Decision making in the frail vascular surgery patient: A scoping review

Abstract

Increasing evidence highlights the adverse impact of frailty and reduced physiologic reserve on surgical outcomes. Therefore, identification of frailty is essential for older adults being evaluated for vascular surgery procedures. Numerous frailty assessment tools are available to quantify the level of frailty and assist in preoperative decision making for these older patients. This review evaluates traditional and novel frailty metrics for their scientific validation, limitations, and clinical utility in vascular surgery decision-making.

1. Introduction

Accurate frailty measurement includes parameters beyond age and comorbidities, as frailty is multifactorial and entails additional domains, such as social, nutritional, cognitive, and functional [1,2]. Both retrospective and prospective frailty screening tools have been developed during the past decade, with expanding literature showing that frailty is associated with a wide variety of adverse perioperative outcomes, including complications, mortality, higher healthcare utilization, and cost [3–6].

Approximately 10% to 20% of individuals older than 65 years are considered frail. It is projected that by 2040, the proportion of vascular surgery patients aged 65 to 84 years will double, likely resulting in a high prevalence of frailty [7,8]. Moreover, compared with other surgical disciplines, vascular surgeons treat patients with a higher burden of comorbidities and functional/cognitive deficits [6,9,10], who are more likely to experience adverse postoperative events [11,12]. Current estimates suggest 20% to 60% of vascular surgery patients are frail, depending on the measures used and the cohort selection [4,13]. Therefore, frailty assessment may serve as a practical tool for vascular surgeons in the preoperative setting to enhance risk stratification of their patients, inform their treatment selection, and guide shared decision making while performing routine surgical workup and evaluation. This scoping review discusses the different components of frailty, provides an updated overview of the commonly used assessment tools, and explores the utility of these instruments in predicting perioperative clinical outcomes in patients undergoing vascular interventions. In addition, we highlight future directions for the role of frailty data in evaluating the appropriateness of care and decision making for vascular surgery patients.

2. Methods

We performed a scoping review of the literature to identify key concepts and sources of evidence that provide a comprehensive overview of frailty assessment in vascular surgery patients [14]. PubMed and Ovid databases were searched to identify relevant English-language articles published from 2010 to July 2023 using the terms previously developed and described by Drudi et al [4]. Reference lists of the included studies were also manually searched. Results were stored in Endnote (version X9, Clarivate Analytics), and duplicate references were removed. Inclusion criteria for studies were evaluation of patients undergoing vascular surgery interventions only, the use of frailty screening tools that have been validated in more than one prior study, and acceptable study design (ie, randomized controlled trials, cohort studies, or case-control studies with sample size > 50 patients). Studies were excluded if patients did not undergo vascular intervention, frailty was not measured in the perioperative period, or if frailty was not one of the exposures of interest in the analyses. There were 98 articles selected for full-text review. The reasons for exclusion were that 49 did not meet inclusion criteria and 17 were conference abstracts or study protocols. Ultimately, 32 observational studies contributed to this review. Of note, given the novelty and heterogeneity of studies exploring the morphometric markers of frailty (eg, radiologic markers of sarcopenia), prior validation was not required.

We evaluated a wide assortment of adverse postoperative outcomes, including in-hospital and postdischarge mortality; length of stay (LOS); rehospitalization; discharge location; and postoperative medical and surgical complications, including myocardial infarction (MI); delirium; stroke; renal failure; or return to the operating room.

3. Results

3.1. The pathophysiology and definition of frailty in medicine

Frailty has been described in two broad models: the deficit accumulation model and the phenotypic model.

3.1.1. Deficit accumulation model

The deficit accumulation model conceptualizes that frailty is driven by an accumulation of deficits across medical co-morbidities, function, cognition, nutrition, social, and psychological domains—leading to increased vulnerability for occurrence of adverse events [15–17]. Rockwood et al originally developed this model through the Canadian Study of Health and Aging (CSHA) and proposed the Frailty Index (FI), measuring 70 different items and calculating a ratio of the number of positive responses divided by the total number of items assessed [5,16]. In this model, the fittest patients have an FI of 0.09, and very frail patients scored > 0.50 [5,16]. In a study of 752 inpatients, the FI was determined to be a better predictor of adverse clinical outcomes than age; there was a > 50% risk of mortality among patients with FI > 0.65 within 120 days of measurement [18]. Given that there are many screening items that must be completed to calculate an FI, this frailty assessment is deemed time consuming and impractical, despite its preciseness [5].

In order to apply the deficit model to clinical settings, two adaptations of the original CSHA-FI have now been developed. Rockwood and colleagues [16] developed the Canadian Study of Health and Aging Clinical Frailty Scale (CSHA-CFS) as an abbreviated version and validated it against the more extensive CSHA-FI assessment tool. The CSHA-CFS requires the clinician to assign the patient to one of seven qualitatively defined categories ranging from very fit (1) to severely frail (7). These seven classifications have shown correlation with the numerical FI on the same patients, as well as with the institutionalization or mortality outcomes among community-living geriatric patients [5,16]. The tool still relies on provider’s assessment of patient frailty and, therefore, may be influenced by the clinician’s expertise and may not be applicable as a patient-reported tool. The second adaptation was based on data showing that an abbreviated modification of the CSHA-FI containing 10 to 15 deficit elements had similar predictive ability and model performance as the 70-item FI scale [19]. This has led to adaptation of the original FI into other measures of frailty, applied either retrospectively to existing databases (eg, modified FI) or prospectively in clinical practice (eg, Risk Analysis Index [RAI]) to measure frailty using the deficit accumulation model.

3.1.2. Phenotypic model

Conversely, the phenotypic model defines frailty as a measure of physiological changes, such as weight loss, weakness, poor endurance, slowness, and low physical activity, which, in turn, are associated with poor outcomes [6]. Originally developed by Fried et al [20], this model led to the Hopkins Frailty Score. This model is based on somatic or physical characteristics that manifest as frailty, commonly used in the geriatric population, and is the basis for the development of multiple frailty measurement tools [3,20]. The Fried Scale is a 5-point system that encompasses gait speed, hand-grip strength, unintentional weight loss of > 10 pounds during 1 year, low physical activity, and exhaustion [20]. In the Cardiovascular Health Study, frailty was defined as three or more of these five criteria, prefrail state as one to two positive domains, and robust state as a score of zero [20]. There were significant differences in mortality rates among all three frailty groups [12]. Frail patients were less likely to recover from acute stressors and more prone to adverse outcomes after major surgical procedures [12]. Furthermore, the mortality difference among all three groups increased with time [20]. The expanded Fried Scale, also known as the Bergman scale, adds cognitive impairment and depressed mood, and requires 3 of 7 criteria for a diagnosis of frailty [21]. Other tools that measure slowness as a manifestation of frailty, such as the Timed Up and Go test, are also part of the phenotypic model.

