Background Patients presenting with minor ischemic stroke frequently have early neurological deterioration (END) and poor final outcome. The optimal management of patients with END has not been determined.
Objective To investigate rescue IA therapy (IAT) when patients with acute minor ischemic stroke develop END.
Methods This was a retrospective study of consecutively registered patients with acute minor stroke and END. ‘END’ was defined as an increase in National Institutes of Health Stroke Scale (NIHSS) scores by 1 or more points (or development of new neurological symptoms) and ‘ΔEND−NIHSS’ was defined as numerical difference between NIHSS scores at the time of END and before END. Rescue IAT following END was adjusted for the covariates to evaluate the association between IAT and favorable outcome at 3 months.
Results Among 982 patients with acute minor ischemic stroke, END occurred in 232 (23.6%). Of the 209 patients with END with full data available, 87 (41.6%) had favorable outcomes at 3 months. Rescue IAT following END was performed in 28 (13.4%). Favorable 3-month outcomes were seen in 50% of patients undergoing rescue IAT, including 8/19 (42.1%) undergoing rescue IAT beyond 8 h. By multivariate logistic regression analysis, rescue IAT following END was independently associated with favorable outcome at 3 months (OR=10.9; 95% CI 3.06 to 38.84; p<0.001).
Conclusions The results suggest that rescue IAT may be safe and effective when END occurs in selected patients with acute minor ischemic stroke. Further prospective and randomized studies are needed to confirm our results.
Statistics from Altmetric.com
Patients with acute ischemic stroke who present with minor deficits, and therefore do not undergo immediate intravenous fibrinolysis or catheter-based reperfusion therapy, often deteriorate after presentation and have a poor final outcome.1–4 Early neurological deterioration (END) occurs more often in patients with minor than moderate to severe presenting deficits. The presence of persistent large vessel occlusion despite minor deficits appears to identify a subgroup of patients with a minor stroke at particularly high risk for END and poor outcome.2 ,5
The success of different treatment strategies for patients with END after presentation with minor stroke deficits has not been well delineated. Rescue treatment with catheter-based reperfusion is a promising potential option. IA therapy (IAT) with recent mechanical devices has a high recanalization rate, and is quite safe.6 ,7 Although IAT certainly would not be considered as a primary treatment strategy for acute minor stroke, we presumed that rescue IAT might result in better outcomes when patients with acute minor stroke with relevant arterial occlusion develop END. Although time from onset to IAT is independently associated with favorable outcomes,8–10 IAT is related to good outcomes in some patients with stroke of unclear onset or wake-up stroke, irrespective of time to IAT.11 ,12
We investigated rescue IAT as a treatment when patients with acute minor ischemic stroke develop END.
Subjects and methods
This was a retrospective study of consecutively registered patients with acute ischemic stroke in our tertiary stroke center between November 2008 and May 2013. This study included patients who presented with acute cerebral ischemia with minor deficits within 6 h of the first known abnormal time (FAT); and developed END within 7 days of admission. Minor stroke was defined as a National Institutes of Health Stroke Scale (NIHSS) score of ≤5.3 We excluded patients with (1) non-thrombotic etiologies of ischemic stroke, such as vasculitis and Moyamoya disease, and patients with cancer-related stroke; and (2) prestroke disability, defined as a prestroke modified Rankin Scale (mRS) score of >1. This study was approved by the institutional review board of Chonnam National University Hospital. Written informed consent was not obtained because of the retrospective design of this study.
The following stroke risk factors were identified: age, sex, recent cigaret smoking (cigaret smoking within the past 5 years), hypertension, diabetes mellitus, dyslipidemia, and a previous history of stroke or transient ischemic attack. We assessed neurological status at admission and on each hospital day using NIHSS scores. Time intervals analyzed included FAT-to-hospital arrival time, FAT-to-END time (time of first END after FAT), FAT-to-IAT time, and END-to-IAT time. IAT time was defined as the time the clot was first accessed. Stroke subtypes were stratified according to the Trial of Org 10172 in Acute Stroke Treatment (TOAST) criteria after complete diagnostic investigations.13 In the primary analysis, ‘END’ was defined as an increase in NIHSS scores by ≥1 points (or development of new neurological symptoms) and ‘delta END−NIHSS (ΔEND−NIHSS)’ was defined as the numerical difference between NIHSS scores at the time of END and before END. In sensitivity analysis, END was defined as an increase in NIHSS scores by ≥4 points. END was regularly diagnosed and registered by a physician while monitoring the patient. Scores of 0–2 on the mRS at 90 days were defined as a favorable outcome.
