The frailty syndrome is associated with adverse health outcomes in very old patients with stable heart failure: A prospective study in six Spanish hospitals

International Journal of Cardiology 236 (2017) 296–303

Contents lists available at ScienceDirect

International Journal of Cardiology journal homepage: www.elsevier.com/locate/ijcard

The frailty syndrome is associated with adverse health outcomes in very old patients with stable heart failure: A prospective study in six Spanish hospitals☆ Carlos Rodríguez-Pascual a,b,⁎, Emilio Paredes-Galán c, Ana-Isabel Ferrero-Martínez a, Jose-Luis Gonzalez-Guerrero d, Mercedes Hornillos-Calvo e, Rocio Menendez-Colino f, Ivett Torres-Torres g, Arturo Vilches-Moraga a, Maria-Concepcion Galán h, Francisco Suarez-Garcia h, Maria-Teresa Olcoz-Chiva a,i, Fernando Rodríguez-Artalejo j a

Geriatric Medicine Department, Complejo Hospitalario Universitario de Vigo, Spain University of Lincoln, Lincoln County Hospital, Lincoln, Lincolnshire, United Kingdom Cardiology Department, Complejo Hospitalario Universitario de Vigo, Spain d Geriatric Medicine Department, Complejo Hospitalario de Cáceres, Spain e Geriatric Medicine Department, Hospital Universitario de Guadalajara, Departamento de Medicina, Universidad de Alcalá de Henares, Madrid, Spain f Geriatric Medicine Department, Hospital Universitario La Paz, Departamento de Medicina, Facultad de Medicina, Universidad Autónoma de Madrid, Spain g Geriatric Medicine Department, Complejo Hospitalario de Albacete, Spain h Geriatric Medicine Department, Complejo Hospitalario de Oviedo, Departamento de Medicina, Universidad de Oviedo, Spain i Department of Care of the Elderly,Lincoln County Hospital, Lincoln, United Kingdom j Department of Preventive Medicine and Public Health, Universidad Autónoma de Madrid/IdiPaz, CIBER of Epidemiology and Public Health (CIBERESP), Madrid, Spain b c

a r t i c l e

i n f o

Article history: Received 15 December 2016 Received in revised form 31 January 2017 Accepted 3 February 2017 Available online 8 February 2017 Keywords: Frailty Heart failure Mortality Readmission Functional decline Elderly

a b s t r a c t Background: Most studies on the association between the frailty syndrome and adverse health outcomes in patients with heart failure (HF) have used non-standard definitions of frailty. This study examined the association of frailty, diagnosed by well-accepted criteria, with mortality, readmission and functional decline in very old ambulatory patients with HF. Methods: Prospective study with 497 patients in six Spanish hospitals and followed up during one year. Mean (SD) age was 85.2 (7.3) years, and 79.3% had LVEF N 45%. Frailty was diagnosed as having ≥3 of the 5 Fried criteria. Readmission was defined as a new episode of hospitalisation lasting N24 h, and functional decline as an incident limitation in any activity of daily living at the 1-year visit. Statistical analyses were performed with Cox and logistic regression, as appropriate, and adjusted for the main prognostic factors at baseline. Results: At baseline, 57.5% of patients were frail. The adjusted hazard ratio (95% confidence interval) for mortality among frail versus non-frail patients was 1.93 (1.20–3.27). Mortality was higher among patients with low physical activity [1.64 (1.10–2.45)] or exhaustion [1.83 (1.21–2.77)]. Frailty was linked to increased risk of readmission [1.66 (1.17–2.36)] and functional decline [odds ratio 1.67 (1.01–2.79)]. Slow gait speed was related to functional decline [odds ratio 3.59 (1.75–7.34)]. A higher number of frailty criteria was associated with a higher risk of the three study outcomes (P trend b 0.01 in each outcome). Conclusions: Frailty was associated with increased risk of 1-year mortality, hospital readmission and functional decline among older ambulatory patients with HF. © 2017 Elsevier B.V. All rights reserved.

Abbreviations: ADL, Activities of Daily Living; CGA, Comprehensive Geriatric Assessment; CHS, Cardiovascular Health Study; IADL, Instrumental Activities of Daily Living; NYHA, New York Heart Association; MDRD, Modified Diet in Renal Disease formula; MEC, Mini-Mental State Examination; NTproBNP, N-terminal pro-brain natriuretic peptide; PASE, Physical Activity Scale for the Elderly. ☆ All authors take responsibility for all aspects of the reliability and freedom from bias of the data presented and their discussed interpretation. ⁎ Corresponding author at: University of Lincoln, Green Lane, Lincoln LN6 7DL, United Kingdom. E-mail addresses: [email protected] (C. Rodríguez-Pascual), [email protected] (E. Paredes-Galán), [email protected] (A.-I. Ferrero-Martínez), [email protected] (J.-L. Gonzalez-Guerrero), [email protected] (M. Hornillos-Calvo), [email protected] (R. Menendez-Colino), [email protected] (I. Torres-Torres), [email protected] (A. Vilches-Moraga), [email protected] (M.-C. Galán), [email protected] (F. Suarez-Garcia), [email protected] (M.-T. Olcoz-Chiva), [email protected] (F. Rodríguez-Artalejo).

http://dx.doi.org/10.1016/j.ijcard.2017.02.016 0167-5273/© 2017 Elsevier B.V. All rights reserved.