3.2. Frailty assessment tools in vascular surgery

Aspects of both deficit and phenotypic frailty models can be incorporated into the preoperative workup of surgical patients. Numerous studies have explored the utility of frailty assessment models for vascular surgery patients, particularly regarding their predictive validity for postoperative outcomes (Table 1). The frailty assessment tools outlined in this review highlight the need for a robust and pragmatic frailty assessment that can integrate into the clinical workflow for decision making.

Table 1 –

Summary of frailty assessment tools and their input variables.

Frailty assessment tool	Encompassing variables
Addenbrooke	Impaired mobility on admission, depression, polypharmacy (> 8 medications), anemia, emergency admission, and
Vascular Frailty Scale	Waterlow pressure ulcer score > 13
Barthel Index	Fecal incontinence, urinary incontinence, and help with grooming, toilet use, feeding, transfers, walking, dressing, climbing stairs, and bathing
Bergman Scale	Gait speed, hand-grip strength, unintentional weight loss of > 10 pounds during 1 year, low physical activity, exhaustion, cognitive impairment, and depressed mood
Clinical Frailty Scale	Evaluator’s global assessment and personal observation of mobility, balance, use of walking aids and independence with activities of daily living
CLTI Frailty Risk Score	CHF, COPD, PAD, mental illness, anemia, weight loss, CKD, and fluid and electrolyte disorders
Edmonton Frail Scale	Cognition, general health status, functional independence, social support, polypharmacy, nutrition, mood, continence, and 3-meter gait speed
Fried Scale	Gait speed, hand-grip strength, unintentional weight loss of > 10 pounds during 1 year, low physical activity, and exhaustion
Groningen Frailty Indicator	Mobility, vision, hearing, nutrition, comorbidity, cognition, and psychosocial and physical fitness
Hospital Frailty Risk Score	Hemiplegia, Alzheimer disease, vascular dementia, fall, intracranial injury, cerebrovascular disease or its sequelae, symptoms/signs involving the nervous and musculoskeletal systems, urinary tract infection, urinary incontinence, hematuria, delirium, convulsions, somnolence, abnormalities of gait and mobility, fracture or injury of shoulder and upper arm, disorders of fluid, electrolyte, and acid-base balance, volume depletion, joint disorders, senility, and cellulitis
Modified Frailty Index (mFI)-5	DM, CHF, hypertension, COPD or pneumonia, and functional status
mFI-11	DM, CHF, hypertension, stroke, functional status, MI, PAD, TIA/CVA, COPD/pneumonia, CABG or PCI impaired sensorium, and neurologic deficit
Risk Analysis Index	Age, sex, weight loss, poor appetite, CHF, dyspnea, renal failure, disseminated cancer, functional status, cognitive decline, and living status
Ruptured Aneurysm Frailty Score	Charlson Comorbidity Index score > 1, polypharmacy (> 4 medications), no statin prescribed, and anemia, disability (Katz score), and geriatric-specific domains (hearing and visual impairments)
VQI-Frailty Score	Hypertension, CHF, coronary artery disease, peripheral vascular disease, diabetes, COPD, renal impairment, anemia, underweight, non-home residence, and nonambulatory status
VQI-derived RAI	Age, sex, renal failure, CHF, COPD, living status, functional status, and BMI < 20 or > 35

Abbreviations, BMI, body mass index; CABG, coronary artery bypass grafting; CHF, congestive heart failure; CKD, chronic kidney disease; CLTI, chronic limb-threatening ischemia; COPD, chronic obstructive pulmonary disease; CVA, cardiovascular accident; DM, diabetes mellitus; MI, myocardial infarction; PAD, peripheral artery disease; PCI, percutaneous coronary intervention; RIA, Risk Analysis Index; TIA, transient ischemic attack; VQI: Vascular Quality Initiative.

3.2.1. Addenbrooke’s Vascular Frailty Scale

Ambler et al [22] developed this frailty score model for 413 patients undergoing both elective and nonelective vascular procedures. The Addenbrooke Vascular Frailty Scale (AVFS) is a 6-point scale that encompasses impaired mobility on admission, depression, polypharmacy (> 8 medications), anemia, emergency admission, and Waterlow score for risk of pressure ulcers > 13 [4,22]. Frail patients had 1-year mortality rates of 39% to 58% and 1-year readmission rates of 70% to 87% among survivors [22].

3.2.2. Barthel Index

The Barthel Index is a 20-point measure of basic activities of daily living with a score of 20 indicating complete functional independence [23]. In a study of 107 patients with chronic limb-threatening ischemia (CLTI), frailty according to the Barthel Index was found to be associated with all-cause mortality [24].

3.2.3. Clinical Frailty Scale

The CSHA-CFS is a validated nine-level scale derived from the Comprehensive Geriatric Assessment (CGA), which classifies patients as (1) fit, (2) well, (3) managing well, (4) vulnerable, (5) mildly frail, (6) moderately frail, (7) severely frail, (8) very severely frail, or (9) terminally ill on the basis of evaluator’s global assessment of the accumulated deficits—or “eye-ball” assessment of frailty [25]. This quick frailty scale does not have specific physical measurements to take, rather it quantifies a patient’s overall health status using a general assessment of their mobility, strength, functional status, and cognition. In a study of 138 patients undergoing a major vascular surgery procedure, frailty diagnosed using the CFS was associated with a significantly increased risk of the composite outcome of 30-day mortality or loss of independence [26]. Furthermore, frail status defined with the CFS predicts functional decline in patient’s ability to complete activities of daily living after major vascular surgery procedures up to 2 years after their procedure [10].

