Intracranial arterial stenosis and recurrence in stroke patients with different risk stratifications by Essen stroke risk score

ABSTRACT Objectives We sought to investigate whether the prognostic value of intracranial arterial stenosis (ICAS) is consistent across different risk stratifications using the Essen Stroke Risk score (ESRS). Methods We derived data from the Clopidogrel in High-Risk Patients with Acute Nondisabling Cerebrovascular Events trial. Patients without complete baseline brain imaging data were excluded. Participants were categorized into different risk groups based on ESRS (low risk, 0–2, and high risk ≥ 3). The main outcome was stroke recurrence within 3 and 12 months. Hazard ratios (HRs) and 95% confidence intervals (95%CIs) of ICAS, and other factors associated with stroke recurrence within 3 and 12 months were estimated using the Cox regression method. Results During the 3-month follow-up, 54 patients (7.9%) had recurrent stroke in the low-risk group, and 39 patients (9.6%) had recurrent stroke in the high-risk group. ICAS was associated with a higher risk of stroke within 3 months (HR = 2.761; 95%CI = 1.538–4.957; P < 0.001) in the low-risk group, but not in the high-risk group (HR = 1.501; 95%CI = 0.701–3.213; P = 0.296). ICAS was independently associated with higher recurrent risk in the low-risk group (HR = 2.540; 95%CI = 1.472–4.381; P < 0.001), but not in the high-risk group (HR = 1.951; 95%CI = 0.977–3.893; P = 0.058) within 12 months. Conclusion ICAS was an independent predictor of both 3- and 12-month stroke recurrence in low-risk but not high-risk patients with minor ischemic stroke or transient ischemic attack according to ESRS stratification.


Introduction
Acute minor ischemic stroke (MIS) and transient ischemic attack (TIA) are common manifestations of ischemic cerebrovascular disease, accounting for approximately 65% of the total incidence [1].The recurrent risk after MIS or TIA has been shown to be concentrated in the first 3 months [2,3], especially in the first week [4,5].Although urgent investigation and treatment have reduced subsequent stroke after MIS or TIA [6,7], there remains a residual recurrent risk of 5.1% [8] within the first year, which gradually increases to 9.5% within the next 4 years [9].MIS and TIA share a similar risk of stroke recurrence [8], mechanism [10], treatment strategy [11,12], and guideline recommendations [13,14].More aggressive secondary prevention strategies are effective in reducing recurrent risk after MIS or TIA in high-risk patients [15,16].Several risk stratification scores have been developed to stratify patients at different levels of recurrent risk and optimize treatment strategies [17][18][19].Using the data of the Clopidogrel versus Aspirin in Patients at Risk of Ischemic Events (CAPRIE) trial, the Essen Stroke Risk Score (ESRS) was used to predict the 1-year risk of stroke recurrence [17].The ESRS was later validated in large-scale registries, including the REduction of Atherothrombosis for Continued Health (REACH) Registry [20] and the China National Stroke Registry (CNSR) [21].However, several previous studies reported the prognostic and stratifying value of the ESRS as modest [22,23].
Adding imaging features into risk stratification scores could improve the discriminative value [24,25], and intracranial arterial stenosis (ICAS) is a well-established factor for predicting stroke recurrence after MIS or TIA [26].
A recent study found that more than 13% of patients with transient or minor persistent neurologic symptoms, who were considered as low risk, had positive magnetic resonance imaging (MRI) findings [27].Whether ICAS could further identify high-risk patients across different risk strata according to ESRS, remains unclear.
We sought to investigate the association between ICAS and stroke recurrence at 3 and 12 months in MIS or TIA patients of different risk subgroups categorized by ESRS in the Clopidogrel in High-Risk Patients with Acute Nondisabling Cerebrovascular Events (CHANCE) trial.We hypothesized that ICAS could identify patients with a high risk of recurrence in different ESRS categories.
We present the following article in accordance with the CONSORT reporting checklist.

Study population and data derivation
The study protocol of the CHANCE trial, a randomized, placebo-controlled, double-blind trial, was approved by the ethics committees of all participating centers and has been published previously (including sample size calculation, randomization and blinding methods) [28].The primary goal was to compare the efficacy of a 90-day regimen of clopidogrel (initiated with a loading dose of 300 mg, followed by 75 mg/day during the first 21 days) combined with aspirin (75 mg/day), and aspirin alone (75 mg/day) to reduce stroke recurrence.
Patients aged ≥40 years with acute MIS (National Institutes of Health Stroke Scale [NIHSS] score of 0-3) or moderate-to-high risk TIA (ABCD2 score ≥ 4), and who could be treated within 24 hours after symptom onset were considered eligible.Participants were recruited from 114 subcenters in China from October 2009 to July 2012.Patients with MIS or TIA with complete baseline magnetic resonance angiography (MRA) evaluation were also included.All enrolled participants or their legal representatives provided written informed consent.

