Objective Stroke patients with good collateral circulation achieve the best recovery after mechanical thrombectomy (MT) but strict imaging selection may result in untreated patients that could benefit from MT. We assessed whether the extent of collaterals had modifying effects on the amount of ischemic tissue saved from infarction with MT over best medical treatment (BMT).
Methods This was a single center cohort of consecutive patients (n=339) with proximal occlusions in the carotid territory. Patients were categorized according to a four point category scale on CT angiography as having good (scores 2–3) or poor (scores 0–1) collaterals. The primary outcome measure was the interaction between collaterals and MT on infarct growth. The secondary outcome assessed the treatment effect of MT over BMT on functional status in relation to collateral status. Safety outcomes were mortality and symptomatic intracranial hemorrhage.
Results Collaterals had a modifying effect of MT on infarct growth (P=0.004), with a greater reduction in 96 patients with poor collaterals (38.8 mL) than in 243 patients with good collaterals (1.9 mL). There was also a significant (P<0.001) interaction between the effect of MT and functional outcome in relation to collateral status, with more benefits of MT in patients with poor collaterals. MT was associated with lower mortality than BMT in patients with poor collaterals only.
Conclusion Compared with BMT, the use of MT in the early time window in large vessel stroke results in a more substantial limitation of infarct growth in patients with poor collaterals.
- ct perfusion
- magnetic resonance angiography
Statistics from Altmetric.com
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.
The fate of ischemic tissue is not uniform in stroke patients, most likely due to variable temporal growth of the ischemic core from the penumbral area, which is modulated by collateral blood flow.1 The presence or absence of collateral flow is a crucial determinant of the evolution of infarct growth and can be evaluated on CT angiography images.2–5 Better collateral circulation has been related to better recanalization and reperfusion and better outcomes after mechanical thrombectomy (MT).6–8 The volume of infarcted tissue (both before and after revascularization) has consistently been associated with worse outcomes in stroke patients.9–11 Advanced neuroimaging techniques, such as CT perfusion (CTP), allow measurement of the impact of this collateral flow on ischemic tissue, identifying severely hypoperfused areas that are assumed to be mainly non-viable tissue (infarct core) and less severely hypoperfused tissue indicating penumbral areas.
The value of endovascular reperfusion in stroke has been established in clinical trials that mainly included patients with small infarct cores. Some of these trials,12 13 including two trials studying the effect of MT in late time windows,14 15 used advanced imaging (MR or CTP) for selecting patients with small infarct cores. However, strict imaging selection strategies may leave many patients that could benefit from MT untreated.16 17 Subgroup analyses of the HERMES collaboration has shown that specific patient subgroups with overall poor outcomes after large vessel stroke, such as elderly patients or patients presenting with severe neurological deficits, benefit as much as other subgroups of patients from MT.18 For the same reason, even if patients with good collateral circulation who have the best recovery after reperfusion, the actual impact of reperfusion might be especially significant in patients with poor collateral flow that typically present severe hypoperfusion and large infarct cores at the first neuroimaging assessment.
The objective of this study was to assess how collateral circulation influences the response to MT in stroke patients with large vessel occlusion. We measured the relative efficacy of MT over medical care alone to limit infarct growth in patients with different collateral grades.
The study population (n=339) was part of a prospectively collected clinical registry of patients with acute ischemic stroke admitted to a comprehensive stroke center between March 2010 and December 2017. We included all consecutive patients with proximal occlusions in the carotid territory in which pretreatment CT angiography (CTA) and whole brain CTP, as well as follow-up MRI, were available. Proximal arterial occlusions included those located at the terminal intracranial carotid artery, the M1–M2 segments of the middle cerebral artery, and the A1 segment of the anterior cerebral artery. Intravenous recombinant tissue plasminogen activator was given within 4.5 hours from stroke onset following current recommendations.19 The treatment allocation was not randomized, and the final decision to perform the endovascular procedure in each case was left to the treating team, but the decisions followed predefined recommendations that changed over time. Figure 1 summarizes the patient flow diagram with the treatment allocation and reasons for exclusion from the study. All patients were treated with stent retrievers and/or contact aspiration. We prospectively collected baseline characteristics, demographics, clinical course, and reperfusion therapy related variables. The local ethics committee at the Hospital Clinic approved the study (reg code HCB/2018/0680).
