Abstract
Objectives:
The purpose of this study was to determine if a progressive, prescribed home-based aerobic exercise program would alter the natural physiological processes that maintain fluid balance stability in patients with New York Heart Association (NYHA) class III/IV heart failure after medical optimization (titration of oral medical therapy with or without the infusion of an intravenous inotrope).
Methods:
A total of 56 men and women from a large tertiary trauma I hospital were enrolled with 56 subjects contributing to baseline analysis and 42 subjects at 24 weeks. Subjects were diagnosed with heart failure via NYHA classification IV or III for at least 6 months and were hospitalized for a current acute decompensation exacerbation in which they were being medically optimized. The exercise intervention was a home-based, prescribed, progressive aerobic exercise program lasting for 24 weeks. The exercise participants had weekly phone calls to gather data and progress the exercise program and one 12-week follow up. The usual care participants received random phone calls to collect data and had one 12-week follow up visit to attain physical assessment values.
Results:
Subjects were primarily female (59%), nonwhite (54%), and NYHA class IV (52%) versus class III (48%). The mean age was 58 years (±11.8 years). The subjects had a mean ejection fraction of 17.7 % (±7%) and mean maximal oxygen consumption of 12.1 (±3.4). Using a hierarchical multiple regression model, it was demonstrated that an exercise prescription (intensity, frequency, duration) significantly predicted 24 h weight fluctuations within a NYHA class III/IV heart failure population after medical optimization (R2 linear = 0.713, F = 3.224, p = 0.015).
Conclusion:
This study demonstrated that exercise is a successful adjunctive therapy to managing the daily weight variability or fluid status instability of patients with NYHA class III/IV heart failure that is often a debilitating aspect of the syndrome.
Introduction
Background
Heart failure (HF) is a significant healthcare concern in the United States with Americans’ lifetime risk of developing HF after the age of 40 currently at 20% and that increases with age [Chun et al. 2012; Lloyd-Jones et al. 2009; Yancy et al. 2013]. In addition, the financial burden is significant as HF healthcare costs are over $30 billion a year with over half of those costs attributed to hospitalizations. Fluid volume overload is considered the major cause of hospitalization of acute decompensated HF (ADHF) in patients with HF [Chun et al. 2012; Fonarow et al. 2008].
The most recent joint guidelines published by the American Heart Association and the American College of Cardiology (2013) [Yancy et al. 2013], the guidelines by the Heart Failure Society of America (2010) [Heart Failure Society of America et al. 2010], and the guidelines by the European Society of Cardiology (2012) [McMurray et al. 2012], attempt to bring guidance to the HF practitioner using the most recent research and categorizing the evidence into the guidelines for best practice. Within the guidelines there is an overriding theme to decongest and fluid optimize the patient while improving symptoms (often precipitated by congestion and fluid overload). The tools currently available are pharmacological interventions aimed at neural hormonal blockades and diuretics aimed to flush the body directly of excess fluid and mechanical tools (ultrafiltration) both used once the patient is in a state of fluid excess. These interventions, although lifesaving, need improvements as the patients are becoming intolerant to the medications, polypharmacy is a long-term disadvantage, and mechanical diuresis is expensive and only available in certain settings.
Current research has supported the use and value of exercise interventions among adults with HF [Piepoli, 2013; Piepoli et al. 2011; Ventura-Clapier et al. 2007]. Exercise training (ET) affects HF by improving endothelial function and increasing peak oxygen consumption, which ultimately improves cardiac function [Antunes-Correa et al. 2012; Piepoli, 2013]. ET has demonstrated physiological effects varying from improved strength, agility, and flexibility stemming from the skeletal muscle changes to improved functional and exercise capacity as a result of vascular and cardiac improvements [Antunes-Correa et al. 2012; Braith and Beck, 2008; Belardinelli et al. 2012; Kobayashi et al. 2003; Piepoli, 2013; Smart et al. 2006, 2007]. ET has also shown a positive effect on the autonomic system as well as neural hormonal activation [Braith et al. 1999; Negrao and Middlekauff 2008; Soares-Miranda et al. 2011]. These physiological effects that ET has had on the patient with HF has translated into decreased HF symptoms while improving perceived quality of life [Dracup et al. 2007; Jankowska et al. 2008; Kobayashi et al. 2003; O’Connor et al. 2009; Piepoli, 2013; Smart et al. 2007; Taylor et al. 2012]. Finally, ET has decreased rehospitalization rates, and findings from the National Institutes of Health funded HF exercise trial, HF-ACTION, demonstrated that ET is safe and, when prognostic factors are accounted for, ET improves all-cause mortality or all-cause hospitalization after 3 years [Belardinelli et al. 2012; O’Connor et al. 2009; Piepoli, 2013].