Data collection and follow-up
Trained investigators or coordinators collected information on neurologic evaluation (modified Rankin Scale, mRS; NIHSS), physical examination (weight, height and vital signs), concomitant medications, and adverse events at baseline and each follow-up visit (randomization day, on day 21, day 90 [28], and at 12 months [29].Demographic characteristics were collected at baseline.

Calculation and categorization of the ESRS
The ESRS was calculated on admission by the investigators.The point allocation method was in accordance with that of the development study of the ESRS.The total score ranged from 0 to 9: 1 point for age 65-75 years; 2 points for age >75 years; 1 point for each of hypertension, diabetes, previous myocardial infarction (MI), other cardiovascular diseases (except for MI and atrial fibrillation), peripheral arterial disease, smoker, and additional TIA or ischemic stroke in addition to qualifying events.In the CAPRIE study, the low-risk group was defined as patients with a score of 0-2, and high-risk was defined as a score ≥ 3 [17].In the REACH study, the patients in the high-risk group had an ESRS > 3 [20].The cut-off value in this study were further determined by the event rate according to a different ESRS score in the CHANCE study.

Imaging interpretation
Two experienced stroke specialists, who were blinded to patients' baseline information, treatment assignment, and outcomes, centrally reviewed the MRI data collected from participating subcenters.According to the Warfarin-Aspirin Symptomatic Intracranial Disease trial criteria [30], ICAS was defined as stenosis of 50%-99% or occlusion of at least one of the major intracranial arteries, including the intracranial portion of the internal carotid arteries, middle cerebral arteries (M1/M2), the intracranial portion of vertebral arteries, and the basilar artery.The degree of stenosis was determined by the maximum intensity projection of the relevant arterial segment on 3D time-of-flight (TOF) MRA.In situations where there was a > 10% disagreement on judging the degree of stenosis, a third reader reviewed the image and the disagreement was resolved by consensus.

Outcome definition
We focused on the primary efficacy outcome (new stroke [stroke recurrence]), including ischemic and hemorrhagic stroke at 3 months.We also analyzed stroke recurrence within 12 months, the percentage of new clinical vascular events (including ischemic stroke, hemorrhagic stroke, MI, or vascular death) and moderate or severe bleeding according to the Global Use of Strategies to Open Occluded Coronary Arteries definition were assessed among different risk groups.All suspected events were reviewed and designated by the central adjudication committee.

Statistical analysis
Categorical variables are presented as percentages, and continuous variables are presented as medians with interquartile ranges or mean values with standard deviations (SD).Univariate analyses were performed to compare the baseline characteristics among different risk groups and were performed using the chisquared test and one-way analysis of variance or Kruskal -Wallis test for categorical and continuous variables, respectively.Hazard ratios (HRs) and 95% confidence intervals (95%CIs) of ICAS and other factors of stroke recurrence (new stroke) within 3 and 12 months were estimated using the Cox regression method by including all potential confounding factors.The Kaplan -Meier method was used to depict cumulative recurrent risk within 3 and 12 months.Twosided P-values <0.05 were considered to indicate statistical significance.SAS software, version 9.4 (SAS Institute Inc.) was used to perform all statistical analyses.

Distribution and baseline characteristics of participants
Among the 5170 patients enrolled in the CHANCE trial, 1342 patients recruited from 45 study sites were included in the imaging sub-study.After excluding 137 patients lacking T1-weighted images, 4 lacking T2-weighted images, 23 lacking diffusion-weighted images, and 89 lacking 3D-TOF-MRA, 1089 patients with all required MRI sequences were included (Figure 1).
The mean (SD) age of the participants was 63.1 (10.7) years, and 65% were men.The included patients were older, had a lower body mass index (BMI), less history of ischemic stroke, longer time interval between symptom onset and randomization, and were more likely to have a minor stroke as a qualifying event.(S1).