The imaging protocol included a baseline multimodal whole brain CT scan with non-contrast CT (NCCT), CTP, and CTA in a Siemens Somatom Definition Flash unit within a median (IQR) of 169 (93–275) min from stroke onset. To ensure complete filling of the collateral circulation, CTA was acquired immediately after CTP. The Alberta Stroke Program Early CT Score (ASPECTS) was assessed on the baseline NCCT. Collaterals were scored on CTA using a four point validated scale: 0=absent collateral supply; 1=collateral supply filling <50% of the occluded arterial territory; 2=collateral supply filling >50% but <100%; and 3=100% collateral supply.20 To increase the statistical power of the study, the scores were categorized as good collaterals (grades 2 and 3) and poor collaterals (grades 0 and 1). CTP maps were processed using singular value decomposition with a delay correction (Apollo MIStar software), and the volumes of hypoperfused and non-viable tissue were established according to a threshold of >3 s in the delay time maps and a relative cerebral blood flow <30% compared with the contralateral hemisphere within the tissue with delay, respectively. Final vessel patency was graded on DSA at the end of MT, according to the modified Thrombolysis in Cerebral Infarction classification, and successful recanalization was defined as grades 2b–3.
Time to recanalization was the delay from symptom onset to recanalization or to the end of the endovascular procedure in patients in whom recanalization was not achieved. Infarct growth was calculated by subtracting baseline CTP infarct core from the final infarct volume measured on a co-registered follow-up MRI performed on a 1.5 T Siemens Magneton Aera unit (Siemens, Erlangen, Germany) at a median (IQR) of 40.3 (25.9–70.1) hours. Diffusion weighted image lesion volumes were calculated using Amira software (Visage Imaging GmbH, Berlin, Germany) through a semiautomated thresholding method to identify regions of interest with a signal intensity exceeding the values in the contralateral hemisphere by >3 SD. Bleeding complications were scored on follow-up imaging according to the European Cooperative Acute Stroke Study criteria, as hemorrhagic infarction and parenchymal hematoma type 1 and type 2.21 Symptomatic intracranial hemorrhage (sICH) was defined as any parenchymal hematoma associated with an increment of at least 4 points in the National Institutes of Health Stroke Scale (NIHSS) score.
Statistical analysis and outcome measures
As treatment allocation was not randomized, the primary efficacy outcome measure of MT over best medical care was the extent of infarct growth in relation to collateral score. The secondary efficacy outcome of MT over best medical care was functional outcome at 90 days, measured using the modified Rankin Scale (mRS) in relation to collateral score. Safety outcomes were the rate of sICH and mortality at 90 days following treatment modality in relation to collateral score.
Continuous variables are reported as mean (SD) or median (IQR), as appropriate. Categorical variables are reported as proportions and differences in proportions were studied with the χ2 test or Fisher’s exact test. We tested the interaction of the quality of collaterals on the effect of MT over best medical care on the outcome measures (infarct growth, mRS) using regression models (linear regression with log transformation for infarct growth and ordinal regression for mRS). The independent effect of MT in each collateral category was assessed in ordinal regression models adjusted for all the covariables identified by the HERMES collaborators (age, sex, baseline NIHSS, ASPECTS, location of occlusion, tandem lesions, delay to imaging, and intravenous tissue plasminogen activator administration),18 variables with P values <0.1 in unilateral analysis, and the period of treatment (up to 2014 vs after 2014), due to the changes in the imaging criteria used to indicate MT. All analyses were performed using SPSS V.22.0, and the level of significance was set at P<0.05, two sided.