C. Rodríguez-Pascual et al. / International Journal of Cardiology 236 (2017) 296–303

1. Introduction Frailty is an age-associated medical syndrome characterised by increased vulnerability to even minor stressors, which manifests as higher risk of adverse health outcomes including disability, hospitalisation and death [1,2]. And heart failure (HF) is the most common cause of hospital admission in individuals aged 65 years or older and shows a high risk of mortality, disability and hospital readmission [3–6]. Of note is that frailty and HF are frequently associated [7]; indeed, it has been suggested that these two syndromes share common pathogenic mechanisms [8] and that some treatment modalities, such us physical exercise, benefit both of them [9–10]. Very old patients with HF show a high prevalence of frailty and disability [11–13], and their mortality depends on the degree of functional and cognitive impairment [14–16] ; moreover frailty has also been linked to greater functional decline and higher risk of hospital readmission and death in these patients [9,11,17–19]. However, knowledge of the prognostic relevance of frailty in HF is rather limited. In fact, some studies have included relatively young patients with a mean age of 66–68 years, who mostly had reduced leftventricle ejection fraction (LVEF) [12,20–21]. In other investigations, patients were recruited during a hospitalisation episode [11,14,19, 22–23], which may contribute to deconditioning and increase the frequency of frailty. Moreover, several studies only selected patients in good functional condition [11], or in specific settings like the community [24] and cardiology clinics [20–21]. And most importantly, a number of investigations used non-standard definitions of frailty, such as disability, functional decline or cognitive impairment [14,20–21,24], and other studies simply focused on indicators or individual components of the frailty phenotype [6,25–27]. Thus, the influence of frailty, based on widely-accepted diagnostic criteria, on the prognosis of very old patients with stable HF is still uncertain. Accordingly, we examined the association of the frailty syndrome with mortality, readmission and functional decline in very old ambulatory patients after a hospital discharge for HF. 2. Methods 2.1. Study design and participants This study was conducted with ambulatory patients discharged from six Spanish hospitals with a main diagnosis of HF from December 1, 2010 to November 30, 2012. Patients were previously admitted to the Geriatric Medicine or Cardiology departments, and referred to the Geriatric Medicine specialised outpatient clinic when they met the following inclusion criteria: a) age ≥ 75 years on recruitment; b) HF diagnosis according to Framingham [28] and European Society of Cardiology [29] criteria; c) having had a hospitalisation with the main discharge diagnosis of HF in the last twelve months; and d) substantial comorbidity, based on a Charlson index ≥ 3 [30]. HF criteria in each patient were assessed by cardiologists and geriatricians. Exclusion criteria were: a) terminal illness with a life expectancy b5 months according to the researcher's opinion; b) functional or cognitive impairment that limited the patient in attending follow-up visits or completing the study questionnaires; c) being on a waiting list for any invasive cardiac procedure; and d) difficulty of follow-up due to other reasons such as moving out of the hospital area. Informed written consent was given by study participants, and the study protocol was approved by the local institutional review board. 2.2. Study variables We collected standard sociodemographic and biomedical variables as well as data from a comprehensive geriatric assessment (CGA). Sociodemographic variables included age, gender, cohabitation, and educational level. Biomedical variables comprised cardiovascular risk

297

factors, comorbidities and the Charlson index, HF aetiology, NYHA functional class, LVEF measured with transthoracic echocardiography, laboratory measures including glomerular filtration rate as estimated with the Modified Diet in Renal Disease formula (MDRD) [31], weight and height, blood pressure, heart rate, and active drug treatment on recruitment. The CGA included: a) Cognitive function as per the Spanish adaptation of the Mini-Mental State Examination (MEC) [32]; b) Depression, with the 15-item Yesavage Geriatric Depression Scale [33]; c) Limitations in activities of daily living (ADL) based on the Katz index [34], and in instrumental activities of daily living (IADL) with the Lawton and Brody index [35]; d) Mobility limitation based on this scale: 1, no walking limitation; 2, use of a walking cane or stick; 3, use of a Zimmer frame or needing help from one person; and 4, wheelchair-bound or needing help from two people for ambulation; e) Health-related quality of life, assessed with the Minnesota Living with Heart Failure Questionnaire [36]; and f) Frailty, assessed with the 5 phenotypic criteria proposed by Fried et al. in the Cardiovascular Health Study (CHS) [2]: 1) muscle weakness, based on the CHS cutpoints of grip strength, measured with a Jamar type dynamometer on the dominant hand; three measurements were performed and the highest value was selected; 2) slowness, according to the CHS cutpoints of slow gait speed measured on a 15-feet distance without acceleration period, and patients unable to walk were considered to meet the slowness criterion; 3) low physical activity, as assessed with the Physical Activity Scale for the Elderly (PASE) questionnaire [37], the cut point was the lowest quintile, which corresponded to a score of 0, that is, less than half an hour walking outside the home on a daily basis; 4) exhaustion, based on a positive answer to any of two questions taken from the Centre for Epidemiologic Studies Depression Scale: “Have you felt that everything you did was a big effort?” and “Have you felt that you could not keep on doing things?” at least 3–4 days a week [38]; and 5) unintentional weight loss ≥4.5 kg or ≥5% of body weight in the last year. Individuals were classified as frail when they had ≥3 criteria, as pre-frail when having 1–2, and as robust when no criterion was present. For this analysis, robust and pre-frail patients were grouped as non-frail. The “timed get-up and go” test [39] was recorded in all patients with the ability to walk. In order to attenuate the influence of hospitalisation-related deconditioning on functional and frailty measures, the recruitment visit was done at least one month after hospital discharge. 2.3. Study outcomes Study participants were prospectively followed up during one year. Follow-up started on the day when the CGA was performed and ended on the 1-year visit, the date of death, or the date of last contact (in those lost to follow-up), whichever came first. Study outcomes were all-cause death, readmission, and incident functional limitation during follow-up. Readmission was defined as an episode of hospitalisation lasting N24 h, and the analysis was done considering the time to the first readmission. Incident functional limitation was any newly developed limitation in ADL assessed at the 1-year visit. Data on study outcomes were collected at the 1-year follow-up visit (where the CGA was performed again), through review of the electronic clinical chart, and from telephone interview with patient and relatives. 2.4. Statistical analysis From the 507 study participants we excluded 10 who lacked complete information on frailty (Supplementary Fig. 1). Thus, analyses were conducted with 497 patients, of whom 286 (57.5%) were frail at baseline. Descriptive analyses were performed using percentages for categorical variables and the mean ± standard deviation (SD) for continuous variables. Differences in sociodemographic, biomedical and CGA variables between frail and non-frail patients were assessed with a chi-square test for categorical variables, and the Student's t-test for continuous variables.

298

C. Rodríguez-Pascual et al. / International Journal of Cardiology 236 (2017) 296–303

Table 1 Baseline characteristics of patients discharged after heart failure, in the total study sample and according to frailty status.

