2019 Volume 83 Issue 2 Pages 334-341
Background: Low body mass index (BMI) is a relevant prognostic factor for heart failure (HF), but HF patients with low BMI are reported to be at risk of not receiving optimal drug treatment. We sought to evaluate the efficacy of cardiac rehabilitation (CR) in patients with low vs. normal BMI.
Methods and Results: We studied 152 consecutive patients (low BMI, n=32; normal BMI, n=119) who participated in a 3-month CR program. Low BMI was defined as <18.5 kg/m2 and normal BMI, as 18.5≤BMI<25 kg/m2. All patients underwent cardiopulmonary exercise testing and muscle strength testing at the beginning and end of the 3-month CR program. After CR, a significantly greater proportion of HF patients with low BMI had a positive change in peak V̇O2 than in the normal BMI group (91% vs. 70%; P=0.010). Average percent change in peak V̇O2 was significantly greater in patients with low vs. normal BMI (17.1±2.8% vs. 7.8±1.5%; P<0.001). In addition, on multivariable logistic regression, low BMI was an independent predictor of a positive change in peak V̇O2 after CR (OR, 3.97; 95% CI: 1.10–14.31; P=0.035).
Conclusions: CR has a greater effect in patients with low than normal BMI, and low BMI is an independent predictor of a positive change in peak V̇O2. Thus, CR should be strongly recommended for HF patients with low BMI.
Heart failure (HF) is an important health-care problem. Despite recent advances in both pharmacological and device therapy, the prognosis of HF remains poor.1 After reaching doses of HF drugs recommended by the American Heart Association/American College of Cardiology (AHA/ACC) guidelines,2 the most promising approach to treating HF seems to be combination therapy that includes drug treatment, nutritional counseling, and exercise.
For patients with HF, low body mass index (BMI), which reflects lower body weight, is a relevant prognostic factor.3–6 Given that HF patients with low BMI are at higher risk of not receiving optimal drug treatment due to hypotension,7 concomitant non-pharmacological therapy such as nutrition and exercise therapy is especially important in this patient population.
Cardiac rehabilitation (CR) is one of most important non-pharmacological treatments for HF. CR improves exercise capacity and quality of life and reduces all-cause and HF-related hospitalization.8–10 Therefore, it is recommended by the AHA/ACC guidelines.2 Whether CR is effective in HF patients with low BMI, however, is not fully understood. We conducted this study to elucidate the efficacy of CR in HF patients with low vs. normal BMI.
We studied 190 consecutive HF patients (mean age, 64±13 years; men, 74%) who were admitted due to HF and underwent the CR program at the National Cerebral and Cardiovascular Center, Japan between November 2007 and May 2014. The diagnosis of HF was based on the Framingham criteria.2 We defined low BMI as BMI <18.5 kg/m2 and normal BMI as 18.5≤BMI<25 kg/m2 based on the World Health Organization classification.11 We focused on the influence of CR in patients with low BMI, therefore we excluded patients with high BMI (BMI ≥25 kg/m2; n=38; 20%), which is a prognostic factor for HF. Ultimately, 152 patients met the enrollment criteria. All patients underwent echocardiography before the CR program, and cardiopulmonary exercise testing (CPX), blood examinations, and muscular strength testing at the beginning and end of the 3-month CR program. The study complied with the Declaration of Helsinki, was approved by the institutional ethics committee, and all patients gave written informed consent.
CR ProgramOur CR program for HF has been described previously.12,13 Before starting the CR program, patients were confirmed to have no evidence of ischemia or severe arrhythmia during a 200-m walking test or other exercise tests. Patients gave written informed consent. The CR program consisted of 3–5 sessions/week that included walking, bicycling on an ergometer, and 40–60 min of calisthenics for 3 months. Exercise intensity was individualized at 30–50% of heart rate (HR) reserve (Karvonen’s equation: k=0.3–0.5)14 obtained on CPX, anaerobic threshold, or at levels 11–13 (“fairly light” to “somewhat hard”) of the 6–20 scale rating of perceived exertion (original Borg’s score15,16). For patients with left ventricular ejection fraction (LVEF) <40%, we used caution and prescribed a slightly lower level of exercise intensity, 30–40% of HR reserve and fewer sessions each week (3 sessions/week). The exercise program usually began with supervised sessions for 2–4 weeks, usually during hospitalization, followed by home exercise combined with 1–2 supervised sessions per week for the remaining 8–10 weeks. Home exercise consisted mainly of brisk walking at the prescribed HR for 30–50 min, 3–5 times per week. In addition, patients diagnosed with physical deconditioning engaged in low-intensity resistance training (LRT) using the patient’s own body weight. Exercises consisted of standing calf raises and sit-to-stand movements. Each repetition was performed for 8 s (3 s of concentric contraction, 1 s of maintaining full flexion, 3 s of eccentric contraction, and 1 s of maintaining a slightly flexed position), based on AHA recommendations.13,17 Subjects were instructed to continue breathing normally during the exercise. For each exercise, 3 sets of 7–10 repetitions were performed each day, 2–3 times per week.
Patients were encouraged to attend education classes, which were held 3 times each week. Classes consisted of lectures on coronary artery disease, secondary prevention of cardiac disease, HF management, diet, smoking cessation, and medicine-taking guidance given by physicians, nurses, dieticians, and pharmacists. In addition, all patients received individual counseling on exercise, secondary prevention, and daily life activities by a physician and nurse at the time of hospital discharge and at the end of the 3-month CR program.
