Frailty Scores Impact the Outcomes of Urgent Carotid Interventions in Acute Stroke Patients: A Comprehensive Analysis of Risk and Prognosis
Cameron St. Hilaire, Jeff Burton, Kundanika Lunkkadi, W Charles Sternbergh, III, Samuel Money, Daniel Fort, Hernan Bazan
Ochsner, New Orleans, LA
Background: Carotid artery disease accounts for up to one-third of all ischemic strokes. With the advent of regional stroke centers, ‘urgent’ carotid interventions including carotid endarterectomy (uCEA) and carotid artery stenting (uCAS) are increasingly performed within two weeks following an acute ischemic stroke. Patient selection for uCEA & uCAS is primarily based on overall health and risk profile as interpreted by the surgeon or interventionalist, in addition to the presenting National Institutes of Health Stroke Scale (NIHSS) severity. Due to the increased risk of major adverse cardiac events, uCAS is offered to patients deemed high risk. However, determining who fits this “high-risk” profile is a topic of debate. The Society for Vascular Surgery has made recommendations for which patients are better suited for uCAS based on anatomy and comorbidities, with the caveat that defining this group in terms of symptomatology and risk assessment is difficult.1 Stroke primarily affects those over the age of 65 years. Based on the 2020 US Census, this group enlarged over the previous 10-year period more rapidly than ever, growing from 40 million to 56 million and now accounting for 17% of the U.S. population.2 Recently, the concept of frailty has come into focus concerning the treatment of the elderly population. A decline in function across multiple organ systems related to aging describes frailty, which results in a decreased physiologic reserve. This puts patients at higher risk of complications or death after interventions. Up to fifteen percent of the American non-nursing home population over 65 years are considered frail, and 45% are “pre-frail”.3 While there are multiple clinical-based frailty calculators, they often include manual assessment components that are not feasible in the acute care setting. Gilbert et al developed a frailty index calculator that uses the ICD-10 diagnosis codes from the electronic medical record (EMR) to calculate Hospital Frailty Risk Scores and identifies patients at higher risk of adverse outcomes.4 This was validated using a cohort of over 1 million patients and compared against two well-established scoring systems. This Frailty Risk Score calculator serves as a more objective measure of an elderly patient’s physiologic reserve than the experienced physician’s “eyeball test” and has the potential to give meaningful insight before carotid intervention. There is a paucity of data on how patient frailty before an urgent carotid intervention following an acute stroke influence outcomes. We aimed to evaluate the impact of a standardized and validated frailty score in patients undergoing uCEA and uCAS following an acute stroke and assess its effects on postoperative clinical outcomes (stroke, death, and myocardial infarction [MI]), as well as neurological functional dependence (modified Rankin scale, mRS).
Methods: From a prospectively maintained database, we identified 307 acute stroke patients treated with either uCEA or uCAS from January 1, 2015, through April 30, 2023. We evaluated ICD-10 codes for each patient from 180 days before admission to 1 day after, allowing for provider documentation of comorbidities in patients new to the health system. The patient frailty score was calculated using the Hospital Frailty Risk Index based on the calculator developed by Gilbert et al.4 A list of 109 frailty-specific ICD-10 codes is used, each assigned a weighted score correlating to impact on frailty. The sum of the scores for all the conditions makes up the frailty index (range: 0 - 100). Since the stroke population of patients primarily falls into the high-risk category defined by this model, we created stroke-specific risk categories by graphing the incidence of stroke, death, and MI with associated frailty scores. Inflection points delineated low-, intermediate-, and high-risk categories. Primary endpoints were 30-day stroke, death, and MI; secondary endpoints were discharge mRS. All analyses were performed using SAS, version 9.4 for Windows (SAS Institute, Cary, NC) with significance level α = 0.05. Mean NIHSS was evaluated via negative binomial regression. Hospital LOS and days from hospital admission to intervention were treated as time-to-event measures and assessed via Kaplan-Meier survival analysis. Two-sample t-tests analyzed the remaining continuous measures, and chi-square or Fisher's exact tests analyzed the categorical measures. Previous research has established a strong association between presenting NIHSS and functional outcomes described by the mRS. To further explore other associated factors, we display mRS as a function of presenting NIHSS broken out by other elements in heat maps (Figure). We divided NIHSS into minor (0-4), moderate (5-15), and moderate to severe (≥16) categories. A moderate stroke severity is further divided based on the previous observation that patients have worse outcomes if receiving urgent intervention after presenting with NIHSS >10.
