Frailty in patients undergoing transcatheter aortic valve replacement: from risk scores to frailty-based management

Aortic stenosis (AS) is the most common valvular disease in the western countries and is associated with aging.^[1,2] Transcatheter aortic valve replacement (TAVR) now offers a therapeutic option for elderly patients who are deemed inoperable and is an excellent alternative for those who are considered high-risk for surgical valve replacement.^[3,4] However, mortality after TAVR is still high with one-year and two-year mortality rates of 24.3% and 33.9%, respectively.^[5] Thus, utilization of effective risk stratification tools before TAVR cannot be overstated.^[5–8] Currently, EUROSCORE II and the Society of Thoracic Surgery score are the recommended risk models to predict postoperative mortality in cardiac surgery, however, they were not designed to be used for TAVR patients.^[6,8] The main limitation in the implementation of these models is that they do not take into account frailty.^[9] Frailty assessment could not only offer incremental prognostic information but could also theoretically improve outcomes with preoperative or postoperative management and modification of specific frailty components.^[10,11]

PATHOGENESIS OF FRAILTY AND AORTIC STENOSIS

The pathogenesis of frailty is likely associated with age-related biologic changes that lead to multisystem impairment.^[12,13] Specifically, frailty has been related with dysregulation of the immune, endocrine and musculoskeletal systems,^[12] with chronic inflammation and neuromuscular dysfunction considered key underlying mechanisms.^[14] In severe AS, left ventricular remodeling and diastolic heart failure predispose to reduced exercise capacity and increased episodes of hospitalization.^[15–17] Thus, the natural course of severe AS leads to immobility and a chronic inflammatory state induced by valvular pathology and the frequent hospitalizations.^[15,18,19] Importantly, TAVR patients are frequently considered inoperable or high risk for surgery, thus frailty is by definition more common in this population compared to surgical patients.

IMPACT OF FRAILTY ON MORTALITY AND QUALITY OF LIFE AFTER TAVR

Frailty has been associated with increased short-term and long-term mortality after TAVR. Afilalo, et al.^[10] evaluated seven different frailty scales in the FRAILTY-AVR study, a prospective multi-center cohort comprising 1,020 elderly patients (minimum age: 70 years) undergoing TAVR or surgical aortic valve replacement. Among the TAVR population subgroup, frailty as measured by all seven scales, was independently associated with increased one-year mortality. Essential Frailty Toolset (EFT) was the strongest predictor of one-year mortality [adjusted odds ratio (OR) = 3.19, 95% CI: 2.03−5.02] (Figure 1). Additionally, five out of seven frailty scales significantly predicted thirty-day mortality in TAVR cohort.^[10] In another analysis of FRAILTY-AVR study comprising 400 TAVR patients, the prognostic value of sarcopenia was evaluated, which is one of the biological substrates of frailty.^[20] Sarcopenia was defined as low psoas muscle area by computed tomographic imaging and low chair-rise performance, and was associated with significantly increased one-year mortality after TAVR (OR = 11.30, 95% CI: 2.51−50.91).^[20]

Figure 1.

Essential frailty toolset.

Reprinted from Journal of the American College of Cardiology. Reference: Afilalo J, Lauck S, Kim DH, et al. Frailty in older adults undergoing aortic valve replacement: the FRAILTY-AVR study. J Am Coll Cardiol 2017; 70: 689–700. DOI: 10.1016/j.jacc.2017.06.024 with permission from Elsevier. EFT: essential frailty toolset; SAVR: surgical aortic valve replacement; TAVR: transcatheter aortic valve replacement.

A similarly-sized multicenter registry from Japan evaluated frailty using a semi-quantitative Clinical Frailty Scale (CFS),^[21] which categorized patients into five groups with increasing frailty grade (Table 1).^[22] Results showed that the CFS grade was associated with increasing thirty-day mortality risk and was assessed as an independent risk factor for one-year death rate [hazard ratio (HR) = 1.28, 95% CI: 1.10−1.49].^[22] In this study, the CFS grade correlated with other markers of frailty such as body mass index < 20 kg/m ², serum albumin < 3.5 mg/dL late gait speed and weak grip strength. Hence, although CFS is a semi-quantitative tool, it accurately reflected the physical component of frailty in TAVR patients and was useful in predicting mortality outcomes after TAVR. Moreover, two additional studies demonstrated that frailty assessment alone accurately predicted one-year mortality after TAVR and its additive value in conventional risk scores of one-year mortality. ^[11,23]

Table 1. Frailty measurement using CFS in the study by Shimura, et al.^[22].

