Research Article

Originally Published 16 March 2017

Impact of the Clinical Frailty Scale on Outcomes After Transcatheter Aortic Valve Replacement

Tetsuro Shimura, MD, Masanori Yamamoto, MD, Seiji Kano, MD, Ai Kagase, MD, Atsuko Kodama, MD, Yutaka Koyama, MD, Etsuo Tsuchikane, MD, … Show All … , Takahiko Suzuki, MD, Toshiaki Otsuka, MD, Shun Kohsaka, MD, Norio Tada, MD, Futoshi Yamanaka, MD, Toru Naganuma, MD, Motoharu Araki, MD, Shinichi Shirai, MD, Yusuke Watanabe, MD, and Kentaro Hayashida, MD On behalf of the OCEAN-TAVI InvestigatorsAuthor Info & Affiliations

Circulation

Volume 135, Number 21

https://doi.org/10.1161/CIRCULATIONAHA.116.025630

PDF/EPUB

Abstract

Background:

The semiquantitative Clinical Frailty Scale (CFS) is a simple tool to assess patients’ frailty and has been shown to correlate with mortality in elderly patients even when evaluated by nongeriatricians. The aim of the current study was to determine the prognostic value of CFS in patients who underwent transcatheter aortic valve replacement.

Methods:

We utilized the OCEAN (Optimized Catheter Valvular Intervention) Japanese multicenter registry to review data of 1215 patients who underwent transcatheter aortic valve replacement. Patients were categorized into 5 groups based on the CFS stages: CFS 1-3, CFS 4, CFS 5, CFS 6, and CFS ≥7. We subsequently evaluated the relationship between CFS grading and other indicators of frailty, including body mass index, serum albumin, gait speed, and mean hand grip. We also assessed differences in baseline characteristics, procedural outcomes, and early and midterm mortality among the 5 groups.

Results:

Patient distribution into the 5 CFS groups was as follows: 38.0% (CFS 1-3), 32.9% (CFS4), 15.1% (CFS 5), 10.0% (CFS 6), and 4.0% (CFS ≥7). The CFS grade showed significant correlation with body mass index (Spearman’s ρ=−0.077, P=0.007), albumin (ρ=−0.22, P<0.001), gait speed (ρ=−0.28, P<0.001), and grip strength (ρ=−0.26, P<0.001). Cumulative 1-year mortality increased with increasing CFS stage (7.2%, 8.6%. 15.7%, 16.9%, 44.1%, P<0.001). In a Cox regression multivariate analysis, the CFS (per 1 category increase) was an independent predictive factor of increased late cumulative mortality risk (hazard ratio, 1.28; 95% confidence interval, 1.10–1.49; P<0.001).

Conclusions:

In addition to reflecting the degree of frailty, the CFS was a useful marker for predicting late mortality in an elderly transcatheter aortic valve replacement cohort.

Introduction

Editorial, see p 2025

The incidence of elderly patients with severe aortic stenosis (AS) has been increasing in recent decades. Transcatheter aortic valve replacement (TAVR) is a safer alternative to open surgery in elderly patients and has become a popular therapeutic choice for AS management.^1,2 TAVR is recommended for patients with a life expectancy >1 year and who are expected to improve their quality of life as a result of the procedure.³ To identify the optimal candidates for TAVR, adequate prescreening or risk stratification should be required before the procedure. Classical surgical prescreening includes an assessment of the patient’s comorbidities; recently, focus on the patient’s clinical frailty and functional status has increased. Frailty is best known as a geriatric syndrome of impaired resiliency to stressors, which confers a high risk for adverse health outcomes.^4,5 This definition and the prognostic value of frailty factors in patients receiving TAVR have not been evaluated, and the majority of clinical frailty scales have been considered too complex and incongruent for routine use.⁶ The Clinical Frailty Scale (CFS), an outcome of the Canadian Study of Health and Aging, is a semiquantitative tool that provides a generally accepted clinical definition of frailty, which can be easily measured even by nongeriatricians.^7,8 However, whether this tool affects clinical outcomes after TAVR has not been previously investigated. To this end, we reviewed data from a Japanese multicenter registry and investigated the usefulness of CFS for predicting the prognosis of patients who underwent TAVR.

Methods

Patient Population and Study Design

We evaluated data for 1215 consecutive patients with symptomatic severe AS who underwent TAVR procedures available from the OCEAN-TAVI registry (Optimized Transcatheter Valvular Intervention-Transcatheter Aortic Valve Implantation).^9,10 The OCEAN-TAVI registry is an ongoing, prospective, multicenter TAVR registry; the current data include information from 9 institutions in Japan reported between October 2013 and April 2016. Patients were selected for TAVR when considered unsuitable or at high risk for surgical aortic valve replacement by consensus between individual centers and after discussions within the cardiology team. The operative risk was calculated using the Logistic European System for Cardiac Operative Risk Evaluation (Logistic EuroSCORE) and the Society of Thoracic Surgeons Predictive Risk of Mortality (STS) score. Clinical data including frailty factors, patient characteristics, echocardiographic data, procedural variables, length of hospital stay, and in-hospital and all-cause mortality rates were prospectively recorded. The CFS scores were determined by a trained medical professional according to the Canadian Study of Health and Aging grading criteria.^7,8 All CFS stages were calculated by face-to-face assessments with patients and families to determine the baseline frailty status before TAVR. The CFS ranged from 1 (very fit) to 9 (terminally ill). The CFS results were categorized into 5 groups as follows: nonfrail (CFS 1–3), vulnerable (CFS 4), mildly frail (CFS 5), moderately frail (CFS 6), and severely frail (CFS ≥7). The combined patient flow consort with detailed CFS criteria is presented in Figure 1. Additional frailty factors, including gait speed (meters/second) calculated by a 15-feet or 5-meter walk and mean grip strength (kg) were also evaluated.