Objectives
The purpose of the study was to determine if regular prescribed exercise would alter the natural physiological processes that maintain fluid balance stability in patients with New York Heart Association (NYHA) class III/IV HF after medical optimization (titration of oral medical therapy with or without the infusion of an intravenous inotrope). The research examined the relationship of an exercise prescription (exercise intensity, duration, and frequency) on the fluid balance instability (daily weight variability) of patients with advanced HF with the hypothesis being that subjects who exercise at a greater intensity, frequency, and duration will demonstrate a greater increase in fluid stability as demonstrated by daily weight variability.
Methods
Sample/setting
A total of 56 men and women from a large tertiary trauma I hospital with a HF and transplant program were enrolled, with 56 subjects contributing to baseline analysis and 42 subjects at 24 weeks. Subjects had been diagnosed with HF via NYHA for at least 6 months and had been hospitalized for a current ADHF exacerbation in which they had been classified as NYHA III or higher and were being medically optimized. The exclusion criteria included receiving coronary artery bypass graft surgery, percutaneous coronary intervention, or having had a myocardial infarction within the past 6 months prior to the hospital admission, having significant peripheral vascular disease, or being listed for heart transplant prior to the hospitalization (unless blood type O). Due to medical optimization within baseline, medications were neither inclusion nor exclusion criteria.
Intervention
The exercise intervention was a home-based exercise protocol lasting 24 weeks. The exercise participants had weekly phone calls to gather data and progress the exercise program with one 12-week follow-up clinic visit. The usual care participants received random phone calls to collect exercise and pedometer data and had one 12-week follow-up clinic visit to attain physical assessment values.
The exercise intervention was a home-based progressive aerobic exercise protocol lasting 24 weeks. It was prescribed by a HF specialist using the results from an exercise test and the physical exam obtained while the subjects were still in the hospital. The initial exercise prescription was a progressive walking prescription tailored to the individual’s capabilities and needs using the three components: intensity, duration, and frequency. Intensity was gauged by the rate of perceived exertion (RPE) scale and subjects were instructed to exercise to an intensity of 11–13 or to use 40% of their heart rate reserve (HRR). The HRR technique, although very reliable in the normal population, is relatively unreliable within this population due to various cardiac medications being taken by the patients, whereas the RPE is not dependent upon physiological indicators. Duration was variable per subject. As subjects were able to tolerate exercise they were instructed to increase their duration. Frequency was also variable, yet most subjects were asked to exercise a minimum of 5 days each week.
In the initial 12 weeks of the intervention, the subjects walked as their form of structured exercise. During the second 12 weeks, if the subjects had tolerated the walking program, the subjects received a stationary bicycle and were asked to both walk and ride the stationary bicycle. The subjects were allowed to choose the amount of time spent on each type of exercise with the exercise prescription remaining the same. The addition of the bicycle was to add variety and to allow an increase in exercise time for those not able to walk for long periods of time.
The exercise participants received weekly phone calls to gather data and progress the exercise prescription. The usual care participants received random phone calls to collect exercise and pedometer data. All subjects recorded data on daily logs as well as having a 12-week visit to attain a physical exam and collect physical data. Any subjects needing additional follow-up visits were seen as needed in the HF Clinic. During the entire intervention, subjects recorded their daily activity on a daily log which included heart rate, daily weight, shortness of breath scale, chest pain scale, RPE scale, and minutes exercised.