3.2.4. CLTI Frailty Risk Score

The CLTI Frailty Risk Score (CLTI-FRS) was constructed using the retrospective National Inpatient Sample database, based on the deficit accumulation model and comprising the following nine coded diseases: diabetes mellitus, congestive heart failure (CHF), chronic obstructive pulmonary disease (COPD), peripheral artery disease (PAD), depression, mental illness history, hypothyroidism, anemia, and weight loss [27]. Hypothyroidism and depression were excluded from the model, given the pathology of CLTI, and chronic kidney disease and fluid and electrolyte disorders were added as clinically relevant components for CLTI. Frailty is defined as a CLTI-FRS score > 0.40 (number of points divided by 8) [27]. Of 1,414,080 CLTI-related hospitalizations, in-hospital mortality, major amputation, LOS, and hospital costs were highest among frail patients [27].

3.2.5. Edmonton Frail Scale

The Edmonton Frail Scale is a 17-point scale that encompasses cognition, general health status, functional independence, social support, polypharmacy, nutrition, mood, continence, and 3-meter gait speed [28]. In a study of 125 patients undergoing arterial vascular procedures, frailty according to the Edmonton Frail Scale was associated with a composite outcome of postoperative infections, medical complications, adverse functional outcomes, and postoperative cognitive impairment [29].

3.2.6. Groningen Frailty Indicator

The Groningen Frailty Indicator is a 12-point scale that encompasses mobility, vision, hearing, nutrition, comorbidity, cognition, and psychosocial and physical fitness; a score of 4 or more is indicative of frailty [4,30]. This score has been validated in vascular surgery patients and found to be helpful in the early identification of patients at high risk of postoperative delirium [31,32]. Frailty based on the Groningen Frailty Indicator was associated with higher postoperative complications and discharge to a nursing home after vascular surgery [32]. Some frailty domains, such as mobility, nutrition, cognition, and psychosocial condition, appear to have a more pronounced impact [32].

3.2.7. Hospital Frailty Risk Score

The HFRS assigns a point value to 109 International Classification of Diseases, 10^th Revision codes associated with frailty, such as falls, gait abnormalities, and vitamin D deficiency [33,34]. A score is assigned whenever an HFRS International Classification of Diseases, 10^th Revision code is recorded in an individual patient record during the index admission or any admissions in the preceding 3 years [33]. Patients are classified into groups with a low (HFRS < 5), intermediate (5–15), or high (more than 15) frailty risk, based on previously validated cut-off points [33]. The proposed benefit of this assessment tool is that it does not require face-to-face assessment and does not include variables that may not be routinely collected in electronic hospital records. In vascular surgery patients, adding the HFRS to a model adjusted for age, sex, comorbidity score, socioeconomic status, previous hospitalization, and vascular procedure type was shown to improve the prediction of 2-year mortality and prolonged LOS [35].

3.2.8. Modified Frailty Index

The modified Frailty Index (mFI) is an 11-point scale retrospectively developed using the American College of Surgeons National Surgical Quality Improvement Program (ACS NSQIP), based on the deficit accumulation model [36]. Nine of the 11 deficits for the mFI are comorbidities and, with a cutoff of ≥0.25 for frailty in an FI (≥25% of potential deficits present) [6]. The mFI is widely evaluated in surgical literature, and was found to be the most frequently used frailty tool in surgery [6]. Nonetheless, studies using mFI were deemed too unreliable to include in a recent meta-analysis due to its overreliance on comorbidity [6]. Although the mFI has been retrospectively validated in NSQIP registries, it has never been validated as a prospective survey instrument. This measure is becoming obsolete, as 6 of the 11 required variables have been phased out of ACS NSQIP [37]. Although some studies have used mFI-5 as an abbreviated measure instead of mFI-11 in NSQIP studies, mFI-5 is not a validated tool and violates the tenets of the deficit accumulation model of at least 10–15 deficits being present from the original FI for a valid and robust frailty assessment tool [19].

3.2.9. Ruptured Aneurysm Frailty Score

The Ruptured Aneurysm Frailty Score is a 9-point scale that encompasses indicators of comorbidity (eg, Charlson Comorbidity Index score > 1, polypharmacy of five or more medications, no statin prescribed, and anemia), disability (Katz score), and geriatric-specific domains (hearing and visual impairments) [38]. The study of 184 patients undergoing ruptured abdominal aortic aneurysm (AAA) repair demonstrated that frailty by the Ruptured Aneurysm Frailty Score is predictive of 1-year mortality [38]. This tool has not been validated in other cohorts.

3.3.1. RAI

The RAI was developed as a pragmatic frailty assessment tool suitable for large-scale screening across various surgical populations by capturing five domains of frailty: functional, cognitive, physical, social, and nutritional [2,39]. The RAI is based on the accumulation of deficits model, comprising 14 variables, including age, sex, weight loss, poor appetite, CHF, dyspnea, renal failure, presence of disseminated cancer, functional status, cognitive decline, and living status. Scores can range from 0 to 81, with higher scores indicating increasing frailty and higher surgical risk. It has been validated in two forms, the administrative RAI, calculated retrospectively from variables contained in the Veterans Affairs Surgical Quality Improvement Project or ACS NSQIP and the clinical RAI, calculated from responses to a survey instrument and validated prospectively [2,40]. Frailty as assessed by RAI was associated with postoperative mortality across all noncardiac surgical specialties regardless of case-mix [41]. Furthermore, implementation of routine frailty screening with the RAI to increase appropriate patient selection for surgery is associated with improved institutional surgical survival [39].

3.3.2. VQI-derived frailty scoring

The VQI is a multicenter registry of the Society for Vascular Surgery Patient Safety Organization that collects demographic, clinical, procedural, and outcome variables for patients undergoing 14 different vascular procedures [42]. There are two frailty measures that have been developed in this database: the VQI-Frailty Score and VQI-RAI. The VQI-Frailty Score is an abbreviated frailty score based on 11 or fewer VQI variables (ie, hypertension, CHF, coronary artery disease, PAD, diabetes, COPD, renal impairment, anemia, underweight, non-home residence, and nonambulatory status) that map to recognized frailty domains in the CGA [42].

