Effect of frailty on outcomes of endovascular treatment for acute ischaemic stroke in older patients

Abstract

Background

frailty has been shown to be a better predictor of clinical outcomes than age alone across many diseases. Few studies have examined the relationship between frailty, stroke and stroke interventions such as endovascular thrombectomy (EVT).

Objective

we aimed to investigate the impact of frailty measured by clinical frailty scale (CFS) on clinical outcomes after EVT for acute ischemic stroke (AIS) in older patients ≥70 years.

Methods

in this retrospective cohort study, we included all consecutive AIS patients age ≥ 70 years receiving EVT at a single comprehensive stroke centre. Patients with CFS of 1–3 were defined as not frail, and CFS > 3 was defined as frail. The primary outcome was modified Rankin Score (mRS) at 90 days. The secondary outcomes included duration of hospitalisation, in-hospital mortality, carer requirement, successful reperfusion, symptomatic intracranial haemorrhage and haemorrhagic transformation.

Results

a total of 198 patients were included. The mean age was 78.1 years and 52.0% were female. Frail patients were older, more likely to be female, had more co-morbidities. CFS was significantly associated with poor functional outcome after adjustment for age, NIHSS and time to intervention (adjusted odds ratio [aOR] 1.54, 95% confidence interval [CI] 1.04–2.28, P = 0.032). There was trend towards higher mortality rate in frail patients (frail: 18.3%; non-frail: 9.6%; P = 0.080). There were no significant differences in other secondary outcomes except increased carer requirement post discharge in frail patients (frail: 91.6%; non-frail: 72.8%; P = 0.002).

Conclusions

frailty was associated with poorer functional outcome at 90 days post-EVT in patients ≥ 70 years.

Key Points

• Frailty has been shown to be a better predictor of clinical outcomes than age alone across many diseases.

• Few studies have examined the relationship between frailty, stroke and stroke interventions such as endovascular thrombectomy (EVT).

• Frailty was associated with poorer functional outcome post-EVT in older patients, independently of age, National Institute of Health Stroke Scale (NIHSS) and time to intervention.

• Elucidating how frailty predicts clinical and functional outcomes in stroke can lead to better prognostication and decision-making.

Introduction

Acute ischaemic stroke (AIS) is the second leading cause of mortality globally, where 1 in 4 adults over the age of 25 will experience a stroke in their lifetime [1]. Frailty is a state of increased vulnerability to poor resolution of homeostasis following a stress, and is associated with adverse outcomes such as increased mortality and hospitalisations [2]. In the last 20 years, the incidence of stroke had increased by 70% and death from stroke increased by 43% [1, 3]. A meta-analysis of 48,009 patients found that ~22% of patients with AIS have frailty, and 49% had prefrailty or frailty [4]. As the age of the population increases, it is expected that the absolute number and proportion of AIS patients with frailty would increase, therefore it is of great importance to understand the outcomes of acute stroke treatment in this susceptible population.

Endovascular thrombectomy (EVT) is the standard of care for AIS due to large vessel occlusion (LVO), and early recanalisation is key to good clinical outcomes in AIS [5, 6]. Recent guidelines by the Society of NeuroInterventional Surgery (SNIS) state that age should not be used as a contraindication for thrombectomy [7]. Despite this, a large multi-national survey of clinicians and interventionalists found that age is one of the most important factors when a decision for EVT is made [8]. Moreover it is well reported that older patients benefit less from EVT in general [6], although, compared to those not treated with EVT, older patients treated with EVT had significantly better functional outcomes [6]. It is therefore important to further risk stratify and identify older patients who may derive the most benefit from EVT, and differentiate them from those where EVT may not be beneficial and may even cause harm.

Although frailty is closely linked with age, age is not the main contributor to frailty [2]. The prevalence of frailty increases with age from 6% of those aged 60–64 years to 30% of those aged > 85 years [9, 10]. Clinically, frailty is associated with increased mortality, complications and longer hospital admissions [11]. One of the most commonly used scoring systems for frailty is the Clinical Frailty Scale (CFS; [12]). Frailty has been shown to be a better predictor of clinical outcomes than age alone [2, 13]. Few studies have examined the relationship between frailty, stroke and stroke interventions such as EVT [14, 15]. Our study aimed to investigate the impact of CFS on clinical outcomes after EVT for AIS in older patients ≥ 70 years.