More than two-thirds of patients (n=243) showed good collateral circulation, and there were several significant differences between patients with good and poor collateral circulation and between treatment arms (table 1), some reflecting the imaging criteria used to select patients for reperfusion, such as larger initial infarct core and lower ASPECTS in patients with poor collaterals not treated with MT.
The outcome variables are summarized in table 1. The effect of MT on infarct growth also varied across collateral categories (P=0.004), with a more significant reduction in infarct growth in patients treated with MT compared with those managed medically in the poor collateral category than in the good collateral category (38.8 mL vs 1.9 mL difference, figure 2). Infarct growth was associated with a significantly less likelihood of achieving functional independence (OR 0.964 per each mL, 95% CI 0.953 to 0.974, P<0.001) after adjusting for potential confounders. We tested for interaction between the time to CT and the category of collaterals on infarct growth, and there was none in the entire cohort or in the treatment arms. However, these analyses were underpowered as the vast majority of patients (84%) in this study were evaluated in the first 6 hours.
Patients treated with MT had better functional outcome at 90 days in both collateral circulation categories, and there was a significant interaction between the effect of MT and functional outcome (P<0.001), with a more significant effect in patients with poor collaterals (figure 3A). Functional independence occurred in 41% of patients treated with MT despite poor collaterals. The independent effect of MT over the functional outcome of patients was more substantial in patients with poor collaterals in ordinal regression models adjusted for potential confounders (age, sex, admission NIHSS, ASPECTS, intravenous tissue plasminogen activator administration, occlusion site, time to imaging, initial infarct core volume, initial delay volume, atrial fibrillation, and treatment period) (figure 3B and see online supplemental table 1). There was a significant interaction between the treatment modality and all cause mortality in relation to collateral status (P=0.007), with significantly reduced mortality after MT in the group of patients with poor collaterals, but not in patients with good collaterals (table 1). The rate of sICH was similar in all groups, with no interaction between treatment modality and sICH in relation to collateral status (P=0.997).
The quality of collateral circulation has consistently been linked with outcome in stroke patients, and also in patients treated with MT. For this reason, the most interesting observation of the present study is the interaction between collateral grade and effect of MT on infarct growth, suggesting that patients with the worst collateral perfusion may benefit from the treatment even more than those patients with good collateral circulation. Infarct growth was a strong predictor of poor recovery and increased with worse collaterals, and in patients treated with MT, infarct growth was not different in patients with good or poor collateral circulation. Given the clinical relevance of infarct growth, salvage of large amounts of tissue with MT in patients with poor cerebral perfusion may have profound clinical benefits.5
Recent studies have shown that good collateral circulation may preserve penumbral volume over time,22 and that the effect of the treatment in patients with poor collaterals may be limited to the early time window,23 with efficacy declining more abruptly than in patients with good collaterals. Although we found no interaction between time to imaging and effect of MT in the different patient subgroups, this study was not well powered to answer this question because it included patients treated mainly in short time windows, and the efficacy of MT in patients with poor collateral circulation in the late time windows remains, therefore, unknown. However, it is reasonable to think that the infarct growth protecting effect of MT may decline over time, especially in patients with poor collaterals.
Previous knowledge on the prognostic relevance of collaterals highlighted better outcomes after MT in patients with good collateral circulation. While trials that demonstrated the superiority of MT compared with medical management alone often used imaging for selecting those patients most likely to achieve excellent functional outcomes with endovascular recanalization, our results suggest that the range of patients that may benefit from MT may be broader. The magnitude of the benefit obtained with MT in our cohort in the groups of patients with good collaterals is in line with those described in these recent trials (OR approximately 1.8 and 2.5, depending on the specific outcome variables). However, the value of MT could be even more significant in the subgroup of patients with poor cerebral perfusion at very high risk of developing large ischemic lesions and severe neurological deficits. Given the strong association between larger infarct cores and worse outcomes, it is reasonable to expect worse outcomes in patients with poor collaterals, but the effect of MT in these patients was so crucial that a significant proportion (>1 in 3) achieved independent functional status after MT, whereas this was unusual with medical management.