Groups	Condition/CFS stage	Description
CFS: Clinical Frailty Scale.
Group 1	CFS 1	Very fit: robust, active, energetic with regular exercise
	CFS 2	Well: no symptomatic disease, occasional exercise, less fit than CFS 1 patients
	CFS 3	Managing well: well-controlled medical problems, no regular exercise beyond routine walking
Group 2	CFS 4	Vulnerable: independent, though common activity limitation due to medical problems, self-report of ‘slowness’ or ‘tiredness’ during the day
Group 3	CFS 5	Mildly frail: more evident slowing, require help in high order instrumental activity of daily living
Group 4	CFS 6	Moderately frail: require help with all outside and home-based activities
Group 5	CFS 7	Severely frail: completely dependent on others for any activity of daily living, albeit stable
	CFS 8	Very severely frail: completely dependent, close to end of life, vulnerable even to minor illnes
	CFS 9	Terminally ill: approaching to the end of life, or life expectancy < 6 months, even if not otherwise frail

A similar relationship between frailty and short-term and long-term mortality has been replicated by two recent meta-analyses evaluating frailty status in 5,876 and 3,159 patients, respectively.^[24,25] Frailty was associated with increased risk of one-year mortality (HR = 2.16, 95% CI: 1.57−3.00),^[24] and increased early (≤ 30 days; HR = 2.35, 95% CI: 1.78−3.09) and late mortality (> 30 days; HR = 1.63, 95% CI: 1.34−1.97).^[25] A recent retrospective cohort of 890 patients from the Alina Health System database demonstrated that along with other comorbidities, excessive frailty significantly increased the three-year mortality risk (HR = 2.48, 95% CI: 1.79−3.44).^[26] Increasing evidence from additional studies demonstrated that the addition of frailty assessment in existing prediction risk models significantly improves discrimination for short-term and mid-term mortality rates following TAVR.^[27–29] Specifically, a retrospective analysis of 266 frail TAVR patients demonstrated that frailty status was correlated with both periprocedural and late mortality, with higher frailty indexes associated with higher risk for one-year all-cause death.^[27] Additionally, Stortecky, et al.^[28] with a prospective cohort of 100 patients undergoing TAVR suggested that the addition of Multidimensional Geriatric Assessment-based information to global risk scores could further improve the risk prediction for all-cause mortality and major adverse cardiovascular and cerebral events at thirty-day and one-year after the index procedure. Details regarding impact of frailty on mortality in selected TAVR studies are presented in Table 2.

Table 2. Impact of frailty on mortality in selected studies.