The imaging protocol of acute ischemic stroke in our hospital has been previously described.5 Briefly, patients underwent emergency MRI at the emergency department immediately after admission. The MRI protocol consisted of diffusion-weighted imaging (DWI), fluid-attenuated inversion recovery, gradient echo (GRE) imaging, time-of-flight MR angiography (MRA), and perfusion-weighted imaging in sequence. In addition, DWI/GRE imaging was also performed if neurological deterioration occurred. For this study, the images were analyzed by two neurologists (J-TK and M-SP) who were blinded to the clinical data. Discrepancies were resolved by consensus conference. The lesion sites on DWI were categorized into anterior, posterior, or both circulations. The patterns of DWI lesions in anterior circulation were classified as perforating artery infarcts (PAI), pial infarcts (PI), borderzone infarcts (BI), territorial infarcts (TI), and lacunar infarcts (LI) by modifying previous studies.14 Hemorrhagic transformation (HT) was categorized as hemorrhagic infarcts (type 1 or 2) or parenchymal hematomas (type 1 or 2) according to a modified definition established by a previous study.15 We used GRE imaging for the categorization of HT. Arterial occlusion was defined as a complete loss of distal flow signal. Moderate to severe arterial stenosis was defined as a >50% narrowing of the lumen and focal signal loss in the presence of a distal flow signal. Arterial occlusion sites that supplied the acute ischemic region were defined as ‘relevant arterial occlusion’ and determined via analysis of the initial MR angiogram. IA recanalization at the end of the IAT procedure was defined as grade 2 or 3 of the Thrombolysis in Myocardial Infarction scale.
Management following END was decided at the discretion of the attending physicians. Addition of antiplatelets, administration of intravenous antithrombotic agents, such as tirofiban or argatroban, or rescue IAT were the management options for END. Decisions on rescue IAT were made on the basis of neurological symptoms, DWI at the time of END, and initial/follow-up MRA. Rescue IAT was generally performed according to the following criteria: (1) catheter-accessible symptomatic arterial occlusion on MRA, (2) small DWI lesions on initial imaging and after END, and (3) the presence of cortical signs or symptoms. However, even if patients met the criteria for rescue IAT, physicians decided to carry out or not to carry out rescue IAT on the basis of their case-specific assessment of risk and benefit of IAT. Informed written consent for rescue IAT was given by each patient or family. Rescue IAT was performed using a variety of methods, including mostly mechanical thrombectomy (Solitaire devices). In addition, clot location was assessed during IAT.
Data are presented as mean±SD or the frequency of categorical variables. Categorical variables were analyzed using the χ2 test and Fisher's exact test, when appropriate. Continuous variables were analyzed using the independent sample t test or the Mann–Whitney U, test when appropriate. Multivariate logistic regression analysis was used to evaluate independent factors associated with favorable outcomes at 3 months (adjusted for age, baseline NIHSS, TOAST classifications, time from FAT to visit, diabetes mellitus, relevant arterial occlusion, ΔEND−NIHSS, and rescue IAT). For the sensitivity analysis, in patients with relevant arterial occlusion on initial MRA (n=92), rescue IAT was adjusted for the following covariates to evaluate the association between rescue IAT and favorable outcome at 3 months: model 1 was adjusted for age, initial NIHSS, and TOAST classifications; and model 2 was adjusted for age, initial NIHSS, TOAST classifications, and ΔEND−NIHSS. ORs and 95% CIs were calculated. A p value of <0.05 was considered statistically significant. All statistical analyses were performed using SPSS for Windows, V.17 (SPSS Inc, Chicago, Illinois, USA).
During the study period, 982 patients with acute minor ischemic stroke presented within 6 h of FAT. Of these, 287 (29.2%) patients had relevant arterial occlusion. END occurred in 232 (23.6%) patients. Of these, 23 were excluded: 10 owing to loss to follow-up within 7 days after early discharge, seven owing to non-thrombotic or cancer-related etiologies, and six owing to prestroke mRS >1. Among the resulting analytic population of 209 patients with minor ischemic patients and END, mean age was 68.0±11.6 years and 117 (56.0%) were male. Of the 209 patients, 87 (41.6%) had favorable outcomes at 3 months, and 92 (44.0%) had relevant arterial occlusion on initial MRA. END occurred a median of 19 h after FAT and 16 h and 3 min after hospital arrival. Rescue IAT was performed on 28 (13.4%).
Characteristics of patients undergoing rescue IAT
Characteristics of patients undergoing and not undergoing rescue IAT after END are shown in table 1. ΔEND−NIHSS was significantly higher in patients undergoing rescue IAT. Patients with atrial fibrillation and relevant arterial occlusion on initial MRA also more frequently underwent rescue IAT. ΔEND−NIHSS values were higher and time from FAT to END was faster in patients treated with rescue IAT. Some patients with no initial steno-occlusion (n=2, 7.1%) or initial symptomatic stenosis (n=5, 17.9%) also underwent rescue IAT. END in the patients with relevant stenosis were treated with angioplasty (and stenting). The patients without relevant arterial steno-occlusion who were treated with rescue IAT had developed new interval occlusions. HT was non-significantly more common in patients undergoing rescue IAT, but most (5/7) were hemorrhagic infarction rather than frank hematoma. On univariate analysis, there were no significant differences in frequency of favorable outcomes at 3 months and death within 3 months between the two groups.
