- Email: [email protected]
Frailty Assessment in Advanced Heart Failure
ARTICLE IN PRESS Journal of Cardiac Failure Vol. ■■ No. ■■ 2016
Brief Report
Frailty Assessment in Advanced Heart Failure SHIVANK A. MADAN, MD, MHA,1 NADIA FIDA, MD,2 POULAMI BARMAN, MS,3 DANIEL SIMS, MD,1 JOOYOUNG SHIN, MD,1 JOE VERGHESE, MD,4 ILEANA PIÑA, MD, MPH,1 ULRICH JORDE, MD,1 AND SNEHAL R. PATEL, MD1 Bronx, New York; Houston, Texas; and Rochester, Minnesota
ABSTRACT Background: Several studies have recently demonstrated the value of frailty assessment in a general heart failure (HF) population; however, it is unknown whether these findings are also applicable in advanced HF. We investigated the utility of frailty assessment and its prognostic value in elderly patients with advanced HF. Methods: Forty consecutive elderly subjects aged ≥65 years, with left ventricular ejection fraction ≤35%, New York Heart Association class III or IV, and a 6-minute walk test <300 m were enrolled from the HF clinic at Montefiore Medical Center between October 2012 and July 2013. Subjects were assessed for frailty with the Fried Frailty Index, consisting of 5 components: hand grip strength, 15-foot walk time, weight loss, physical activity, and exhaustion. All subjects were prospectively followed for death or hospitalization. Results: At baseline, the mean age of the cohort was 74.9 ± 6.5 years, 58% female, left ventricular ejection fraction 25.6 ± 6.4%, 6-minute walk test 195.8 ± 74.3 m and length of follow-up 454 ± 186 days. Thirtyfive percent were prefrail and 65% were frail. Frailty status was associated with the combined primary endpoint of mortality and all-cause hospitalization (hazard ratio [HR] 1.93, 95% confidence interval [CI] 1.15– 3.25, P = .013). On individual analysis, frailty was associated with all-cause hospitalizations (HR 1.92, 95% CI 1.12–3.27, P = .017) and non-HF hospitalizations (HR 3.31, 95% CI 1.14- 9.6, P = .028), but was not associated with HF hospitalizations alone (HR 1.31, 95% CI 0.68–2.49, P = .380). Conclusions: Frailty assessment in patients with advanced HF is feasible and provides prognostic value. These findings warrant validation in a larger cohort. (J Cardiac Fail 2016;■■:■■–■■) Key Words: Frailty, heart failure, elderly, advanced heart failure, hospitalization, mortality.
in the general heart failure (HF) population have demonstrated that frailty is associated with increased health care utilization, hospitalizations, mortality,3–5 and incident HF.6 When HF reaches more advanced states, it manifests similar to frailty, as a biological syndrome where the primary insult is cardiac dysfunction but leads to systemic consequences. Because of this significant overlap, the utility of frailty assessment in advanced HF is unclear. The aim of the current study was to investigate the utility of frailty assessment and its prognostic value in elderly patients with advanced HF.
Frailty is a biological syndrome defined as a decreased homeostatic reserve leading to an increased vulnerability to stressors and adverse outcomes.1 Frailty manifests clinically as a disproportionate change in health status in response to a physical or psychological stress.2 Several recent studies From the 1Division of Cardiology, Department of Medicine, Montefiore Medical Center, Albert Einstein College of Medicine, Bronx, New York; 2Division of Cardiology, Department of Medicine, Houston Methodist Hospital, Houston, Texas; 3Department of Health Sciences Research, Division of Biomedical Statistics and Informatics Mayo Clinic, Rochester, Minnesota and 4 Division of Geriatrics, Department of Medicine, Montefiore Medical Center, Albert Einstein College of Medicine, Bronx, New York. Reprint requests: Snehal R. Patel, MD, Division of Cardiology, Heart Failure, Cardiac Transplantation and Mechanical Circulatory Support, 3400 Bainbridge Avenue, Medical Arts Pavilion- 7th floor, Bronx, New York 10467. Tel: +1 718 920 2248; Fax: +1 718 652 1833. E-mail: SNEPATEL@ montefiore.org. Manuscript received November 20, 2015; revised manuscript received February 3, 2016; revised manuscript accepted February 5, 2016. See page ■■ for disclosure information. 1071-9164/$ - see front matter © 2016 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.cardfail.2016.02.003
Methods In this single-center pilot study, consecutive patients from Montefiore Medical Center HF Clinic, between October 2012 and July 2013, aged ≥65 years, New York Heart Association (NYHA) class III or IV, and left ventricular ejection fraction (LVEF) ≤ 35% (measured by echocardiogram within 30 days of the study visit) were screened to participate. Key 1
ARTICLE IN PRESS 2 Journal of Cardiac Failure Vol. ■■ No. ■■ ■■ 2016 and combined for a composite score of 0 to 5. Subjects with a score of 0 were classified as “not frail,” 1–2 as “prefrail,” and 3 or higher as “frail.” Assessment of the individual domains is described in the supplementary Table S1. All subjects were followed from the time of testing through August 2014 for the combined primary endpoint of allcause hospitalizations or death. Secondary endpoints included all-cause mortality, non-HF hospitalizations, and HFrelated hospitalizations only. Outcomes were ascertained through review of medical records and confirmed by telephone follow-up. Statistical Analysis Baseline characteristics were described as frequencies for categorical and as mean ± standard deviation for continuous variables, and compared using chi-squared test and unpaired t test, respectively. Cox proportional hazards ratio (HR) modified for Andersen-Gill modeling was calculated to test the effect of frailty on hospitalizations and mortality, and adjusted for covariates. Unlike the traditional Cox model, which accounts for only the first hospitalization, the AndersenGill model takes into account multiple hospitalizations8,9 and treats each hospitalization for each subject as a separate observation. Statistical analysis was performed using R statistical software and 2-tailed P values of <.05 were considered significant. Results
Fig. 1. (A) Flowchart describing the patient selection process towards achieving the final study cohort of 40 patients. (B) Spectrum of frailty: percentage of patients with different frailty scores (range 0–5). (C) Percentage of patients meeting frailty criteria for each of the individual components of the Fried Frailty Index. LVEF, left ventricular ejection fraction; 6MWT = 6 minute walk test.
