variables that were predictive of 90-day mRS scores for patients treated with direct aspiration as a first-line therapy. At the univariate level, the time from groin puncture to revascularization, the number of passes with the primary technique, postintervention ICH, postprocedure mRS score, and postprocedure NIHSS score were significant predictors of 90-day mRS scores. Interestingly, despite a strong trend toward significance, neither the final TICI scores nor the preprocedure mRS scores were significant predictors of 90-day mRS scores in this patient cohort.
Multivariate ordinal logistic regression models were constructed using variables that were determined to be statistically significant at the univariate level. Immediate postintervention mRS and NIHSS scores were omitted from multivariate models because they are covariate with 90-day mRS scores and are less designed to evaluate clinical progress. At the multivariate level, only the number of passes with the primary technique and postintervention ICH remained statistically significant predictors of 90-day mRS scores (table 4). The time from groin puncture to revascularization was not significant at the multivariate level.
Table 5 reports the results from univariate and multivariate ordinal logistic regression modeling of the relationships between certain periprocedural variables and 90-day mRS scores for patients treated with stent retriever as a first-line therapy. At the univariate level, age, preprocedure mRS score, pre-procedure NIHSS score, final TICI score, postprocedure mRS score, and postprocedure NIHSS score were significant predictors of 90-day mRS scores. At the multivariate level, only age and preprocedure mRS scores remained significant predictors of 90-day mRS scores for patients treated with stent retriever as a first-line therapy.
Discussion
Results from recent RCTs and the HERMES meta-analysis provide conclusive evidence that mechanical thrombectomy is an effective treatment for large vessel AIS in the anterior circulation.1–6 The 2017 ASTER trial followed up on these results by comparing the differences in revascularization between first-line direct aspiration and first-line stent retriever thrombectomy for AIS in the anterior circulation. Though it was powered to show superiority of first-line direct aspiration, the ASTER trial concluded that first-line direct aspiration was non-inferior to first-line stent retriever.9 10 Preliminary results from the multicenter COMPASS RCT have also demonstrated the non-inferiority of ADAPT as a first-line strategy in comparison with stent retriever.12 13 These findings carry additional relevance for healthcare systems as direct aspiration and stent retriever techniques differ in their practicality, cost, and efficiency.14
Given that certain study design factors inherent to RCTs may introduce differences that prevent the results being applicable in a ‘real-world’ setting, it is important for subsequent studies to assess how well certain findings from RCTs translate into everyday clinical practice. For example, some of the key exclusion criteria in the ASTER trial included a pre-stroke mRS score >3, time from stroke onset >6 hours, and stroke location anywhere other than M1, M2, and intracranial internal carotid artery.10 Since the reporting of the HERMES study, few studies have assessed the ‘real-world’ translatability of mechanical thrombectomy techniques.15–17 Notably, since the ASTER trial, no studies, to our knowledge, have compared the outcomes of direct aspiration and stent retriever as first-line mechanical thrombectomy techniques for the everyday management of patients with AIS. The aim of our study was to obtain detailed, prospectively collected data from multiple high-volume stroke centers across the United States to determine how well the key findings of the ASTER trial translate into the real-world setting. To achieve this we compared metrics for revascularization and patient outcomes following direct aspiration or stent retriever thrombectomy as first-line therapies for emergent large vessel occlusion AIS in the anterior circulation.
In this study, patients within the pooled population were assigned to one of two cohorts according to the first-line mechanical thrombectomy technique employed. Analysis of the demographic and baseline characteristics of these cohorts showed that their composition was remarkably similar (table 1). Certain differences were found in the characteristics of imaging and endovascular procedures (table 2). Some of these differences are well known and have been reported among the preliminary COMPASS results, such as the longer times between groin puncture and revascularization in the stent retriever cohort.13 Other cohort differences may be better understood after considering the nature of the data underlying this study. Because data were obtained equally from 15 geographically diverse, high-volume stroke centers across the United States, individual center-level differences exist in patient accessibility to the hospital, the established protocols in place for triaging patients, types of imaging used for diagnosis, and even the preferred primary and secondary techniques for performing mechanical thrombectomy. The significant differences between the cohorts in the imaging modalities used for diagnosis, secondary techniques most often implemented, and times from stroke onset to hospital arrival and from hospital arrival to groin puncture may be partially explained by these factors. These factors are important to consider as they reflect real-world differences seen in everyday clinical practice.