Study	Number of patients	TAVR patients	Design	Frailty scales	Outcome
					Thirty-day mortality	One-year mortality	Long-term mortality
*Presented as estimated using EFT in TAVR and surgical aortic valve replacement cohorts.**Presented as estimated using EFT in TAVR cohort. ***Presented as frailty index combined with EUROSCORE. ****Presented as frailty index combined with STS score. ADL: activities of daily living; BADL: basic activity of daily living; BMI: body mass index; CI: confidence interval; EFT: essential frailty toolset; HR: hazard ratio; IADL: instrumental activities of daily living; LR: likelihood ratio; MMSE: mini-mental state examination; MNA: mini-nutritional assessment; NCT: national clinical trial; NRI: net reclassification improvement; OCEAN-TAVI: Optimized CathEtervAlvulariNtervention-TAVR; OR: odds ratio; PARTNER: placement of aortic transcatheter valves; SPPB: short physical performance battery; STS: Society of Thoracic Surgeons; TAVR: transcatheter aortic valve replacement; TUG: time up and go; TVT: transcatheter valve therapy registry; 5MGS: 5-m gait speed.
Afilalo, et al.[10]	1,020	646	Prospective multicenter cohort	Fried, Fried+, SPPB, Rockwood, Bern, Columbia, EFT	OR = 3.29*, 95% CI: 1.73–6.26	OR = 3.19**, 95% CI: 2.03–5.02
Shimura, et al.[22]	1,215	1,215	Multicenter registry (OCEAN-TAVI)	Clinical frailty scale stage	HR = 1.42, 95% CI: 1.04–1.95	HR = 1.28, 95% CI: 1.10–1.49
Rogers, et al.[11]	544	544	Prospective single-center cohort	≥ 3 out of 5: BMI < 20 kg/m 2, Serum albumin < 3.5 g/dL, Katz index of independence in ADL score ≤ 4/6, Low grip strength, Slow 15-foot walk time	OR = 5.06, 95% CI: 1.36–18.8; NRI = 73%, 95% CI: 31–41.1 (after adding frailty to STS prediction model)	OR = 2.75, 95% CI: 1.55–4.87; NRI = 52.7%, 95% CI: 28–77 (after adding frailty to STS prediction model)
Schoenenberger, et al.[23]	330	330	Prospective single-center cohort	MMSE, TUG, MNA, BADL, IADL, MOBILITY DISABILITY		HR = 3.29, 95% CI: 1.98–5.47; LR chi-square test statistic = 38.27, C-statistic = 0.72, P < 0.001 ***; LR chi-square test statistic = 28.71, C-statistic = 0.68, P < 0.001 ****
Tang, et al.[26]	890	890	Retrospective single-center cohort	≥ 3 out of 5: BMI < 20 kg/m 2, Serum albumin < 3.5 g/dL, Katz index of independence in ADL score ≤ 4/6, Low grip strength, Slow 15-foot walk time	HR = 2.48, 95% CI: 1.79–3.44
Steinvi, et al.[27]l	498	498	Retrospective single-center cohort	≥ 3 out of 5: BMI < 20 kg/m 2, Serum albumin < 3.5 g/dL, Katz index of independence in ADL score ≤ 4/6, Low grip strength, 5-meter walk test		HR = 2.2, 95% CI: 1.25–3.96 (frailty status); HR = 2.0, 95% CI: 1.08–3.7 (for 3 frailty criteria); HR = 3.07, 95% CI: 1.4–6.7 (for 4–5 frailty criteria)
Stortecky, et al.[28]	100	100	Prospective cohort	Bern scale, BADL, IADL, MMSE, MNA, Poor mobility, TUG	OR = 8.33, 95% CI: 0.99–70.48	OR = 3.68, 95% CI: 1.21–11.19
Hermiller, et al.[29]	2,482 (derivation cohort)	2,482 (derivation cohort)	Secondary analysis of Core Valve Trial (NCT01240902)	5MGS, Albumin, Assisted living, BADL, Falls, Grip strength, MMSE, Weight loss	HR = 1.60, 95% CI: 1.04–2.47 (Albumin < 3.3 g/dL); HR = 1.68, 95% CI: 1.05–2.69 (Assisted living); HR = 1.82, 95% CI: 1.17–2.82 (Katz ADL ≥ 2 deficits)	HR = 1.94, 95% CI: 1.54–2.45 (Katz ADL ≥ 2 deficits); HR = 1.90, 95% CI: 1.52–2.38 (Assisted living); HR = 1.40, 95% CI: 1.04–1.91 (Albumin < 3.3 g/dL); HR = 1.43, 95% CI: 1.10–1.85 (Weight loss); HR = 1.36, 95% CI: 1.03–1.81 (Falls); HR = 1.26, 95% CI: 1.03–1.55 (Grip strength); HR = 1.42, 95% CI: 1.06–1.91 (5MGS);
Green, et al.[30]	244	244	Secondary analysis of PARTNER trial (NCT00530894)	Albumin, Grip strength, Gait speed, Katz ADL		HR = 2.5, 95% CI: 1.40–4.35
Arnold, et al.[31]	8,039	8,039	Observational cohort from the STS/ACC TVT Registry	5MGS	OR = 1.11, 95% CI: 1.01–1.22
Puls, et al.[34]	300	300	Prospective single-center cohort	Katz ADL < 6			HR = 2.67, 95% CI: 1.7–4.3
Van Mourik, et al.[24]	5,876 (9 studies)	5,876 (9 studies)	Meta-analysis	Different frailty scales among the included studies		HR = 2.16, 95% CI: 1.57–3.00
Anand, et al.[25]	1,900 (4 studies)	1,900 (4 studies)	Meta-analysis	Four studies using objective frailty assessment	HR = 2.35, 95% CI: 1.78–3.09
Anand, et al.[25]	3,159 (7 studies)	3,159 (7 studies)	Meta-analysis	Seven studies using objective frailty assessment			HR = 1.63, 95% CI: 1.34–1.97