exclusion criteria included duration of HF < 6 months, acute decompensation within the previous 30 days, or inability to walk. Patients who met these inclusion/exclusion criteria underwent a 6-minute walk test (6MWT) and were enrolled in the final study cohort only if they walked <300 m (Fig. 1a). Hence, for the purposes of this study, advanced HF was defined as “NYHA class III or IV, LVEF ≤35%, and a 6MWT of <300 m.” The cutoff of <300m on 6MWT was chosen because it has been shown to define an increased risk of adverse events in HF.7 The study was approved by the Albert Einstein College of Medicine Institutional Review Board. Frailty was assessed using a modified version of the Fried Frailty Index as defined in the Cardiovascular Health Study.2 Five domains were assessed for frailty: weight loss, exhaustion, weakness, slow gait, and reduced physical activity. Each domain was scored 0 or 1 based on its absence or presence
Seventy-three subjects met the initial criteria and underwent 6MWT. Of these, 40 walked <300 m and formed the final study cohort (Fig. 1a). Overall characteristics included, mean age of 74.9 ± 6.5 years, 58% female, 37.5% NYHA class IV, LVEF 25.6 ± 6.4%, 6MWT 195.8 ± 74.3 m, and Charlson Comorbidity Index (CCI) score of 4.9 ± 1.9. There was a spectrum of frailty scores, ranging from 0 to 5 (Fig. 1b). The proportion of subjects meeting frailty criteria for each of the individual components is shown in Fig. 1c. When graded according to the prespecified Fried criteria: 0 subjects were not frail, 14 (35%) were prefrail, and 26 (65%) were frail. The baseline demographics of the overall cohort and as analyzed by the prefrail and frail groups are shown in Table 1. There was a higher prevalence of diabetes in the prefrail group; otherwise, there were no significant differences in the baseline characteristics. Both groups were equally well medicated and the length of follow up was similar. Frailty and Outcomes During follow-up, 10 patients died and 26 were hospitalized for any cause, including 20 for an exacerbation of HF. Including repeated events, there were a total of 69 all-cause hospitalizations, 45 (65%) of which were due to HF exacerbations. Pneumonia, infections and gastrointestinal bleeding were the major causes of non-HF hospitalizations. Of the 10
ARTICLE IN PRESS Frailty Assessment in Advanced Heart Failure
•
Madan et al
3
Table 1. Baseline Demographics of the Overall Cohort and as analyzed by the Prefrail and Frail Groups
Age (y) Sex (F) Race: Whites Blacks Hispanics Asians/others LVEF (%) NYHA class (IV) LVEDD (cm) HTN DM Atrial fibrillation PAD GFR HF etiology (ICM) HF duration (mo) COPD Depression Medications Beta-blockers ACEI/ARBs Aldosterone antagonists Hydralazine/nitrates 6MWT (m) Follow-up (d) Charlson Comorbidity Index
Total (n = 40)
Pre-frail (n = 14)
Frail (n = 26)
P Value: Pre-frail vs Frail
74.9 ± 6.5 23 (57.5%) 4 (10%) 16 (40%) 17 (42.5%) 3 (7.5%) 25.6 ± 6.4% 15 (37.5%) 5.96 ± 0.87 35 (87.5%) 22 (55%) 17 (42.5%) 4 (10%) 46.9 ± 22.4 19 (47.5%) 88.1 ± 79.8 9 (22.5%) 8 (20%)
76.0 ± 5.3 6 (42.8%) 2 (14%) 3 (21.4%) 8 (57.1%) 1 (7.1%) 26 ± 5.5 5 (35.7%) 5.82 ± 0.76 13 (92.8%) 11 (78.5%) 4 (28.6%) 2 (14.3%) 47.7 ± 16.9 8 (57.1%) 106.9 ± 87.4 5 (35.7%) 1 (7.1%)
74.4 ± 7.2 17 (65.4%) 2 (7.7%) 13 (50.0%) 9 (34.6%) 2 (7.7%) 25.3 ± 6.9 10 (38.4%) 6.01 ± 0.93 22 (84.6%) 11 (42.3%) 13 (50%) 2 (7.7%) 46.5 ± 25.2 11 (42.3%) 75.3 ± 74.1 4 (15.4%) 7 (26.9%)
.475* .169† .342†
.750* .738† .600* .452† .028† .191† .507† .873* .37† .320* .142† .136†
39 (97.5%) 32 (80%) 20 (50%) 18 (45%) 195.8 ± 74.3 454 ± 186 4.9 ± 1.9
13 (92.8%) 10 (71.4%) 8 (57.1%) 6 (42.9%) 192.4 ± 68.06 469.8 ± 179.2 4.2 ± 0.7
26 (100%) 22 (84.6%) 12 (46.2%) 12 (46.2%) 197.7 ± 78.71 435.7 ± 190.8 5.2 ± 2.3
.168† .32† .507† .842† .833* .585* .128*