Figure 1. **The combined patient flow consort with detailed CFS criteria.** CFS indicates Clinical Frailty Scale; and TAVR, transcatheter aortic valve replacement.

TAVR Procedures

TAVR procedures have already been described in detail.^9,10 The balloon-expandable Edwards SAPIEN-XT valve (Edwards Lifesciences) and the self-expandable Medtronic CoreValve Revalving System (Medtronic) are clinically approved in Japan. The prosthesis size (20, 23, 26, or 29 mm for Edwards SAPIEN-XT valve and 26, 29 mm for CoreValve Revalving System) was determined on the basis of preprocedural echocardiographic and multidetector computed tomographic findings. The devices were delivered by the femoral artery, iliac artery, apical routes, aortic routes, or subclavian approach. Criteria for selection of the transfemoral (TF), iliac, apical, direct aorta, and subclavian approach included the size, calcification, and tortuosity of the aortoiliofemoral arterial tree and the ratio between sheath size and minimum femoral size. Procedural and other complications during TAVR were evaluated according to the VARC-2 criteria (Valve Academic Research Consortium-2).¹¹ Information on the possible occurrence or cause of death was obtained from the treating hospital or by calling the patient’s family directly. The medical ethics committee at each hospital approved this study protocol, and written informed consent was obtained from all patients before undergoing TAVR.

Statistical Analysis

All statistical analyses were performed using IBM SPSS statistics 22 (SPSS, Inc.) and MedCalc software package (version 12.1.4, MedCalc Software). Continuous variables are expressed as mean ± standard deviation (SD) or median and interquartile range (IQR; 25%-75%), depending on variable distribution. Categorical data are expressed as percentages of the total. Comparisons among the 5 groups were made using χ² tests for categorical covariates and 1-way analysis of variance for continuous covariates that were listed as mean value and SD and the Kruskal-Wallis test for continuous variables that were listed as medians with interquartile ranges. Comparisons of CFS grading and each frailty parameter (body mass index [BMI], albumin, gait speed, and mean grip strength) were analyzed by using Spearman’s correlation. Subgroup analysis of CFS grading predicting late mortality in each frailty component (albumin, BMI, gait speed, and grip strength) and sex was evaluated using Cox regression multivariate modeling. The quantitative and qualitative interactions of CFS grading and each frailty component were also tested in these groups. The interactions of event rate among the subgroups of CFS 1-3, CFS 4-6, and CFS ≥7 in each frailty were analyzed using generalized linear model. The receiver operating characteristic curve analysis was used to examine the discriminating ability of the variables for mortality. The usefulness of CFS grading and traditional surgical risk models, such as the Logistic EuroSCORE and the STS score for predicting mortality, was evaluated by c-statistics; thereafter, additional measurements of these combined factors were compared in change of each c-statistic.¹² A univariate Cox regression analysis was performed to obtain the hazard ratio (HR) for 30-day and late mortality after TAVR. Thereafter, a multivariate analysis was performed using the baseline clinical characteristics and the variables with P-values <0.05 in the univariate analysis to examine the independent associations of CSF grading with late mortality. To perform a sensitivity analysis, the prognostic value of the CFS subgroup was assessed in Cox regression multivariate model 1 (CFS per 1 category increase, excluding CFS ≥7) and model 2 (CFS 1–4 versus CFS ≥5). The Kaplan-Meier method was used to estimate cumulative mortality rates in the 5 groups. Survival differences in each group were compared using log-rank tests. Statistical tests were all 2-sided, and values of P<0.05 were considered statistically significant.

Results

Baseline Patient and Procedural Characteristics

Baseline patient characteristics are summarized in Table 1. Of the 1215 patients included in this study (age 84.4±5.0 years of age, Logistic EuroSCORE 17.0±13.1%), 38.0% were categorized as CFS 1 to 3 (n = 462), 32.9% as CFS 4 (n = 400), 15.1% as CFS 5 (n = 183), 10.0% as CFS 6 (n = 122), and 4.0% as CFS ≥7 (n = 48 [CFS 7: n = 44, CFS 8: n = 4]). No patients were graded as CFS 9. The distributions of CFS grading of the 9 institutions are presented in Figure I in the online-only Data Supplement. Among the 5 study groups, significant differences were observed in the mean age, sex, height, weight, BMI, and body surface area (all P<0.001). The serum albumin and hemoglobin levels were significantly lower in the severely frail group (all P<0.05). The levels of brain natriuretic peptide and the number of patients with peripheral artery disease and previous stroke also differed among groups (all P<0.05). As a result, we observed significant differences among groups in the Logistic EuroSCORE and STS score (all P<0.001). Echocardiographic data did not significantly differ among groups. The associations among albumin, BMI, gait speed, grip strength, and CFS grading are shown in Figure 2. The associations among each of the 4 components and CFS differed significantly among the 5 groups (all P<0.001). We identified significant correlations between CFS grade and BMI (Spearman’s ρ = −0.077, P=0.007), albumin (ρ = −0.22, P<0.001), gait speed (ρ = −0.28, P<0.001), and mean grip strength (ρ = −0.26, P<0.001). Unfortunately, we were unable to obtain data for gait speed and mean grip strength for all patients (76.5% [n = 930] and 71.0% [n = 863], respectively). The availabile gait speed and grip strength data from the 9 centers are presented in Figure II in the online-only Data Supplement. The baseline characteristics of the study patients with nonmissing and missing gait speed and grip strength are presented in Table I in the online-only Data Supplement.