Of the 307 patients undergoing either uCEA or uCAS following a stroke, the average age was 65.9 years, and 37.5% were female (Table). Hypertension (88.6%), tobacco use (67.8%), and hyperlipidemia (76.5%) were the most common comorbidities. Twenty-six percent had a prior stroke. Peripheral artery disease was more common in uCAS (21.1% vs. 11.1%, p=0.021). There was no significant difference in frailty scores between uCEA and uCAS. Twenty-three percent of patients had no symptoms at presentation (NIHSS = 0), 34.2% had a minor stroke (NIHSS = 1-4), 30.9% had a moderate stroke (NIHSS = 5-15), and the minority had a moderate to severe (NIHSS = 16-20) (6.2%) or severe stroke (NIHSS >20) (4.6%). Ninety patients (29.3%) underwent uCAS, and 217 (70.6%) underwent uCEA. In 21.2% of patients, stroke neurologists, vascular surgeons, and interventionalists used tissue plasminogen activator (tPA), and in 13.7%, they performed thrombectomy. uCAS patients were more likely to receive tPA (30% vs. 17.5%, p=0.015) and to have a thrombectomy (38.9% vs. 13.7%, p=<0.001). The average days to intervention were 3.1, with shorter intervals in uCAS than in uCEA patients (1.3 vs. 3.9, p=<0.001). uCAS patients had higher average presenting NIHSS than uCEA (10 vs. 3.8, p<0.001).
The 30-day composite stroke, death, and MI was 8.1% for the entire cohort. uCAS patients had higher composite scores (12.2% vs. 6.5%). However, this was insignificant and primarily reflected a higher death rate (6 patients, 6.7%). Hemorrhagic conversion was seen in 3.6% of the patients, significantly more common in uCAS than in uCEA patients (8.9% vs. 1.4%, p=0.003) and not influenced by tPA use (6.1% vs 2.9%, p= 0.11). On average, only uCEA after minor stroke resulted in functional independence whereas those receiving uCAS after minor or moderate stroke had this outcome. There was no clear correlation between time to intervention and discharge mRS score.
There was a direct relationship between increasing frailty score and composite stroke, death, and MI incidence with inflection points at scores of 10 and 30 (Fig, Panel A). The majority of patients in this study were in the intermediate- and high-risk frailty groups (50.5% and 41.7%, respectively). High-risk frailty score (>30) was associated with significantly higher risk of stroke, death, and MI compared to the remaining cohort (11.7% vs 5.6%, p<0.001). The low-risk stroke, death, and MI rate was falsely elevated by a young patient with no preexisting comorbidities who died. As the presenting NIHSS increased, the discharge mRS increased; those presenting with NIHSS greater than 15 had an average mRS of 2.4, noting functional dependence (Fig, Panel B). Patients who presented with a minor stroke on average discharged functionally independent. Patients with high-risk frailty score (>30) presenting with a moderate stroke were more likely to be functionally dependent (mRS>2) on discharge (67% vs 41.3%, p<0.001).
Conclusions: Americans are aging, and those presenting with ischemic strokes are more likely to be frail than in years past. Frailty describes the broad physiologic changes that occur with advancing age and decreased ability to tolerate significant physiologic stressors. Previous work established the influence of stroke severity (NIHSS), intervention time, and tPA use on overall stroke, death, MI, and mRS outcomes. Here we demonstrate the adverse effects of frailty on this patient population. Frailty is associated with higher 30-day adverse clinical outcomes (stroke, death and MI). Additionally, frailty bears a significant role in overall neurologic outcome and functional independence.
We demonstrate that frailty is an independent risk factor for clinical and neurologic outcomes in addition to the observed relationship to presenting stroke severity (NIHSS), as demonstrated in the Figure, Panel B. Importantly, frailty affects all patients regardless of the intervention they underwent. This knowledge may assist with anticipating prognosis following acute carotid-related stroke, educating patients and families, and helping guide the choice of therapeutic intervention or medical only treatment.
Table: Patient characteristics, Frailty Risk Scores, Stroke Severity (NIHSS), Carotid Intervention, Thrombolysis Use, Time to Intervention, and Neurologic Outcomes (mRS).
Figure: Clinical Outcomes following urgent carotid interventions. A) Frailty index score related to clinical outcomes (Stroke, Death, MI). B) Heat maps with a correlation between presenting stroke severity (NIHSS) and discharge neurologic outcome (mRS) based on frailty score and intervention type.
Back to 2024 Abstracts