ACEI, angiotensin-converting enzyme inhibitor; ARBs, angiotensin II receptor blockers; COPD, chronic obstructive pulmonary disease; DM, diabetes mellitus; F, female; GFR, glomerular filtration rate; HF, heart failure; HTN, hypertension; ICM, ischemic cardiomyopathy; LVEDD, left ventricle end diastolic diameter; LVEF, left ventricular ejection fraction; 6MWT, 6-minute walking test; NYHA, New York Heart Association; PAD, peripheral arterial disease. Boldface type indicates p value of <.05. *Unpaired t test. † Chi square test.
patients that died, 6 were attributable to cardiovascular causes, with no significant difference in the prefrail and frail groups (2 vs 4, P = 1.0). Compared with the prefrail group, frail subjects were at an approximately 2-fold increased risk for the primary endpoint of all-cause hospitalization or death (cumulative HR 1.93, 95% CI 1.15–3.25, P = .013, Fig. 2a). This association remained significant after adjusting for diabetes (cumulative HR 2.02, 95% CI 1.17–3.49, P = .011), race (Black vs others) (cumulative HR 2.23, 95% CI 1.29–3.84, P = .004) and also after adjusting for diabetes, age, sex, and CCI (cumulative HR 1.95, 95% CI 1.06–3.59, P = .031). There were more allcause hospitalizations in the frail vs prefrail group: median (interquartile range) 2 (0.75–2) vs 0 (0–2.25) with a cumulative HR of 1.92 (95% CI 1.12–3.27, P = .017, Fig. 2b). Although there was a higher proportion of deaths in the frail group (n = 8 [31%] vs n = 2 [14%]), this was not statistically significant (cumulative HR 2.18, 95% CI 0.46–10.27, P = .324). In subgroup analysis, frailty status was associated with increased risk of non-HF–related hospitalizations: median (interquartile range): 0.5 (0–1) vs 0 (0-0), cumulative HR 3.31 (95% CI 1.14–9.64, P = .028, Fig. 2d), but not HF-related hospitalizations: median (interquartile range) 1 (0– 2) vs 0 (0–2.25), cumulative HR 1.31, 95% CI 0.68–2.49, P = .382 (Fig. 2c). Finally, the 6MWT was not predictive of either the primary or any of the secondary end points in our cohort, and there was no correlation between the 6MWT and frailty (R2 = 0.005, P = .675).
Discussion The prevalence and prognostic value of frailty in a general HF population is well documented.3–5 In the current study, we evaluated the role of frailty in advanced HF patients. We found that frailty testing via the Fried Frailty Index was able to describe a spectrum in which 35% of subjects were prefrail and 65% were frail. Moreover, the frail group was associated with an approximately 2-fold increased risk of death or hospitalization when compared with the prefrail group. To our knowledge, this is the first report describing the utility of frailty assessment in advanced HF. Our elderly HF cohort was representative of the real world with significant comorbidities: average age was 75 years, LVEF 26%, extremely poor endurance with 6MWT of 196 m, and a CCI of almost 5. Not surprisingly, the event rate was high in the overall cohort, with 73% either dying or hospitalized during an average follow-up of 1.25 years. Nonetheless, frailty testing identified those at highest risk of death or hospitalization. The predictive value was maintained for the individual endpoints of all cause hospitalization and trended toward significance for mortality. Importantly, frailty status was predictive of non-HF–related hospitalizations, but was not associated with HF hospitalizations. This last finding is critical and may support the assumption that frailty assesses a different metric than other traditional HF tools. The prognostic value of the 6MWT in HF is well established10 and has been linked to frailty in prior analysis
ARTICLE IN PRESS 4 Journal of Cardiac Failure Vol. ■■ No. ■■ ■■ 2016
Fig. 2. Kaplan-Meier survival analysis modified for Andersen Gill modeling. Probability of freedom from (A) all-cause hospitalization and death, (B) all-cause hospitalization, (C) heart failure hospitalizations only, and (D) non-heart failure hospitalizations in prefrail vs frail subjects. Corresponding P values are shown in the graphs.
