Association Between Premorbid Frailty Status and Functional Independence in Acute Ischemic Stroke Patients Following Endovascular Treatment in the Late Window

Introduction

Randomized clinical trials have demonstrated that endovascular treatment (EVT) is the first-line therapeutic strategy for acute ischemic stroke (AIS) caused by large vessel occlusion (LVO).^1,2 However, even if treated by EVT, less than 40% of AIS patients in the late time window (6 to 24 hours from the last known well) achieved functional independence.^3,4 Specifically, as reported in a nationally representative study in 2020, the estimated prevalence and mortality rate of stroke in China were 2.6% and 3.43‰, respectively.⁵ Over the past two decades, the absolute number of stroke incidents and mortality has increased by 70% and 43%,⁶ underscoring the necessity of novel prognostic markers to improve clinical decision-making and periprocedural management strategies. Although the predictive roles of cerebral tissue impairment have been extensively investigated in AIS patients undergoing EVT in the late window, the impacts of peripheral organs on outcomes in these patients remain largely unknown.

Frailty is a syndrome characterized by increased vulnerability to stressors caused by a cumulative decline across multiple physiological systems.⁷ In several clinical settings, including neurosurgery and brain tumor resection, frailty has been reported to increase the incidence of mortality and complications.⁸ Stroke, a prototypical stressor event, is often associated with multiple comorbidities, such as hypertension and diabetes.⁹ However, frailty was rarely evaluated in recent practice guidelines on stroke and was not yet routinely measured in stroke patients.¹⁰ Recent pilot studies recruiting AIS patients within 6 hours of stroke onset reported that frailty status was associated with mortality and poor functional outcomes in stroke patients.¹¹ Due to differences in clinical features and outcomes between AIS in the early and late time window,^4,12 the impact of frailty on AIS patients undergoing EVT in the late window remains unclear.

Among existing frailty assessment tools, the five-item modified frailty index (mFI-5) from the American College of Surgeons National Surgical Quality Improvement Program is widely recommended and applied in various emergency conditions.^13–15 The mFI-5, consisting of five easily accessible parameters, enables a rapid and objective frailty assessment by integrating key dysfunctions from multiple peripheral organs¹³ which is aligned with the need to conserve time and resources in the emergency procedure.¹⁶ In contrast, tools like the 33-item cumulative deficit frailty index and the Hospital Frailty Risk Score require extensive clinical data and complex calculations, making them impractical for rapid evaluations in emergencies.¹⁷ Additionally, previous literature has shown that mFI-5 demonstrates comparable predictive value to other assessment tools across multiple clinical scenarios.^18,19 Therefore, based on our national multicenter trial of RESCUE-BT, the current study aimed to investigate the impacts of frailty evaluated by the mFI-5 on clinical outcomes in AIS patients undergoing EVT in the late window and to validate these findings in an external cohort.

Materials and Methods

Study Design and Participants

The RESCUE-BT was a prospective, double-blind, randomized clinical trial of intravenous tirofiban plus EVT versus placebo plus EVT for patients presenting with an occlusion of the internal carotid artery (ICA) or middle cerebral artery (MCA) within 24 hours of symptom onset. The trial protocol and patient eligibility criteria have been reported.²⁰ The study was registered with the Chinese Clinical Trial Registry (http://www.chictr.org.cn, ChiCTR-INR-17014167). This study was conducted in compliance with the Declaration of Helsinki. All recruited patients or legal representatives signed a written informed consent before randomization. Among 948 patients randomized in the RESCUE-BT trial, patients meeting time from stroke onset to groin puncture greater than 360 minutes and with pre-stroke modified Rankin Scale (mRS) score 0–2 were included in the current cohort.

For the external validation cohort, we included 223 acute ischemic stroke (AIS) patients with large vessel occlusion (LVO), who met the above inclusion criteria, from the Second Affiliated Hospital of Chongqing Medical University and the Chongqing University Three Gorges Hospital between January 2019 and September 2023. The study protocol was approved by the Human Research Ethics Committee of the Second Affiliated Hospital of Chongqing Medical University (NO.2023–43).

