Time–outcome relationship in acute large-vessel occlusion exists across all ages: subanalysis of RESCUE-Japan Registry 2

Early reperfusion after endovascular thrombectomy is associated with an improved outcome in ischemic stroke patients; however, the time dependency in elderly patients remains unclear. We investigated the time–outcome relationships in different age subgroups. Of 2420 patients enrolled in the RESCUE-Japan Registry 2 study, a study based on a prospective registry of stroke patients with acute cerebral large-vessel occlusion at 46 centers, we analyzed the data of 1010 patients with successful reperfusion after endovascular therapy (mTICI of 2b or 3). In 3 age subgroups (< 70, 70 to < 80, and ≥ 80 years), the mRS scores at 90 days were analyzed according to 4 categories of onset-to-reperfusion time (< 180, 180 to < 240, 240 to < 300, and ≥ 300 min). In each age subgroup, the distributions of mRS scores were better with shorter onset-to-reperfusion times. The adjusted common odds ratios for better outcomes per 1-category delay in onset-to-reperfusion time were 0.66 (95% CI 0.55–0.80) in ages < 70 years, 0.66 (95% CI 0.56–0.79) in ages 70 to < 80 years, and 0.83 (95% CI 0.70–0.98) in ages ≥ 80 years. Early reperfusion was associated with better outcomes across all age subgroups. Achieving early successful reperfusion is important even in elderly patients.


Subjects.
This study is a post-hoc analysis of RESCUE-Japan Registry 2 15 , which was a prospective multicenter registry that enrolled 2420 patients with acute cerebral large-vessel occlusion at 46 centers in Japan between October 1, 2014 and September 30, 2016. RESCUE-Japan Registry 2 was designed to clarify the generalizability of the effectiveness of endovascular therapy in real-world patients. We enrolled consecutive patients aged ≥ 20 years who were hospitalized within 24 h of the onset of acute cerebral large-vessel occlusion.
Imaging and endovascular therapy. The diagnostic and treatment modalities were not unified in the RESCUE-Japan Registry 2. The Alberta Stroke Program Early Computerized Tomography Score (ASPECTS) was derived from computed tomography (CT) or magnetic resonance diffusion-weighted imaging (DWI) 16,17 . In patients with stroke in the posterior circulation, we measured the posterior circulation ASPECTS (pc-ASPECTS) using DWI 18 . The treatment modalities were determined by the attending physician. In this study, endovascular therapy included thrombectomy using stent retrievers and/or aspiration catheters, balloon angioplasty, stenting, intra-arterial fibrinolysis, piercing using guidewires and/or microcatheters, or a combination of these treatments, all of which have been approved in Japan. The stent retrievers used in this study were the Solitaire ™ 2 revascularization device (Covidien, Irvine, CA), the Trevo ® ProVue retriever/Trevo ® XP ProVue retriever (Stryker, Fremont, CA), and the Revive ® retriever (Codman, Raynham, MA). The aspiration catheter used was the Penumbra ® system (Penumbra, Alameda, CA). Other devices for endovascular therapy procedures such as stenting or angioplasty were selected by the physicians in charge.
Variables and measurements. We obtained clinical information of patients from hospital charts. Followup information up to 90 days was collected primarily through a review of the hospital charts or by contacting the patients, relatives, and/or physicians. We used the following clinical data for the analyses in the current study: age, sex, vascular risk factors (i.e., atrial fibrillation, hypertension, and diabetes mellitus), pre-stroke modified Rankin Scale (mRS) score 19 , National Institute of Health Stroke Scale (NIHSS) score 20 , blood glucose level on admission, systolic blood pressure on admission, location of occlusion, use of intravenous recombinant tissue plasminogen activator (IV-rtPA), ASPECTS, onset-to-puncture time (OPT), ORT, modified thrombolysis in cerebral infarction (mTICI) score (0, no perfusion; 1, minimal perfusion; 2a, reperfusion of less than half of the previously occluded territory; 2b, reperfusion of more than half of the previously occluded territory; 3, complete reperfusion) 21 , symptomatic intracranial hemorrhage (sICH), and mRS score at 90 days after stroke. sICH was defined as intracranial hemorrhage within 72 h after stroke with neurological worsening of ≥ 4 points in the NIHSS score 22 . OPT and ORT were defined as the duration from the time that the patient was last seen to be well to the groin puncture and the end of the endovascular therapy, respectively. We defined successful reperfusion as an mTICI score of ≥ 2b.
We divided the patients into 3 age subgroups: < 70, 70 to < 80, and ≥ 80 years. Since the time-outcome relationship was demonstrated in patients with OPT of 6 h or less 7-10 , but not in patients with OPT of more than 6 h 23,24 , we focused on the early time window and divided ORT into 4 categories with thresholds of 180, 240, and 300 min. We compared the clinical background characteristics and outcome measurements according to the ORT categories in each age subgroup.
The primary outcome was set as the ordinal score on the mRS score at 90 days after stroke. The secondary outcomes were set as a good outcome (defined as mRS score ≤ 2) and mortality at 90 days after stroke. The safety outcome was defined as the occurrence of an sICH within 72 h after stroke. www.nature.com/scientificreports/ Statistical analyses. We analyzed the database of patients with successful reperfusion after endovascular therapy (mTICI score 2b or 3). We analyzed continuous variables using the Mann-Whitney U test or Kruskal-Wallis test, and expressed them as median values and interquartile ranges. We analyzed categorical data using the chi-square test, and expressed them as numbers and percentages. In each age subgroup, the trends in the distributions of mRS scores were analyzed according to the ORT categories using the Jonckheere-Terpstra test.
We developed multivariate logistic regression models to assess the association between ORT and outcomes by adjusting for the following clinically relevant factors: ASPECTS (≥ 6 or < 6), NIHSS score, pre-stroke mRS score, use of IV-rtPA, site of the main occlusions (anterior or posterior circulation), and vessels of the main occlusions (internal carotid artery [ICA] and M1 segment of the middle cerebral artery occlusion or other arteries). ORT was entered as a categorical or a continuous variable. Because the threshold of 6 for both anterior circulation ASPECTS and pc-ASPECTS was reported to show good discrimination ability 25,26 , we dichotomized both scores according to whether they were above or below 6. If both CT and DWI were performed before endovascular therapy, we used ASPECTS on DWI in patients with anterior circulation occlusion. We used pc-ASPECTS in patients with stroke in the posterior circulation. To assess potential selection bias, we performed sensitivity analyses. First, we developed multivariate logistic regression models using the multiple imputation method (Supplementary Methods). Second, we analyzed the database of patients with an OPT of ≤ 6 h. Third, we developed multivariate logistic regression model using ASPECTS on CT in patients with both ASPECTS on DWI and CT. Statistical significance was set at P < 0.05. We conducted all analyses with R software using the rms package (version 3.3.3; F Foundation for Statistical Computing, Vienna, Austria) 27 .

