Abstract
Whether risk of death interacts with change in 6-minute walking test (6MWT) distance after cardiac rehabilitation (CR) on heart failure (HF) prognosis is unknown. We studied 1,569 patients with HF admitted to 6 inpatient CR units. A 6MWT was performed at admission to and at discharge from CR. The primary outcome was 3-year mortality. The interaction of baseline risk and 50-meter increase in 6MWT distance on the primary outcome was tested using a Cox regression model with an interaction term between baseline risk and 50-meter increase. To further illustrate the interaction between baseline risk and change in 6MWT distance, participants were stratified into low- and high-risk subgroups. The p value for the interaction was 0.588. Three-year mortality was 16.1% in the low-risk and 53.7% in the high-risk subgroup. 43.0% of the low-risk and 43.4% of the high-risk patients achieved an increase ≥50 meters in 6MWT distance. The adjusted hazard ratio of 3-year mortality for the patients who achieved a 50-meter increase in 6MWT distance was 0.67 (0.47–0.96; p = 0.030) in the low-risk and 0.75 (0.61–0.91; p = 0.005) in the high-risk subgroup. These findings suggest that improvement in functional capacity after CR is associated with improved survival regardless of baseline risk of death.
Introduction
Despite the therapeutic advances in the last few decades, heart failure (HF) still has a grim prognosis. Functional disability, impaired quality of life, and loss of independence are hallmark features of HF, with a substantial proportion of patients having moderate to severe difficulty with basic activities of daily living. Persistently severe or worsening functional disability contributes to worsen prognosis. Cardiac rehabilitation (CR) has been regarded as an evidence-based pillar of HF management and is rated as a class I recommendation by current guidelines. Meta-analyses showed that CR is effective in improving functional capacity and quality of life and reducing rehospitalization rates. In addition, large observational studies suggested an association between participation in CR and reduced risk of mortality. ,, The 6-minute walking test (6MWT) a safe, inexpensive, reliable, and reproducible tool for assessing functional capacity and prognosis in HF. , It is noteworthy that, for many patients with HF, the 6MWT entails a maximal effort, as documented by the achieved percentage of peak aerobic capacity and respiratory exchange ratio in the last part of exercise. A 50-meter increase in 6MWT distance is regarded as a clinically meaningful improvement in functional capacity, is associated with a significant improvement in NYHA class and health-related quality of life, and has been used as a surrogate end-point in clinical trials. Previous studies also demonstrated that a 50-meter increase in 6MWT distance after CR or during up-titration of guideline-directed medical therapy independently predicts improved long-term survival. Ferreira et al suggested that an increase in 6MWT distance is of much more prognostic value in severely symptomatic patients. Myhre et al also observed a significant effect modification for shorter 6MWT distances at baseline and older age on the prognostic value of improvement in 6MWT distance, with improvement in 6MWT distance having a larger prognostic value for older patients and patients with shorter baseline walking distances. This finding leads to hypothesize that the prognostic value of 50-meter increase in 6MWT distance may be conditional on the risk of death; both old age and short walking distances are indeed established, powerful markers of increased mortality. In this study, we investigated whether baseline risk of death modifies the association between improvement in 6MWT distance and mortality in HF.
Methods
This was a multicenter observational retrospective study. The study population consisted of 1,732 patients with chronic HF admitted to 6 specialized inpatient CR units of a nationwide Research Institute in the field of rehabilitation medicine in Italy between January 2013 and December 2016, who had available data for NT-proBNP and 6MWT at admission to CR. This data set had been used to develop the Severe Functional Impairment/NT-proBNP (SFI/NT-proBNP) risk score. According to the national regulatory rules governing admissions to inpatient CR for HF in Italy, patients were admitted after a hospitalization for HF or because of declining functional capacity and/or clinical status. All participating centers are part of a single department of CR and share a common formal rehabilitation program. In each participating center, the multidisciplinary CR rehabilitation team comprises the following professionals: cardiologist, physiatrist, physiotherapist, psychologist, dietitian, and nurse. Our formal multidisciplinary cardiac rehabilitation program is led by cardiologists and is designed to promote stable clinical conditions, improve physical function through a supervised exercise training plan tailored to the individual level of functional ability at presentation, provide specialized medical assistance, and optimize medical treatment. The exercise program consists of a supervised training program including active/passive mobilization; assisted ambulation; respiratory, musculoskeletal flexibility, movement coordination and/or callisthenic exercises, and training on a (unloaded) bedside/upright cycle ergometer. The types of exercises and exercise intensity are gradually progressed throughout the rehabilitation period, according to the individual functional and clinical conditions.
The patients who died in the inpatient rehabilitation facilities (N. 17, 1%) or were transferred back to an acute care facility (N. 37, 2.1%), those with missing data for 6MWT at discharge from CR (N. 73, 4.2%), and those lost to follow-up (N. 36, 2.1%) were excluded, leaving 1,569 patients available for analysis.