Sociodemographic variables Age (years) Women Less than primary education Cohabitation Alone Spouse Family other than spouse Nursing-home Biomedical variables Medical history and comorbidities Charlson comorbidity index Dementia COPD Chronic renal failure Diabetes Anaemia Hypertension Stroke Peripheral vascular disease Atrial fibrillation on admission Previous admission due to HF Previous MI Previous HF diagnosis Aetiology of heart failure Hypertensive Ischemic Valvular Idiopathic myocardiopathy Symptoms NYHA functional classes III–IV Echocardiography Left ventricle ejection fraction ≤45% Severe aortic stenosis Treatment on discharge ACEI/ARB Beta-blockers Aldosterone inhibitors Laboratory measures in serum Haemoglobin (g/dl) Glycosylated haemoglobin (g/l) Creatinine (mg/dl) Urea (mg/dl) MDRD-estimated GFR (ml/min) Cholesterol (per each mg/dl) LDL-c (per each mg/dl) Albumin (serum) (per each g/dl) H-s CRP (per each mg/dl) NTpBNP on admission (ng/dl) NTpBNP at discharge (ng/dl) Outcomes One-year mortality One-year hospital readmission One-year incident ADL limitation

Total (n = 497)

Frail (n = 286)

Non-frail (n = 211)

P-value

85.2 ± 7.3 61% 61.7%

85.7 ± 5.1 67.5% 38.9%

84.4 ± 9.4 52.1% 37.4%

0.05 0.001 0.74

14.3% 27.7% 53.3% 5.7%

12.7% 21.1% 59.9% 6.3%

16.6% 36.5% 42.2% 4.7%

3.2 ± 1.8 7.4% 32.2% 25.2% 32% 28.4% 82.3% 14.5% 12.5% 61% 39.3% 13.9% 89.9%

3.2 ± 1.8 9.4% 30% 25.5% 31.8% 30.4% 84.6% 16.1% 13.6% 62.2% 38.4% 15.4% 89.5%

3.1 ± 1.8 4.7% 35.2% 24.6% 32.1% 25.6% 79% 12.4% 10.9% 59.4% 40.4% 11.9% 90.5%

0.51 0.05 0.24 0.83 0.92 0.26 0.12 0.24 0.41 0.56 0.70 0.27 0.76

47.3% 23.5% 26.8% 4.2%

54.2% 24.8% 29% 3.5%

50.7% 21.8% 23.7% 5.1%

0.46 0.45 0.22 0.37

27.9%

31.4%

23.2%

0.04

20.7% 10.9%

21.2% 12.7%

20% 8.5%

0.81 0.20

56.6% 35.9% 24.0%

55.7% 36.8% 29.0%

57.85% 34.6% 17.2%

0.64 0.63 b0.005

12.16 ± 2.06 6.5 ± 1.31 1.28 ± 0.62 71.4 ± 37.7 46.3 ± 16.8 157.6 ± 40.9 96.1 ± 35.7 6.73 ± 10.87 7.51 ± 12.77 7416 ± 8709 4355 ± 5899

12.0 ± 1.9 6.4 ± 1.2 1.29 ± 0.61 72.2 ± 37.6 45.1 ± 16.6 158.3 ± 40.8 96.1 ± 37.4 6.71 ± 11.21 9.04 ± 14.2 8089 ± 8654 5129 ± 6838

12.3 2.1 6.5 ± 1.4 1.27 ± 0.64 70.2 ± 38.0 47.8 ± 17 156.7 ± 41.0 96.1 ± 33.4 6.75 ± 10.40 5.46 ± 10.19 6532 ± 8737 3228 ± 3942

0.10 0.51 0.77 0.56 0.10 0.67 0.99 0.97 0.005 0.14 b0.01

20.1% 39.4% 36%

26.9% 44.6% 41.6%

10.5% 32.8% 20.5%

b0.001 b0.05 b0.05

b0.001

Data are mean ± standard deviation for continuous variables and percentage for qualitative variables. ACEI/ARB: Angiotensin converting enzyme inhibitors/angiotensin receptor blockers, ADL: activities of daily living, COPD: chronic obstructive pulmonary disease, GFR: glomerular filtration rate, HF: heart failure, MDRD: Modified Diet in Renal Disease formula, MI: myocardial infarction, NTpBNP: N-terminal fraction of brain natriuretic peptide. SD: standard deviation.

The association of baseline frailty and other variables with the time to all-cause death or readmission was summarised with hazard ratios (HR) and their 95% confidence interval (CI), obtained from Cox regression. Given that information on incident ADL limitation was available only at the 1-year visit, the association between frailty and the risk of ADL limitation was summarised with odds ratios (OR) and their 95% CI, estimated from logistic models. Frailty and other variables associated with the outcomes in the univariate analysis (p b 0.10) were selected for a multivariate analysis, where a backward stepwise procedure was used to identify those variable independently associated with the outcomes. We also ran a secondary analysis in which models were further adjusted for variables that, regardless of statistical significance in our sample, have been shown in the literature to be associated with both frailty and the study outcomes.

To identify the individual frailty criteria associated with death, readmission or functional decline, we replicated the analyses using each frailty criterion as the main independent variable; these analyses were adjusted for the variables related with each outcome in the multivariate analysis and additionally for the rest of the frailty criteria. Statistical significance was based on a 2-tailed P-value b 0.05. Statistical analyses were performed with IBM SPSS 21. 3. Results 3.1. Baseline characteristics of study participants Tables 1 and 2 show the baseline characteristics of participants according to frailty status. In comparison with non-frail patients, those

C. Rodríguez-Pascual et al. / International Journal of Cardiology 236 (2017) 296–303 Table 2 Baseline results of comprehensive geriatric assessment among patients discharged after heart failure, in the total study sample and according to frailty status. Variable

Total Frail Non-frail P(N = 497) (N = 286) (N = 211) value

ADL limitation IADL limitation MEC (points)

54.7% 80.6% 24.0 ± 6.75 41.6%

66.1% 39.2% b0.001 89.2% 68.9% b0.001 23.1 ± 6.6 24.9 ± 6.5 0.005 46.3%

35.1%

0.01

5.2 ± 3.4 39.6%

5.7 ± 3.5 44.6%

4.5 ± 3.1 33%

b0.001 0.01

15 ± 7.23

12.64 ± 5.89 0.40 ± 0.16 12.28 ± 20.4 35.3% 70% 28.3 ± 15.9

18.21 ± 7.63 0.63 ± 0.25 38.44 ± 32 11% 17.1% 18.3 ± 11.4

b0.001

– – – 56.6% 36% 7.4% 43.3 ± 21.1

5.7% 31.4% 62.9% – – – 36.3 ± 18.8

MEC b 24 Yesavage depression scale (points) Depression Frailty criteria Grip strength (kg) Gait speed (m/s)