Table 1. Baseline Characteristics of Study Patients

Baseline Clinical Characteristic	Nonfrail CFS 1–3	Vulnerable CFS 4	Mildly Frail CFS 5	Moderately Frail CFS 6	Severely Frail CFS ≥7	P Value
Total patients, n (%)	462 (38.0)	400 (32.9)	183 (15.1)	122 (10.0)	48 (4.0)
Mean age (SD), y	83.8±5.08	84.5±4.98	85.1±4.62	85.4±4.59	85.1±5.73	0.001 (*P for trend 0.11)
≤74, n (%)	21 (4.5)	15 (3.8)	4 (2.2)	2 (1.6)	2 (4.2)	0.46 (*P for trend 0.14)
75–84, n (%)	218 (47.2)	163 (40.8)	70 (38.3)	44 (36.1)	21 (43.8)	0.088 (*P for trend 0.026)
≥85, n (%)	223 (48.3)	222 (55.5)	109 (59.6)	76 (62.3)	25 (52.1)	0.016 (*P for trend 0.006)
Male sex, n (%)	188 (40.7)	96 (24.0)	43 (23.5)	24 (19.7)	10 (20.8)	<0.001
Height, cm	151.7±8.77	148.9±8.88	147.4±8.98	147.4±9.03	146.9±9.41	<0.001
Weight, kg	52.0±10.2	49.2±10.0	49.0±9.00	47.1±10.7	45.5±9.20	<0.001
Body mass index, kg/m²	22.5±3.37	22.1±3.49	22.6±4.06	21.6±3.89	21.0±3.40	0.005
Body mass index <20, kg/m², n (%)	104 (22.5)	120 (30.0)	48 (26.2)	46 (37.7)	20 (41.7)	0.001
Body surface area, m²	1.46±0.17	1.41±0.17	1.40±0.15	1.37±0.17	1.35±0.16	<0.001
Logistic EuroSCORE, %	14.8 (7.40–18.4)	16.7 (8.37–21.7)	19.6 (9.0–25.6)	20.0 (8.37–23.3)	23.2 (11.0–29.5)	<0.001
STS score, %	7.37 (4.20–8.60)	7.84 (4.79–9.22)	10.3 (5.48–11.1)	10.4 (5.56–11.1)	11.7 (6.00–15.1)	<0.001
NYHA class (III/IV), n (%)	179 (38.7)	206 (51.5)	108 (59.0)	73 (59.8)	27 (56.3)	<0.001
15-ft walk gait speed (m/s, n=930)	0.92±0.32	0.79±0.36	0.74±0.33	0.64±0.31	Unable to walk	<0.001
Mean grip strength, kg, n=927	18.7±7.16	16.1±6.25	15.0±4.84	14.1±5.73	13.2±5.80	<0.001
Preprocedural laboratory data
B-type natriuretic peptide, pg/mL	388.0 (102.3–467.9)	499.6 (132.1–599.3)	547.6 (157.0–697.3)	717.2 (168.0–932.5)	831.8 (191.5–744.0)	<0.001
Creatinine, mg/dL	1.06±0.58	0.99±0.46	1.11±0.69	1.09±0.55	1.00±0.46	0.14
Albumin, g/dL	3.86±0.42	3.81±0.46	3.65±0.57	3.54±0.50	3.29±0.67	<0.001
Albumin <3.5, g/dL (%)	66 (14.3)	79 (19.8)	61 (33.3)	51 (41.8)	27 (56.3)	<0.001
Hemoglobin, g/dL	11.5±1.62	11.1±1.56	10.9±1.57	10.8±1.59	10.2±1.51	<0.001
Comorbidity, n (%)
Peripheral artery disease	52 (11.3)	63 (15.8)	40 (21.9)	15 (12.3)	12 (25.0)	0.002
Previous myocardial infarction	34 (7.4)	34 (8.5)	14 (7.7)	10 (8.2)	4 (8.3)	0.98
Previous stroke	51 (11.0)	52 (13.0)	31 (16.9)	23 (18.9)	11 (22.9)	0.031
Previous percutaneous coronary intervention	125 (27.1)	116 (29.0)	49 (26.8)	31 (25.4)	14 (29.2)	0.93
Previous coronary artery bypass graft	39 (8.4)	24 (6.0)	9 (4.9)	9 (7.4)	3 (6.3)	0.49
Diabetes mellitus	125 (27.1)	105 (26.3)	52 (28.4)	31 (25.4)	11 (22.9)	0.94
Hypertension	364 (78.8)	317 (79.3)	137 (74.9)	87 (71.3)	34 (70.8)	0.22
Chronic kidney disease	291 (63.0)	247 (61.8)	96 (52.5)	76 (62.3)	31 (64.6)	0.15
Pulmonary disease	140(30.3)	113 (28.2)	51 (27.9)	39 (32.0)	18 (37.5)	0.65
Liver disease	10 (2.2)	12 (3.0)	9 (4.9)	4 (3.3)	4 (8.3)	0.11
Cancer active	29 (6.3)	24 (6.0)	3 (1.6)	2 (1.6)	3 (6.3)	0.044
Echocardiographic data
Pre–left ventricle ejection fraction, mean ± SD	62.0±12.9	62.7±11.7	61.9±14.5	60.8±13.9	59.2±12.8	0.34
Aortic valve area, cm²	0.65±0.16	0.63±0.16	0.63±0.17	0.62±0.18	0.58±0.19	0.10
Indexed aortic valve area, cm²/m²	0.44±0.11	0.45±0.11	0.45±0.11	0.45±0.13	0.43±0.13	0.75
Aortic valve mean gradient, mm Hg	50.1±17.4	52.1±18.3	50.3±17.7	49.5±18.1	18.6±2.69	0.34
Aortic valve peak gradient, mm Hg	85.8±28.3	89.5±30.0	84.9±28.7	83.3±29.5	89.7±29.3	0.14
Aortic valve peak velocity, m/sec	4.56±0.77	4.66±0.79	4.53±0.79	4.51±0.78	4.65±0.75	0.16
Aortic regurgitation ≥ grade 2, n (%)	53 (11.5)	33 (8.3)	11 (6.0)	12 (9.8)	6 (12.5)	0.20
Mitral regurgitation ≥ grade 2, n (%)	42 (9.1)	38 (9.5)	23 (12.6)	8 (6.6)	9 (18.8)	0.11