Among the 28 patients undergoing rescue IAT, nine received IAT within 8 h from FAT to IAT. Six of these nine patients had favorable outcomes at 3 months. The other 19 patients received IAT beyond the 8 h time window (median 30 h; IQR 16) and 8/19 (42.1%) patients had favorable outcomes at 3 months (see online supplementary table S1). Some patients with no initial steno-occlusion (n=2) or initial symptomatic stenosis (n=5) also underwent rescue IAT following END. END in the patients with relevant stenosis were treated with angioplasty (and stenting). The patients without relevant arterial steno-occlusion who were treated with rescue IAT had developed new interval occlusions. HT was non-significantly more frequent in patients undergoing rescue IAT following END, but most (5/7) were hemorrhagic infarction rather than frank hematoma. Initial lesions or lesion patterns were 13 cortical small lesions (PI), nine scattered borderzone lesions, six basal ganglia lesions (PAI or LI), and six small territorial lesions (partial TI). On follow-up DWI after END, four no lesion changes, three new lesions in different territories beyond the initial lesions, 13 new pial or borderzone lesions in the initial lesion territories, growth of two small lesions in the initial territory and five edematous changes were shown.
Factors associated with favorable outcomes at 3 months
Online supplementary table S2 presents the general characteristics of patients with and without favorable outcomes at 3 months. On univariate analysis, patients with favorable outcomes had lower baseline NIHSS and ΔEND−NIHSS than those without, and less often had a relevant large artery steno-occlusion. Multivariate analysis indicated that rescue IAT was independently associated with increased odds of favorable outcome at 3 months (OR=10.9; 95% CI 3.06 to 38.84; p<0.001). Factors independently associated with reduced odds of favorable outcome at 3 months were higher NIHSS scores, longer interval from FAT to hospital, diabetes mellitus, relevant arterial occlusion, and higher ΔEND−NIHSS (table 2).
In sensitivity analysis, when END was defined as NIHSS worsening by ≥4 points rather than ≥1 points, END occurred in 115 patients, and rescue IAT remained independently associated with favorable outcome after adjustment for age, initial NIHSS, diabetes mellitus, and ΔEND−NIHSS (OR=6.66; 95% CI 2.04 to 21.7; p=0.002).
Characteristics of patients with relevant arterial occlusion on initial MRA
In patients with END with relevant arterial occlusion on initial MRA (n=92), rescue IAT was performed in 23% and not performed in 77% following END (table 3). Patients with middle cerebral artery or internal carotid artery occlusion more frequently underwent rescue IAT following END than those with other arterial occlusion. In patients with lesions in the anterior circulation, the lesion patterns of BI and PAI on DWI had a high frequency of rescue IAT. ΔEND−NIHSS values were higher in patients treated with rescue IAT.
On univariate analysis, rescue IAT and low baseline NIHSS were related to favorable outcomes at 3 months (see online supplementary table S3). Large artery atherosclerosis according to TOAST classifications and higher ΔEND−NIHSS were related to unfavorable outcomes at 3 months. By multivariable logistic regression analysis, rescue IAT, low baseline NIHSS, and low ΔEND−NIHSS were independently associated with favorable outcomes at 3 months (table 4).
The mean time from FAT to IAT was 1277±1684 min (median 670 min; IQR 1099 min). The mean time from END to IAT was 103±98 min (median 80 min; IQR 57 min). In patients undergoing rescue IAT, there was no significant difference in IAT times (from FAT to IAT and from END to IAT) between patients with favorable outcomes and those without (median time from FAT to IAT, 1080 min (IQR 843 min) vs 580 min (IQR 1250 min); p=0.62 and median time from END to IAT, 90 min (IQR 78.75 min) vs 60 min (IQR 59 min); p=0.40).
The results of this study indicate that rescue IAT among patients with acute minor ischemic stroke and END is independently associated with favorable outcome at 3 months. Other factors associated with favorable outcome were lower presenting stroke severity, less severe deficit progression at time of deterioration, shorter time from onset to hospital arrival, absence of diabetes, and absence of a relevant large vessel occlusion. Substantial rates of good outcome were obtained among patients treated with rescue IAT within 8 h of initial ischemia onset, and also among patients treated beyond this time window, at a median of 21 h and 30 min after initial stroke onset.