0.50 ± 0.25 Physical activity (points in PASE 23.4 ± 29 questionnaire) Unintentional weight loss 25% Exhaustion 47.5% Timed get-up and go test 23.8 (seconds) ± 14.9 Number of frailty criteria None 2.4% One 13.3% Two 26.8% Three 32.5% Four 20.8% Five 4.2% MLWHFQ (points) 40.3 ± 20.4

b0.001 b0.001 b0.001 b0.001 b0.001

b0.001

b0.001

Data are mean ± standard deviation for continuous variables and percentage for qualitative variables. ADL: activities of daily living, IADL; instrumental activities of daily living, MEC: Mini-Mental State Examination. MLWHFQ: Minnesota Living With Heart Failure Questionnaire.

with frailty were older, were more frequently women, and were more often living with family members other than the spouse. They were also more likely to suffer from dementia and to be treated with aldosterone antagonist drugs, and showed worse NYHA functional class and higher NTproBNP (Table 1). Lastly, frail patients more frequently showed limitations in ADL and IADL, and had worse scores on cognitive, depression and quality of life assessments (Table 2).

299

weakness (HR 1.56, 95% CI 0.75–3.2). Since most frailty criteria showed mortality HRs above 1, we analysed mortality according to the number of frailty criteria. In this analysis, a higher number of criteria was linked to a higher mortality (P linear trend = 0.003); however, a statistically significant higher mortality was observed only in those with N 3 frailty criteria (Table 4). 3.2.2. Hospital readmission Readmission data were available in 419 (84.3%) patients. Of them, 165 (39.4%) had one or more readmissions during the 1-year followup. As shown above for mortality, on univariate analysis many variables were associated with increased risk of readmission. However, on multivariate Cox regression, only chronic renal failure, serum albumin and frailty showed a statistically significant association; the HR (95% CI) for frailty was 1.66 (1.17–2.36) (Supplementary Table 1). When additionally adjusted for literature-based relevant variables, the corresponding HR (95% CI) for hospital readmission associated with frailty was 1.65 (1.11–2.46) (Fig. 1). In analyses adjusted as in the multivariable model in Supplementary Table 1 and also for the rest of frailty criteria, a tendency to increased risk of readmission was found for each frailty criterion, though none of them reached statistical significance (Table 4). A higher number of frailty criteria was progressively associated with higher risk of readmission (P linear trend = 0.004), and a statistically significant higher risk was observed among those with N2 frailty criteria (Table 4). 3.2.3. Functional decline Among the 398 patients who survived during the 1-year follow-up, the CGA could be performed at the 1-year visit in 325 (81.6%). Among them, 117 (36%) suffered functional decline with incident limitation in at least one ADL. Supplementary Table 2 shows the many variables associated with functional decline on univariate analyses. On multivariate logistic models, only age, the MEC score and frailty remained associated with functional decline; for frailty, the OR (95% CI) of functional decline was 1.67 (1.01–2.79). After additional adjustment for literature-based relevant variables, the OR (95% CI) of functional decline associated with frailty was 1.55 (0.91–2.66). Table 4 shows that, after multivariate adjustment as in supplementary Table 2 and also for the rest of the frailty criteria, slow gait speed was significantly related with increased risk of functional decline (OR 3.59, 95% CI 1.75–7.34). Also, the risk of functional decline augmented with the number of frailty criteria (P linear trend = 0.005); the increased risk of functional decline achieved statistical significance in patients with N1 frailty criteria (Table 4).

3.2. Frailty and adverse health outcomes 4. Discussion 3.2.1. Mortality During a 1-year follow-up, 99 (20%) patients died. Univariate analysis identified many factors associated with increased mortality risk, but on multivariate analysis only a diagnosis of dementia, dependence in IADL, NYHA III–IV functional class, serum creatinine and frailty remained as independent predictors of all-cause death; specifically, the HR (95% CI) of mortality associated with frailty was 1.98 (1.20–3.27) (Table 3). After additional adjustment for literature-based clinically relevant variables, including age, sex, Charlson index, LVEF ≤ 45%, previous HF-related hospitalisation other than the index admission, angiotensin-converting enzyme inhibitors or angiotensin receptor blockers, and beta-blockers treatment, frailty still showed an association with increased mortality (HR 2.15, 95% CI 1.23–3.76) (Fig. 1). Table 4 shows the association between each frailty criterion at baseline and 1-year subsequent mortality. Analyses were adjusted as in the multivariable model in Table 3 and also for the rest of the frailty criteria. An increased risk of death was found in patients with low physical activity (HR 1.64, 95% CI 1.10–2.45) or exhaustion (HR 1.83, 95% CI 1.21–2.77); also, some tendency to higher mortality was observed in those with slow gait speed (HR 1.86, 95% CI 0.95–3.65) or muscle

In this multicentre study in Spain, the frailty syndrome was associated with an increased risk of 1-year mortality, hospital readmission and functional decline among very old ambulatory patients with HF. When frailty criteria were considered separately, low physical activity and exhaustion were linked to increased mortality, and slow gait speed to functional decline. Moreover, a higher number of frailty criteria at baseline was associated with a progressively higher risk of the three study outcomes. 4.1. Frailty and health outcomes in HF Previous studies have shown an increased risk of mortality in frail HF patients [9], but some of them used disability as a proxy for frailty [14, 20–21,24]; unfortunately they are not equivalent because frailty is considered a pre-disability state [40]. Our results show that frailty is an independent prognostic factor for mortality, readmission and functional decline after adjusting for disability, cognitive impairment, comorbidity and other relevant HF covariates. Other studies have used frailty measures other than the Fried phenotype. Volpato et al. showed that in elderly patients admitted with HF

300

C. Rodríguez-Pascual et al. / International Journal of Cardiology 236 (2017) 296–303

Table 3 Association between baseline characteristics and 1-year mortality among patients discharged after heart failure. Univariate analysis