CFS, Clinical Frailty Scale; EuroSCORE, European System for Cardiac Operative Risk Evaluation; NYHA, New York Heart Association; and STS, Society of Thoracic Surgeons Predictive Risk of Mortality.

Analyzed by Armitage Cochran trend test.

Values are numbers (%) or mean±SD. The form X (Y-Z) indicates median (IQR; 25%-75%).

Figure 2. **Association among 4 major frailty markers and CFS grading.** A, Association between albumin and CFS grading. A mean albumin value in each CFS group tends to become lower across the 5 groups. In addition, we identified significant correlations between CFS grade and albumin (Spearman’s ρ = −0.22, P<0.001). B, Association between BMI and CFS grading. A mean BMI in each CFS group tends to become lower across the 5 groups. The highest incidence of hypoalbuminemia was observed in the CFS ≥7 group. Also, CFS grade and the incidence of BMI <20 kg/m² showed significant correlations (Spearman’s ρ = −0.077, P=0.007). C, Association between gait speed and CFS grading. A mean gait speed in each CFS group tends to become late across the 5 groups. Furthermore, significant correlations were identified between CFS grade and gait speed (Spearman’s ρ = −0.28, P<0.001). D, Association between grip strength and CFS grading. The mean grip strength tended to become progressively weaker with advancing CFS stage. Moreover, significant correlations between CFS grade and mean grip strength were revealed. CFS indicates Clinical Frailty Scale.

Peri- and Postprocedural Patient Characteristics

Peri- and postprocedural patient characteristics are shown in Table 2. The TF approach was used in 969 patients (79.8%), and non-TF approaches were used in 246 patients (20.2%). Significant differences were observed among groups with respect to TF approach, use of local anesthesia, procedure time, and contrast volume use (P=0.006, P=0.003, P=0.005, P=0.005, respectively). Significant differences occurred among the 5 CFS stages with respect to the total number of patients with acute kidney injury (32 [6.9%] versus 35 [8.8%] versus 21 [11.5%] versus 17 [13.9%] versus 9 [18.8%], P=0.015), major vascular complications (17 [3.7%] versus 40 [10.0%] versus 9 [4.9%] versus 5 [4.1%] versus 4 [8.3%], P=0.002), and life-threatening/major bleeding (19 [4.1%] versus 28 [7.0%] versus 15 [8.2%] versus 12 [9.8%] versus 6 [12.5%], P=0.036). The total length of stay in hospital and intensive care unit significantly increased in parallel with increasing CFS scale severity (P<0.001 and P<0.001, respectively).

Table 2. Peri- and Postprocedural Patient Characteristics and 30-Day In-Hospital Outcomes