The association between frailty and quality of life (QoL) with TAVR was shown in a secondary analysis of the PARTNER trial, where frailty, as assessed by the Kansas City Cardiomyopathy Questionnaire overall summary (KCCQ-OS) score was an independent predictor of death or poor QoL at one year after TAVR.^[30] Additionally, a substudy of Transcatheter Valve Therapy Registry among 6,151 patients with available 5-m walk test data revealed that patients with slower gait speed (an index of physical frailty) reported worse one-year KCCQ-OS compared to those with normal gait speed.^[31]

FRAILTY AND POST-TAVR COMPLICATIONS

Data are conflicting regarding the relationship of frailty and postoperative adverse events (as defined by the Valve Academic Research Consortium-2 consensus guidelines).^[32] Shimura, et al.^[22] showed a significant relationship between escalating CFS grade and the incidence of postoperative vascular complications, acute kidney injury (AKI), life threatening major bleeding and length of hospital and intensive care unit stays. The same group of authors in another secondary analysis of Optimized CathEter vAlvular iNtervention-TAVI registry demonstrated an association between frailty and procedural complications, such as need for transfusion, AKI, bleeding and cardiac tamponade using serum albumin level of 3.5 g/dL as a cutoff to dichotomize patients in frail and non-frail.^[33] Finally, Puls, et al.^[34] in a prospective single center cohort of 300 patients concluded that the incidence of minor bleeding, transfusion requirement, stage 3 AKI and renal replacement therapy and prolonged hospitalization was significantly greater in the subgroup of frail patients.

On the other hand, Green, et al.^[30,35] did not identify significant associations between frailty markers and postprocedural stoke, vascular complications or major bleeding events in two secondary analyses of the PARTNER trial (NCT00530894). However, the small sample of frail patients in those studies was probably a limiting factor for detection of such associations. Proper selection of TAVR eligible patients can improve the postoperative outcomes.^[35] Similarly, Honda, et al.^[36] studying 150 patients undergoing TAVR suggested that although objective malnutrition status assessment could predict one-year mortality after TAVR, malnutrition was not associated with increased incidence of in-hospital adverse events following TAVR. Furthermore, a retrospective study enrolling 157 patients undergoing preoperative chest computed tomography examinations to evaluate sarcopenia by measuring skeletal muscle index at two thoracic levels, demonstrated that skeletal muscle index was not associated with surgical complications (postoperative renal failure, stroke, bleeding) following TAVR.^[37] Considering the small datasets and the heterogeneous outcomes, more studies will be needed in the future to better characterize the ideal frailty components whose assessment can improve and predict outcomes in AS patients undergoing TAVR.

FRAILTY-GUIDED MANAGEMENT IN THE PERI-TAVR SETTING AND FUTURE DIRECTIONS

Frailty evaluation provides prognostic information that is additive to existing risk scores. However, it could also provide the basis for therapeutic interventions that could improve outcomes. Experience gained from interventional studies involved frail patients has shown that frailty is potentially reversible.^[38,39] Regular exercise intervention and physical activity may reduce frailty, especially in individuals at greater risk for disability.^[40–42] Additionally, nutritional supplementation and follow-up after discharge in malnourished independent, geriatric patients significantly benefits QoL.^[43,44] To maximize benefits, frailty and disability can be successfully treated using an interdisciplinary multifaceted treatment strategy addressing simultaneously multiple frailty components.^[45–48]