Sociodemographic variables Male Age (years) Not living alone or with partner Less than primary education Biomedical variables Medical history and comorbidities Dementia COPD Chronic renal failure Diabetes Anaemia Hypertension Stroke Dyslipidaemia Peripheral arterial disease Charlson index (per point) Atrial fibrillation on presentation Previous admission due to HF Previous MI Previous diagnosis of HF Ischaemic heart disease Depression Neoplasia Echocardiography LVEF (per point) LVEF ≤ 45% Severe aortic stenosis Geriatric Assessment MEC (per point) Yesavage depression scale (per point) ADL limitation IADL limitation MLWHFQ (per point) Frailty Gait speed (m/s) Grip strength (kg) Physical activity (per point in PASE questionnaire) Weight loss Exhaustion Number of frailty criteria Timed get-up and go test (seconds) Symptoms NYHA III–IV Laboratory measures in serum Haemoglobin (g/dl) Glycosylated haemoglobin (g/l) Serum creatinine (mg/dl) Urea (mg/dl) MDRD-estimated GFR (ml/min) Total cholesterol (mg/dl) LDL-cholesterol (mg/dl) Albumin (g/dl) H-s CRP (mg/dl) NTpBNP on admission (per 5000 ng/dl)a NTpBNP at discharge (per 5000 ng/dl)a Drug treatment at discharge ACEI-ARB at discharge Beta blockers Aldosterone inhibitors

Multivariate analysis

Hazard ratio

95% confidence interval

P-value

1.00 0.97 1.30 0.83

0.67–1.50 0.93–1.01 0.86–1.97 0.54–1.27

0.99 0.21 0.20 0.40

2.46 1.00 1.56 0.81 1.50 1.16 1.25 0.81 1.07 1.12 1.21 1.53 1.62 0.89 1.07 0.90 1.13

1.39–4.33 0.65–1.53 1.02–2.37 0.52–1.26 0.99–2.26 0.67–2.01 0.74–2.11 1.52–1.25 0.60–1.93 0.02–1.24 0.78–1.87 1.02–2.29 0.99–2.64 0.48–1.68 0.64–1.79 0.51–1.55 0.58–2.17

b0.005 0.99 0.03 0.37 0.05 0.60 0.39 0.34 0.80 0.02 0.38 0.03 0.05 0.73 0.78 0.70 0.71

0.99 1.56 2.25

0.97–1.00 0.95–2.56 1.27–3.97

0.43 0.07 0.005

0.96 1.07 1.83 1.25 1.01 2.63 0.85 0.94 0.98 0.97 1.91 1.55 1.01

0.93–0.99 1.01–1.14 1.20–2.80 1.11–1.20 1.00–1.02 1.65–4.20 0.77–0.94 0.91–0.97 0.97–0.99 0.61–1.54 1.27–2.88 1.29/1.85 1.00–1.03

0.01 0.02 0.01 b0.001 0.01 b0.001 b0.005 b0.001 b0.005 0.91 b0.005 b0.001 b0.005

2.08

1.38–3.12

b0.001

0.93 0.87 1.37 1.00 0.98 0.99 0.99 0.99 1.01 1.29 1.42

0.84–1.02 0.66–1.15 1.05–1.80 1.00–1.01 0.96–0.99 0.99–1.00 0.98–1.00 0.97–1.01 1.00–1.03 1.17–1.43 1.23–1.63

0.14 0.34 0.02 b0.005 0.007 0.12 0.13 0.64 0.006 b0.001 b0.001

0.89 1.04 1.46

0.60–1.33 0.68–1.58 0.95–2.25

0.59 0.84 0.08

Hazard ratio

95% confidence interval

P-value

1.97

1.06–3.63

0.03

1.15

1.04–1.28

0.01

1.98

1.20–3.27

0.007

1.82

1.21–2.75

0.004

1.50

1.12–2.01

0.006

ACEI-ARB: Angiotensin converting enzyme inhibitors/angiotensin receptor blockers, ADL: activities of daily living, IADL: instrumental activities of daily living, COPD: chronic obstructive pulmonary disease, GFR: glomerular filtration rate, HF: heart failure, H-s CRP: high-sensitivity C-reactive protein, LDL-c: LDL-cholesterol, LVEF: left ventricle ejection fraction, MDRD: Modified Diet in Renal Disease formula, MEC: mini-mental state examination, MI: myocardial infarction, MLWHFQ: Minnesota Living with heart failure, NTpBNP: N-terminal fraction of brain natriuretic peptide, NYHA: New York Heart Association functional class, PASE: physical activity questionnaire in the elderly. a These variables were not included in the multivariate analysis because they were recorded in only 278 cases.

(among other medical conditions), low scores on the Short Portable Physical Battery (SPPB) at admission and at discharge were related with increased length of stay and higher risk of 12-month readmission and mortality [19,27]. They also showed that low SPPB scores were linked to higher 30-month mortality [22]. These studies were based on hospitalised patients, so deconditioning due to hospitalisation and

acute disease could affect the associations of interest. Our findings are in line with those from these studies, but were obtained in patients in a stable situation to attenuate the effect of hospitalisation on adverse outcomes. Only a few studies have used the Fried phenotype to assess the influence of frailty on health outcomes in HF patients. Dominguez-Rodriguez