The variables of interest were fluid status, as measured by 24 h weight, and measures of exercise (intensity, duration, and frequency) as measured via self-report. Fluid status was assessed as daily body weight variability. Daily body weight was attained from four separate sources: hospital records during the initial hospital stay, self-report daily activity logs attained during the 24 weeks of intervention, weights taken intermittently during the study at follow-up visits, and weights from clinical records. The daily weight variability was assessed by taking the standard deviation of one week’s daily weights. The standard deviation yields a variance of daily weight for the subject during a 7-day time frame, indicating either a great amount of fluid shift or a static fluid shift. This technique was done to alleviate any increase or decrease in body composition that was attributable to exercise (i.e. gain in muscle mass or loss of body fat) or lack of exercise (loss of muscle mass or gain of body fat). Body composition shifts are much slower and are not so easily detectable in variance within daily weight, whereas intracellular to extracellular fluid shifts are rapid and therefore more easily detectable with such mathematic techniques.
Measures of actual exercise performed were grouped according to measures of intensity, duration, or frequency and were self-reported. Exercise intensity was measured by RPE scale while exercising. Exercise duration was measured via minutes of exercise each exercise session. Finally, exercise frequency was measured by the number of days exercised each week. Daily and weekly pedometer readings (steps taken) during the 24-week time frame were recorded for each subject as an objective measure of user accuracy (objective measure of physical activity) and correlated with each exercise variable (self-report measure of exercise activity) (see Table 1). There were no outlying correlates for these factors (i.e. all subjects were within acceptable compliance range, exercise factors correlated within range of each other).
Baseline demographics.
HF, heart failure; Max VO2, maximal oxygen consumption; NSS, not statistically significant; NYHA, New York Heart Association; SD, standard deviation.
Statistical analysis
A univariate analysis was conducted at baseline, 12, and 24 weeks to assess variables of interest. A bivariate analysis was conducted at baseline, 12, and 24 weeks on all variables of interest to determine if any multicollinearity existed and to establish if any other important relationships existed prior to running the multivariate analysis. Assumptions for the multivariate analysis were tested prior to running the full model. Finally, a hierarchical multiple regression was run to determine if exercise intensity (as measured by RPE), exercise frequency (as measured by days exercised per week), and exercise duration (as measured by minutes of exercise per session) directly predicted daily weight fluctuations (as measured by standard deviations of daily weights during a week time frame) when baseline weight fluctuations and event causing exit from the study were controlled within the model.
Results
Subjects were primarily female (59%) and nonwhite (54%). The mean age was 58 years (±11.8 years). The subjects were primarily NYHA class IV (52%) versus class III (48%) and had a mean ejection fraction of 17.7 % (±7%). Baseline demographics are summarized in Table 1.
Within this sample, medical optimization during hospitalization for the acute decompensation event was significant in the treatment and stabilization of their HF. As presented in Table 2, all of the assessed signs and symptoms decreased from admission to discharge with the presence of jugular venous distention (JVD), peripheral edema, murmur, and cardiac S3 dropping significantly (from 68% to 34%). In addition, classification improved by 41% with systolic and diastolic blood pressure also improving. Finally, these subjects lost an average of 6.7 lb between admission and discharge.
Baseline characteristics: from admission to discharge.
BP, blood pressure; Max VO2, maximal oxygen consumption; NYHA, New York Heart Association; PAR, physical activity recall; SD, standard deviation.
Italic text in the table represents clinically significant findings.
At baseline both groups had high weekly standard deviations for daily weights, whereas over time the standard deviation of weekly daily weight decreased for those who exercised yet those who did not exercise maintained a higher standard deviation (Figure 1). The weekly standard deviation is the change in daily weight over the week and was the most indicative of fluid shifts over shorter periods (i.e. 24–48 h). As the study progressed exercising subjects became more stable and 24–48 h standard deviation dropped to zero for almost all subjects whereas weekly standard deviations still demonstrated variation, which is why weekly standard deviations were used instead of 24 or 48 h standard deviations.