Results
Baseline characteristics. Among 2420 enrolled patients, 21 patients were excluded because 12 patients did not meet the eligibility criteria and 9 patients were lost to follow-up. Of the remaining 2399 patients, 1278 patients underwent endovascular therapy. Among them, 177 patients with unsuccessful reperfusion (mTICI 0-2a) and 7 patients without reperfusion time data were excluded. Patients without data of ASPECTS, NIHSS score, or mRS score before stroke were also excluded. Thus, 1010 patients were included in the analysis (Fig. 1).
Clinical characteristics according to the age subgroups are shown in Table 1. The proportions of patients with female sex, hypertension, atrial fibrillation, and pre-stroke mRS score > 2 were higher in the elderly subgroups. The NIHSS score and systolic blood pressure were higher in elderly patients. The proportion of patients with diabetes mellitus, IV-rtPA use, and ASPECTS ≥ 6 were not different among the age subgroups. The blood glucose level, location of the occlusion, OPT, and ORT were not different among the age subgroups.
The clinical characteristics according to the ORT categories in each age subgroup are shown in Supplementary Tables S1-S3. The proportion of patients with atrial fibrillation and an NIHSS score were higher in the shorter ORT categories in patients aged ≤ 80 years. The proportion of patients with IV-rtPA use was lower in those with an ORT > 300 min. The proportion of patients with ASPECTS ≥ 6 in patients aged < 70 years and the proportion of patients with ICA or M1 segment occlusion in patients aged 70 to < 80 years was different among the ORT categories.  (Table 2). These adjusted common odds ratios were significantly different among age subgroups (interaction P = 0.031).