In the VQI-RAI, the nutritional domain is represented by extremes of body mass index (calculated as kg / m²; underweight: < 20; morbidly obese: > 35) and shortness of breath is replaced with COPD [44,45]. Because the VQI data set does not contain variables for cancer or cognitive status, these variables are not included in the VQI-RAI. This tool includes age, sex, body mass index, chronic kidney disease, CHF, COPD, living status, and functional status [44,45].

The VQI-based assessments, like most abbreviated tools, are based on the deficit accumulation model and easy to use in large registry data analyses. However, they are population-specific, as they were developed in retrospective data of patients undergoing vascular surgery. Prospective validation and comparison with other frailty measures should be performed for these VQI-based tools in future research.

3.4. Physiologic or phenotypic markers of frailty in vascular surgery

3.4.1. Nutritional markers

There have been recent attempts to include other phenotypic assessments in frailty scoring, including nutritional status using laboratory values (eg, serum albumin or pre-albumin levels), triceps skinfold thickness, and imaging-based assessment of sarcopenia [46]. Albumin is a biomarker of systemic inflammation and nutritional status, both of which contribute to the frailty phenotype; low preoperative albumin level was associated with a higher risk of postoperative functional decline and adverse outcomes [47,48]. In a study of 1,089 patients undergoing fenestrated and branched endovascular aortic repair, hypoalbuminemia was associated with increased 30-day and 2-year mortality rates [49].

3.4.2. Morphometric sarcopenia and body composition

Sarcopenia, loss of lean muscle mass and quality, is highly prevalent in vascular surgical patients and may have prognostic clinical value [43,50]. Although frailty and sarcopenia are frequently overlapping health states, not all individuals with sarcopenia are frail [6,51]. Studies have shown the utility of computed tomography measurement of total psoas muscle area (TPA) or skeletal muscle area (SMA; abdominal wall, paraspinal, and psoas) as a marker of sarcopenia. For TPA, the axial slice at the level of the most caudal aspect of the L3 vertebra is identified and a region of interest is traced manually around the psoas muscle of each side. The combined total area is normalized for patient height, with sarcopenia defined as a TPA of < 500 mm²/m² [51–53]. Decreased psoas muscle area, measured at the mid-body of the third lumbar vertebral body (L3) on axial imaging [54], was associated with higher risk of mortality after surgery [55,56]. Sarcopenia, defined as having an SMA < 114.0 cm² in men or < 89.8 cm² in women, has been used to assess sarcopenia in patients with CLTI and showed sarcopenia is an independent predictor of mortality [57,58]. Preoperative sarcopenia is associated with poorer survival and longer LOS in patients undergoing elective endovascular aortic aneurysm repair [53].

Because both sarcopenia and myosteatosis are components of body composition, some models have also included the latter in their frailty imaging markers. Myosteatosis is defined as decreased skeletal muscle density relative to body mass index, measured as low Hounsfield units on computed tomography [59,60]. In a study of 210 patients admitted to a vascular unit, decreased subcutaneous fat depth was significantly associated with mortality, readmission within 12 months, and increased cost of health care [61]. More than one-half of all patients undergoing AAA repair are sarcopenic, and sarcopenia with myosteatosis is associated with double the mortality of sarcopenia alone in this patient population [59]. In a study of 314 patients admitted to a tertiary vascular unit in a single year, the rates of sarcopenia, measured by SMA [54,57,58], did not differ between occlusive and aneurysmal diagnoses, and was not associated with mortality [43]. Similarly, another study assessed morphometric sarcopenia as a method of risk stratification in patients undergoing elective AAA intervention and found TPA [51–53] was not associated with mortality and was thus not a suitable risk stratification tool for this patient population [62]. Differences in association of sarcopenia with outcomes are likely due to heterogeneity in study designs. For example, the two previously mentioned studies, by Waduud et al [62] and Heard et al [43], had larger sample sizes and better accounted for confounders, whereas the study by Thurston et al [53]—which showed a strong association—used the same measure of sarcopenia as the study by Waduud et al, but a lower threshold to define sarcopenia. Further studies are necessary to better define and standardize the cutoff for sarcopenia.

3.4.3. Mobility measures

3.4.3.1. Gait speed

In the 5-meter gait speed test, patients position their feet behind and just touching the 0-meter start line and are instructed to “walk at their comfortable pace” until a few steps past the 5-meter mark [63]. Three trials are repeated, allowing sufficient time for recuperation between trials and average gait speed to be calculated. Gait speed is an independent predictor of adverse outcomes after cardiac surgery, with each 0.1-m/s decrease conferring an 11% relative increase in mortality [63,64]. In a prospective study of 131 patients with PAD undergoing lower extremity revascularization, low gait speed independently predicted adverse outcomes [65].

3.4.3.2. Walking test

The 6-minute walking test (6MWT) is a commonly used test for the assessment of the functional capacity of patients with CHF. In this assessment, participants walk 100 feet for 6 minutes with instructions to cover as much distance as possible. Although the exercise stress test remains the gold standard, the 6MWT is less physically strenuous and may provide reliable information about the patient’s daily activity [66]. A cutoff of ≤500 steps has been found to correlate with a “frail” or high-risk state [67,68]. In patients with PAD, the 6MWT predicts mortality risk and mobility loss, is sensitive to the natural history of declines in walking endurance, and detects improved walking endurance in response to therapeutic interventions [67,69,70]. A smartphone-based 6MWT has shown clinical reproducibility in patients with cardiovascular disease [71]. This allows for measurement of real-world functional capacity in a continuous manner [71].

3.4.3.2. Grip strength

Frailty measures that are walking-based tests may have limited utility for patients with impairment from PAD. Grip strength has been used for frailty assessment among older adults, but its use within patients with cardiovascular disease has not been rigorously validated [72]. In a prospective cross-sectional cohort study of 311 patients with vascular disease, dominant hand–grip strength identified frailty in 27.7% [73]. Frail status based on grip strength was associated with comorbidity, cardiac risk, and sarcopenia in this population [73]. Grip strength may serve as an inexpensive risk screening tool that is easily implemented in ambulatory clinics, and forgoes the need for imaging or walking-based measures [74].