C. Rodríguez-Pascual et al. / International Journal of Cardiology 236 (2017) 296–303

301

et al. reported that the frailty phenotype was associated with a higher risk of admission after cardiac resynchronisation therapy in advanced HF [41]. Very recently, in a single centre study, Vidan et al. found that frailty predicts 1-year mortality and readmission in elderly patients hospitalised for HF in good functional status (no dependence in ADL or cognitive impairment) [11]; however, by contrast with our results, frailty was not associated with incident ADL disability. Also of note is that in this study, the prevalence of frailty was very high (76%), which could partly result from recruiting hospitalised patients. Thus, our study is the first to show that frailty, measured with the Fried criteria, predicts several adverse outcomes among very old ambulatory patients with stable HF, who represent the majority of patients with HF. We did not find a statistically significant relationship between LVEF and any study outcome (Table 3 and Supplementary Tables 1 and 2). However, only 20.7% of patients showed LVEF ≤ 45%, so statistical power could be insufficient to detect a significant association; thus, we decided to adjust all outcome analyses for reduced LVEF (Table 4, and Fig. 1). 4.2. Individual components of frailty and health outcomes in HF Our findings are in line with those of Vidan et al., who reported that low physical activity was associated with increased 1-year mortality [11]. However, our study is the first to find that exhaustion predicts mortality. Physical exhaustion is an indicator of HF severity and cardiac cachexia, but given that our analyses were adjusted for disease-severity indicators like NYHA functional class, and that weight loss was not associated with increased mortality, the association between exhaustion and higher mortality does not seem to be entirely due to HF severity and cachexia. Although physical activity and exhaustion may be associated, our results were adjusted for the rest of the frailty criteria, so it is unlikely that the impact of low physical activity on health outcomes is explained by exhaustion and vice versa. In our study, slow gait speed showed a clear association with functional decline and a non-statistically significant tendency to increased risk of mortality and readmission. Slow gait speed and ADL limitations measured before HF diagnosis have shown synergistic impact on mortality among older patients with incident HF in the CHS [18]. Recently, Pulignano et al. also found that gait speed improves the mortality prediction by the Cardiac and Comorbid Conditions Heart Failure risk score (3C–HF) [26]. However, the methods in these studies differed from ours in that the CHS gait speed was measured before patients suffered from HF, and in the study by Pulignano et al. the patients were 10 years younger than ours and only 20% had preserved LVEF, possibly because they were selected from cardiology clinics. Lastly, in patients with similar mean age to ours and mostly with preserved LVEF recently discharged from hospital, Vidan et al. did not find an association between slowness and mortality [11]. Like Vidan et al. [11], we found a progressive increase in the risk of mortality with a higher number of frailty criteria, but we extend knowledge in this field by showing a similar continuous relationship with hospital readmission and functional decline. We also found that patients presenting only two frailty criteria were more likely to develop ADL limitation. This could be relevant because, in these patients, frailty could be prevented with appropriate interventions [42–44], which may lead to reduced risk of mortality and readmission.

Fig. 1. Adjusted 1-year survival curves for mortality (A) and first hospital readmission (B) in frail and non-frail patients discharged after heart failure. (*) Adjusted for age, sex, dementia, serum creatinine level, limitation in IADL, NYHA III–IV functional class, Charlson comorbidity index, LVEF ≤ 45%, previous admission due to HF, treatment with beta-blockers, and treatment with ACEI/ARB.

disability, might be partially driven by sarcopenia and related frailty criteria, such as muscle weakness. 4.4. Strengths and limitations of the study

4.3. Mechanism of the association between frailty and health outcomes This association likely results from alteration of the multiple physiological systems that characterise frailty, including increased chronic inflammation, oxidative stress, and altered hormone signalling or poor immunological response, among others [45]. Sarcopenia (loss of muscle mass and strength) plays a central pathogenic role in frailty [2,45]. Thus, future research should establish if the association of low physical activity and exhaustion with mortality, and of slow gait speed with ADL

The main strengths were the multicentre design, which increases generalizability, the use of standard criteria to diagnose frailty, and that the same staff in each centre collected frailty measures and performed CGA, which favours reliability. In addition, losses to follow-up were very small. Our study also had some limitations. Given that the frailty phenotype entails a dichotomous diagnosis, the severity of frailty could not be adequately graded; future research using the deficit accumulation approach to frailty [1] should establish if finer categories of frailty might improve

302

C. Rodríguez-Pascual et al. / International Journal of Cardiology 236 (2017) 296–303

Table 4 Association between individual frailty criteria at baseline and 1-year mortality, hospital readmission and functional decline among patients discharged after heart failure. Mortality

Frailty criteriad Muscle weakness Slow gait speed Low physical activity Exhaustion Unintentional weight loss Number of frailty criteria 0–1 2 3 4–5 P for linear trend

Readmission

Functional decline

N/events

Hazard ratioa (95% CI)

P-value

N/events

Hazard ratiob (95% CI)

P-value

N/events

Odds ratioc (95% CI)

P-value

433/91 351/75 167/44 233/61 123/25

1.56 (0.75–3.2) 1.86 (0.95–3.65) 1.64 (1.10–2.45) 1.83 (1.21–2.77) 1.00 (0.64–1.58)

0.22 0.07 0.01 b0.01 0.97

363/136 297/117 140/57 194/82 108/42

1.19 (0.67–2.13) 1.57 (0.98–2.52) 1.04 (0.73–1.50) 1.36 (0.96–1.92) 1.01 (0.68–1.48)

0.54 0.06 0.82 0.08 0.95

277/106 218/93 100/40 133/49 80/21

1.87 (0.80–4.38) 3.59 (1.75–7.34) 0.98 (0.56–1.70) 1.03 (0.61–1.72) 0.76 (0.41–1.41)

0.14 b0.001 0.94 0.91 0.39

78/6 133/17 161/34 124/42

Ref. 1.51 (0.59–3.85) 1.96 (0.81–4.77) 3.51 (1.46–8.42) 0.003

0.38 0.13 0.005

72/19 115/36 130/46 10,353

Ref. 1.51 (0.82–2.78) 1.80 (0.99–3.25) 2.69 (1.50–4.82) 0.004

0.18 0.05 0.001

64/10 100/40 104/44 57/23

Ref. 4.12 (1.70–9.96) 4.70 (1.95–11.27) 3.80 (1.44–9.98) 0.005

0.002 0.001 0.007

CI: Confidence interval. a Adjusted for diagnosis of dementia, IADL dependence, NYHA III–IV functional class and serum creatinine, according to multivariate model in Table 3. b Adjusted for chronic renal failure and serum albumin, according to multivariate model in Supplementary Table 1. c Adjusted for age, and MEC score, according to multivariate model in Supplementary Table 2. d Analyses for each separate frailty criterion were additionally adjusted for the number of the rest of frailty criteria (range 0–4).

prognostic assessment in HF. Another limitation was the use of internationally recognised cut-points for gait speed and strength, because these might require adaptation to the physical characteristics of each population. Also, we have been unable to explain why frail patients received aldosterone inhibitors more frequently than non-frail patients (17.2 vs 29.0%, p b 0.005). However, this finding was not due to a different frequency between frail and non-frail patients in reduced LVEF, NYHA advanced functional class, right-sided HF or tricuspid regurgitation, which are factors potentially linked to prescription of aldosterone inhibitors. Finally, another limitation was that many important variables were self-reported. Although in most cases we used validated questionnaires or instruments widely available in clinical practice, future research should be based on objective measures when possible. 4.5. Practical implications Frailty is associated with higher risk of mortality, readmission and functional decline in very old patients with HF. Since frailty identifies high risk patients, it might be used for patient stratification and treatment selection, although the benefit of this strategy should be assessed in clinical trials. Frailty, and especially prefrailty, among patients with HF is potentially reversible with appropriate interventions, which include physical exercise, reduction of polypharmacy, increased intake of protein and vitamin D, and educational programs involving patients, family and carers. Moreover, there is evidence that disease care managers working within HF management programs improve self-management skills and readiness to make changes in health behaviours, which may support the above interventions [47]. 5. Conclusion Frailty is a frequent finding in very old patients with stable HF, which substantially increases the risk of adverse outcomes including mortality, hospital readmission and functional decline. Because of this, treatment of HF must be optimised in frail patients. Moreover, future research should establish if well-accepted treatments of frailty (physical exercise, vitamin D, caloric and protein support, reduction of polypharmacy) [46] improve prognosis in these patients. Funding This work has been partially funded by the Instituto de Salud Carlos III, C/ Sinesio Delgado, 4, 28029 Madrid, Spain (grants PI08/1280, PI09/ 91064, and PI14/01044).