Cut-off score of risk stratification according to the ESRS
The cut-off value was determined by comparing the recurrent risk of each score and the risk of the total participants in the CHANCE trial: The low-risk group was defined as participants with an ESRS of 0-2, and the high-risk group was defined as participants with an ESRS score ≥ 3. (S1)

Three-month outcome
During the 3-month follow-up, 54 patients (7.9%) had a new stroke in the low-risk group, and 39 patients (9.6%) had a new stroke in the high-risk group.Patients who had a new stroke within 3 months in the high-risk group, according to the ESRS, were more likely to have a higher BMI and lower blood pressure.In univariate analysis, ICAS was associated with recurrence in the low-risk group, but not in the high-risk group (Table 1).As shown in the Kaplan -Meier curves of time-to-event for the primary efficacy outcome (recurrence within 3 months), the recurrent risk was higher in patients with ICAS than in those without, both in the < 3 ESRS and ≥ 3 ESRS groups.Similar risks of stroke recurrence were seen in patients with ICAS in the low-and high-risk groups based on the ESRS (Figure 2).In multivariate analysis, after adjusting for all potential confounding factors and study treatment assignment, ICAS was associated with a higher risk of new stroke within 3 months (HR = 2.761; 95%CI, 1.538-4.957) in the low-risk group, but not in the highrisk group (HR = 1.501; 95%CI, 0.701-3.213)(Figure 3).

Twelve-month outcome
Sixty

Discussion
In this subgroup analysis of the CHANCE trial, we observed that ICAS was an independent predictor of 3-month and 12-month stroke recurrence in the lowrisk group of MIS or TIA patients but not in the highrisk group, based on ESRS stratification.The high recurrent risk was observed in MIS or TIA patients with ICAS, even in those recognized as at low-risk by ESRS.
The ESRS predicted the 12-month prognosis after ischemic stroke, mainly based on conventional cardiovascular risk factors and previous clinical manifestations.The ESRS was derived from a dataset of 6431 ischemic stroke patients in the CAPRIE trial [17], and was validated in randomized controlled trials (ESPS-II [31] and outpatient registries (REACH [20], CNSR [21].A significant increase in recurrent risk was seen in these studies from lower to higher ESRS score categories. The predictive value of the ESRS for short-term (3-month) outcome is limited [24].More recent scores incorporate baseline imaging features for association with etiologic mechanisms.The ABCD2+MRI scoring system also includes imaging parameters, including large vessel occlusion and diffusion weighted imaging (DWI) positivity, to enhance the predictability of recurrent stroke after MIS or TIA.The area under the curve (AUC) of 3-month recurrence increased from 0.78 to 0.88 after including imaging parameters.Noticeably, imaging parameters on baseline MRI alone had better predictability than the traditional score, with an AUC of 0.84 [32].The ABCDE+ score added large-artery atherosclerosis etiology and DWI positivity to the ABCD2 system, which improved the AUC for predicting 3-month recurrence after TIA from 0.48 to 0.67 [33].Furthermore, by incorporating ipsilateral carotid stenosis, the ABCD3-I score improved the identification of highest-risk TIA patients [25].
Among the etiologies of ischemic stroke or TIA, the large-vessel atherosclerosis subtype has the highest recurrent risk [34].Previous studies found that ICAS, one of the essential imaging parameters of the large-vessel atherosclerosis subtype, is prevalent in MIS or TIA patients [35,36], and is associated with higher recurrent risk [26,37].Even with aggressive medical treatment, the risk of stroke recurrence within 12 months after the index ischemic event was up to 12%-15% in patients with ICAS [38,39].In our study, after ESRS stratification, inconsistent predicting values of ICAS were observed among different risk categories.
Potential explanations of our findings are multifactorial.First, the conventional cut-off (<3/≥3) score of ESRS has not been used to discriminate high-risk patients from low-risk patients in some validation studies [23,40].In a previous validation study based on four large-scale national registries, 12-month stroke recurrence in patients with an ESRS of 0-2 was 2.6%, while the recurrence rate in patients with an ESRS ≥ 3 was 3.6% [23].In a genetic subgroup analysis of the CHANCE trial, the 12-month recurrence was 11.2% and 13.7% of the low-risk group (ESRS of 0-2) and the high-risk group (ESRS ≥3), respectively [41]; only slight differences in 12-month recurrence were observed in both studies.The limited ability of discriminating patients with a higher recurrence risk from those with lower risk may partly explain the additive predictive performance of ICAS in the low-risk group according to ESRS.Second, the ESRS included demographic characteristics, medical history, and smoking habits; without baseline imaging parameters, the ESRS could not provide indications of etiology, as specific etiologic subtype implied a particularly unsteady state of recurrent risk [41].Inconsistency between clinical manifestations and DWI positivity has been reported in traditionally considered lower-risk MIS or TIA patients [27].Therefore, we further investigated the role of vessel imaging in patients with different risk strata according to ESRS.Although the rate of ICAS increased with the ESRS risk strata, certain proportion of low-risk patients were diagnosed with ICAS (38.6%).