Objective The rate of progression of the ischemic lesion is variable in patients with stroke. We tested the hypothesis that the tissue saving effect of mechanical thrombectomy (MT) is greater in fast progressors.
Methods A single-center cohort of consecutive patients (n=242) with occlusions of the terminal internal carotid or M1 segment of the middle cerebral artery treated with MT (n=195) or best medical treatment (n=47), known time from onset, and full imaging (baseline CT perfusion and follow-up MRI) available was studied. The estimated infarct progression rate (eIPR) was calculated at baseline and patients were categorized as fast/slow progressors according to the median eIPR of 4.8 mL/hour. The primary outcome measure was the interaction between eIPR category and MT on infarct growth. The secondary outcomes assessed the effect of MT on final infarct volume and functional status in relation to the eIPR category. The safety outcomes were mortality and symptomatic intracranial hemorrhage.
Results The eIPR category had a modifying effect (Pi=0.017) of MT on infarct growth that was significantly reduced with MT only in fast progressors (median (IQR) 3.8 mL (−11–55) vs 41 mL (11–107) with medical treatment; p=0.009, adjusted p=0.045). There was also a significant interaction on final infarct volume (Pi=0.005), with a greater reduction after MT in fast progressors. The functional status improved with MT both in fast and slow progressors, with no significant modifying effect of eIPR category (Pi=0.201). There were also no significant interactions on safety outcomes.
Conclusion MT in stroke patients with large vessel occlusion limits infarct growth more significantly in fast progressors.
- CT perfusion
Statistics from Altmetric.com
The infarct progression rate1 2 and the volume of infarcted tissue3–5 have been consistently associated with the functional outcomes in patients with large vessel stroke. The velocity of infarct growth is variable in stroke patients,6 most likely due to uneven expansion of the ischemic core from the penumbral area in relation to the efficacy of the collateral blood flow. While the value of endovascular reperfusion was established in clinical trials that mostly included patients with small infarct cores, non-randomized studies suggest that patients with larger infarcts may also benefit from mechanical thrombectomy (MT)7 8 and further studies are needed on the efficacy of thrombectomy in stroke patients with poor natural history.9 The experience obtained in previous non-randomized studies has been useful to design three ongoing trials that are testing the value of endovascular thrombectomy compared with best medical care alone in patients with acute ischemic stroke with large lesions (IN EXTREMIS, TENSION, and TESLA).10–12
The objective of this study was to assess whether the pretreatment infarct progression rate influences the response to MT in stroke patients with large vessel occlusion. For this, we compared the relative efficacy of MT over best medical care to limit infarct growth in patients with fast or slow infarct progression, with the primary hypothesis that the tissue saving effect of MT may be more significant in the former group.
The study population (n=242) was part of a prospectively collected clinical registry of patients with acute ischemic stroke admitted to a Comprehensive Stroke Center. As shown in the flow diagram (figure 1), we included all consecutive patients with occlusions of the terminal internal carotid or M1 segment of the middle cerebral artery, known time from onset, and full imaging information. The treatment allocation (MT in 195 patients and best medical treatment only in 47) was not randomized but decided by the treating physician, and the reasons for exclusion from thrombectomy are summarized in the flow diagram, the most frequent being participation in the REVASCAT trial.13 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 a non-contrast CT, CT perfusion, and CT angiogram acquired in a Siemens Somatom Definition Flash unit, and a follow-up MRI performed on a 1.5T Siemens Magneton Aera unit (Siemens, Erlangen, Germany). The volume of non-viable tissue was calculated using Apollo MIStar software according to a threshold of relative cerebral blood flow (rCBF) <25% of the contralateral hemisphere within the tissue with delay >3 s. Assuming a constant infarct growth, the infarct progression rate was estimated by dividing the non-viable tissue volume by the time delay to imaging.6 14 The relationship between estimated infarct progression rate (eIPR) and infarct growth was analyzed, and the best model fit corresponded to a quadratic model in patients not treated with MT whereas there was no relationship in patients treated with MT (figure 2). The median eIPR value adequately differentiated two groups with widely different infarct growths, and for that reason, the patients were categorized according to the median eIPR of the cohort as fast (≥4.8 mL/hour) or slow (<4.8 mL/hour) progressors. Although the categorization of patients would be mostly the same, we chose the 25% threshold of rCBF instead of the often used threshold of 30% to avoid overestimating the eIPR and minimizing negative infarct growth values. Diffusion-weighted image (DWI) lesion volumes were calculated using Amira software (Visage Imaging, Berlin, Germany) through a semiautomated thresholding method. Infarct growth was calculated by subtracting the baseline infarct core from the final infarct volume. Symptomatic intracranial hemorrhage (sICH) was defined as any parenchymal hematoma associated with an increment of at least 4 points on the National Institutes of Health Stroke Scale (NIHSS) score.
Statistical analysis and outcome measures
The primary outcome measure was the interaction between eIPR category and MT on infarct growth. The secondary outcomes included the interaction between eIPR category and MT on the final infarct volume and the functional outcome at 90 days measured using the modified Rankin Scale (mRS) by certified neurologists. The safety outcomes were mortality and sICH. We tested the interactions in linear (infarct growth, final infarct), ordinal (mRS shift), or binary logistic regression (mortality and sICH) models. Multivariate models also tested the independent effect of MT on infarct growth and final infarct volume, including eIPR and variables with intergroup differences (p<0.1) as covariates to account for baseline differences between treatment arms. Infarct growth and final infarct volume did not follow a normal distribution, and we compared them using the Mann–Whitney test. We did not apply normality transformations because the primary objective of the linear regression models was not inference. We performed all the analyses using SPSS Version 22.0, setting the level of significance at p<0.05 (two-sided).