Methods

Study population

In this retrospective cohort study, we included all consecutive AIS patients age ≥ 70 years receiving EVT at a single tertiary centre from 2017 to 2020. Data were extracted from a centralised, prospectively collected database, which included all cases of AIS patients presenting to the tertiary hospital. The inclusion criteria are: (i) adults who had neuroimaging confirmed AIS, (ii) with anterior and/or posterior circulation AIS large or medium vessel occlusion and (iii) underwent EVT within 6 h of symptom onset. Only patients with a premorbid modified Rankin Scale (mRS) score of 0–2 qualified for EVT. We classified the site of arterial occlusion as one of the following: internal carotid artery (ICA), terminal ICA (T-lesion; T-occlusion), first and second segments of the middle cerebral artery (MCA M1, M2), tandem occlusion, basilar artery (BA), vertebral artery (VA) or proximal posterior cerebral artery (PCA; [16]). The following data were collected: demographics, co-morbidities (presence of hypertension, diabetes mellitus, dyslipidaemia, ischemic heart disease, previous stroke and atrial fibrillation), and clinical variables including clinical severity of stroke as represented by the National Institute of Health Stroke Scale (NIHSS), mRS, subtype of ischemic stroke as per the Trial of Org 10,172 in Acute Stroke Treatment (TOAST) classification and site of occlusion. All patients underwent acute neuroimaging using a computed tomography (CT) scan and CT angiography.

Frailty

Premorbid clinical frailty score (CFS) was collected for all patients age ≥ 70 years admitted with AIS. The CFS is a well-validated, widely-used frailty tool in multiple clinical settings, and has been associated with clinical outcomes including mortality, functional decline, mobility and cognitive decline [11]. It is an ordinal scale from 1 (very fit) to 9 (terminally ill), that mixes items such as co-morbidity, cognitive impairment and disability. All patients in this study underwent routine geriatric assessment, which included the collection of information on cognition, mobility and function. CFS was scored retrospectively [17, 18] by an independent stroke nurse who was trained by a geriatrician. Retrospective CFS scoring by trained personnel based on data from comprehensive geriatric assessments has been validated with good reliability and accuracy [17, 18]. Patients with CFS of 1–3 were defined as not frail, and CFS > 3 was defined as frail in this study [12].

Primary and secondary outcomes

The primary outcome in this study was measured using the mRS at 90 days, which is a widely used clinical outcome measure of functional independence post-stroke [6]. A mRS of 0–2 was considered as good functional outcome. This was independently assessed by a qualified stroke neurologist at 90-day follow-up. The secondary outcomes included the following hospital outcomes: duration of hospitalisation, in-hospital mortality and carer requirement, and the following stroke outcomes: successful reperfusion, symptomatic intracranial haemorrhage (ICH) and haemorrhagic transformation. The extent of angiographic reperfusion was evaluated according to the modified Thrombolysis in Cerebral Infraction (mTICI) grading system, and successful reperfusion was defined as mTICI ≥ 2b. The presence of ICH was determined by repeat brain imaging 22–36 h after treatment.

Statistical analysis

Continuous variables were presented as mean ± standard deviation (SD) and compared using Mann–Whitney U test (non-parametric data) or Student’s t-test (parametric data). Categorical variables were presented as frequency and percentages and compared using the Chi-square test or Fisher’s exact test where appropriate. Adjusted analyses were performed using logistic regression to identify if frailty status was independently associated with poor functional outcomes in older stroke patients. The multivariable model included adjustment for age, NIHSS and median onset to puncture time, which were decided a priori to be significant predictors of functional outcomes from previous literature [19, 20]. All statistical analyses were performed using SPSS version 25 (IBM Corp., Armonk, NY, USA). A P value of <0.05 was considered statistically significant.

Ethics approval for this study was obtained from the local ethics review board (DSRB Reference 2021/00623).