Data collection
The data were extracted from the electronic Hospital Information System shared between the participating centers and entered into a REDCap database. All patients provided informed written consent to the use of their data in an anonymous form for scientific purposes. Any identifying information was removed from the database and replaced with an identification number. The study was undertaken as part of a companion research project, which was approved by the Ethics Committee of Istituti Clinici Scientifici Maugeri on 27 July 2021 (approval number: 2576-CE). Survival status was ascertained by linkage to the national Health Information System. The patients were followed-up until death or November 30, 2019.
Six-minute walking test
As part of our rehabilitation program, a 6MWT was performed at admission to and at discharge from CR by trained therapists, according to a standardized protocol. For patients unable to walk without assistance, the 6MWT distance was set to 0 meters. Change in the 6MWT distance was measured as the absolute difference in meters between discharge and admission tests. We also calculated change in percent-of-predicted distance walked on the 6MWT using the Enright reference equation. According to the original article, percent-of-predicted distance was not calculated for 140 patients with body mass index >35.
Exposure variable
The exposure variable was improvement in functional capacity defined as a 50-meter or greater increase in 6MWT distance.
Primary outcome
The primary outcome was 3-year all-cause mortality after discharge from CR.
Baseline risk assessment
We used the SFI/NT‑proBNP risk score to estimate the baseline risk of death. The SFI/NT-proBNP is a multiparametric point-based risk score originally developed to predict 1-year mortality in HF. Briefly, the SFI/NT-proBNP risk score is based on 8 variables collected at admission to CR and assigns 1, 2, 3, or 4 points to age 60–69, 70–79, 80–89, or ≥90, respectively; 3 points to male sex; 2 points to moderate-to-severe anemia (Hb <11 g/dL); 2 points to systolic blood pressure <100 mm Hg; 2 points to estimated glomerular filtration rate <30 mL/min/1.73 m 2; 2 points to sodium <136 mEq/L; 3, 4, or 7 points to NT-proBNP levels of 800–1,599, 1,600–3,199, or ≥3,200 pg/mL, respectively; and 4 points to severe functional impairment defined as the inability to perform a 6MWT or a 6MWT distance <300 m.
Statistical analysis
Data are reported as mean and standard deviation (SD) or median with interquartile range (IQR) for continuous variables and as number and percentage for categorical variables. We used the Student’s t-test or the Mann–Whitney test to compare continuous variables and the χ 2 test to compare categorical variables. The normality of the 6MWT data distribution was assessed using the Shapiro-Wilk test. The incidence rate of death/100 person-year with 95% confidence intervals was estimated using the Poisson distribution. The median survival time was calculated as the time point at which the probability of survival equalized 50%. Cumulative mortality rates were estimated using the Kaplan-Meier method and a log-rank test was used to compare groups.
The association of the SFI/NT-proBNP risk score with the primary outcome was assessed using Cox proportional hazards analysis. Hazard ratios (HR) with 95% confidence intervals (CI) were calculated. Predicted risks of death at 3 year from the Cox model were calculated by combining the baseline survival function at 3 years with the linear predictor of the SFI/NT-proBNP risk score. Discrimination was assessed by calculating the C index. We used the bootstrap resampling technique with 200 replicates to obtain optimism-corrected C index and reported the mean results and bootstrap estimated 95% confidence intervals. Calibration was assessed by calibration plots of observed versus predicted mortality across quintiles of predicted probabilities. Linear regression was used to calculate the intercept and slope of the calibration lines and to compare them to the ideal (intercept = 0, and slope = 1).
We used multivariable logistic regression analysis to identify the baseline characteristics that independently predicted an increase in 6MWT distance ≥50 meters. The association of 50-meter increase in 6MWT distance with mortality was examined using Cox proportional hazards regression models. The risk score was used as a covariate in all regression models. The interaction between risk score and 50-meter increase in distance walked on the occurrence of death was tested in a Cox regression model with an interaction term between the risk score and 50-meter increase. The interaction analysis was conducted with the risk score as a continuous variable. To further illustrate the interaction between baseline risk and change in 6MWT distance, the prognostic value of 50-meter increase was also examined stratifying the patients into 2 risk categories (low-risk and high-risk), based on the median value of the SFI/NTproBNP risk score. Statistical analyses were conducted using STATA software, version 14 (Stata-Corp LP, College Station, Tex).
Results
Table 1 displays patients’ baseline characteristics stratified by the predicted 3-year risk of death. Compared with low-risk patients, high-risk patients were approximately eleven years older; had higher burden of comorbidities and lower body mass index; had poorer functional capacity, lower hemoglobin levels, and poorer renal function; and had higher blood urea nitrogen and NT-proBNP levels.