Conflict of interest The authors report no relationships that could be construed as a conflict of interest. Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.ijcard.2017.02.016. References [1] K. Rockwood, D. Hogan, C. Macknight, Conceptualisation and measurement of frailty in elderly people, Drugs Aging 17 (2000) 295–302. [2] L.P. Fried, C.M. Tangen, J. Walston, for the Cardiovascular Health Study Collaborative Research Group, Frailty in older adults: evidence for a phenotype, J. Gerontol. A Biol. Sci. Med. Sci. 56 (2001) M146–M156. [3] H. Bueno, J.S. Ross, Y. Wang, et al., Trends in length of stay and short-term outcomes among Medicare patients hospitalized for heart failure, 1993–2006, JAMA 303 (2010) 2141–2147. [4] K. Dharmarajan, A.F. Hsieh, Z. Lin, et al., Diagnoses and timing of 30-day readmissions after hospitalization for heart failure, acute myocardial infarction, or pneumonia, JAMA 309 (2013) 355–363. [5] E. Sánchez, M.T. Vidán, J.A. Serra, F. Fernández-Avilés, H. Bueno, Prevalence of geriatric syndromes and impact on clinical and functional outcomes in older patients with acute cardiac diseases, Heart 97 (2011) 1602–1606. [6] S.I. Chaudhry, G. McAvay, S. Chen, et al., Risk factors for hospital admission among older persons with newly diagnosed heart failure: findings from the Cardiovascular Health Study, J. Am. Coll. Cardiol. 61 (2013) 635–642. [7] H. Khan, A.P. Kalogeropoulos, V.V. Georgiopoulou, et al., Frailty and risk for heart failure in older adults: the health, aging, and body composition study, Am. Heart J. 166 (2013) 887–894. [8] D.S. Goldwater, S.P. Pinney, Frailty in advanced heart failure: a consequence of aging or a separate entity? Clin. Med. Insights Cardiol. 9 (Suppl. 2) (2015) 39–46. [9] M. Goldfarb, R. Sheppard, J. Afilalo, Prognostic and therapeutic implications of frailty in older adults with heart failure, Curr. Cardiol. Rep. 17 (2015) 92. [10] D.W. Schopfer, D.E. Forman, Cardiac rehabilitation in older adults, Can. J. Cardiol. 32 (2016) 1088–1096. [11] M.T. Vidán, V. Blaya-Novakova, E. Sánchez, J. Ortiz, J.A. Serra-Rexach, H. Bueno, Prevalence and prognostic impact of frailty and its components in non-dependent elderly patients with heart failure, Eur. J. Heart Fail. 18 (2016) 869–875. [12] S. Altimir, J. Lupon, B. Gonzalez, et al., Sex and age differences in fragility in a heart failure population, Eur. J. Heart Fail. 7 (2005) 798–802. [13] N.F. Woods, A.Z. LaCroix, S.L. Gray, et al., Women's Health Initiative. Frailty: emergence and consequences in women aged 65 and older in the Women's Health Initiative observational study, J. Am. Geriatr. Soc. 53 (2005) 1321–1330. [14] S.I. Chaudhry, Y. Wang, T.M. Gill, H.M. Krumholz, Geriatric conditions and subsequent mortality in older patients with heart failure, J. Am. Coll. Cardiol. 55 (2010) 309–316. [15] C. Rodríguez-Pascual, A. Vilches-Moraga, E. Paredes-Galán, A.I. Ferrero-Martinez, M. Torrente-Carballido, F. Rodríguez-Artalejo, Comprehensive geriatric assessment and hospital mortality among older adults with decompensated heart failure, Am. Heart J. 164 (2012) 756–762. [16] C. Rodriguez-Pascual, E. Paredes-Galan, A. Vilches-Moraga, A.I. Ferrero-Martinez, M. Torrente-Carballido, F. Rodriguez-Artalejo, Comprehensive geriatric assessment and 2-year mortality in elderly patients hospitalized for heart failure, Circ. Cardiovasc. Qual. Outcomes 7 (2014) 251–258.