of less sick HF patients.11 In contrast, we did not find a significant correlation between 6MWT and frailty. This is likely because our cohort was predefined based on 6MWT (<300 m), limiting the variability of this measurement in the study population. Regardless of this discrepancy, our findings highlight that frailty assessment can have added prognostic value in HF patients with a very poor 6MWT. There are several limitations to the current study. First, this is a single-center pilot study with a small sample size and the findings need validation in a larger cohort. Second, we chose to use the most accepted frailty tool—the Fried Frailty Index; however, several other frailty assessment tools have been validated, which were not incorporated. Although the original Fried Frailty Index used a weighted score of kilocalories per week for assessment of low physical activity, we used a questionnaire to measure physical activity levels. This modified version has been validated in previous studies.12 Third, we did not compare the predictive value of frailty to other prognostication tools such as cardiopulmonary stress testing or the Seattle HF Model.13 Finally, because of the small sample size, we did not include many covariates that could affect hospitalization rates (like NT pro-BNP levels, chronic obstructive pulmonary disease, depression) into our multivariate model; however, we did include the CCI as a measure of the global burden of comorbidity14 in the multivariate analysis. In conclusion, our data demonstrate that frailty testing in elderly patients with advanced HF is feasible and provides
clinical prognostic value. These are intriguing findings that require validation in a larger cohort. Disclosures There are no financial disclosures or conflict of interest for any of the authors in regards to this publication. Acknowledgments This research was funded by Division of Cardiology, Montefiore Medical Center, Albert Einstein College of Medicine, Bronx, NY. Appendix: Supplementary material Supplementary data to this article can be found online at doi:10.1016/j.cardfail.2016.02.003. References 1. Fried LP, Ferrucci L, Darer J, Williamson JD, Anderson G. Untangling the concepts of disability, frailty, and comorbidity: implications for improved targeting and care. J Gerontol A Biol Sci Med Sci 2004;59:M255–63.
ARTICLE IN PRESS Frailty Assessment in Advanced Heart Failure 2. Fried LP, Tangen CM, Walston J, Newman AB, Hirsch C, Gottdiener J, et al. Frailty in older adults evidence for a phenotype. J Gerontol A Biol Sci Med Sci 2001;56:M146–57. 3. Cacciatore F, Abete P, Mazzella F, Viati L, Della Morte D, D’Ambrosio D, et al. Frailty predicts long-term mortality in elderly subjects with chronic heart failure. Eur J Clin Invest 2005;35:723–30. 4. Chaudhry SI, McAvay G, Chen S, Whitson H, Newman AB, Krumholz HM, 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 2013;61:635–42. 5. McNallan SM, Singh M, Chamberlain AM, Kane RL, Dunlay SM, Redfield MM, et al. Frailty and healthcare utilization among patients with heart failure in the community. JACC Heart Fail 2013;1:135–41. 6. Khan H, Kalogeropoulos AP, Georgiopoulou VV, Newman AB, Harris TB, Rodondi N, et al. Frailty and risk for heart failure in older adults: the health, aging, and body composition study. Am Heart J 2013;166:887–94. 7. Rostagno C, Olivo G, Comeglio M, Boddi V, Banchelli M, Galanti G, et al. Prognostic value of 6-minute walk corridor test in patients with mild to moderate heart failure: comparison with other methods of functional evaluation. Eur J Heart Fail 2003;5:247–52.
•
Madan et al
5
8. Guo Z, Gill TM, Allore HG. Modeling repeated time-to-event health conditions with discontinuous risk intervals: an example of a longitudinal study of functional disability among older persons. Methods Inf Med 2008;47:107. 9. Andersen PK, Gill RD. Cox’s regression model for counting processes: a large sample study. Ann Stat 1982;1100–20. 10. Cahalin LP, Mathier MA, Semigran MJ, Dec GW, DiSalvo TG. The six-minute walk test predicts peak oxygen uptake and survival in patients with advanced heart failure. CHEST J 1996;110:325–32. 11. Boxer R, Kleppinger A, Ahmad A, Annis K, Hager D, Kenny A. The 6-minute walk is associated with frailty and predicts mortality in older adults with heart failure. Congest Heart Fail 2010;16:208–13. 12. Verghese J, Holtzer R, Lipton RB, Wang C. Mobility stress test approach to predicting frailty, disability, and mortality in high-functioning older adults. J Am Geriatr Soc 2012;60:1901–5. 13. Levy WC, Mozaffarian D, Linker DT, Sutradhar SC, Anker SD, Cropp AB, et al. The seattle heart failure model prediction of survival in heart failure. Circulation 2006;113:1424–33. 14. Henkel DM, Redfield MM, Weston SA, Gerber Y, Roger VL. Death in heart failure a community perspective. Circ Heart Fail 2008;1: 91–7.