Table 1
Patients’ baseline characteristics stratified by the predicted 3-year risk of death
| Variables |
All Patients
(N = 1,569) |
Low Risk
(N = 844) |
High Risk
(N = 725) |
p Value |
|---|---|---|---|---|
| Demographics | ||||
| Age (years), mean (SD) | 67.9 (12.8) | 62.8 (12.3) | 73.7 (10.7) | < 0.001 |
| Females, N (%) | 429 (27.3) | 238 (28.2) | 191 (26.3) | 0.411 |
| Body mass index (kg/m 2), mean (SD) | 27.2 (6.1) | 28.2 (6.5) | 26.1 (5.3) | < 0.001 |
| Comorbidities | ||||
| Obesity (body mass index >30), N (%) | 400 (25.5) | 256 (30.3) | 144 (19.9) | < 0.001 |
| Hypertension, N (%) | 747 (47.6) | 314 (37.2) | 433 (59.7) | < 0.001 |
| Diabetes mellitus, N (%) | 473 (30.1) | 226 (26.9) | 247 (34.1) | 0.002 |
| Chronic obstructive pulmonary disease, N (%) | 313 (19.9) | 136 (16.1) | 177 (24.4) | < 0.001 |
| Moderate-to-severe anemia (hemoglobin <11 g/dL), N (%) | 282 (18.0) | 55 (6.5) | 227 (31.3) | < 0.001 |
| Renal dysfunction (eGFR <60 mL/min/1.73 m 2), N (%) | 918 (58.5) | 367 (43.5) | 551 (76.0) | < 0.001 |
| Atrial fibrillation, N (%) | 578 (36.8) | 246 (29.1) | 332 (45.8) | < 0.001 |
| Clinical findings | ||||
| NYHA Class, N (%) | ||||
| I/II | 533 (34.0) | 292 (34.6) | 241 (33.2) | 0.571 |
| III | 892 (56,9) | 503 (59.6) | 389 (53.7) | 0.017 |
| IV | 85 (5.4) | 23 (2.7) | 62 (8.6) | < 0.001 |
| Undefined | 59 (3.8) | 26 (3.1) | 33 (4.6) | 0.126 |
| ICD in patients with EF ≤0.40, N (%) | 533 (48.8) | 312 (52.7) | 221 (44.2) | 0.005 |
| Systolic blood pressure (mm Hg), mean (SD) | 112.0 (6.9) | 113.6 (16.7) | 110.2 (17.1) | < 0.001 |
| Systolic blood pressure <100 mm Hg, N (%) | 259 (16.5) | 96 (11.4) | 163 (22.5) | < 0.001 |
| Left ventricular ejection fraction, mean (SD) | 35.9 (12.5) | 36.2 (12.1) | 35.6 (13.0) | 0.365 |
| Left ventricular ejection fraction ≤0.40, N (%) | 1092 (69.6) | 592 (70.1) | 500 (69.0) | 0.613 |
| Six-minute walking test | ||||
| Distance walked at admission (meters), median (IQR) | 271 (88-368) | 352 (275-412) | 140 (0-255) | < 0.001 |
| Severe functional impairment at admission, N (%) | 862 (54.9) | 236 (28.0) | 626 (86.3) | < 0.001 |
| Laboratory findings | ||||
| Hemoglobin (g/dL), mean (SD) | 12.9 (2.0) | 13.5 (1.7) | 12.1 (2.0) | < 0.001 |
| Creatinine (mg/dL), mean (SD) | 1.38 (0.57) | 1.18 (0.38) | 1.61 (0.67) | < 0.001 |
| eGFR (mL/min/1.73 m 2), mean (SD) | 56.5 (23.6) | 64 (21) | 48 (23) | < 0.001 |
| ≥60 mL/min/1.73 m 2 | 651 (41.5) | 477 (56.5) | 174 (24.0) | < 0.001 |
| 45-59 mL/min/1.73 m 2 | 384 (24.5) | 207 (24.5) | 177 (24.4) | |
| 30-44 mL/min/1.73 m 2 | 352 (22.4) | 135 (16.0) | 217 (29.9) | |
| <30 mL/min/1.73 m 2 | 182 (11.6) | 25 (3.0) | 157 (21.7) | |
| Blood urea nitrogen (mg/dL), mean (SD) | 34 (18) | 25.4 (11.4) | 38.4 (20.5) | < 0.001 |
| Sodium (mEq/L), mean (SD) | 138.8 (3.7) | 139.0 (3.3) | 138.4 (4.0) | 0.001 |
| Sodium <136 mEq/L, N (%) | 274 (17.5) | 111 (13.2) | 163 (22.5) | < 0.001 |
| NT-proBNP (pg/mL), median (IQR) |
2024
(811-4256) |
910
(417-1932) |
4231
(2478-6927) |
< 0.001 |
| <800, N (%) | 387 (24.7) | 382 (45.3) | 5 (0.7) | < 0.001 |
| 800–1,599, N (%) | 291 (18.5) | 195 (23.1) | 96 (13.2) | |
| 1,600–3,199, N (%) | 357 (22.8) | 216 (25.6) | 141 (19.4) | |
| ≥3,200, N (%) | 534 (34.0) | 51 (6.0) | 483 (66.6) | |
| SFI/NT-proBNP risk score, median (IQR) | 11 (7-15) | 7 (5-9) | 15 (13-17) | < 0.001 |
| Treatments at discharge in patients with EF ≤0.40 | ||||
| Beta-blockers, N (%) | 1,033 (94.6) | 574 (97.0) | 459 (91.8) | 0.001 |
| RAAS-Is, N (%) | 988 (90.5) | 574 (97.0) | 414 (82.8) | < 0.001 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