Assessment of Frailty

Frailty was evaluated by the mFI-5 index, as described by the American College of Surgeons National Surgical Quality Improvement Programs database.¹⁴ This index integrated functional status and complications, with scores ranging from 0 to 5. Patients were assigned one point for each of the following: non-independent functional status before stroke, diabetes mellitus, chronic obstructive pulmonary disease (COPD), congestive heart failure (CHF), and hypertension requiring medication. Frailty was defined as mFI-5 scores ≥2.²¹

Data Collection and Assessment

We retrospectively collected baseline characteristics, including patient age, sex, diabetes mellitus, hypertension, atrial fibrillation, chronic obstructive pulmonary disease (COPD), congestive heart failure (CHF), pre-stroke mRS score, occlusion site, the Trial of ORG 10172 in Acute Stroke Treatment (TOAST) classification, National Institutes of Health Stroke Scale (NIHSS) score, and Alberta Stroke Program Early CT Score (ASPECTS). The time from stroke onset to puncture (OTP) and time from stroke onset to recanalization (OTR) were also recorded. The American Society of Interventional and Therapeutic Neuroradiology/Society of Interventional Radiology (ASITN/SIR) collateral vessel grading system was used to assess collateral status. The quality of reperfusion was evaluated using the extended Thrombolysis in Cerebral Infarction (eTICI) score at final angiography.

Outcomes

The primary outcome was functional independence at 90 days, defined as an mRS score of 0–2. For the derivation set from the RESCUE-BT trial, the mRS assessment was independently evaluated by two mRS-certified neurologists. Their evaluations were based on structured video or voice recordings. In cases of disagreement, a third neurologist adjudicated the final score.²⁰ For the external validation cohort, the same procedure was followed, with adjudication by two certified neurologists (L.H. and C.G), and disagreements resolved by a third neurologist (C.L). All neuroradiologists were blinded to treatment allocation during mRS data analysis. The secondary outcome included the mRS score at 90 days, ranging from 0 (asymptomatic) to 6 (death), and the change in NIHSS score from baseline to 24 hours or 5–7 days post-recanalization. Safety outcomes included 90-day mortality, the incidence of any intracranial hemorrhage (ICH), and symptomatic ICH (sICH) within 48 hours after EVT. Any ICH and sICH were assessed using follow-up CT or MRI. The sICH was diagnosed according to the Heidelberg Bleeding Classification.²²

Statistical Analysis

Among patients in this study, the relationship between frailty and functional independence was evaluated using marginal effects from the sjPlot package (https://CRAN.R-project.org/package=sjPlot). After identifying the threshold of mFI-5 as 2, patients were classified into the frail group (mFI-5 of 2–5) and the non-frail group (mFI-5 of 0–1). The normality of data distribution was assessed using the Kolmogorov–Smirnov test (S-K test). Chi-squared or Fisher exact tests were used for categorical variables. Based on the distribution of normality, continuous variables were compared using the Student’s t-test (P≥ 0.050 in the S-K test) or the Mann–Whitney U-test (P<0.050 in the S-K test) to compare baseline characteristics, imaging data, and treatment characteristics between the two groups. Significant factors (P<0.100) and variables previously associated with functional outcomes were selected via univariable analysis and subsequently included in the multivariable logistic regression model. The results were reported as odds ratios (OR) and 95% confidence intervals (CI). In the external validation cohort, logistic analysis was applied to verify the impact of frailty status in predicting outcomes.

For sensitivity purposes, we assessed the heterogeneity of the effect of frailty within subgroups based on patient age, baseline NIHSS score, ASITN/SIR, baseline ASPECTS, OTR, and eTICI score. Given the potential influence of frailty on successful recanalization (eTICI≥2b), multiple logistic regression models were employed to evaluate efficacy and safety outcomes among patients with successful recanalization. Moreover, we used the nearest-neighbor matching algorithm to perform 1:2 propensity score matching. In the analysis, we took the above confounding variables as covariates. Additionally, receiver operating characteristic (ROC) curves were used to examine the relationships between frailty status and functional independence, with DeLong’s test applied to test for significant differences between c-statistics. All tests were two-sided, and P<0.05 was considered statistically significant. All analyses were performed using SPSS software (version 26.0; IBM SPSS Statistics) and R software (version 4.2.2).

Data Availability Statement

Datasets acquired and analyzed during the study are available from the corresponding author on request.

Results

Patient Characteristics

Among the 948 patients randomized in the RESCUE-BT trial, 532 AIS patients receiving EVT in the 6-to-24-hour window were enrolled in this study (Supplementary Figure S1A). Based on the same inclusion and exclusion criteria, we included 223 AIS patients in an external validation cohort from two medical centers in Chongqing (Supplementary Figure S1B). The median age was 66 years, and 64.3% were men (Supplementary Table S1). Other baseline characteristics between the groups were shown in Supplementary Table S1. A total of 247 (46.4%) patients achieved functional independence (90-day mRS≤2).