Secondary outcomes.
Of the 1010 patients with successful reperfusion, we analyzed 897 patients with pre-stroke mRS scores ≤ 2. In patients aged < 70 years and 70 to < 80 years, the proportion of good outcomes was higher in the shorter ORT categories (Fig. 3) Fig. 3). Although a marginal association between delayed ORT and mortality was observed in patients aged 70 to < 80 years, no associations between ORT and mortality were found in the other age subgroups. The adjusted odds ratios for mortality per 1-category delay in ORT were 1.36 (95% CI 0.82-2.26) in patients aged < 70 years, Table 1. Clinical characteristics according to age. ACA anterior cerebral artery, ASPECTS Alberta Stroke Program Early Computed Tomography Score, BA basilar artery, CT computed tomography, DWI diffusionweighted imaging, ICA internal carotid artery, IV-rtPA intravenous recombinant tissue plasminogen activator, IQR interquartile range, MCA middle cerebral artery, mRS modified Rankin Scale, NIHSS National Institutes of Health Stroke Scale, PCA posterior cerebral artery, VA vertebral artery. a ASPECTS was derived from CT or DWI. If both CT and DWI were performed before endovascular therapy, ASPECTS on DWI was used for the analysis. In patients with stroke in the posterior circulation, posterior circulation ASPECTS on DWI was used for the analysis.  Table 2). There was no heterogeneity with respect to age subgroups (interaction P = 0.75). No difference in the rate of sICH among the ORT categories was observed in any age subgroup (Supplementary Table S4).

Sensitivity analysis.
After the imputation of missing data using the multiple imputation method, the adjusted common odds ratios for better outcomes per 1-category delay in ORT were 0.66 (95% CI 0.55-0.79) in patients aged < 70 years, 0.67 (95% CI 0.56-0.79) in patients aged 70 to < 80 years, and 0.79 (95% CI 0.67-0.93) in patients aged ≥ 80 years (Supplementary Table S5), which also showed a significant time-outcome relationship in all age subgroups. These adjusted common odds ratios were not different among the age subgroups (interaction P = 0.087).
In patients who underwent both CT and DWI before endovascular therapy in patients with anterior circulation occlusion, we used ASPECTS on CT for sensitivity analysis instead of ASPECTS on DWI. The adjusted common odds ratios for better outcomes per 1-category delay in ORT were 0.64 (95% CI 0.54-0.78) in patients Figure 2. Distributions of the modified Rankin Scale scores at 90 days according to onset-to-reperfusion time categories in each age subgroup. In each age subgroup, onset-to-reperfusion time was associated with better outcomes.  Table S5). These adjusted common odds ratios were significantly different among the age subgroups (interaction P = 0.015).