Brief Report
Frailty Assessment in Advanced Heart Failure SHIVANK A. MADAN, MD, MHA,1 NADIA FIDA, MD,2 POULAMI BARMAN, MS,3 DANIEL SIMS, MD,1 JOOYOUNG SHIN, MD,1 JOE VERGHESE, MD,4 ILEANA PIÑA, MD, MPH,1 ULRICH JORDE, MD,1 AND SNEHAL R. PATEL, MD1 Bronx, New York; Houston, Texas; and Rochester, Minnesota
ABSTRACT Background: Several studies have recently demonstrated the value of frailty assessment in a general heart failure (HF) population; however, it is unknown whether these findings are also applicable in advanced HF. We investigated the utility of frailty assessment and its prognostic value in elderly patients with advanced HF. Methods: Forty consecutive elderly subjects aged ≥65 years, with left ventricular ejection fraction ≤35%, New York Heart Association class III or IV, and a 6-minute walk test <300 m were enrolled from the HF clinic at Montefiore Medical Center between October 2012 and July 2013. Subjects were assessed for frailty with the Fried Frailty Index, consisting of 5 components: hand grip strength, 15-foot walk time, weight loss, physical activity, and exhaustion. All subjects were prospectively followed for death or hospitalization. Results: At baseline, the mean age of the cohort was 74.9 ± 6.5 years, 58% female, left ventricular ejection fraction 25.6 ± 6.4%, 6-minute walk test 195.8 ± 74.3 m and length of follow-up 454 ± 186 days. Thirtyfive percent were prefrail and 65% were frail. Frailty status was associated with the combined primary endpoint of mortality and all-cause hospitalization (hazard ratio [HR] 1.93, 95% confidence interval [CI] 1.15– 3.25, P = .013). On individual analysis, frailty was associated with all-cause hospitalizations (HR 1.92, 95% CI 1.12–3.27, P = .017) and non-HF hospitalizations (HR 3.31, 95% CI 1.14- 9.6, P = .028), but was not associated with HF hospitalizations alone (HR 1.31, 95% CI 0.68–2.49, P = .380). Conclusions: Frailty assessment in patients with advanced HF is feasible and provides prognostic value. These findings warrant validation in a larger cohort. (J Cardiac Fail 2016;■■:■■–■■) Key Words: Frailty, heart failure, elderly, advanced heart failure, hospitalization, mortality.
in the general heart failure (HF) population have demonstrated that frailty is associated with increased health care utilization, hospitalizations, mortality,3–5 and incident HF.6 When HF reaches more advanced states, it manifests similar to frailty, as a biological syndrome where the primary insult is cardiac dysfunction but leads to systemic consequences. Because of this significant overlap, the utility of frailty assessment in advanced HF is unclear. The aim of the current study was to investigate the utility of frailty assessment and its prognostic value in elderly patients with advanced HF.
Frailty is a biological syndrome defined as a decreased homeostatic reserve leading to an increased vulnerability to stressors and adverse outcomes.1 Frailty manifests clinically as a disproportionate change in health status in response to a physical or psychological stress.2 Several recent studies From the 1Division of Cardiology, Department of Medicine, Montefiore Medical Center, Albert Einstein College of Medicine, Bronx, New York; 2Division of Cardiology, Department of Medicine, Houston Methodist Hospital, Houston, Texas; 3Department of Health Sciences Research, Division of Biomedical Statistics and Informatics Mayo Clinic, Rochester, Minnesota and 4 Division of Geriatrics, Department of Medicine, Montefiore Medical Center, Albert Einstein College of Medicine, Bronx, New York. Reprint requests: Snehal R. Patel, MD, Division of Cardiology, Heart Failure, Cardiac Transplantation and Mechanical Circulatory Support, 3400 Bainbridge Avenue, Medical Arts Pavilion- 7th floor, Bronx, New York 10467. Tel: +1 718 920 2248; Fax: +1 718 652 1833. E-mail: SNEPATEL@ montefiore.org. Manuscript received November 20, 2015; revised manuscript received February 3, 2016; revised manuscript accepted February 5, 2016. See page ■■ for disclosure information. 1071-9164/$ - see front matter © 2016 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.cardfail.2016.02.003
Methods In this single-center pilot study, consecutive patients from Montefiore Medical Center HF Clinic, between October 2012 and July 2013, aged ≥65 years, New York Heart Association (NYHA) class III or IV, and left ventricular ejection fraction (LVEF) ≤ 35% (measured by echocardiogram within 30 days of the study visit) were screened to participate. Key 1
ARTICLE IN PRESS 2 Journal of Cardiac Failure Vol. ■■ No. ■■ ■■ 2016 and combined for a composite score of 0 to 5. Subjects with a score of 0 were classified as “not frail,” 1–2 as “prefrail,” and 3 or higher as “frail.” Assessment of the individual domains is described in the supplementary Table S1. All subjects were followed from the time of testing through August 2014 for the combined primary endpoint of allcause hospitalizations or death. Secondary endpoints included all-cause mortality, non-HF hospitalizations, and HFrelated hospitalizations only. Outcomes were ascertained through review of medical records and confirmed by telephone follow-up. Statistical Analysis Baseline characteristics were described as frequencies for categorical and as mean ± standard deviation for continuous variables, and compared using chi-squared test and unpaired t test, respectively. Cox proportional hazards ratio (HR) modified for Andersen-Gill modeling was calculated to test the effect of frailty on hospitalizations and mortality, and adjusted for covariates. Unlike the traditional Cox model, which accounts for only the first hospitalization, the AndersenGill model takes into account multiple hospitalizations8,9 and treats each hospitalization for each subject as a separate observation. Statistical analysis was performed using R statistical software and 2-tailed P values of <.05 were considered significant. Results
Fig. 1. (A) Flowchart describing the patient selection process towards achieving the final study cohort of 40 patients. (B) Spectrum of frailty: percentage of patients with different frailty scores (range 0–5). (C) Percentage of patients meeting frailty criteria for each of the individual components of the Fried Frailty Index. LVEF, left ventricular ejection fraction; 6MWT = 6 minute walk test.