Based on the margin effects of logistic regression analysis, Figure 1A illustrated the probability of achieving functional independence (mRS≤2) versus poor outcome (mRS≥3) at 90 days with higher mFI-5. The probability distributions of functional independence and poor outcome intersected between mFI-5 scores of 0 and 1, suggesting that a higher mFI-5 of 1 (dotted line) might be an appropriate threshold value to discriminate between functional independence and poor outcomes. To further validate this finding, we analyzed the association between mFI-5 and favorable functional outcomes among patients stratified by mFI-5 scores of 1 or ≥2. Logistic regression analysis showed patients with mFI-5≥2 had significantly lower odds of functional independence (adjusted OR (aOR) 0.31, 95% CI 0.17–0.56, P<0.001) (Figure 1B) and favorable outcome (aOR 0.28, 95% CI 0.15–0.51, P<0.001) (Figure 1C). These above analyses might demonstrate that the optimal threshold for determining frailty should be an mFI-5 score of 2.

Figure 1 Marginal effects to exhibit the relationship between mFI-5 and 90-day functional Independence. Patients with mFI-5 scores greater than 1 exhibited an increased likelihood of poor outcomes (A). Those with mFI-5≥2 had significantly lower rates of functional independence (mRS 0–2) (B) and favorable outcome (mRS 0–3) (C). Abbreviations: NIHSS, National Institutes of Health Stroke Scale; ASPECTS, Alberta Stroke Program Early CT Score, occlusion site; ASITN/SIR, the American Society of Interventional and Therapeutic Neuroradiology/Society of Interventional Radiology, eTICI, extended thrombolysis in cerebral infarction; OTR, time from stroke onset to recanalization. Notes: The regression for marginal effects was adjusted by age, sex, atrial fibrillation, pre-stroke mRS score, baseline NIHSS score, baseline ASPECTS, ASITN/SIR grade, eTICI grade, occlusion site, and OTR. Adjusted ORs with 95% CIs are reported, with a mFI-5 score of 0 as the reference category.

With the cut-off value of mFI-5, 443 patients were assigned to the non-frail group and 89 to the frail group. Compared with the non-frail group, the proportion of females (46.1% vs 33.6%, P=0.025) was significantly higher in the frail group. There were 235 (53.0%) patients with mFI-5 of 1 in the non-frailty group, and 86 (96.6%) patients with mFI-5 of 2 in the frailty group. Other baseline characteristics were well-balanced between the groups (P>0.05) (Table 1).

Table 1 Baseline Characteristics According to Groups of Non-Frail/Frail

The Impact of Frailty on Clinical Outcomes

The median [IQR] mRS at 90 days was 4 [2–5] in the frail group and 3 [1–4] in the non-frail group (aOR 2.48, 95% CI 1.62–3.81, P <0.001) (Table 2). The frail group had a significantly lower frequency of functional independence than the non-frail group (29.2% vs 49.9%, aOR 0.37, 95% CI 0.21–0.65, P<0.001) (Table 2 and Figure 2). After adjusting for center stratification, frailty remained independently associated with lower functional independence (29.2% vs 49.9%, aOR 0.31, 95% CI 0.17–0.58, P<0.001) (Supplementary Table S2). Frailty was also negatively correlated with the NIHSS score change at 24 hours (aOR 2.19, 95% CI 1.45–3.30, P<0.001) and 5–7 days after recanalization (aOR 2.03, 95% CI 1.35–3.06, P<0.001).

Table 2 Clinical Outcomes According to Groups of No-Frail/Frail

Figure 2 Distribution of the modified Rankin Scale (mRS) scores at 90 days in the frail and the non-frail groups.

However, there was no correlation between frailty and sICH (aOR 1.19, 95% CI 0.53–2.71, P=0.663) (Table 2). Similarly, the incidence of HI1, HI2, PH1, and PH2 hemorrhagic subtypes did not significantly differ between frail and non-frail patients (P>0.05) (Supplementary Table S3). Regarding non-hemorrhagic complications, frail patients had a significantly higher likelihood of developing pneumonia after EVT (65.2% vs 51.7%, aOR 1.77, 95% CI 1.07–2.94, P=0.025) (Supplementary Table S4).

The Association Between Frailty and Traditional Risk Factors

In multivariable logistic regression analysis, frailty (aOR 0.37, 95% CI 0.21–0.65, P<0.001), age (aOR 0.96, 95% CI 0.94–0.98, P<0.001), baseline NIHSS score (aOR 0.90, 95% CI 0.86–0.94, P<0.001), baseline ASPECTS (aOR 1.21, 95% CI 1.06–1.36, P=0.003) and eTICI score (aOR 1.55, 95% CI 1.31–1.83, P<0.001) were associated with functional independence (Supplementary Table S5).