Discussion
In this current post-hoc analysis of a large clinical registry of patients with acute cerebral large-vessel occlusion, earlier reperfusion was associated with better outcomes across all age subgroups. To the best of our knowledge, this is the first report to show the time-outcome relationship in elderly patients using a large clinical registry. Our findings confirmed that achieving early successful reperfusion is important in patients of all ages. Previous studies have investigated the association between ORT and outcomes in different age subgroups; however, the results in the elderly subgroup differed between studies 13,14 . One of two studies showed that ORT was independently associated with a favorable outcome at 90 days in 34 patients aged ≥ 80 years 13 , whereas the other study did not find an association between ORT and a good functional outcome in 78 patients aged > 80 years with mTICI of 2b or 3 14 . In elderly patients, because of the severe stroke outcome 11 , it is difficult to show a small difference according to ORT. Given the retrospective nature of the study, an age-related selection bias for endovascular therapy may have influenced the difficulty in demonstrating the differences. The present study is important in that it showed the time-outcome relationship in patients of all ages, although the relationship was small and partial in elderly patients. Logistic regression models for primary outcome (mRS at 90 days) and good outcome (mRS ≤ 2 at 90 days) showed heterogeneity among age subgroups; however, the time-outcome relationship existed across all age subgroups (Table 2). In patients aged ≥ 80 years, the proportion of good outcomes according to ORT categories was not significantly different, but was significantly smaller in patients with ORT ≥ 300 min after adjustment for confounders (Fig. 3). In patients with OPT of 6 h or less, shorter time to reperfusion was consistently associated with better outcomes across all age subgroups without heterogeneity (Supplementary Table S5). Thus, a shorter time to reperfusion is important for all ages.
We assessed the effects of early reperfusion in consecutive patients who were hospitalized within 24 h of the onset of acute large-vessel occlusion. However, randomized clinical trials using imaging-based selection criteria showed that there was no time-outcome relationship in patients with late hospital arrival 23,24 . Furthermore, a large database showed that the time-outcome relationships were stronger in patients with an OPT of ≤ 270 min than in those with an OPT of > 270 min 10 . Based on these previous reports, we set ORT categories with thresholds of 180, 240, and 300 min, and performed a sensitivity analysis in patients with an OPT of ≤ 6 h. Consequently, we focused on the early time window and observed consistent time-outcome relationships across all ages. However, we did not intend to determine the time limit for endovascular therapy. Our registry previously revealed that endovascular reperfusion therapy is effective, regardless of the onset-to-door time 15 .
Although a marginal association of early reperfusion with survival was observed in patients aged 70 to < 80 years, no associations were noted between ORT and mortality in the other age subgroups. In addition, Table 2. Adjusted odds ratios for outcomes according to ORT. CI confidence interval, mRS modified Rankin Scale, mTICI modified thrombolysis in cerebral infarction, ORT onset-to-reperfusion time, sICH symptomatic intracranial hemorrhage. a Adjustment was made for the Alberta Stroke Program Early Computed Tomography Score (ASPECTS) (≥ 6 or < 6), National Institutes of Health Stroke Scale score, pre-stroke modified Rankin Scale score, use of intravenous recombinant tissue plasminogen activator, site of the main occlusions (anterior or posterior circulation), and vessels of the main occlusions (internal carotid artery and M1 segment of the middle cerebral artery occlusion or other arteries). ASPECTS was derived from computed tomography (CT) or magnetic resonance diffusion-weighted imaging (DWI). If both CT and DWI were performed before endovascular therapy, ASPECTS on DWI was used for the analysis. In patients with stroke in the posterior circulation, posterior circulation ASPECTS on DWI was used for the analysis. b ORT was divided into four categories: < 180, 180 to < 240, 240 to < 300, and ≥ 300 min. www.nature.com/scientificreports/ no association between ORT and sICH was found in any age subgroup. Previous reports from the meta-analysis of the HERMES collaborators and a large practical database documented that both mortality and sICH were not associated with the onset-to-randomization time or OPT 1,12 , which is consistent with our current results. However, in patients aged ≥ 80 years, the odds ratios for mortality and sICH tended to be low. This may also be due to an age-related selection bias for endovascular therapy.
The clinical background characteristics in the present study were not different from those in previous reports. Male sex 13,14 , hypertension 13 , cardiogenic embolism 14 , and high NIHSS score 14 have been reported to be associated with elderly patients undergoing endovascular thrombectomy. Early hospital arrival was reported to be associated with atrial fibrillation, a high NIHSS score, and ambulance use in ischemic stroke patients 28 . This may be because atrial fibrillation is associated with stroke severity 29 . In our cohort, the proportion of patients with ASPECTS ≥ 6 in patients aged < 70 years and the proportion of those with ICA or M1 segment occlusion in patients aged of 70 to < 80 years was different among the ORT categories, although there were no trends according to ORT.
Several limitations of the current study should be noted. First, this was an observational study; thus, the selection of endovascular therapy depended on the practicing physician. Moreover, patients with unsuccessful reperfusion were excluded from the analysis. Furthermore, 8% of patients were excluded from the analysis owing to missing data. These factors may have led to a potential selection bias, although we conducted sensitivity analyses. Second, although we systematically registered stroke patients with acute large-vessel occlusion over a 2-year period at 46 centers, the sample size was not large enough to fully evaluate the time-outcome effect in each age subgroup. Third, the imaging methods were not standardized and perfusion imaging was not analyzed in this study. We used two types of ASPECTS: ASPECTS on CT, an ASPECTS on DWI in patients with anterior circulation occlusion. Although we conducted a sensitivity analysis, this was not sufficient to adjust for the impact of imaging on the outcome. Therefore, a study with modern standardized imaging methods is needed to evaluate the impact of time to reperfusion. Fourth, the assessments of the mRS scores could be biased, as acute treatment was not completely masked although the assessments were performed by independent physicians. Fifth, the assessments of the mTICI scores could be biased. A previous report revealed a discrepancy in the assessment of the mTICI score between local operators and independent core laboratories 30 . This may have led to a selection bias.

Conclusions
Our large prospective registry of acute large-vessel occlusion revealed that early reperfusion was associated with better outcomes across all age subgroups. Achieving early successful reperfusion is important in patients of all ages, even in octogenarians.

Figure 3.
Proportions of a good outcome and adjusted odds ratios for a good outcome according to onset-toreperfusion time (ORT) categories in each age subgroup. The proportion of a good outcome (defined as mRS score ≤ 2) was lower in the delayed ORT categories than in those with ORT < 180 min in each age subgroup, although the association was marginal in patients aged ≥ 80 years.