exclusion criteria included duration of HF < 6 months, acute decompensation within the previous 30 days, or inability to walk. Patients who met these inclusion/exclusion criteria underwent a 6-minute walk test (6MWT) and were enrolled in the final study cohort only if they walked <300 m (Fig. 1a). Hence, for the purposes of this study, advanced HF was defined as “NYHA class III or IV, LVEF ≤35%, and a 6MWT of <300 m.” The cutoff of <300m on 6MWT was chosen because it has been shown to define an increased risk of adverse events in HF.7 The study was approved by the Albert Einstein College of Medicine Institutional Review Board. Frailty was assessed using a modified version of the Fried Frailty Index as defined in the Cardiovascular Health Study.2 Five domains were assessed for frailty: weight loss, exhaustion, weakness, slow gait, and reduced physical activity. Each domain was scored 0 or 1 based on its absence or presence
Seventy-three subjects met the initial criteria and underwent 6MWT. Of these, 40 walked <300 m and formed the final study cohort (Fig. 1a). Overall characteristics included, mean age of 74.9 ± 6.5 years, 58% female, 37.5% NYHA class IV, LVEF 25.6 ± 6.4%, 6MWT 195.8 ± 74.3 m, and Charlson Comorbidity Index (CCI) score of 4.9 ± 1.9. There was a spectrum of frailty scores, ranging from 0 to 5 (Fig. 1b). The proportion of subjects meeting frailty criteria for each of the individual components is shown in Fig. 1c. When graded according to the prespecified Fried criteria: 0 subjects were not frail, 14 (35%) were prefrail, and 26 (65%) were frail. The baseline demographics of the overall cohort and as analyzed by the prefrail and frail groups are shown in Table 1. There was a higher prevalence of diabetes in the prefrail group; otherwise, there were no significant differences in the baseline characteristics. Both groups were equally well medicated and the length of follow up was similar. Frailty and Outcomes During follow-up, 10 patients died and 26 were hospitalized for any cause, including 20 for an exacerbation of HF. Including repeated events, there were a total of 69 all-cause hospitalizations, 45 (65%) of which were due to HF exacerbations. Pneumonia, infections and gastrointestinal bleeding were the major causes of non-HF hospitalizations. Of the 10
ARTICLE IN PRESS Frailty Assessment in Advanced Heart Failure
•
Madan et al
3
Table 1. Baseline Demographics of the Overall Cohort and as analyzed by the Prefrail and Frail Groups
Age (y) Sex (F) Race: Whites Blacks Hispanics Asians/others LVEF (%) NYHA class (IV) LVEDD (cm) HTN DM Atrial fibrillation PAD GFR HF etiology (ICM) HF duration (mo) COPD Depression Medications Beta-blockers ACEI/ARBs Aldosterone antagonists Hydralazine/nitrates 6MWT (m) Follow-up (d) Charlson Comorbidity Index
Total (n = 40)
Pre-frail (n = 14)
Frail (n = 26)
P Value: Pre-frail vs Frail
74.9 ± 6.5 23 (57.5%) 4 (10%) 16 (40%) 17 (42.5%) 3 (7.5%) 25.6 ± 6.4% 15 (37.5%) 5.96 ± 0.87 35 (87.5%) 22 (55%) 17 (42.5%) 4 (10%) 46.9 ± 22.4 19 (47.5%) 88.1 ± 79.8 9 (22.5%) 8 (20%)
76.0 ± 5.3 6 (42.8%) 2 (14%) 3 (21.4%) 8 (57.1%) 1 (7.1%) 26 ± 5.5 5 (35.7%) 5.82 ± 0.76 13 (92.8%) 11 (78.5%) 4 (28.6%) 2 (14.3%) 47.7 ± 16.9 8 (57.1%) 106.9 ± 87.4 5 (35.7%) 1 (7.1%)
74.4 ± 7.2 17 (65.4%) 2 (7.7%) 13 (50.0%) 9 (34.6%) 2 (7.7%) 25.3 ± 6.9 10 (38.4%) 6.01 ± 0.93 22 (84.6%) 11 (42.3%) 13 (50%) 2 (7.7%) 46.5 ± 25.2 11 (42.3%) 75.3 ± 74.1 4 (15.4%) 7 (26.9%)
.475* .169† .342†
.750* .738† .600* .452† .028† .191† .507† .873* .37† .320* .142† .136†
39 (97.5%) 32 (80%) 20 (50%) 18 (45%) 195.8 ± 74.3 454 ± 186 4.9 ± 1.9
13 (92.8%) 10 (71.4%) 8 (57.1%) 6 (42.9%) 192.4 ± 68.06 469.8 ± 179.2 4.2 ± 0.7
26 (100%) 22 (84.6%) 12 (46.2%) 12 (46.2%) 197.7 ± 78.71 435.7 ± 190.8 5.2 ± 2.3
.168† .32† .507† .842† .833* .585* .128*