Baseline Clinical Characteristic	Nonfrail CFS 1–3	Vulnerable CFS 4	Mildly Frail CFS 5	Moderately Frail CFS 6	Severely Frail CFS ≥7	P Value
Total patients, n (%)	462 (38.0)	400 (32.9)	183 (15.1)	122 (10.0)	48 (4.0)
Periprocedural variable
Transfemoral approach, n (%)	370 (80.1)	314 (78.5)	160 (87.4)	94 (77.0)	31 (64.6)	0.006
Nontransfemoral approach, n (%)	92 (19.9)	86 (21.5)	23 (12.6)	28 (23.0)	17 (35.4)	0.006
Transapical approach, n (%)	81 (17.5)	70 (17.5)	20 (10.9)	24 (19.7)	14 (29.2)	0.79
Transilliac approach, n (%)	9 (1.9)	13 (3.3)	1 (0.5)	4 (3.3)	1 (2.1)	0.51
Transaortic approach, n (%)	0 (0.0)	2 (0.5)	2 (1.1)	0 (0.0)	2 (4.2)	0.011
Transsubclavian approach, n (%)	2 (0.4)	1 (0.3)	0 (0.0)	0 (0.0)	0 (0.0)	0.82
Total time of procedure, min	84.9±42.2	90.5±49.7	92.9±40.5	101.5±74.6	100.9±49.9	0.005
Total time of anesthesia, min	154.7±53.3	155.2±60.1	152.6±58.5	163.2±82.6	161.7±60.2	0.56
Local anesthesia, n (%)	67 (14.5)	39 (9.8)	9 (4.9)	8 (6.6)	5 (10.4)	0.003
Contrast medium volume, mL	124.8±64.2	136.3±62.9	138.7±63.6	118.8±52.9	135.7±60.3	0.005
Postprocedural variable, n (%)
Coronary obstruction	6 (1.3)	3 (0.8)	1 (0.5)	2 (1.6)	0 (0.0)	0.73
Disabling stroke	9 (1.9)	7 (1.8)	2 (1.1)	4 (3.3)	0 (0.0)	0.80
Acute kidney injury	32 (6.9)	35 (8.8)	21 (11.5)	17 (13.9)	9 (18.8)	0.015
Major vascular complication	17 (3.7)	40 (10.0)	9 (4.9)	5 (4.1)	4 (8.3)	0.002
Minor vascular complication	32 (6.9)	16 (4.0)	8 (4.4)	8 (6.6)	1 (2.1)	0.25
Life-threatening/disabling bleeding	19 (4.1)	28 (7.0)	15 (8.2)	12 (9.8)	6 (12.5)	0.036
Major bleeding	40 (8.7)	70 (17.5)	43 (23.5)	22 (18.0)	12 (25.0)	<0.001
Minor bleeding	48 (10.4)	56 (14.0)	29 (15.8)	25 (20.5)	8 (16.7)	0.038
2-valve implantation	5 (1.1)	3 (0.8)	4 (2.2)	5 (4.1)	0 (0.0)	0.05
Cardiac tamponade	4 (0.9)	7 (1.8)	7 (3.8)	3 (2.5)	1 (2.1)	0.15
Conversion to open surgery	2 (0.4)	5 (1.3)	4 (2.2)	3 (2.5)	1 (2.1)	0.23
Outcomes at 30 days, n (%)
All-cause mortality	6 (1.3)	7 (1.8)	3 (1.6)	4 (3.3)	3 (6.3)	0.13
Cardiovascular mortality	5 (1.1)	7 (1.8)	2 (1.1)	4 (3.3)	2 (4.2)	0.27
In-hospital outcomes
All-cause mortality, n (%)	8 (1.7)	10 (2.5)	6 (3.3)	5 (4.1)	4 (8.5)	0.065
Cardiovascular mortality, n (%)	4 (0.9)	6 (1.5)	0 (0.0)	4 (3.3)	0 (0.0)	0.083
Days in hospital, d	12.0 (9.0–19.0)	15.0 (9.0–25.0)	19.0 (12.0–31.0)	18.0 (10.0–34.3)	23.0 (13.0–40.0)	<0.001
Days in intensive care unit, d	1.0 (1.0–2.0)	1.0 (1.0–3.0)	2.0 (1.0–3.0)	2.0 (1.0–3.0)	2.0 (1.0–5.0)	<0.001

CFS indicates Clinical Frailty Scale.

Values are numbers (%) or mean±SD. The form X (Y-Z) indicates median (IQR; 25%-75%).

30-Day and Cumulative 1-Year Mortality

Clinical 30-day follow-up was obtained in all patients enrolled in the OCEAN-TAVI registry. The 30-day and in-hospital mortality rates were not significantly different (P=0.13 and P=0.27, respectively). However, the CFS grade increment was significantly associated with increasing risk of 30-day mortality (HR, 1.42; 95% confidence interval [CI], 1.04–1.95; P=0.029). In all, 143 patients died, and 113 of them died within 1 year; of these patients, 28 were in the CFS 1-3 group, 29 patients in CFS 4, 22 patients in CFS 5, 17 patients in CFS 6, and 17 patients in CFS ≥7. Kaplan-Meier analysis of cumulative mortality in the 5 groups on the basis of CFS is presented in Figure 3. Cumulative 1-year mortality rates in each group were 7.2%, 8.5%, 13.4%, 17.6%, and 45.1%, respectively.

Figure 3. **Kaplan-Maier analysis of all-cause mortality of patients in CFS 1–3, CFS 4, CFS 5, CFS 6, and CFS ≥7 groups.** The cumulative 1-year mortality rates slightly increased across the CFS 1-3, CFS 4, CFS 5, and CFS 6 groups, and a significant increase was observed in patients with CFS ≥7 compared with all other groups. CFS indicates Clinical Frailty Scale.

Prognostic Value of Baseline CFS After TAVR Procedure

The c-statistics for the Logistic EuroSCORE and STS score per unit increment were 0.63 (95% CI, 0.57–0.68; P<0.001) and 0.61 (95% CI, 0.56–0.66; P<0.001), respectively. When CFS grading was added to the scoring system, the c-statistics increased significantly (0.66 [95% CI, 0.63–0.69], P=0.012; and 0.64 [95% CI, 0.61–0.67], P=0.026, respectively). Results of the Cox regression analysis for the association between late mortality and clinical findings are presented in Table 3. In the univariate analysis, HR for CFS, BMI <20 kg/m², Logistic EuroSCORE, New York Heart Association III/IV, brain natriuretic peptide, creatinine, albumin <3.5 mg/dL, hemoglobin, peripheral artery disease, prior coronary artery bypass graft, previous stroke, comorbidity of pulmonary disease, liver disease, and TF approach were independently associated with cumulative late mortality. The multivariate Cox regression model indicated that CFS (HR, 1.28; 95% CI, 1.10–1.49; P<0.001), New York Heart Association III/IV (HR, 1.59; 95% CI, 1.08–2.34; P=0.019), creatinine (HR, 1.21; 95% CI, 1.10–1.38; P=0.001), albumin <3.5 (HR, 1.68; 95% CI, 1.14–2.47; P=0.008), hemoglobin (HR, 0.82; 95% CI, 0.73–0.92; P=0.001), comorbidity of pulmonary disease (HR, 1.55; 95% CI, 1.08–2.22; P=0.017), liver disease (HR, 2.36; 95% CI, 1.12–4.94; P=0.023), and TF approach (HR, 0.64; 95% CI, 0.42–0.97; P=0.035) were independent predictors of late mortality. The subgroup analysis of CFS grading (excluding CFS ≥7 and CFS 1–4 versus CFS ≥5) also revealed a trend or significant increased risk of late mortality after TAVR (Table II in the online-only Data Supplement). Event rate distributions among the CFS 1-3, CFS 4-6, and CFS ≥7 were described in each subgroup of frailty (Figure 4). The interactions of CFS subgroups and 4 frail components did not show any statistical differences (BMI: P=0.33, albumin: P=0.78, gait speed: P=0.34, grip strength: P=0.078), although the grip strength and CFS category tended to show a difference. Table III in the online-only Data Supplement shows the results of the subgroup analysis of CFS grading that predicted late mortality in each frailty component and by sex. The interactions of these factors with CFS grading did not show any statistical differences, although the grip strength and CFS category tended to show a difference.