Results

A total of 198 patients age ≥ 70 years who underwent EVT for AIS were included. The mean age was 78.1 (SD 5.7) years and 103 (52.0%) were female. The median CFS of this cohort was 4 (interquartile range [IQR] 3–4), and 94 (47.5%) had CFS 1–3, whereas 104 (52.5%) had CFS > 3. All patients had a premorbid mRS of 0–2 to qualify for EVT, and the distributions of mRS and CFS are shown in Supplementary Figures 1 and 2, Supplementary data are available in Age and Ageing online.

Baseline characteristics and laboratory findings

The mean age of frail patients was 80.4 (SD 5.9) years and that of non-frail patients was 75.7 (SD 4.3) years (P < 0.001). More frail patients were female than non-frail patients (frail: 61.5%, 64 patients; non-frail 41.5%, 39 patients; P = 0.005). Frail patients were more likely to have hypertension (frail: 90.4%, 94 patients; non-frail 77.7%, 73 patients; P = 0.014), atrial fibrillation (frail: 71.2%, 74 patients; non-frail: 52.1%, 49 patients; P = 0.006), previous stroke or transient ischemic attack (TIA; frail: 21.2%, 22 patients; non-frail: 8.6%, 8 patients; P = 0.014) and cognitive impairment (frail: 18.3%, 19 patients; non-frail: 4.3%, 4 patients; P = 0.002) than non-frail patients. There were no significant differences in racial distribution, diabetes mellitus, dyslipidaemia, ischemic heart disease and smoking between the two groups. Patients who were frail had significantly lower haemoglobin (frail: 12.2 g/dl, SD 2.2; non-frail: 13.0 g/dl, SD 1.7; P = 0.005) and albumin (frail: 32.3 g/l, SD 4.6; non-frail: 34.0 g/l, SD 4.0; P = 0.018). The baseline characteristics and laboratory findings are shown in Table 1.

Stroke characteristics and treatment

There were no significant differences in median NIHSS on arrival between patients who were frail and not frail (frail: median NIHSS 20, IQR 16–23; non-frail median NIHSS 18, IQR 13–23, P = 1.000; Table 1). According to the TOAST classification, large artery atherosclerosis was less common in frail patients (frail: 17.6%, 18 patients; non-frail 33.0%, 30 patients, P = 0.014) and cardioembolic stroke were more common in frail patients (frail: 70.6%, 72 patients; non-frail 51.6%, 47 patients, P = 0.007). The distribution in the sites of vessel occlusion were similar between patients who were frail and those who were not. There were no significant differences in the proportion of patients who received bridging thrombolysis between the two groups (frail: 57.7%, 60 patients; non-frail: 62.8%, 59 patients; P = 0.467). There were no significant differences in procedure times, such as the median time from onset to arrival, time from door to needle or time taken for clot retrieval between the two groups (Table 1).

Primary outcome

At 90 days, significantly fewer frail patients had mRS of 0 (frail: 4.8%, 5 patients; non-frail: 14.9%, 14 patients; P = 0.016; Table 1). Achieving a good functional outcome at 90 days (mRS 0–2) was significantly less common in frail patients (16.3%, 17 patients) than non-frail patients (30.9%, 29 patients; P = 0.016). CFS as an ordinal variable was significantly associated with poor functional outcome (mRS 3–6) on univariate analysis (odds ratio [OR] 1.73; 95% confidence interval [CI] 1.22–2.44; P = 0.002) and multivariable analysis adjusted for age, onset to puncture time and NIHSS on arrival (adjusted OR [aOR] 1.54, 95% CI 1.04–2.28; P = 0.032). On ordinal shift analyses, frailty was associated with an unfavourable shift in the mRS outcomes (OR 2.38; 95% CI 1.44–3.95; P = 0.001), which remained significant after adjusting for age, NIHSS and onset to puncture time (aOR 2.09; 95% CI 1.20–3.65; P = 0.009). This is demonstrated in Figure 1.