Slow and fast progressors showed several differences both at baseline and during follow-up (table 1). There were also some baseline differences between patients treated with MT and those receiving medical management within each eIPR category (table 1).
The eIPR category had a modifying effect (Pi=0.017) of MT on infarct growth that was significantly reduced with MT only in fast progressors (median (IQR) 3.8 mL (−11–55) vs 41 mL (11–107) with best medical treatment; p=0.009, adjusted p=0.045; table 1, figure 3). There was also a significant interaction on final infarct volume (Pi=0.005), with a greater reduction after MT in fast progressors (median (IQR) 26 mL (11–74) vs 78.5 mL (51–144) with best medical treatment; p=0.002, adjusted p=0.052; table 1, figure 3). The functional status improved with MT both in fast and slow progressors (table 1), with no significant modifying effect of the eIPR category (p=0.201). The overall adjusted OR for improvement across the mRS with MT was 3.7 (95% CI 2.0 to 6.9, p<0.001). There were also no significant interactions of eIPR and MT on mortality (Pi=0.144) and symptomatic intracranial hemorrhage (Pi=0.998).
Our study shows that patients who are fast progressors benefit greatly from MT compared with best medical care in terms of less infarct expansion, lower infarct volume, and better functional outcomes. As in previous reports, the velocity at which ischemic penumbral tissue is converted into infarct was variable in our population, and the results of this study reinforce the idea that MT may be beneficial in patients with severe stroke by showing that MT resulted in a more substantial limitation of infarct growth in patients with fast eIPRs. According to previous estimations,6 compared with medical management, MT would save from infarction as many as 783 million neurons, 5.7 trillion synapses, and 4914 km of myelinated fibers in a fast progressor patient.
Patients with large infarcts at baseline were mostly excluded from the clinical trials of MT because of the belief that reperfusion could be even detrimental due to reperfusion injury, high risk of sICH, and malignant edema.15 However, subgroup analyses of the patients included in these trials16 17 and other non-randomized studies18 19 suggest that more patients with stroke due to proximal vessel occlusion may benefit from MT. As expected, fast progressors had more severe strokes with more severe neurological deficits, larger infarct cores, and poorer collateral circulation. Despite similar rates and successful endovascular reperfusion, fast progressors had larger final infarct volumes and worse functional outcomes. However, the benefits of MT on radiological outcomes were more pronounced in fast progressors, and the benefits in terms of functional outcome were similar in both eIPR categories.
There were no safety concerns and mortality was even reduced in patients with fast eIPRs. A recent meta-analysis has observed that MT contributes to significant benefits in survival during the first 90 days after acute ischemic stroke compared with medical therapy alone (15% vs 18.7%), with a number needed to treat of 27 to prevent one death.20 In our study, the effect of thrombectomy on survival at 90 days was much more significant in fast progressors, with a number needed to treat of just 6 to prevent one death, whereas there was no survival benefit in slow progressors. The observed reduction in mortality does not seem to originate from differences in adverse events such as severe hemorrhagic complications but from functional benefits of MT in patients who are at high risk of developing catastrophic infarcts under medical treatment alone. Overall, these results add to previous evidence showing that, at least in the early time window, patients with large vessel stroke benefit from MT despite having poor prognostic factors such as large lesions,7 8 poor collateral circulation,21 or systemic biomarkers of poor outcome.22 23 Although the results of ongoing randomized trials studying whether MT has a therapeutic effect in patients with large established lesions will be of utmost importance, these results suggest that the severity of the hypoperfusion should not be used to exclude patients from revascularization, at least in the early therapeutic window. A better understanding of the clinical and neuroimaging parameters by which to select these patients is imperative, and ongoing clinical trials will hopefully help us to serve this devastated patient population better.
The main limitations of the study are the lack of randomized treatment allocation and the existence of baseline differences between patients treated with MT or best medical treatment. For that reason, we chose as the primary outcome measure a quantitative radiological measure like infarct growth that may be less affected by clinical confounders and might even result in underestimation of the effect of MT due to some of the baseline differences (note that patients with larger infarct cores might be closer to the ceiling effect for further infarct growth). Most of the patients in this study were treated in the early time window and we cannot extrapolate our findings to later time windows. As the trials that showed similar benefits from thrombectomy in late time windows24 25 used advanced imaging to select patients with small infarct cores, further studies are needed to assess whether patients with faster IPRs also benefit from MT in later time windows.
MT in stroke patients with large vessel occlusion limited infarct growth more significantly in patients with fast infarct progression and there were no safety concerns of the endovascular procedure, with even improved survival compared with patients managed medically.
ARJ and XU contributed equally.
Contributors AR and CL made substantial contributions to the conception and design of the study; collected and analyzed clinical data; acquired, processed and analyzed radiological data; and drafted the manuscript for intellectual concept. CM made substantial contributions to the collection and analysis of clinical data. SR, LL, SA, VO, and RT made a substantial contribution to the collection and analysis of clinical data. JB and JM made substantial contributions to the acquisition, processing, and analysis of radiological data. AC and XU made substantial contributions to the conception and design of the study; collected and analyzed clinical data; drafted the manuscript for intellectual content and revised the draft critically.
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.
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.