Table 3. Cox Regression Analysis for the Association Between Cumulative Mortality and Clinical Findings

Variable	Univariate Analysis			Multivariate Analysis
Variable	Hazard Ratio	95% Confidence Interval	P Value	Hazard Ratio	95% Confidence Interval	P Value
Clinical Frailty Scale
CFS (per 1 group increment)	1.51	1.33–1.72	<0.001	1.28	1.10–1.49	0.001
Adjusting factors
Age (per 4.99 y [=1-SD] increase)	0.93	0.79–1.08	0.34
Male (for female)	1.35	0.96–1.90	0.088
Body mass index <20 (for body mass index ≥0)	1.44	1.02–2.03	0.036	1.23	0.85–1.77	0.27
Current smoker (for nonsmoker)	1.57	0.58–4.24	0.37
Logistic EuroSCORE (per 1-point increase)	1.03	1.02–1.04	<0.001	1.01	1.00–1.02	0.19
NYHA class III/IV (for I/II)	2.41	1.70–3.44	<0.001	1.59	1.08–2.34	0.019
B-type natriuretic peptide (per 695.2 pg/mL [1-SD] increase)	1.25	1.12–1.39	<0.001	1.02	0.88–1.17	0.74
Creatinine (per 0.56 mg/dL [1-SD] increase)	1.37	1.23–1.49	<0.001	1.21	1.10–1.38	0.001
Albumin <3.5 g/dL (for albumin ≥3.5)	3.10	2.22–4.33	<0.001	1.68	1.14–2.47	0.008
Hemoglobin (per 1.0 g/dL increase)	0.71	0.64–0.79	<0.001	0.82	0.73–0.92	0.001
Diabetes mellitus, n	1.35	0.95–1.91	0.094
Hypertension, n	1.19	0.79–1.79	0.40
Peripheral artery disease, n	1.58	1.06–2.34	0.024	0.93	0.59–1.48	0.77
Previous myocardial infarction, n	1.46	0.88–2.42	0.15
Previous coronary artery bypass graft, n	1.71	1.01–2.87	0.044	1.54	0.87–2.75	0.14
Previous stroke, n	1.56	1.03–2.35	0.036	1.30	0.84–2.01	0.24
Previous percutaneous coronary intervention, n	1.05	0.73–1.50	0.81
Pulmonary disease, n	1.51	1.08–2.12	0.016	1.55	1.08–2.22	0.017
Liver disease, n	3.02	1.63–5.60	<0.001	2.36	1.12–4.94	0.023
Cancer active, n	1.41	0.74–2.68	0.29
Pre–left ventricle ejection fraction (per 1.0% increase)	0.99	0.98–1.00	0.20
Transfemoral (for nontransfemoral)	0.56	0.39–0.79	0.001	0.64	0.42–0.97	0.035

CFS indicates Clinical Frailty Scale; EuroSCORE, European System for Cardiac Operative Risk Evaluation; NYHA, New York Heart Association; and STS, Society of Thoracic Surgeons Predictive Risk of Mortality.

Figure 4. **Event rate distributions among CFS 1–3, CFS 4–6, and CFS ≥7 in each subgroup of frailty.** A, Mortality rates among CFS 1-3, CFS 4-6, and CFS ≥7 between albumin <3.5 and ≥3.5. B, Mortality rates among CFS 1-3, CFS 4-6, and CFS ≥7 between BMI <20 and ≥20. C, Mortality rates among CFS 1-3, CFS 4-6, and CFS ≥7 between low and high gait speed. D, Mortality rates among CFS 1-3, CFS 4-6, and CFS ≥7 between low and high grip strength. BMI indicates body mass index; and CFS, Clinical Frailty Scale.