In the multivariable model, NIHSS was independently associated with poor functional outcome (aOR 1.13; 95% CI 1.07–1.20; P < 0.001), but age (adjusted OR 1.04; 95% CI 0.96–1.12; P = 0.331) and onset to puncture time (adjusted OR 1.00; 95% CI 1.0–1.0; P = 0.817) was not independently associated with the primary outcome. The results of the multivariable analysis performed are shown in Table 2.

Secondary outcomes

The median duration of hospitalisation (frail: 15 days, IQR 8–24; non-frail: 15 days, IQR 8–28; P = 0.539) were similar between frail and non-frail patients. The in-hospital mortality trended towards being higher in frail patients than non-frail patients, but did not reach statistical significance (frail: 18.3%, 19 patients; non-frail: 9.6%, 9 patients; P = 0.080). Other secondary outcomes including successful reperfusion (mTICI ≥ 2b), symptomatic ICH and haemorrhagic transformation were similar between frail and non-frail patients. Frail patients were more likely to require carers on discharge. Of the frail patients, 91.6% (76 patients) had carer requirement, whereas only 72.8% (59 patients) of non-frail patients required carers (P = 0.002). The results of the secondary outcomes are shown in Table 1.

Discussion

Our study demonstrated that (i) frailty as measured with CFS was associated with higher mRS scores at 90 days post-EVT and (ii) the association with poorer functional outcomes remained significant after adjusting for age. As only patients with premorbid mRS of 0–2 qualified for EVT, this suggested that frail patients were more likely to have increased 90-day mRS post-EVT, and non-frail patients were more likely to show preserved functional outcome post-EVT.

Frailty as a state of vulnerability is a clinical syndrome that overlaps with but is distinct from comorbidity and disability [21]. Although these conditions frequently co-exist, not all frail patients are disabled, not all disabled patients are frail, and not all comorbidities result in frailty or disability. This distinction is important, as it may be possible to attenuate or reverse frailty trajectories in order to reduce its burden on health outcomes. Disability as measured by mRS is a familiar concept in stroke and is frequently used to determine eligibility for participation in stroke clinical trials and interventions. In our study, the distribution of baseline CFS was different from mRS, showing that they are distinct clinical entities (Supplementary Figures 1 and 2, Supplementary data are available in Age and Ageing online). Therefore, pre-stroke mRS as a measure of disability is not a substitute for frailty assessment. There may be a segment of stroke patients who have no functional disability based on mRS, but who are frail and may have poorer outcomes with stroke interventions.

The relationship between frailty and stroke is poorly understood with no clear consensus on the best methods to evaluate frailty in individuals with stroke or undergoing stroke interventions [15]. Clinical frailty has been found to be well associated with poor functional outcomes, complications and mortality in multiple healthcare settings and specialties. Clinical frailty is associated with worse outcomes post-stroke or TIA. In a study of 15,468 patients, increase in the Hospital Frailty Risk Score was associated with increased length of stay, 90-day mortality, readmission and worse health-related quality of life [22]. Other surrogate markers of frailty, such as grip strength and walking speed were also associated with poor recovery and survival post-stroke [23]. Frail patients may also derive less benefit from hyperacute reperfusion therapies [15]. In 433 patients ≥ 75 years, Evans et al. using frailty on CFS found that frailty was associated with higher 28-day mortality and reduced improvement in NIHSS post thrombolysis. Frailty measured by the Hospital Frailty Risk Score was associated with increased 90-day mortality and worse mRS after EVT in LVO stroke in two recent studies based in Germany [14, 24]. In this study, we used the cut off of 70 years old, as expert consensus from 6 major international, European and US societies recommend that all patients over the age of 70 years should be screened for frailty [2]. This age threshold was also used in various other studies on frailty across different clinical settings [11, 25]. Similarly, we found that CFS was a significant predictor of worse 90-day mRS post-EVT, and trended towards increased mortality, but did not reach statistical significance. This was likely due to our study being underpowered for this outcome.