Receiver operating characteristic (ROC) curve analysis demonstrated that adding frailty into the conventional predictive model significantly improved its performance in predicting 90-day functional independence (P=0.036, by DeLong’s test) (Figure 3).

Figure 3 Receiver operating characteristic (ROC) curve analysis for predicting 90-day functional Independence. Abbreviations: mRS, modified Rankin Scale; NIHSS, National Institutes of Health Stroke Scale; ASPECTS, Alberta Stroke Program Early CT Score; ASITN/SIR, the American Society of Interventional and Therapeutic Neuroradiology/Society of Interventional Radiology; eTICI, extended thrombolysis in cerebral infarction; OTR time from stroke onset to recanalization. Notes: Model 1 represented the baseline conventional model, which included age, sex, atrial fibrillation, pre-stroke mRS score, baseline NIHSS score, baseline ASPECTS, ASITN/SIR grade, eTICI grade, occlusion site, and OTR. Model 2 incorporated frailty into the conventional model. Model 2(AUC:0.77) exhibited higher predictive performance than model 1(AUC:0.75) for predicting functional independence (mRS 0–2) (ΔAUC = 0.02, p =0.036, by DeLong’s test).

Subgroup and Sensitivity Analysis

We did not detect any evidence of heterogeneity in the efficacy of frailty status among prespecified subgroups (defined according to age, baseline NIHSS, ASITN/SIR grades, baseline ASPECTS, OTR and eTICI, P for interaction > 0.05) (Figure 4).

Figure 4 Subgroup analysis of functional Independence (mRS 0–2) at 90 days based on frailty status. Abbreviations: NIHSS, score, ASPECTS, ASITN/SIR, and OTR were chosen at the median.; TICI was categorizeSd according to clinical characteristics. Adjusted variables included age, sex, atrial fibrillation, pre-stroke mRS score, baseline NIHSS score, baseline ASPECTS, ASITN/SIR grade, eTICI grade, occlusion site, and OTR.; NIHSS, National Institutes of Health Stroke Scale; ASPECTS, Alberta Stroke Program Early CT Score, occlusion site; ASITN/SIR, the American Society of Interventional and Therapeutic Neuroradiology/Society of Interventional Radiology; eTICI, extended thrombolysis in cerebral infarction; OTR, time from stroke onset to recanalization. Notes: The forest plot illustrated the odds ratios (ORs) for functional independence at 90 days in the prespecified subgroups. The thresholds for age, baseline.

Baseline characteristics were well balanced between the frail and non-frail groups after 1:2 propensity score matching (PSM) (Table 1). In PSM analysis, frail patients still demonstrated a significantly lower proportion of 90-day functional independence compared to non-frail patients (28.9% vs 41.7%, aOR 0.44, 95% CI 0.22–0.85, P = 0.015) (Table 2).

The Impact of Frailty on Patients with Successful Recanalization

Among 482 AIS patients with successful recanalization (eTICI≥2b), frailty was inversely associated with functional independence (aOR 0.39, 95% CI 0.22–0.69, P=0.001). The median [IQR] mRS at 90 days in the frail group was 4 [2–5], significantly higher than 2 [1–4] in the non-frail group (aOR 2.41, 95% CI 1.54–3.76, P<0.001) (Table 3).

Table 3 Clinical Outcomes According to Groups of No-Frail/Frail in Patients with Successful Recanalization

Validation in the External Cohort

The baseline characteristics of the external validation cohort were presented in Supplementary Table S6. The median (IQR) mRS at 90 days was significantly higher in the frail group than in the non-frail group (3 [2–5] vs 2 [1–4], aOR 2.05, 95% CI 1.08–3.88, P=0.027). Additionally, frail patients had significantly lower rates of functional independence (aOR 0.38, 95% CI 0.16–0.89, P=0.028) (Table 4).

Table 4 Clinical Outcomes According to Groups of No-Frail/Frail in the External Validation Cohort

Discussion

To our knowledge, this multicenter cohort study is the first to explore the association between premorbid frailty status and functional independence among AIS patients treated with EVT in the late window. Our findings suggest that: 1) frailty, as assessed by the mFI-5 index, was inversely associated with 90-day functional independence in patients undergoing EVT in the extended window (6–24 hours from stroke onset). 2) the negative impact of frailty on clinical outcomes persisted in an external cohort and was not modified by traditional risk factors. 3) these comorbidities encompassed by mFI-5 might have a cumulative effect on clinical outcomes. 4) adding frailty status to the baseline traditional variable model significantly enhanced the prediction of 90-day functional independence.