The conclusions of the study, especially those related to clinical outcomes, are limited by its non-randomized nature but interestingly, a recent analysis of the HERMES collaboration on the imaging features of patients included in a randomized trial of MT suggested that a broad range of patients may benefit from MT, including those with poor collaterals.24 This evidence adds to previous reports of favorable outcomes after MT in patients not meeting the inclusion criteria of clinical trials.25 26 The analysis of collaterals in HERMES did not reveal an interaction between the effects of MT and collateral grade, but it did not include information on infarct growth and it was potentially limited by selection bias, because one of the trials (Endovascular Treatment for Small Core and Anterior Circulation Proximal Occlusion with Emphasis on Minimizing CT to Recanalization Times (ESCAPE)) excluded patients with poor collaterals, and others (Solitaire With the Intention For Thrombectomy as Primary Endovascular Treatment (SWIFT PRIME) and Extending the Time for Thrombolysis in Emergency Neurological Deficits-Intra-Arterial (EXTEND-IA)) used perfusion studies to select patients. While it would be desirable to confirm the benefit of MT in patients with poor collateral circulation in randomized clinical trials, in the meantime, these results suggest that simple imaging protocols may be enough for the selection of patients that benefit from MT, at least in the early time window after stroke.
The results of the primary outcome measure of infarct growth are especially relevant because they describe a potential mechanism explaining the protective effect of MT in patients with severe stroke, and also because they may be less affected by confounding factors and, as mentioned above, the main limitation of this study is the baseline differences between groups due to the non-randomized allocation of treatment. Although the selection criteria for MT in our center did not explicitly include evaluation of collaterals, it seems evident that they influenced the choice of treatment, as demonstrated by the baseline differences in infarct core volumes between treatment groups. However, it is interesting to note that greater infarct cores at baseline in patients with poor collaterals not treated with MT might actually limit infarct growth in those patients and therefore underestimate the effect of MT on the primary outcome measure of the study. The finding of similar benefit in radiological and clinical outcomes in patients at the highest risk of developing large disabling strokes indicates that this benefit is due to the vast capacity of endovascular MT to savage tissue and therefore improve outcomes in these patients. Despite previous studies showing that patients with low ASPECTS scores may benefit from MT,27 we do not know how the quality of collateral circulation influences the effectiveness of MT in patients with low ASPECTS because most of the patients in this cohort had small infarct cores and no signs of large lesions on NCCT. However, in our experience, most of the patients arriving at hospitals in regions with well organized stroke alert protocols show high ASPECTS on first imaging23 24 28 and may, therefore, benefit from treatment even in the presence of poor collateral circulation.
Poor collateral circulation is a potent predictor of infarct growth and inferior functional outcomes in patients with ischemic stroke due to large vessel occlusion, and this association is stronger in patients not treated with MT. In the early time window, the benefit of MT compared with best medical treatment regarding infarct growth limitation may be even more substantial in patients with poor collaterals than in patients with good collaterals.
AR and CL contributed equally.
Contributors ARJ and CL made substantial contributions to the conception and design of the study, collected and analyzed the clinical data, acquired, processed and analyzed the radiological data, and drafted the manuscript for intellectual concept. CM made substantial contributions to the collection and analysis of the clinical data. YZ made substantial contributions to the acquisition, processing, and analyzing of the radiological data. SR, LL, SA, and VO made substantial contributions to the collection and analysis of the clinical data. NM, FZ, MW, and JM made substantial contributions to the acquisition, processing, and analysis of the radiological data. AC and XU made substantial contributions to the conception and design of the study, collected and analyzed the clinical data, drafted the manuscript for intellectual content, and revised the draft critically.
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 None declared.
Ethics approval The local ethics committee at the Hospital Clinic approved the study (reg code HCB/2018/0680).
Provenance and peer review Not commissioned; externally peer reviewed.
Patient consent for publication Not required.