C. Rodríguez-Pascual et al. / International Journal of Cardiology 236 (2017) 296–303 [17] P. Le Corvoisier, S. Bastuji-Garin, B. Renaud, et al., Functional status and comorbidities are associated with in-hospital mortality among older patients with acute decompensated heart failure: a multicentre prospective cohort study, Age Ageing 44 (2015) 225–231. [18] S.I. Chaudhry, G. McAvay, Y. Ning, H.G. Allore, A.B. Newman, T.M. Gill, Risk factors for onset of disability among older persons newly diagnosed with heart failure: the Cardiovascular Health Study, J. Card. Fail. 17 (2011) 764–770. [19] S. Volpato, M. Cavalieri, F. Sioulis, et al., Predictive value of the short physical performance battery following hospitalization in older patients, J. Gerontol. A Biol. Sci. Med. Sci. 66 (2011) 89–96. [20] P. Gastelurrutia, J. Lupón, S. Altimir, et al., Fragility is a key determinant of survival in heart failure patients, Int. J. Cardiol. 175 (2014) 62–66. [21] J. Lupon, B. Gonzalez, S. Santaeugenia, et al., Prognostic implication of frailty and depressive symptoms in an outpatient population with heart failure, Rev. Esp. Cardiol. 61 (2008) 835–842. [22] D. Chiarantini, S. Volpato, F. Sioulis, et al., Lower extremity performance measures predict long-term prognosis in older patients hospitalized for heart failure, J. Card. Fail. 16 (2010) 390–395. [23] R. Rozzini, T. Sabatini, G.B. Frisoni, M. Trabucchi, Frailty is a strong modulator of heart failure-associated mortality, Arch. Intern. Med. 163 (2003) 737. [24] F. Cacciatore, P. Abete, F. Mazzella, et al., Frailty predicts long-term mortality in elderly subjects with chronic heart failure, Eur. J. Clin. Investig. 35 (2005) 723–730. [25] A.X. Lo, J.P. Donnelly, G. McGwin Jr., V. Bittner, A. Ahmed, C.J. Brown, Impact of gait speed and instrumental activities of daily living on all-cause mortality in adults N65 years with heart failure, Am. J. Cardiol. 115 (2015) 797–801. [26] G. Pulignano, D. Del Sindaco, A. Di Lenarda, et al., IMAGE-HF study investigators. Incremental value of gait speed in predicting prognosis of older adults with heart failure, J. Am. Coll. Cardiol. Heart Fail. 4 (2016) 289–298. [27] S. Volpato, M. Cavalieri, G. Guerra, et al., Performance-based functional assessment in older hospitalized patients: feasibility and clinical correlates, J. Gerontol. A Biol. Sci. Med. Sci. 63 (2008) 1393–1398. [28] P.A. McKee, W.P. Castelli, P.M. McNamara, W.B. Kannel, The natural history of congestive heart failure: the Framingham study, N. Engl. J. Med. 285 (1971) 1441–1446. [29] K. Dickstein, A. Cohen-Solal, G. Filippatos, et al., ESC Committee for Practice Guidelines (CPG). ESC guidelines for the diagnosis and treatment of acute and chronic heart failure 2008: the Task Force for the Diagnosis and Treatment of Acute and Chronic Heart Failure 2008 of the European Society of Cardiology. Developed in collaboration with the Heart Failure Association of the ESC (HFA) and endorsed by the European Society of Intensive Care Medicine (ESICM), Eur. Heart J. 29 (2008) 2388–2442. [30] M.E. Charlson, P. Pompei, K. Ales, C.R. MacKenzie, A new method of classifiying prognostic comorbidity in longitudinal studies: development and validation, J. Chronic Dis. 40 (1987) 375–383. [31] A.S. Levey, J. Coresh, T. Greene, et al., Chronic kidney disease epidemiology collaboration. Using standardized serum creatinine values in the modification of diet in renal disease study equation for estimating glomerular filtration rate, Ann. Intern. Med. 145 (2006) 247–254.

303

[32] A. Lobo, P. Saz, G. Marcos, J.L. Día, et al., Revalidation and standardization of the cognition mini-exam (first Spanish version of the Mini-Mental Status Examination) in the general geriatric population, Med. Clin. 112 (1999) 767–774. [33] A. Shah, R. Herbert, S. Lewis, R. Mahendran, J. Platt, B. Bhattacharyya, Screening for depression among geriatric inpatients with short versions of the geriatric depression scale, Age Ageing 26 (1997) 217–221. [34] S. Katz, T.D. Downs, H.R. Cash, R.C. Grotz, Progress in development of the index of ADL, Gerontologist 10 (1970) 20–30. [35] M.P. Lawton, E.M. Brody, Assessment of older people: self-maintaining and instrumental activities of daily living, Gerontologist 9 (1969) 179–186. [36] T.S. Rector, S.H. Kubo, J.N. Cohn, Patients' self-assessment of their congestive heart failure, part 2: content, reliability and validity of a new measure, the Minnesota living with Heart Failure Questionnaire, Heart Fail. 3 (1987) 198–209. [37] R.A. Washburn, E. McAuley, J. Katula, S.L. Mihalko, R.A. Boileau, The Physical Activity Scale for the Elderly (PASE): evidence for validity, J. Clin. Epidemiol. 52 (1999) 643–651. [38] M.V. Zunzunegui-Pastor, M. Delgado, E. Pérez, A.I. Yagüe-Alonso, M.L. Illescas, V. Galat-León, Validación de la Escala CES-D para la medida de la sintomatología depresiva en una población de personas mayores española, Rev. Multidiscip. Gerontol. 8 (1998) 156–161. [39] D. Podsialo, S. Richardson, The timed “Up & Go”: a test of basic functional mobility for frail elderly persons, J. Am. Geriatr. Soc. 39 (1991) 142–148. [40] L.P. Fried, L. Ferrucci, J. Darer, J.D. Williamson, G. Anderson, Untangling the concepts of disability, frailty, and comorbidity: implications for improved targeting and care, J. Gerontol. 59 (2004) 255–263. [41] A. Dominguez-Rodriguez, P. Abreu-Gonzalez, A. Jimenez-Sosa, et al., The impact of frailty in older patients with non-ischaemic cardiomyopathy after implantation of cardiac resynchronization therapy defibrillator, Europace 17 (2015) 598–602. [42] G. Abellan van Kan, Y. Rolland, H. Bergman, J.E. Morley, S.B. Kritchevsky, B. Vellas, The I.A.N.A. Task Force on frailty assessment of older people in clinical practice, J. Nutr. Health Aging 12 (2008) 29–37. [43] L. Bibas, M. Levi, M. Bendayan, L. Mullie, D.E. Forman, J. Afilalo, Therapeutic interventions for frail elderly patients: part I. Published randomized trials, Prog. Cardiovasc. Dis. 57 (2014) 134–143. [44] M. Bendayan, L. Bibas, M. Levi, L. Mullie, D.E. Forman, J. Afilalo, Therapeutic interventions for frail elderly patients: part II. Ongoing and unpublished randomized trials, Prog. Cardiovasc. Dis. 57 (2014) 144–151. [45] A. Clegg, J. Young, S. Iliffe, M.O. Rikkert, K. Rockwood, Frailty in elderly people, Lancet 381 (2013) 752–762. [46] J.E. Morley, B. Vellas, G.A. Van Kan, et al., Frailty consensus: a call to action, J. Am. Med. Dir. Assoc. 14 (2013) 392–397. [47] M.M. Ciccone, A. Aquilino, F. Cortese, et al., Feasibility and effectiveness of a disease and care management model in the primary health care system for patients with heart failure and diabetes (Project Leonardo), Vasc. Health Risk Manag. 6 (2010) 297–305.