ACEI, angiotensin-converting enzyme inhibitor; ARBs, angiotensin II receptor blockers; COPD, chronic obstructive pulmonary disease; DM, diabetes mellitus; F, female; GFR, glomerular filtration rate; HF, heart failure; HTN, hypertension; ICM, ischemic cardiomyopathy; LVEDD, left ventricle end diastolic diameter; LVEF, left ventricular ejection fraction; 6MWT, 6-minute walking test; NYHA, New York Heart Association; PAD, peripheral arterial disease. Boldface type indicates p value of <.05. *Unpaired t test. † Chi square test.
patients that died, 6 were attributable to cardiovascular causes, with no significant difference in the prefrail and frail groups (2 vs 4, P = 1.0). Compared with the prefrail group, frail subjects were at an approximately 2-fold increased risk for the primary endpoint of all-cause hospitalization or death (cumulative HR 1.93, 95% CI 1.15–3.25, P = .013, Fig. 2a). This association remained significant after adjusting for diabetes (cumulative HR 2.02, 95% CI 1.17–3.49, P = .011), race (Black vs others) (cumulative HR 2.23, 95% CI 1.29–3.84, P = .004) and also after adjusting for diabetes, age, sex, and CCI (cumulative HR 1.95, 95% CI 1.06–3.59, P = .031). There were more allcause hospitalizations in the frail vs prefrail group: median (interquartile range) 2 (0.75–2) vs 0 (0–2.25) with a cumulative HR of 1.92 (95% CI 1.12–3.27, P = .017, Fig. 2b). Although there was a higher proportion of deaths in the frail group (n = 8 [31%] vs n = 2 [14%]), this was not statistically significant (cumulative HR 2.18, 95% CI 0.46–10.27, P = .324). In subgroup analysis, frailty status was associated with increased risk of non-HF–related hospitalizations: median (interquartile range): 0.5 (0–1) vs 0 (0-0), cumulative HR 3.31 (95% CI 1.14–9.64, P = .028, Fig. 2d), but not HF-related hospitalizations: median (interquartile range) 1 (0– 2) vs 0 (0–2.25), cumulative HR 1.31, 95% CI 0.68–2.49, P = .382 (Fig. 2c). Finally, the 6MWT was not predictive of either the primary or any of the secondary end points in our cohort, and there was no correlation between the 6MWT and frailty (R2 = 0.005, P = .675).
Discussion The prevalence and prognostic value of frailty in a general HF population is well documented.3–5 In the current study, we evaluated the role of frailty in advanced HF patients. We found that frailty testing via the Fried Frailty Index was able to describe a spectrum in which 35% of subjects were prefrail and 65% were frail. Moreover, the frail group was associated with an approximately 2-fold increased risk of death or hospitalization when compared with the prefrail group. To our knowledge, this is the first report describing the utility of frailty assessment in advanced HF. Our elderly HF cohort was representative of the real world with significant comorbidities: average age was 75 years, LVEF 26%, extremely poor endurance with 6MWT of 196 m, and a CCI of almost 5. Not surprisingly, the event rate was high in the overall cohort, with 73% either dying or hospitalized during an average follow-up of 1.25 years. Nonetheless, frailty testing identified those at highest risk of death or hospitalization. The predictive value was maintained for the individual endpoints of all cause hospitalization and trended toward significance for mortality. Importantly, frailty status was predictive of non-HF–related hospitalizations, but was not associated with HF hospitalizations. This last finding is critical and may support the assumption that frailty assesses a different metric than other traditional HF tools. The prognostic value of the 6MWT in HF is well established10 and has been linked to frailty in prior analysis
ARTICLE IN PRESS 4 Journal of Cardiac Failure Vol. ■■ No. ■■ ■■ 2016
Fig. 2. Kaplan-Meier survival analysis modified for Andersen Gill modeling. Probability of freedom from (A) all-cause hospitalization and death, (B) all-cause hospitalization, (C) heart failure hospitalizations only, and (D) non-heart failure hospitalizations in prefrail vs frail subjects. Corresponding P values are shown in the graphs.