The notion of frailty and comorbidities in patients with cerebrovascular disorders has attracted increasing attention.²³ Several reports identified a significant association between frailty and adverse events after cerebrovascular intervention.^11,17 Unlike previous studies, our analysis focused on AIS patients receiving EVT in the late window and demonstrated the predictive value of frailty in this specific population. Notably, our post-hoc analysis of the RESCUE-BT trial revealed that nearly 70% of frail patients undergoing EVT during the late window failed to achieve functional independence. This elevated failure rate might be attributed to two key factors: First, patients in the late window had a longer onset-to-recanalization (OTR) time, resulting in more severe brain tissue damage and a higher risk of reperfusion injury after revascularization.²⁴ Second, frail patients have increased vulnerability and reduced brain tissue tolerance.²⁵ Therefore, prolonged OTR time might exacerbate their low tolerance to reperfusion stressors and increase the incidence of poor prognosis. In addition, our sensitivity analyses demonstrated frailty was significantly associated with poor outcomes even after successful recanalization, underscoring prior brain injury and brain tissue fragility. Lastly, patients with higher frailty scores often have greater pre-stroke disability or higher pre-stroke mRS, suggesting that functional independence may not be a realistic endpoint. Instead, a more appropriate treatment goal for these patients could be returning to their baseline pre-stroke mRS.

The components of mFI-5 include hypertension, diabetes mellitus, chronic obstructive pulmonary disease (COPD), and congestive heart failure (CHF), which are factors that can impact the outcomes of patients with AIS. First, in addition to promoting atherosclerosis and carotid artery stenosis, hypertension was an independent determinant of adverse outcomes after EVT in AIS.²⁶ Second, diabetes mellitus was associated with less salvageable tissue at admission and worse infarct progression, significantly affecting the 90-day clinical outcomes after EVT.²⁷ Third, COPD might act as a surrogate for a clinically significant smoking burden.²⁸ A case-control study demonstrated a strong dose-response relationship between daily smoking and ischemic stroke, regardless of gender.²⁹ Fourth, a fall in arterial blood pressure due to low cardiac output in AIS with CHF could lead to impaired autoregulation of cerebral blood flow, making patients more susceptible to ischemic injury.³⁰ Thus, mFI-5 represents well-known vascular risk factors, which are of relevance to cerebrovascular pathology. In clinical practice, multiple comorbidities are widely observed among AIS patients.⁹ However, few studies have reported their combined impact on prognosis. In this study, we observed that patients with frailty based on the cumulative effect of comorbidities, showed worse outcomes.²¹

The findings of this study expand the applicability of mFI-5 to AIS patients in the late time window. Current guidelines for AIS patients to receive EVT are written by the patient’s age, occlusion site, NIHSS score, and ASPECTS at admission.^31,32 However, these factors focus mainly on the baseline stroke severity and degree of cerebral tissue damage while neglecting the premorbid general physical status. Frailty characterizes the patients’ systemic comorbidities before stroke. In clinical practice, mFI might serve as a valuable risk-stratification tool to identify frail patients with lower potential for recovery, thereby optimizing periprocedural management strategies such as tailoring rehabilitation intensity and strengthening the monitoring.^33,34 Therefore, while frailty may attenuate the absolute benefit of EVT, existing evidence suggests that the rate of favorable outcomes among AIS patients in the late window could be improved by EVT, which remains a recommended treatment option at present.²

Several limitations should be noticed when interpreting our findings. Firstly, this study was a post-hoc analysis of the RESCUE-BT trial, carrying a risk of type I error. To minimize the effects on the results of this study, we conducted a multicenter external cohort to confirm our results. Secondly, as the RESCUE-BT trial excluded patients who had undergone pre-operative thrombolysis and patients in the late time window were not recommended for thrombolysis, we were unable to assess the impact of frailty on the outcomes of patients receiving bridging therapy. Thirdly, the small number of people with mini-stroke subgroup (NIHSS ≤ 5 at admission) limited the statistical power to determine the effects of frailty.³⁵ Future prospective studies with larger cohorts are required to assess the role of frailty in patients suffering mini-strokes or undergoing bridging therapy. Lastly, medical history and prior treatments were recorded from physician notes, which often included patient-reported information, introducing a potential recall bias. Future studies should investigate the role of frailty in identifying patients in decision-making for EVT and predicting outcomes by enrolling patients who receive EVT and those who do not.

Conclusion

This multicenter cohort study demonstrated that pre-stroke frailty, as assessed by the mFI-5 index, was independently associated with poor prognosis in AIS patients undergoing EVT in the late window. As a complementary prognostic marker to existing traditional indicators, frailty assessment may aid in refining risk stratification and optimizing perioperative patient management.