Weekly standard deviation of daily weight.
A hierarchical multiple regression model significantly predicted 24 h weight fluctuations within a NYHA class III/IV HF population after medical optimization (R2 linear = 0.713, F = 3.224, p = 0.015). The model determined that exercise intensity (RPE), exercise frequency (days exercised per week), and exercise duration (minutes of exercise per session) directly predicted daily weight fluctuations (standard deviations of daily weights during a week time frame) when baseline weight fluctuations and event causing exit from the study were controlled. Severity of illness at onset (NYHA class III or IV) did not predict daily weight fluctuations. Neither did any cardiovascular physiological parameter (i.e. baseline peak VO2, mean blood pressure, resting heart rate, peak heart rate) predict daily weight fluctuations. Overall, the model was predictive as theoretically indicated.
Discussion
The subjects in this study were a diverse group of patients with advanced HF. Although all patients were medically optimized in the hospital during an acute decompensation event, the patients ranged in etiology of HF. The facility was the location for all heart transplants within the region; some of the participants qualified for transplants during the study and this altered how the subjects approached exercise. The severity of illness trigger was a significant motivator for many participants. In addition, many subjects were disqualified from transplantation for being cancer survivors. This mental challenge (dealing with a second chronic life illness) was often an individual motivator independent of the severity of illness. Therefore, for this group of diverse subjects, severity of illness was not a predictor of motivation to exercise and was most likely the determinant for severity of illness at baseline not being a predictor within the model. Motivation to exercise within this very ill population was unique with all exercise participants who were able to exercise throughout the 24 weeks of the intervention.
Fluid status was viewed as a physiological expression of movement of fluid from the intracellular space to the extracellular space and as such was a difficult variable to assess easily and without undue burden to the subject. Therefore, an indirect assessment using multiple daily weight measures and using the variability of these measures to determine the subjects’ fluid shift was used to assess this fluid movement. Although weight in this circumstance will be influenced by long-term exercise, it was the day-to-day change in weight (i.e. fluid stability) that was of interest to the author. Therefore, total weight gained or lost over the course of the study was not included in the variables nor considered to be relevant to the fluid status of the subjects. Using the variable of fluid status and exercise in a multiple regression model demonstrated that when baseline daily weight fluctuations (i.e. if a subject is prone to extreme fluctuations) and how the subject left the study (i.e. death) were controlled, an exercise prescription (intensity, frequency, duration) directly predicts daily weight fluctuations. This is significant in the symptom management of patients with advanced HF who are often on multiple medications. Using nonpharmacological methods that are safe and effective, in addition to current guideline medications, to manage symptoms is key in the total medical management of these patients.
Conclusion
This pilot study demonstrated that exercise has potential as a successful adjunctive therapy to managing the 24 h weight variability or fluid status instability of patients with NYHA class III/IV heart failure, which is often a debilitating aspect of the disease. The statistical model failed to yield other cardiovascular physiological predictors in addition to the exercise intervention with an initial baseline profile of daily weight fluctuation of the subject and what caused the subject to leave the study. The failure of the model to identify a complete exercise responder profile warrants further research into exactly what it is that enables an exercising patient with advanced HF to stabilize their daily weights. Additional research examining more homogeneous HF patient populations (i.e. patients with HF and preserved ejection fraction compared with those with reduced ejection fraction) with a more structured protocol (structured supervised exercise sessions with weekly exams) to determine the exact physiological mechanisms at play is warranted. Finally, more research into measurement of fluid shifts and noninvasive, accurate measures is warranted to ensure precise variable measurement is being conducted.
Footnotes
Acknowledgements
The author would like to acknowledge the Dorn VAMC Research Service for their support and assistance in this author’s research.
Conflict of interest statement
The author declares that there is no conflict of interest.
Funding
The following institutions provided funding for data collection, data analysis, and manuscript development and preparation: Veterans Health Administration, University of Florida and Georgia Regents University.