Although these interventions have not been studied adequately in the TAVR population, existing evidence suggests that the early implementation of cardiac rehabilitation program following TAVR enhances patients’ exercise tolerance and functional status as well as QoL.^[49] Proponents of frailty assessment in TAVR in fact support frailty assessment at baseline and after TAVR to allow for the detection of functional changes in the elderly.^[25] Further studies have demonstrated significant benefits regarding functional and emotional status of post-TAVR patients who followed multicomponent rehabilitation programs including early mobilization, physical therapy and psychological support.^[50,51] A multi-disciplinary approach to these patients requires the collaboration of the Heart Valve team with extracardiac specialties including nutritionists, physiotherapists, occupational therapists and psychologists, as soon as the frailty is identified in the preoperative phase. Considering the results of Mamane, et al.^[20] evaluation of sarcopenia, which can be objectively assessed with a limited computed tomography scan or bioelectrical impedance testing along with a clinical muscle test, may provide additional benefits in patients prior to TAVR. The reason is that sarcopenic patients can be effectively managed with intensive nutritional supplementation and exercise to increase their functional reserve for TAVR.

These results will be confirmed by the ongoing PERFORM-TAVR trial (NCT03522454) where the investigators evaluate a multi-faceted intervention in 250 frail participants with Short Physical Performance Battery score ≤ 8.^[52] The intervention consists of a protein-rich oral nutritional supplement for four weeks before-TAVR and a home-based exercise program for twelve weeks post-TAVR. Similarly, results from two ongoing trials (TAVR-FRAILTY/NCT02597985, TAVR-Prehab/NCT03107897) evaluating the efficacy of rehabilitation prior to TAVR are eagerly anticipated.

FRAILTY ASSESSMENT IN CURRENT TAVR PRACTICE

Although multiple studies have shown the importance of frailty assessment in predicting clinical outcomes following TAVR, there is not a single, consensus tool to evaluate frailty in clinical practice.^[53] The 2017 ACC/AHA guidelines for the management of patients with heart valve disease recommend the use of Katz Activities of Daily Living (ADL) to categorize patients to none, mild, or moderate to severe frailty status.^[54] However, they recognize that alternative frailty scores may be acceptable to evaluate frailty stage.^[54] Expert Consensus Decision Pathway for TAVR recommend six-minute walking test and dependence on ADL for physical functioning evaluation; Mini Mental State Exam to address cognitive function; screening for depression with history followed by a validated tool such as the Center for Epidemiologic Studies Depression Scale; and addressing risk for malnutrition (body mass index < 20 kg/m ², ten-pound weight loss in past year, serum albumin < 3.5 mg/dL or score ≤ 11 on Mini Nutritional Assessment). ^[55] Additionally, the role of the Heart Valve team consisting of a multidisciplinary group of experts in different cardiology subspecialties in collaboration with the referral physician and cardiac surgeon is considered critical for effective frailty assessment and risk stratification of TAVR patients.^[55] The 2019 Canadian Cardiovascular Society TAVI position statement supports the use of the EFT and the KCCQ for frailty and QoL assessment, respectively.^[56] However, the wide and heterogenous definition of frailty hinders the comparability of the plethora of frailty tools. Future research into the biological basis of frailty will enable a better understanding of this clinical syndrome and its subsequent standardized, objective incorporation in clinical practice. We propose the incorporation of frailty assessment in clinical practice using EFT or CFS as an initial screening test for patients undergoing TAVR along with further diagnostic and therapeutic considerations to address the key components of frailty (Figure 2).

Figure 2.

Recommended frailty-based clinical management pathway of elderly TAVR patients.

BMI: body mass index; CFS: clinical frailty scale; CT: computed tomography; EFT: essential frailty toolset; GDS: geriatric depression scale; MNA: mini-nutritional assessment; SPPB: short physical performance battery; TAVR: transcatheter aortic valve replacement.

CONCLUSIONS

Frailty is a major risk factor for mortality and functional decline following TAVR and should be a standard evaluation in the heart valve clinic. Baseline and follow-up frailty-guided management could provide additional support to patients and improve outcomes. Useful information from ongoing trials and properly designed future prospective cohorts will provide evidence regarding the appropriate interventions in frail TAVR patients in the preoperative and postoperative period with an ultimate goal to improve survival and functional status. Future efforts should also focus on standardizing frailty assessment, given the large spectrum of current frailty scales and tools.

ACKNOWLEDGMENTS

All authors had no conflicts of interest to disclose.