3.5. Association of frailty with procedure-specific outcomes after vascular surgery

3.5.1. Mixed vascular surgery cohorts

Fourteen studies investigated validated frailty tools in patients who collectively underwent a wide range of vascular procedures for carotid disease, aortic aneurysm, PAD, and/or dialysis access creation (Table 2) [10,26,29,31,32,35,42,74–80]. Frailty, based on Groningen Frailty Indicator, mFI-11, Edmonton Frail Scale, CFS, RAI, HFRS, and CGA, was associated with various metrics of mortality and postoperative complications. Three studies investigated phenotypic measures. Sarcopenia is a predictor of mortality in all vascular admissions and patients undergoing procedures for carotid disease, aortic aneurysm, PAD, and/or dialysis access creation [43,50]. Grip strength is associated with comorbidity, cardiac risk, and sarcopenia in patients with vascular disease [73].

Table 2 –

Association of frailty with outcomes after vascular surgery in mixed vascular surgery cohorts.

First author	Year	Country	Design	Surgical procedure	Frailty assessment tool	Outcomes	Sample size	Findings
Pol [31]	2011	Netherlands	Retrospective cohort	All elective vascular procedures	GFI	Primary: delirium. Secondary: surgical complication and LOS	142	GFI was associated with postoperative delirium but not prolonged LOS.
Karam [74]	2013	United States	Retrospective cohort	CEA, PVI, LEBP, EVAR, OAR, LEA	mFI-11	Primary: 30-d mortality. Secondary: Surgical site infection, MI, Calvin IV complication	67,308	As the mFI increased, postoperative wound infection, all occurrences, and mortality increased.
Partridge [29]	2015	United Kingdom	Prospective cohort	Elective and emergency OAR, EVAR, any foot/leg amputation, or any arterial vascular surgery	EFS	Primary: LOS. Secondary: Postoperative morbidity	125	EFS was associated with longer LOS, postoperative complications, and adverse functional outcomes. This association between EFS and LOS was strengthened with the addition of MoCA.
Ehlert [75]	2016	United States	Retrospective cohort	CAS, CEA, OAR, EVAR, PVI, LEBP	mFI-11	Primary: 30-d mortality. Secondary: Clavien-Dindo class IV complications	72,106	mFI was a better discriminator of mortality than Lee Cardiac Risk Index and ASA class in open procedures.
Arya [76]	2016	United States	Retrospective cohort	Elective PVI, LEBP, OAR, EVAR, CAS, and CEA	mFI-11	Primary: Non-home discharge. Secondary: 30-d morbidity and mortality, LOS	15,843	Frail patients were nearly twice as likely to not return home.
Takeji [77]	2018	Japan	Prospective cohort	All surgery for PAD	CFS	Primary end point: 2-y survival. Secondary end point: 2-y AFS	643	Frailty is associated with overall survival and AFS in patients aged ≤75 y and > 75 y, those who underwent endovascular or bypass surgery, and those with/without chronic renal failure.
Donald [26]	2018	United States	Retrospective cohort	All surgery for elective AAA, TAA, and PAD procedures	CFS	Discharge to a nonhome location or 30-d mortality	134	CFS can be used to discharge to a facility or morality after elective major vascular surgery.
Heard [43]	2018	United Kingdom	Retrospective cohort	All vascular admissions	SMA/height2	Mortality and non-home discharge	314	Sarcopenia is not independently associated with mortality.
Reeve [73]	2018	United States	Prospective cohort	All patients seen in vascular surgery clinic	Grip strength	Charlson Comorbidity Index, Revised Cardiac Risk Index, and sarcopenia	311	Grip strength is associated with comorbidity, cardiac risk, and sarcopenia in patients with vascular disease.
Ghaffarian [50]	2019	United States	Retrospective cohort	OAR, EVAR, CEA, PVI and LEBP, HD access creation and intervention, amputation	CFS and SMA	Overall survival	415	Frailty and sarcopenia overlap to varying degrees and can be used alone or in combination to predict long-term survival of older patients. Only frailty is an independent predictor of mortality.
Visser [32]	2019	Netherlands	Prospective cohort	CEA, CAS, OAR, PVI, LEBP, amputation	GFI	Primary: 30-day morbidity. Secondary: 30-day mortality, hospital readmission, facility after discharge	630	Frailty is associated with a higher risk of post-operative complications and discharge to a nursing home. Some frailty domains (mobility, nutrition, cognition and psychosocial condition) appear to have a more pronounced impact.
Rothenberg [78]	2020	United States	Retrospective cohort	All elective vascular procedures	RAI	Mortality, major complications, and LOS	139,569	RAI is a good predictor of mortality after vascular surgery but performed poorly for TAA and TAAA. RAI was not a strong predictor of major complications or prolonged LOS.
Aitken [35]	2021	Australia	Retrospective cohort	PVI, LEBP, EVAR, OAR, CEA, CAS	HFRS	Primary: mortality at 30 d and 2 y after index admission. Secondary: LOS, emergency readmission	9,752	Adding the HFRS to a model adjusted for comorbidities improved the prediction of 2-y mortality and prolonged hospital stay, but there was minimal improvement for 30-d mortality and readmission.
Gilbertson [79]	2021	United States	Retrospective cohort	EVAR, TEVAR, LEBP, PVI, CEA, OAR	CFS	1-y major complication, non-home living status or death	163	Frail patients who undergo high-stress vascular procedures have a significantly higher rate of complications leading to loss of functional independence and mortality
Ayyash [80]	2022	United Kingdom	Prospective cohort	OAR, EVAR, ABF, fem-fem, fem-pop	CFS	30-d readmission, in-hospital mortality, major morbidity, LOS, and readmission	97	The CFS is more practical than the Edmonton Frailty Scale.
DeAngelo [10]	2023	United States	Retrospective cohort	Elective minor and major vascular procedures	CFS	Primary: decline in functional status. Secondary: 2-y all-cause mortality	126	Frailty was associated with an increased likelihood of decline in ADLs at 2 y.
Khan [42]	2022	United States	Retrospective cohort	CAS, CEA, OAR, EVAR, PVI, LEBP	VQI-FS	Primary: 9-mo mortality	265,632	VQI-FS using 7 variables in addition to procedure-specific risk has strong correlation with 15-mo mortality.