Despite these cohort-level differences, analysis of the characteristics of revascularization showed that there were no statistically significant differences between patients who received first-line direct aspiration and those receiving first-line stent retriever therapy. Though a slightly greater proportion of patients receiving first-line stent retriever achieved revascularization with the primary technique alone, this difference was not significant, and this trend was reversed when revascularization was assessed at the end of treatment after multiple techniques had been implemented, if necessary. A breakdown of the individual final TICI scores revealed a similar score distribution between the two groups, with the vast majority of patients achieving final TICI scores of either 2b or 3 (figure 1). These findings generally reflect the efficacy outcomes from the ASTER trial, which reported successful revascularization (TICI ≥2b) in 84.9% of the first-line direct aspiration group and 86.2% of the first-line stent retriever group determined at the study sites at the end of all techniques (p=0.71).10
Similar results were found when various metrics of clinical outcome were examined in this study. The mean mRS and NIHSS scores immediately following the intervention and 90 days later did not differ between the two cohorts (figure 2). This finding is particularly interesting given that the only baseline characteristic that differed between the patient cohorts was the preprocedure mRS score. Patients who received first-line direct aspiration had significantly higher preprocedure mRS scores on average than those in the stent retriever group. However, after the interventions, mRS scores between the cohorts were not different in the short- or long-term. These findings confirm the functional outcome results of the ASTER trial, which reported that there was no significant difference between the 90-day mRS scores of patients who received first-line direct aspiration versus first-line stent retriever (p=0.15).10 The findings from our study also support the primary efficacy results of the COMPASS trial, which showed that first-line direct aspiration was non-inferior to first-line stent retriever as indicated by 90-day mRS scores (p=0.0014).13
The rates of postintervention ICH and procedure-related adverse events were not different between the two first-line techniques. These results mirror those of the ASTER trial, which by comparison, reported lower rates of symptomatic ICH (direct aspiration: 5.3%, stent retriever: 6.5%) but higher rates of procedure-related adverse events (direct aspiration: 16.2%, stent retriever: 15.9%) than this present study.10 Although there were no statistically significant differences between the techniques for these metrics, it is important to highlight that the rates of serious complications were not low for either study. Patients must understand that a certain amount of risk is assumed when undergoing treatment with any mechanical thrombectomy technique, which must be balanced against the poor natural history of untreated large vessel occlusion.18
Following descriptive analysis of these data, we turned to univariate and multivariate ordinal logistic regression in both treatment cohorts to determine which patient and periprocedural variables were most predictive of patient outcomes after 90 days. For patients who received first-line direct aspiration, results from multivariate analyses pointed to postintervention ICH and number of passes taken with the primary technique as variables that significantly increased the odds of 90-day mRS scores.19 Both of these variables have been identified in previous analyses as significant predictors of poor outcomes following mechanical thrombectomy.20 In addition, a recent meta-analysis has shown that the risk of any ICH is significantly increased for patients receiving mechanical thrombectomy compared with standard medical therapy.18 Taken together, these findings suggest that further work should be done to identify factors that lead to ICH following mechanical thrombectomy as several studies, including our study, have shown it to be related to poorer long-term functional outcomes in patients with AIS.
For patients who received first-line treatment with a stent retriever, age and pre-procedure mRS scores were most predictive of 90-day mRS scores. Although relatively few studies have examined the effect of age on mechanical thrombectomy outcomes for patients with AIS, age was among the patient variables examined in the HERMES study. This study used pooled data from RCTs to show that, despite mechanical thrombectomy providing a clear benefit over standard medical therapy in patients over 80, age served as a strong negative independent predictor of outcomes.6 More recently, a single-center study found that patients over 80 who underwent first-line direct aspiration for AIS were significantly less likely to achieve a 90-day mRS score≤2.20
Our study has some important limitations that must be acknowledged. First, because this was a retrospective study, temporal relationships between variables cannot be assessed and selection bias may exist within the results. Second, it was challenging to obtain complete data points for each patient encounter as this was a multicenter study. As such, applicable analyses were conducted using only patient encounters with complete data. Third, a limited sample of patients was included in this study, which may mean that certain analyses lacked the power to detect significance for variables with small effect sizes. Finally, as mentioned previously, certain aspects of these data must be considered in the context of the study design. Stroke centers tend to have preferences for certain mechanical thrombectomy techniques. Possibly, certain trends in these data are the result of institution-level factors, such as geography, infrastructure, and established protocols for patient management, which might influence the results of the entire cohort. These structural differences may be difficult to control, but they do reflect some of the intricacies of real-world clinical practice that might not be captured in the controlled setting of an RCT.