The mechanisms behind the association of frailty with poor post-EVT outcomes are manifold. Older patients had poorer EVT outcomes, and the prevalence of frailty increases with age [9]. The HERMES meta-analysis of 5 randomised controlled trials found that the rates of good function outcome at 90 days and mortality after EVT were 46 and 15% respectively in the general trial population, in contrast to 30 and 28% in those above the age of 80 years [6]. A meta-analysis of 16 studies on those aged >80 years found lower success of therapy (OR 0.72), worse functional outcomes (OR 0.40) and higher mortality (OR 2.26) than those <80 years old [19]. Although we found that frail patients were significantly older, frailty remained significantly associated with poor functional outcomes after adjustment for age, confirming the findings of Evans et al. in a population post-thrombolysis [15]. Age was associated with poor functional outcomes on univariate analysis but not after adjusting for CFS, NIHSS and time to puncture in our study, suggesting that CFS may be a better predictor of worse mRS than age alone. Further prospective studies are needed to validate this finding and investigate the performance of CFS as one of the factors in determining suitability for EVT in the clinical setting.

Frailty has been associated with increased stroke severity measured by the NIHSS. A prospective study on 232 patients found that frailty was not associated with short-term outcomes after adjustments for NIHSS and length of stay, therefore stroke severity may mediate the effect of frailty on stroke outcomes. However, this study excluded patients with pre-existing disability (mRS ≥ 3), and therefore reported a lower proportion of frail patients of 12%, which may have introduced significant selection bias. NIHSS is an independent predictor of poor stroke outcomes in older adults, and NIHSS ≥ 16 was correlated with poor prognosis post-EVT in one study [20]. NIHSS may be combined with age in the SPAN index (Stroke Prognostication using Age and NIHSS Stroke Scale), and a combined score of ≥100 had significantly lower favourable outcomes after EVT [26]. In contrast, other studies found an independent effect of frailty after adjustment for NIHSS, age and vascular risk factors [14, 15]. In our study, CFS remained independently associated with poor functional outcomes after adjusting for NIHSS, age and onset to puncture time, although the effect size was reduced. Furthermore, the NIHSS on arrival was similar between frail and non-frail patients, which confirmed the findings of another study using CFS in patients ≥ 75 years [15].

Complications of EVT and AIS include symptomatic ICH and haemorrhagic transformation, however studies on these outcomes and frailty are limited. There were no significant differences in risk of symptomatic ICH in octogenarians post-EVT in a meta-analysis of 16 studies [19]. One study using the Hospital Frailty Risk Score found no significant differences in ICH and periprocedural subarachnoid haemorrhage among different frailty groups [24]. Similarly, our study found no significant differences in symptomatic ICH and haemorrhagic transformation between frail and non-frail patients, suggesting that frailty may have smaller impacts on these complications, but further research is needed to confirm these findings.

Limitations

There were several limitations to this study. First, this was a single-centre retrospective cohort study. Hence, causation could not be established, and there may be unidentified confounding factors in the association between CFS and clinical outcomes post-EVT. As data was collected from a single tertiary centre, results may not be generalisable to other populations. CFS was also computed retrospectively based on routine geriatric assessment that included information on cognition, mobility and function, which may be susceptible to inaccuracies in the interpretation of data not primarily intended for CFS calculation. However, retrospective CFS scoring has been validated with good reliability and accuracy [17, 18]. Second, this study included 198 patients and may be underpowered for the survival outcome, therefore concrete conclusions cannot be drawn. Third, we only investigated short-term outcomes up to 90 days, therefore the differences in long-term outcomes are unknown. We adopted the commonly used clinical outcomes in stroke trials, including 90-day mRS [6], facilitating comparisons with previous literature.

Conclusions

In this study, we demonstrated that frailty was associated with higher mRS scores at 90 days post-EVT in those ≥70 years, independently of age, NIHSS and time to intervention. Frail patients were more likely to have increased 90-day mRS post-EVT, and non-frail patients were more likely to show preserved functional outcome post-EVT. Elucidating how frailty predicts clinical and functional outcomes in stroke can lead to better prognostication and decision-making. Further prospective studies are needed to evaluate the use of frailty screening tools in stroke patients and optimising patient selection for EVT.

Declaration of Conflicts of Interest

None.

Declaration of Sources of Funding

None.

References

Author notes