Our findings of END in nearly one-quarter of patients with minor ischemic stroke are consonant with the many prior studies that have shown a high frequency of END among patients presenting with minor deficits.2 There have been few formal studies evaluating management strategies following END. Medical treatment includes continuing supportive care, hypervolemic therapy, induced hypertension, intensified antiplatelet therapy, or acute anticoagulation. Previous END series have not fully specified the frequency and outcomes of each of these strategies, but have shown the outcome to be a high rate of disability among patients in whom END is treated medically. This study is worthwhile in systematically investigating endovascular recanalization in patients with END and showing that rescue IAT is associated with substantially improved final outcomes.
In our cohort, rescue IAT for END was pursued more often when patients had more severe neurological deficit worsening, had a relevant arterial occlusion present on imaging, and had atrial fibrillation. In addition, although volumes of infarction were not measured in this study, most patients receiving rescue IAT had small new lesions on follow-up DWI with initially small lesion patterns. These selection factors probably reflect a greater willingness to intervene when the progression is striking and a target arterial occlusion is available. The association of rescue IAT with improved outcome suggests that patients with these factors will benefit from rescue intervention. It is striking in our cohort that rescue IAT averted poor outcomes in patients with more severe deterioration and with relevant arterial occlusions, features that usually predict worse outcome following END. However, because we did not use a formal protocol to trigger the use of rescue IAT, such a conclusion can only be tentative. It is also possible that selection factors might have magnified the apparent treatment benefit of rescue IAT.
Of note, about 70% of patients who underwent rescue IAT in our study were treated beyond the conventional 8 h time window for IAT for presenting deficits, with an overall median time to treatment of 11 h 10 min. Despite the late intervention, approximately half of patients who underwent rescue IAT had favorable outcomes at 3 months. These findings are consistent with those of the DEFUSE 2 (Diffusion and Perfusion Imaging Evaluation For Understanding Stroke Evolution) prospective observational study, which found that patients with substantial penumbral tissue had equally good outcomes from endovascular reperfusion whether they were treated early (≤6 h) or late (>6 h).16 This finding suggests that patients who can maintain a penumbra longer, represent a subpopulation who will most benefit from reperfusion.17 In our study, patients with relevant arterial occlusion on initial MRA had only minor symptoms, suggesting that cerebral blood flow to substantial tissue volumes were mildly reduced, but sufficient to support synaptic function and not initially place the tissue in immediate jeopardy—that is, the benign oligemic state.18 However, at the time of END, these patients probably had delayed collateral failure (cases with borderzone infarct patterns), clot migration or propagation or early recurrent stroke, subjecting new regions to reduced blood flow and imminent infarction.19
Although rescue IAT was performed at late time intervals in this series, because the time of recurrence to emergency thrombus extraction might be much shorter, the rates of hemorrhage were low. In addition, the good safety profile probably primarily reflects the fact that only the small brain regions associated with the initial minor deficits developed severe ischemia for prolonged durations before the intervention and that mechanical intervention potentially has less hemorrhagic risk than pharmacologic fibrinolysis. Furthermore, patients with END following initial minor symptoms may have reduced vulnerability to reperfusion injury because of ischemic preconditioning. In this series the frequency of parenchymal hematoma and death was similar between patients treated or not treated with rescue IAT.
The degree of neurological deficit progression required to consider that a patient has evidence of neurological deterioration has varied in the literature, from 0 points on the NIHSS (any worsening even if not indexed by the scale) to 1, 2, 4, and other point thresholds. We defined END in this study as worsening by ≥1 point in order to be broadly inclusive. However, in the sensitivity analysis, the association of rescue IAT with improved clinical outcome was robust, and present also when END was defined as a ≥4-point increase on the NIHSS.
The results of this study are subject to some limitations. First, this is a single-center, retrospective study. Second, intervention was at the physician's discretion (potential selection bias), and outcome assessments were not blinded (potential assessment bias). Because of these limitations, the results of our study should be interpreted with caution. The findings need to be confirmed in a randomized controlled trial. Further studies are warranted to confirm our results and to delineate optimal criteria for rescue IAT in patients with END.
In conclusion, the results of this study suggest that rescue IAT may be safe and effective when END occurs in selected patients with acute minor ischemic stroke. Further prospective and randomized studies are needed to confirm our results.
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Files in this Data Supplement:
Contributors Substantial contributions to the conception or design of the work: J-TK, S-HH, JLS or the acquisition, analysis, or interpretation of data for the work: J-TK, K-HChoi, M-SP. Drafting the work or revising it critically for important intellectual content: J-TK, JLS, S-HH. Final approval of the version to be published and agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved: J-TK, S-HH, WY, M-SP, K-HChoi, KHCho, JLS.
Funding This work was supported by a grant (CRI13041-21) from Chonnam National University Hospital Biomedical Research Institute and a grant from the Korea Healthcare technology R&D Project, Ministry of Health and Welfare, Republic of Korea (HI10C2020).
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.
Data sharing statement We agree to share the data from this study; there are no additional unpublished data.
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.