Conclusions
Using prospectively collected data from multiple high-volume stroke centers across the United States, this retrospective study aimed to conduct a ‘real-life’ comparison of the revascularization and functional clinical outcomes of patients receiving either direct aspiration or stent retriever as first-line treatments for AIS in the anterior circulation. Our data suggest that first-line direct aspiration results in similar revascularization rates and 90-day functional outcomes to those achieved with first-line stent retriever thrombectomy. This study supports the results of the ASTER trial and the preliminary results from the COMPASS trial. Findings from these studies are important and may shape the way in which stroke is managed across the United States. The choice of mechanical thrombectomy technique may have tangible implications for practicality and cost savings in everyday clinical practice.
References
Footnotes
Contributors All authors have made a substantial contribution to the conception and design of the studies and/or the acquisition and/or the analysis of the data and/or the interpretation of the data; drafted the work or revised it for significant intellectual content; approved the final version of the manuscript; and agree to be accountable for all aspects of the work, including its accuracy and integrity.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests AHS: financial interest/investor/stock options/ownership: Imperative Care, Q’Apel Medical, Inc; consultant/advisory board: Cerenovus, Imperative Care, Medtronic, MicroVention, Penumbra, Q’Apel Medical, Inc, Stryker. JE: consultant: Stryker Neurovascular, Medtronic. AT: consultant: BALT/Blockade, Cerenovus, Medtronic, Microvention-Terumo, Penumbra, Imperative Care, Stryker. KW0: consultant: Cerenovus, Medtronic, Microvention, Penumbra, Stryker. AR: consultant: Stryker, Cerenovus, Microvention. BB: consultant: Penumbra, Medtronic, Stryker; shareholder: Penumbra. RH: grant/research support: Medtronic, Stryker, Microvention; consultant: Medtronic, Stryker, Cerenovous, Microvention, Balt. JEDA: consultant: Penumbra, Medtronic, Microvention. ASA: consultant: Cerenovus, Medtronic, Microvention, Penumbra, Stryker; research support: Microvention, Cerenovus, Penumbra. AA: consultant: Cerenovus, Medtronic, Microvention, Penumbra, Stryker; research support: Microvention, Cerenovus, Penumbra. MK: consultant: Penumbra, Medtronic. LP: consultant: Medtronic. DF: consultant/speakers bureau: Penumbra; Stryker; research support: Cerenovus, Medtronic, Microvention, Penumbra, Stryker; stock ownership: Penumbra. IL: consultant: Medtronic, Stryker, Cerenovus. RDL: consultant: Cerenovus, Penumbra, Q’Apel Medical Inc, Imperative Care, Mivi Neuroscience; research support: Medtronic, Asahi Intec. JM: consultant: Cerebrotech, Rebound Therapeutics, TSP Inc, Lazarus Effect, Medina, Pulsar Vascular; investor: Blockade, Medina, Lazarus Effect, TSP Inc; research support: Penumbra
Ethics approval Approval was obtained from the institutional review boards of Mount Sinai, Medical University of South Carolina, State University of New York at Buffalo, Stony Brook School of Medicine, Lyerly Baptist Medical Center, Fort Sanders Regional Medical Center, West Virginia University Hospitals, Swedish Medical Center, Abbott Northwestern Hospital, University of Saskatchewan College of Medicine, University of Tennessee Health Sciences Center and Semmes Murphey Clinic, Erlanger Health System, California Pacific Medical Center, and Miami Cardiac and Vascular Institute.
Provenance and peer review Not commissioned; externally peer reviewed.
Patient consent for publication Not required.