Discussion

The current Japanese multicenter registry study demonstrated the potential prognostic value of the simple CFS grading tool for risk stratification before TAVR in elderly patients. The CFS grade was associated with increasing risk of 30-day and cumulative 1-year mortality. Frailty is accepted as a general indicator of a patient’s vulnerability that is highly associated with adverse health outcomes in the geriatric field.^4,5 However, no consensus exists on the most appropriate definition of frailty, and no consensus exists regarding the most clinically useful tool to assess frailty. The CFS grading derived by the Canadian Study of Health and Aging committee is 1 of the most reliable methods to assess frailty among elderly patients.^6–8 Although the CFS grading tool is disadvantaged by its semiquantitative classification, our study identified a positive correlation between CFS grade and several other indicators of frailty defined by the VARC-2 criteria, including BMI <20, serum albumin <3.5, gait speed, and grip strength.¹¹ It should be mentioned that although information regarding BMI and albumin levels were obtained for all patients in this study, we were unable to assess gait speed and grip strength in certain patients, for example those unable to walk or to maintain a grip. In particular, there were no available data concerning gait speed and limited data of grip strength in patients categorized with CFS ≥7. As a result, event rate comparison between CFS grading and other frailty components were not sufficient. The CFS grading method is a simple tool which can be used by nongeriatricians for the purpose of assessing patients’ frailty, and the positive correlation between CFS and other markers of frailty described above supports the feasibility and utility of semiquantitative CFS grading.

In this study, the 30-day mortality rate was only 1.9% as only 23 of 1215 patients died within this time period. Given the limited data set, we were unable to accurately assess the predictive power of CFS grading for 30-day mortality, but it is notable that mortality rates increased in proportion with increasing CFS grade. According to the previous reports, the worse prognosis within 30-day in patients underwent TAVI were investigated concerning the frailty with modified specific frailty criteria defined by the multidimensional geriatric assessments.^13,14 The complex frailty criteria were reliable on the basis of geriatric specialist’s assessments although it was difficult to adapt the daily clinical practice. In addition, relatively simple frailty scores using the Katz index and gait speed also revealed worse early clinical outcomes after TAVI.^15,16 In this study, gait speed was significantly correlated with CFS grade. The results described above illustrate that multiple frailty components, including CFS grade, can be used to risk stratify patients undergoing TAVR. These results were informative because the traditional surgical risk models used to predict early mortality did not account for patients’ frailty.

The cumulative 1-year mortality risk increased progressively with advancing CFS grade (7.2%, 8.6%, 15.7%, 16.9%, and 44.1%, respectively), and CFS was identified as an independent predictor of cumulative 1-year mortality after TAVR. The PARTNER US study is a pivotal study in transcatheter heart valve research which consists of two patient cohorts.¹ In Cohort B, patients who were considered inoperable for surgical aortic valve replacement were randomized to TAVR versus medical treatment, and the 1-year mortality rate was 50.7% in the medical group.¹ This result was comparable to the 1-year mortality rate observed in this study for patients with CFS ≥7, although we were unable to extrapolate what would have been the 1-year mortality rate of these patients if they had been treated medically. Therefore, the practical implication of these findings should be strictly discussed prior to TAVR intervention in patients considered too sick or categorized as CFS ≥7. Before making a final decision, clinicians should review the indications for TAVR by taking into account factors such as multiple comorbidities reflected by traditional surgical risk scores, as well as frailty component scores such as the semi quantitative CFS grading system. The multivariate Cox regression analysis indicated that New York Heart Association class III/IV, creatinine, albumin < 3.5, hemoglobin, coexisting pulmonary disease, liver disease, and non-TF approach were independent predictors of late mortality. These results are in keeping with previous research which has identified these clinical variables as significant predictors of higher mortality rates after TAVR.^17–20 With regards to albumin <3.5, the VARC-2 consensus document suggested that both of these variables reflect a clinically frail state,¹¹ and this suggestion was corroborated by a recent report which demonstrated that albumin <3.5 were associated with worse prognosis after TAVR.²¹ Taken together, these results highlight the importance of careful decision making regarding the appropriateness of TAVR in patients with a high CFS grade. This decision should be made after a thorough assessment of an individual patient’s life expectancy and likelihood of physiological functional recovery. Furthermore, they demonstrate the importance of recognizing that frailty should be an important variable in preprocedure screening for patients undergoing TAVR. The c-statistics analysis indicates that when combined with traditional surgical risk models (eg, EuroSCORE or STS score), CFS grading provides improved predictive value for mortality after TAVR. The availability of CFS grading along with a surgical risk score might be useful to identify those patients who are too sick to benefit from TAVR.

Limitations

Several study limitations should be addressed. First, the CFS tool is semiquantitative and subjective in nature, and therefore it is predisposed to interobserver variability. However, this fact does not appear to reduce the power of this frailty score to predict outcomes. Second, the CFS grading at each institution varied, and the percentages of CFS 1-3 scores increased across the 9 institutions. Thus, these differences may have affected the clinical outcomes. However, the result of stratified Cox regression analysis adjusted for institutional differences also confirmed the result of our conclusion. The CFS grading was an independent predictive factor of increased mortality after TAVR (HR, 1.27; 95% CI, 1.08–1.48; P=0.004). Third, the number of 30-day deaths in this study (n = 23, 1.9%) was too small to analyze using the multivariate regression model. It is possible that the results described earlier may have been different if we had been able to evaluate data for a larger number of events. Fourth, outcomes beyond 1 year are not sufficient in this report. Further investigations are needed to reveal the relationship between the CFS grading and long-term clinical outcomes. Fifth, data regarding gait speed and grip strength were not available for all patients, and the prevalence of multiple comorbidities was higher in the group with missing data compared with the group with nonmissing data. This study lacked gait speed and grip strength data for severely frail patients with CFS ≥7 or an advanced CFS stage. The sample volume of nonmissing gait speed and grip strength were powerless to evaluate for the prognostic implication of CFS grading in relation to existing frail markers.

Conclusions

The CFS is a useful predictor of short- and midterm mortality in patients undergoing TAVR. The CFS is a simple tool to assess patient’s frailty and can be used by nongeriatric specialists, thereby increasing its clinical utility. Assessing patients’ CFS grade before TAVR will positively affect their clinical care and improve postoperative outcomes.