of less sick HF patients.11 In contrast, we did not find a significant correlation between 6MWT and frailty. This is likely because our cohort was predefined based on 6MWT (<300 m), limiting the variability of this measurement in the study population. Regardless of this discrepancy, our findings highlight that frailty assessment can have added prognostic value in HF patients with a very poor 6MWT. There are several limitations to the current study. First, this is a single-center pilot study with a small sample size and the findings need validation in a larger cohort. Second, we chose to use the most accepted frailty tool—the Fried Frailty Index; however, several other frailty assessment tools have been validated, which were not incorporated. Although the original Fried Frailty Index used a weighted score of kilocalories per week for assessment of low physical activity, we used a questionnaire to measure physical activity levels. This modified version has been validated in previous studies.12 Third, we did not compare the predictive value of frailty to other prognostication tools such as cardiopulmonary stress testing or the Seattle HF Model.13 Finally, because of the small sample size, we did not include many covariates that could affect hospitalization rates (like NT pro-BNP levels, chronic obstructive pulmonary disease, depression) into our multivariate model; however, we did include the CCI as a measure of the global burden of comorbidity14 in the multivariate analysis. In conclusion, our data demonstrate that frailty testing in elderly patients with advanced HF is feasible and provides
clinical prognostic value. These are intriguing findings that require validation in a larger cohort. Disclosures There are no financial disclosures or conflict of interest for any of the authors in regards to this publication. Acknowledgments This research was funded by Division of Cardiology, Montefiore Medical Center, Albert Einstein College of Medicine, Bronx, NY. Appendix: Supplementary material Supplementary data to this article can be found online at doi:10.1016/j.cardfail.2016.02.003. References 1. Fried LP, Ferrucci L, Darer J, Williamson JD, Anderson G. Untangling the concepts of disability, frailty, and comorbidity: implications for improved targeting and care. J Gerontol A Biol Sci Med Sci 2004;59:M255–63.
ARTICLE IN PRESS Frailty Assessment in Advanced Heart Failure 2. Fried LP, Tangen CM, Walston J, Newman AB, Hirsch C, Gottdiener J, et al. Frailty in older adults evidence for a phenotype. J Gerontol A Biol Sci Med Sci 2001;56:M146–57. 3. Cacciatore F, Abete P, Mazzella F, Viati L, Della Morte D, D’Ambrosio D, et al. Frailty predicts long-term mortality in elderly subjects with chronic heart failure. Eur J Clin Invest 2005;35:723–30. 4. Chaudhry SI, McAvay G, Chen S, Whitson H, Newman AB, Krumholz HM, 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 2013;61:635–42. 5. McNallan SM, Singh M, Chamberlain AM, Kane RL, Dunlay SM, Redfield MM, et al. Frailty and healthcare utilization among patients with heart failure in the community. JACC Heart Fail 2013;1:135–41. 6. Khan H, Kalogeropoulos AP, Georgiopoulou VV, Newman AB, Harris TB, Rodondi N, et al. Frailty and risk for heart failure in older adults: the health, aging, and body composition study. Am Heart J 2013;166:887–94. 7. Rostagno C, Olivo G, Comeglio M, Boddi V, Banchelli M, Galanti G, et al. Prognostic value of 6-minute walk corridor test in patients with mild to moderate heart failure: comparison with other methods of functional evaluation. Eur J Heart Fail 2003;5:247–52.
•
Madan et al
5
8. Guo Z, Gill TM, Allore HG. Modeling repeated time-to-event health conditions with discontinuous risk intervals: an example of a longitudinal study of functional disability among older persons. Methods Inf Med 2008;47:107. 9. Andersen PK, Gill RD. Cox’s regression model for counting processes: a large sample study. Ann Stat 1982;1100–20. 10. Cahalin LP, Mathier MA, Semigran MJ, Dec GW, DiSalvo TG. The six-minute walk test predicts peak oxygen uptake and survival in patients with advanced heart failure. CHEST J 1996;110:325–32. 11. Boxer R, Kleppinger A, Ahmad A, Annis K, Hager D, Kenny A. The 6-minute walk is associated with frailty and predicts mortality in older adults with heart failure. Congest Heart Fail 2010;16:208–13. 12. Verghese J, Holtzer R, Lipton RB, Wang C. Mobility stress test approach to predicting frailty, disability, and mortality in high-functioning older adults. J Am Geriatr Soc 2012;60:1901–5. 13. Levy WC, Mozaffarian D, Linker DT, Sutradhar SC, Anker SD, Cropp AB, et al. The seattle heart failure model prediction of survival in heart failure. Circulation 2006;113:1424–33. 14. Henkel DM, Redfield MM, Weston SA, Gerber Y, Roger VL. Death in heart failure a community perspective. Circ Heart Fail 2008;1: 91–7.