Abbreviations: AAA, abdominal aortic aneurysm; ABF, Aortobifemoral bypass; ADL, activities of daily living; AFS, amputation-free survival; ASA, American Society of Anesthesiology; CAS, carotid artery stenting; CEA, carotid endarterectomy; CFS, Clinical Frailty Scale; CGA, Comprehensive Geriatric Assessment; EFS, Edmonton Frail Scale; EVAR, endovascular aortic aneurysm repair; fem-fem, Femoro-femoral bypass; fem-pop, femoro-popliteal bypass; FS, Frailty Scale; GFI, Groningen Frailty Indicator; HD, hemodialysis; HFRS, Hospital Frailty Risk Score; LEA, Lower Extremity Amputation; LEBP, lower extremity bypass; LOS, length of stay; MoCA, Montreal Cognitive Assessment; mFI, modified Frailty Index; MI, myocardial ischemia; OAR, Open Aneurysm repair; PAD, peripheral artery disease; PVI, peripheral vascular intervention; SMA, skeletal muscle area; TAA, thoracic aortic aneurysm; TAAA, throracoabdominal aortic aneurysm; TEVAR, thoracic endovascular aortic repair; VQI, Vascular Quality Initiative.

3.5.2. Lower extremity amputation

Four studies investigated validated frailty tools in patients who underwent amputations only (Table 3) [81–84]. Frailty based on RAI, CGA, and mFI-5 was associated with postoperative comorbidities; however, mFI-11 did not predict postoperative morbidity or mortality [81–84]. In a recent meta-analysis, the difference in postdischarge mortality between the frail and nonfrail groups was more significant in patients undergoing amputation compared with patients who underwent all other types of vascular surgeries [3].

Table 3 –

Association of frailty with outcomes after lower extremity amputation.

First author	Year	Country	Design	Surgical procedure	Frailty assessment tool	Outcomes	Sample size	Findings
Fang [81]	2017	United States	Retrospective cohort	BKA and AKA for CLTI	mFI-5	Primary: mortality. Secondary: 30-d readmission, unplanned revision, composite adverse events	379	Preoperative clinical frailty is associated with an increased 30-d readmission rate.
D’Cruz [82]	2020	Singapore	Retrospective cohort	BKA and AKA for CLTI	mFI-11	Primary: morality and perioperative complications	211	The mFI did not predict outcome after major amputation.
Tse [83]	2021	United States	Retrospective cohort	Nontraumatic BKA and AKA	RAI	Primary: 30-d mortality. Secondary: postoperative complications, LOS	47,197	RAI is associated with several postoperative outcomes in a dose-dependent manner.
Maltese [84]	2022	United Kingdom	Prospective cohort	Amputation for DFU	FI-CGA	Primary: 6-mo DFU healing. Secondary outcome: 6-mo rehospitalization	76	Frailty is a risk factor for DFU nonhealing and rehospitalization.

Abbreviations: AKA, above-knee amputation; BKA, below-knee amputation; CGA, Comprehensive Geriatric Assessment; CLTI, chronic limbthreatening ischemia; DFU, diabetic foot ulcer; FI, Frailty Index; LOS, length of stay; mFI, modified Frailty Index; RAI, Risk Analysis Index.

3.5.3. Lower extremity bypass and peripheral intervention

Five studies investigated validated frailty tools in patients who underwent open surgical and/or endovascular lower extremity revascularization procedures only (Table 4) [27,85–90]. Frailty, as measured by CLTI-FI and mFI-11, was associated with postoperative mortality and complications in patients with CLTI. However, in a small study of 53 patients undergoing revascularization for acute limb ischemia, frailty predicted discharge site and nonambulation at follow-up, but was not associated with amputation or death [85]. Furthermore, sarcopenia has been found to be a prognostic factor for patients with CLTI undergoing revascularization [57]. Similarly, patients with PAD with declining functional performance on the 6-minute walk are at increased risk for later mobility loss and mortality [67].

Table 4 –

Association of frailty with outcomes after lower extremity bypass and peripheral intervention.

First author	Year	Country	Design	Surgical procedure	Frailty assessment tool	Outcomes	Sample size	Findings
McDermont [70]	2011	USA	Prospective cohort	Patients with PAD	6-min walk test	Mortality and mobility loss	102	Patients with declining functional performance are at increased risk for mobility loss and mortality.
Matsubara [57]	2015	Japan	Retrospective cohort	All surgery for CLTI	SMA	Overall survival	64	Sarcopenia is a prognostic factor for patients with CLTI.
Brahmbhatt [86]	2016	United States	Retrospective cohort	PVI and LEBP	mFI-11	Primary: 2-year AFS Secondary: Clavien-Dindo class IV complications and 30-d or hospital mortality	24,645	mFI was associated with mortality and major complications.
Ali [88]	2018	United States	Retrospective cohort	Infrainguinal bypass	mFI-11	Primary: 30-d mortality Secondary: postoperative MI, stroke, renal failure, graft failure	4,704	With mFI score 0.54–0.63, the risk of 30-d mortality and postoperative complications increases significantly.
Morisaki [87]	2019	Japan	Retrospective cohort	PVI and LEBP	CLTI-FI	Primary: 2-y AFS Secondary: Clavien-Dindo class IV complications and 30-d or hospital mortality	266	CLTI-FI was a predictor of 2-y AFS and 30-d hospital mortality or morbidity.
Sarkar [85]	2021	United States	Retrospective cohort	Procedure for acute limb ischemia	mFI-11	Primary: In-hospital mortality, major amputation, site of discharge, and ambulatory status	53	Although frailty predicts discharge site and nonambulation at follow-up, it is not associated with amputation or death.
Karim [27]	2022	United States	Retrospective cohort	PVI and LEBP	CLTI-FI	Primary: In-hospital mortality and major complications Secondary: hospital cost, LOS, in-hospital complications	1,414,080	In patients with CLTI, frailty is associated with mortality and amputation. Malnourished and frail patients were observed to have a mortality benefit with a less invasive revascularization.
Al-Damluji [90]	2022	United States	Retrospective cohort	PVI and LEBP	VQI-RAI and VQI-FI	perioperative mortality and major amputation	134,081	There is significant variation in frailty detection by VQI-FI and VQI-RAI. Frailty is associated with mortality. Association between frailty and unplanned amputations was only noted in the bypass cohort using VQI-RAI.
Majmundar [89]	2023	United States	Retrospective cohort	PVI and LEBP	HFRS	In-hospital mortality and major amputation	64,338	In patients with CLTI, frailty predicts in-hospital mortality and 6-mo major amputation, regardless of procedure type.