(Supplementary Figure 4) Third, patients who were stratified as high-risk group by ESRS had a more complex risk profile (including multiple risk factors) than patients in the low-risk group.Furthermore, patients with persistent residual inflammatory risk (levels of inflammatory biomarkers) tended to be those with multiple risk factors [42].Inflammation biomarkers were independently associated with stroke recurrence, especially in the large artery atherosclerosis subtype [43].The combined effect of ESRS components [44] and residual inflammatory risk may lead to a less apparent predictive value of ICAS in the high-risk group by ESRS.Forth, causes of ICAS other than atherosclerosis may, in part, explain the decremental predictive value of ICAS from low-risk group to high-risk group.Previous study showed that carriers of RNF213 p.R4810K variant with ICAS had significantly lower ESRS than the noncarriers [45].And the 3-month and 12-month prognosis after ischemic stroke or TIA were not significantly different between carriers and noncarriers of RNF213 p.R4810K variant [46].However, we did not perform the sequencing of RNF213 p.R4810K variant in the current study.Future studies with genetic data focusing on patients of minor ischemic stroke or TIA with ICAS may help to reveal the underlying mechanisms [47].
The short-term (3-month) risk of stroke recurrence in patients with ICAS in the low-risk group (12.9%) by ESRS was comparable to that in the high-risk group (11.9%), with similar results observed in the long-term (12-month) risk (13.7% in the low-risk group versus 14.7% in the high-risk group).These findings are in favor of the hypothesis that the predictive value of ICAS on stroke recurrence was independent of ESRS, a score primarily based on clinical information.Vessel imaging evaluation of intracranial stenosis might be imminent supplementary to traditional risk stratification, especially in patients with a low risk of recurrence stratified by ESRS.
There were certain limitations to the current study.First, we included only 21.1% of all enrolled patients from 45 out of 114 study centers in this sub-study, introducing a potential selection bias.Second, the CHANCE trial only enrolled Chinese patients, which limits the generalization of the findings; indeed, the prevalence of ICAS in Chinese patients was higher than that in other populations (30%-50% [48][49][50] vs. 8%-19% [35,51].Third, the study centers chose to conduct 1.5T or 3.0T MRI according to their MRI machine settings, which might introduce heterogeneity in judging the maximum intensity projection of the relevant arterial segment on 3D-TOF-MRA. ICAS was an independent predictor of 3-month and 12-month stroke recurrence in the low-risk group of MIS or TIA patients but not in the highrisk group, based on ESRS stratification.High recurrent risk was observed in MIS or TIA patients with ICAS, even in those who were recognized as low-risk by ESRS.This indicates the prognostic importance of imaging parameters, which could optimize the risk stratification system based purely on clinical features.Our findings may be useful for researchers or clinicians in future trial design to justify more intensive treatment in certain subgroups (ICAS detected by imaging examination) within those MIS or TIA patients stratified as low risk by ESRS.
Science Foundation of China [81870905]; the Capital's

Figure 2 .
Figure 2. Cumulative occurrence of recurrent stroke according to the presence of ICAS and ESRS using Kaplan -Meier survival analysis within 3 months.A: cumulative risk of recurrence at 3 months with or without ICAS in the low-risk group according to ESRS.B: cumulative risk of recurrence at 3 months with or without ICAS in the high-risk group according to ESRS.C: cumulative risk of recurrence at 3 months categorized by ICAS and ESRS (a: low-risk with ICAS; b: high-risk with ICAS; c: low-risk without ICAS; d: high-risk without ICAS).

Figure 3 .
Figure 3. Multivariate analysis of predictors for 3-month stroke recurrence based on the ESRS stratification.Multivariate model included ESRS components, sex, days from event to enrolment, hours from onset to randomization, other medical history, body mass index, blood pressure, treatment group, concomitant medication, index event and ICAS.HR: Hazard ratio, CI: Confidence interval, ICAS: Intracranial arterial stenosis, MI: Myocardial infarction, PAD: Peripheral artery disease.

Table 1 .
Baseline characteristics in patients stratified by ESRS based on 3-month recurrence.

Table 2 .
Baseline characteristics in patients stratified by ESRS based on 12-month recurrence.