Clinical Perspective

What Is New?

•

In this study, we evaluated the predictive utility of the Clinical Frailty Scale (CFS) in elderly patients who had undergone transcatheter aortic valve replacement and compared this semiquantitative tool against other well-known markers of frailty.

•

We found that cumulative 1-year mortality rates increased proportionally with increasing CFS stage, each of which represented a progressive increase in frailty.

What Are the Clinical Implications?

•

Before receiving transcatheter aortic valve replacement, all patients must undergo preassessment screening to evaluate the safety of the procedure in light of patients’ comorbidities and other clinical variables.

•

The CFS correlated with other markers of frailty, including body mass index, albumin, gait speed, and grip strength.

•

The CFS accurately reflects clinical frailty and is a useful predictor of mortality in elderly patients after transcatheter aortic valve replacement.

•

The CFS should form part of the preoperative assessment as a means to further improve patient outcomes after transcatheter aortic valve replacement.

Supplemental Material

File (025630_supplemental_material.pdf)

Download
231.25 KB

Appendix

OCEAN-TAVI Investigators

The OCEAN-TAVI investigators are as follows: Kentaro Hayashida, Shun Kohsaka, Fumiaki Yashima, Taku Inohara, Yuki Kakefuda, Takahide Arai, Ryo Yanagisawa, Makoto Tanaka, Takashi Kawakami, Yuichiro Maekawa, Kohno Takashi, Akihiro Yoshitake, Yasunori Iida, Masataka Yamazaki, Hideyuki Shimizu, Yoshitake Yamada, Masahiro Jinzaki, Hikaru Tsuruta, Yuji Itabashi, Mitsushige Murata, Michiyuki Kawakami, Shogo Fukui, Motoaki Sano, Keiichi Fukuda (Keio University School of Medicine, Tokyo, Japan); Masanori Yamamoto, Tetsuro Shimura, Ai Kagase, Soh Hosoba, Atsuko Kodama, Hirotomo Sato, Tomohiko Teramoto, Masashi Kimura, Mitsuru Sago, Tatsuya Tsunaki, Shoko Watarai, Masanao Tsuzuki, Keisuke Irokawa, Kazuki Shimizu, Etsuo Tsuchikane, Takahiko Suzuki (Toyohashi Heart Center, Toyohashi, Japan); Seiji Kano, Yutaka Koyama, Toshihiro Kobayashi, Yasuhide Okawa (Nagoya Heart Center, Nagoya, Japan); Norio Tada, Masaki Miyasaka, Yusuke Enta (Senedai Kousei Hospital, Sendai, Japan); Futoshi Yamanaka, Koki Shishido, Tomoki Ochiai, Tsuyoshi Yamabe, Kenichiro Noguchi, Shigeru Saito (Shonan Kamakura General Hospital, Kamakura, Japan); Toru Naganuma, Hiroyoshi Kawamoto, Hirokazu Onishi, Hiroto Yabushita, Satoru Mitomo, Sunao Nakamura (New Tokyo Hospital, Chiba, Japan); Motoharu Araki, Masahiro Yamawaki, Yui Akatsu, Yosuke Honda, Takuro Takama (Saiseikai Yokohama City Eastern Hospital, Yokohama, Japan); Shinichi Shirai, Akihito Isotani, Masaomi Hayashi, Norihiko Kamioka, Mizuki Miura, Takashi Morinaga, Tomohiro Kawaguchi, Mariko Yano, Michiya Hanyu, Yoshio Arai, Hideki Tsubota, Masafumi Kudo, Yuki Kuroda (Kokura Memorial Hospital, Kokura, Japan); Yusuke Watanabe, Akihisa Kataoka, Hitofumi Hioki, Yugo Nara, Hideyuki Kawashima, Fukuko Nagura, Makoto Nakashima, Kazuya Sasaki, Junichi Nishikawa, Tomoki Shimokawa, Tadanori Harada, Ken Kozuma (Teikyo University School of Medicine, Tokyo, Japan).

References

Leon MB, Smith CR, Mack M, Miller DC, Moses JW, Svensson LG, Tuzcu EM, Webb JG, Fontana GP, Makkar RR, Brown DL, Block PC, Guyton RA, Pichard AD, Bavaria JE, Herrmann HC, Douglas PS, Petersen JL, Akin JJ, Anderson WN, Wang D, Pocock S; PARTNER Trial Investigators. Transcatheter aortic-valve implantation for aortic stenosis in patients who cannot undergo surgery. N Engl J Med. 2010;363:1597–1607. doi: 10.1056/NEJMoa1008232.

Abstract

Background:

Methods:

Results:

Conclusions:

Introduction

Methods

Patient Population and Study Design

TAVR Procedures

Statistical Analysis

Results

Baseline Patient and Procedural Characteristics

Peri- and Postprocedural Patient Characteristics

30-Day and Cumulative 1-Year Mortality

Prognostic Value of Baseline CFS After TAVR Procedure

Discussion

Limitations

Conclusions

What Is New?

What Are the Clinical Implications?

Supplemental Material

Appendix

OCEAN-TAVI Investigators

References

eLetters(0)

Information

Published In

Copyright

Versions

History

Permissions

Keywords

Subjects

Authors

Affiliations

Notes

Disclosures

Sources of Funding

Metrics

Citations

View options

PDF and All Supplements

PDF/EPUB

Login options

Purchase Options

Restore your content access

Share

Share article link

Share