Abbreviations: AFS, amputation-free survival; CLTI, chronic limb-threatening ischemia; FI, Frailty Index; HFRS, Hospital Frailty Risk Score; LEBP, lower extremity bypass; LOS, length of stay; mFI, modified Frailty Index; MI, myocardial infarction; PAD, peripheral arterial disease; PVI, peripheral vascular intervention; RAI, Risk Analysis Index; SMA, skeletal muscle area; VQI, Vascular Quality Initiative.

3.5.4. Carotid intervention

Three studies investigated validated frailty tools in patients who underwent carotid endarterectomy or stenting (Table 5) [45,91]. Frailty measured by VQI-RAI was not associated with postoperative stroke, but was associated with decreased long-term survival in patients undergoing carotid endarterectomy [45]. Interestingly, when studied in the ACS NSQIP database, increased RAI score correlated with increased risk for postoperative stroke [91]. This discrepancy may be attributed to the differences in variable collection and the crosswalk between RAI and VQI frailty variables.

Table 5 –

Association of frailty with outcomes after carotid intervention.

First author	Year	Country	Design	Surgical procedure	Frailty assessment tool	Outcomes	Sample size	Findings
Melin [91]	2015	United States	Retrospective cohort	CEA	RAI	Primary: mortality, LOS, non-home discharge Secondary: in-hospital stroke, TIA, MI	44,832	Frailty is a predictor of increased stroke, mortality, MI, and LOS after CEA for both symptomatic and asymptomatic stenosis.
Rothenberg [45]	2020	United States	Retrospective cohort	CEA	RAI	Primary: mortality, LOS, non-home discharge Secondary: in-hospital stroke, TIA, MI	42,869	RAI score is not associated with postoperative stroke, but frailty is associated with decreased long-term survival.
Khan [42]	2022	United States	Retrospective cohort	TCAR	mFI-5	Primary: mortality Secondary: TIA, MI, stroke/death, stroke/TIA, stroke/death/MI, discharge to nonhome facility, and LOS	17,983	Frail patients are at higher risk of in-hospital mortality, TIA, MI, stroke/death, stroke/TIA, stroke/death/MI, discharge to nonhome facility, and extended LOS.

Abbreviations: CAS, carotid artery stenting; CEA, carotid endarterectomy; LOS, length of stay; mFI, modified Frailty Index; MI, myocardial infarction; RAI, Risk Analysis Index; TCAR, trans-carotid artery stenting; TIA, transient ischemic attack.

3.5.5. AAA repair

Four studies investigated validated frailty tools in patients who underwent elective open and/or endovascular AAA repair (Table 6) [44,92–94]. Frailty, measured by mFI, CFS, VQI-RAI, and RAI, was associated with mortality and various postoperative complications in patients undergoing repair of AAA. Interestingly, frailty, but not procedure type, was associated with long-term mortality [94]. Five retrospective studies of phenotypic metrics showed that sarcopenia measured by TPA or SMA were predictive of morbidity and/or mortality in patients undergoing open and/or endovascular AAA repair [53–56,59]. However, in one single-center study, sarcopenia based on TPA was not associated with mortality in a similar cohort of 382 patients [62]. The difference in results may be due to difference in sample size and threshold values to define sarcopenia.

Table 6 –

Association of frailty with outcomes after aortic aneurysm repair.

First author	Year	Country	Design	Surgical procedure	Frailty assessment tool	Outcomes	Sample size	Findings
Lee [55]	2011	United States	Retrospective cohort	Elective OAR	TPA	All-cause mortality	262	TPA was a significant predictor of mortality.
Arya [92]	2015	United States	Retrospective cohort	Elective EVAR or OAR	mFI	Primary: 30-d mortality Secondary: 30-d morbidity and failure to rescue	23,207	Higher mFI, independent of other risk factors, is associated with higher mortality and morbidity in patients undergoing elective EVAR and OAR.
Hale [54]	2016	United States	Retrospective cohort	Elective EVAR	SMA	Overall survival	200	Sarcopenia is a predictor of long-term mortality in patients treated with EVAR.
Drudi [56]	2016	Canada	Retrospective cohort	Elective OAR or EVAR	TPA	All-cause mortality	149	TPA is independently associated with all-cause mortality.
Thurston [53]	2018	Australia	Retrospective cohort	Elective EVAR	TPA/height2	Primary: Overall survival Secondary: Overall postoperative complication	191	Sarcopenia is associated with poorer survival and of longer LOS.
Kays [59]	2018	United States	Retrospective cohort	OAR or EVAR	SMA/height2	Primary: Overall survival Secondary: 30-d morbidity	505	One-half of patients undergoing AAA repair are sarcopenic. Sarcopenia with myosteatosis is associated with double the mortality of sarcopenia alone.
Waduud [62]	2019	United Kingdom	Retrospective cohort	Elective EVAR or OAR	TPA/height2	Primary: Overall survival Secondary: 30-d morbidity	380	TPA was not associated with mortality.
Al Shakarchi [93]	2020	United States	Retrospective cohort	Elective OAR	CFS	Primary: 30-d mortality Secondary: 30-d reoperation, LOS, readmission	184	CFS predicts mortality and morbidity.
George [94]	2020	United States	Retrospective cohort	Elective EVAR or OAR	VQI-RAI	Long-term mortality and non-home discharge	15,803	Frailty, but not procedure type, was associated with long-term mortality; however, frailty and procedure type were both associated with nonhome discharge.
George [41]	2021	United States	Retrospective cohort	Elective EVAR	RAI	Non-home discharge and adverse outcomes	11,694	Frailty was associated with mortality, complications, and non-home discharge

Abbreviations: AAA, abdominal aortic aneurysm; CFS, Clinical Frailty Scale; EVAR, endovascular aortic aneurysm repair; LOS, length of stay; mFI, modified Frailty Index; OAR, open aortic aneurysm repair; RAI, Risk Assessment Index; SMA, skeletal muscle area; TPA, total psoas area; VQI, Vascular Quality Initiative.

4. Discussion

Unique to vascular surgery, patients are evaluated for surgical treatment across a continuum of symptomatology. Invasive surgical therapy is offered to reduce future severe morbidity and mortality in otherwise asymptomatic disease, while other disease presentations have progressive effects on patient health status, symptoms, and function. Therefore, frailty represents a vital sign to predict postoperative outcomes that greatly impact quality of life. The goal of frailty research is to develop and implement an accurate, understandable, and feasible metric that can be used for perioperative optimization and shared decision making to provide goal-concordant care.

Although there are a variety of tools available to measure frailty in patients undergoing vascular surgery, the predictive value of frailty scoring varies by the scale used and the target population. Not surprisingly, regardless of how frailty is measured, it is associated with increased risks for postoperative mortality and morbidity in vascular surgery patients, as evidenced in the literature summarized in this review. A recent meta-analysis of 24 studies entailing a total of 1,886,611 vascular surgery patients demonstrated that frailty is associated with increased in-hospital mortality and postdischarge mortality [3]. Frailty is also found to be significantly associated with a longer LOS, higher rehospitalization, and higher likelihood of non-home discharge [3]. Sub-analyses further confirmed that frailty is significantly associated with the commonly evaluated clinical comorbidities [3]. Preoperative risk assessment tools that incorporate more domains, such as the RAI, may outperform simpler comorbidity measures in predicting long-term mortality [95]. Despite the heterogenous results from the evaluation of different frailty tools, all validated approaches carry predictive value when applied to large cohorts of patients with a wide range of vascular pathology. Therefore, it is important that when researchers and clinicians select a frailty tool for prospective collection that it is feasible to be integrated in the clinical workflow, deliver a confident risk estimate that can impact decision on appropriate care, validated in the applicable clinical setting, and acceptable to various clinical stakeholders who will be involved in the care of these patients [96].

The utility of frailty assessment tools lies in the ability to perform preoperative assessment in patients considering vascular surgery interventions and optimizing them (Fig. 1). Studies have found that the prevalence of frailty in vascular surgery patients can vary between 10% and 20% [41]. Our group’s approach has been to use frailty assessment as a screening tool to identify robust patients that can follow the usual preoperative care workflow. It is the frail and high-risk group that need a “surgical pause” to do additional assessments and follow a multidisciplinary optimization intervention based on resources available. Surgeons have begun to use the electronic medical record for frailty screening and identified ways to integrate collection into the preoperative evaluation workflow [97]. Frailty screening tools can be implemented efficiently within multispecialty and multicenter health care systems [98,99]. In a recent study, the RAI was integrated into the EPIC electronic record system for a large university health care system. Systematic RAI assessment was implemented on 36,261 patients presenting to surgical clinics across five hospitals [98]. The median time required to collect the RAI was 33 seconds (interquartile range 23–53 seconds) [98]. Overall clinic compliance with the recommendation for RAI assessment increased from 58% in the first month of the study period to 84% in the sixth and final month [98]. Moreover, studies have demonstrated that institution-wide frailty screening and a “surgical pause” reduces short- and mid-term mortality [100–102]. Currently, in a multicenter prospective clinical trial, the role of a structured multidisciplinary workflow as an intervention is being studied in the preoperative care of frail veteran patients (ClinicalTrials.gov ID: NCT05037292) [101]. Although many multidisciplinary stakeholders could be incorporated into the care of these patients (Fig. 1), institutional resources and availability of personnel would dictate how the concept of “surgical pause” is adopted for any particular care setting. For example, in our institutional quality improvement study, we created a pathway to involve anesthesia, nutrition, and case management preoperative to optimize frail patients, as other partners, such as geriatrics/palliative care or rehabilitation, did not have the capacity to be included in this pathway at the current time.

Fig. 1 –

Decision-making algorithm in treatment of older or clinically high-risk patients being considered for vascular surgery.

Within vascular surgery, future research needs to focus on identifying different treatment thresholds for common conditions. For example, analysis of VQI data indicate that 45% of patient with AAA undergo repair sizes smaller than the guideline-recommended 55-mm diameter, including 20% of open AAA repairs in frail patients and 35% of endovascular aortic aneurysm repairs in frail patients [94]. Although the robust patients with AAA carry up to a 5% risk of mortality at 1 year, the frail subpopulation has a 15% to 20% risk of 1-year mortality regardless of operative approach [94]. This warrants examination of possibly a higher threshold of AAA repair in frail patients, as natural history studies have shown lower risk of rupture (approximately 5% to 10%, up to a size of 6 cm) compared with 1-year mortality of 20% in these patients after endovascular aortic aneurysm repair/open repair [94]. Asymptomatic carotid disease and claudication are low acuity presentations and interventions for these indications in high-risk frail patients should be examined carefully, as any complications from invasive procedures may significantly alter quality of life and functional independence.

The utility of frailty measures may be further expanded to track longitudinal changes in patients’ function, with their past responses to physiological stressors as an estimate for future response [1]. Comorbidity-based frailty indices, such as the mFI, lack granularity of dynamic metrics over time, whereas tools like the RAI, which capture granular information on functional, social, cognitive, and nutritional status, allow for tracking frailty across a patient’s lifespan [1]. This can be especially important when considering longitudinal care for vascular patients and choosing a vascular access, staging interventions for PAD or aneurysmal disease, or pursuing medical management over invasive therapy, as decision making may change over time to preserve quality of life.

5. Conclusions

This review highlights that frailty—as ascertained by validated assessment tools—is associated with adverse postoperative outcomes in patients who undergo various types of minor or major vascular surgery interventions. Validated frailty assessment tools should be preferred clinically while further study is done on identifying and integrating morphometric markers into existing risk-assessment models. Frailty screening tools can be integrated into the routine clinical workflow of surgeons as a supplement to the subjective assessment of patient fitness to identify high-risk patients who might benefit from perioperative optimization. Furthermore, preoperative frailty assessment can also be used to counsel patients who might not live long enough to benefit from vascular surgery interventions. Ongoing clinical trials and future prospective studies will shed light on how best to incorporate frailty assessment into perioperative care at a larger population level.