CPXPatients underwent symptom-limited CPX using a cycle ergometer with respiratory gas exchange analysis when they were clinically stable at the beginning and at the end of the 3-month CR program. Testing consisted of an initial 2 min of rest, 1 min of warm-up (0-W load), and full exercise with an individualized ramp protocol at increments of 10–15 W/min. Expired gas analysis was performed throughout testing on a breath-by-breath basis. Minute ventilation (V̇E), oxygen uptake (V̇O2), and carbon dioxide production (V̇CO2) data were stored on a computer hard disk every 6 s for off-line analysis (AE-300S, Minato Medical Science, Osaka, Japan). V̇E was plotted against V̇CO2 to represent the V̇E/V̇CO2 slope, excluding the part after the respiratory compensation point when the slope starts to increase. Peak V̇O2 was determined as the highest V̇O2 during exercise (smoothed after a 5-point moving average) or the average V̇O2 of the last 3 data points (18 s) before termination of exercise, and is expressed as a value adjusted to body weight (mL/kg/min). Peak V̇O2 expressed as percentage of the age- and sex-predicted value (peak V̇O2 % predict) was calculated using the following formulae based on Japanese normal subjects:18,19
Peak V̇O2 % predict=measured absolute peak V̇O2 (mL/min)/body weight (kg)/(−0.38×age (years)+52.1)×100 (in male subjects); and
Peak V̇O2 % predict=measured absolute peak V̇O2 (mL/min)/body weight (kg)/(−0.23×age (years)+40.4)×100 (in female subjects).
Other Clinical VariablesAll patients underwent echocardiography before the start of CR while clinically stable under optical medical therapy. LV end-diastolic and end-systolic diameters were determined on 2-D echocardiography. LVEF was measured on echocardiography, radionuclide ventriculography, or LV angiography. Blood samples were drawn for standard measurements of serum creatinine, total cholesterol, albumin, hemoglobin, C-reactive protein (CRP), and plasma B-type natriuretic peptide (BNP) ≤3 days after CPX. BNP was measured using a validated commercially available immunoassay kit (Tosoh, Tokyo, Japan). A digital dynamometer (TKK 5101 Grip; Takei, Tokyo, Japan) was used to measure hand grip strength as an index of upper extremity muscle strength. Knee extensor strength was measured as an index of lower extremity muscle strength with a handheld dynamometer (μTas F-1; ANIMA, Tokyo, Japan). Patients sat on a bench and the dynamometer was fixed to a rigid bar. Three 5-s maximum isometric voluntary contractions of the quadriceps were performed. Muscle strength on the left and right sides was tested consecutively. A rest period of 30 s was provided between measurements.13 The maximum strength values on the left and right sides were averaged and expressed as an absolute value (kg). Values were then adjusted for body weight. After October 2009, we also evaluated body composition. Skeletal muscle mass and body fat mass were measured with a body composition analyzer (InBody720; Biospace, Seoul, Korea) using multi-frequency bioelectrical impedance analysis. A tetra-polar 8-point tactile electrode system was used. The system separately measured the impedance of the patient’s right arm, left arm, trunk, right leg, and left leg at 6 different frequencies (1, 5, 50, 250, 500, and 1,000 kHz) for each body segment. In accordance with the manufacturer’s guidelines, subjects wiped the bottoms of their feet with a proprietary electrolyte tissue before standing on the electrodes embedded in the scale platform of the analyzer. The subjects were instructed to stand upright and to grasp the handles of the analyzer, thereby providing contact with a total of 8 electrodes (2 for each foot and hand).20 The measured skeletal muscle mass (kg) and body fat mass (kg) were automatically displayed.
Data AnalysisFirst, we compared the clinical characteristics and CPX data of the low vs. normal BMI groups. Second, we calculated net change and percent change in peak V̇O2 (∆peak V̇O2 and %∆peak V̇O2) after CR. To investigate the characteristics of patients with a positive change in peak V̇O2 (%∆peak V̇O2 ≥0), we divided patients into the 2 groups according to whether peak V̇O2 increased or not, and compared their clinical characteristics. Finally, we investigated factors associated with a positive change in peak V̇O2.
Statistical AnalysisContinuous variables are expressed as mean±SD while variables with skewed distribution are expressed as median (IQR). Categorical variables are expressed as n (%). Groups were compared using Pearson’s chi-squared test for categorical variables and the unpaired t-test or Mann-Whitney U-test for continuous variables. Logistic regression was used to calculate OR and 95% CI for a positive change in peak V̇O2. We used BMI 18.5 kg/m2 as the cut-off to create a binary explanatory variable. Univariate logistic regression was used to identify parameters with a significant association with a positive change in peak V̇O2. Next, multiple logistic regression with a forced entry procedure was used with covariates with P<0.1 on univariate analysis to assess independent predictive factors for a positive change in peak V̇O2. Two-tailed P<0.05 was considered statistically significant. We use SPSS version 24.0 (IBM, Armonk, NY, USA) for all statistical analyses.
After excluding patients with high BMI (n=38; 20%), there were 152 patients in the outcome analysis (Figure 1), of whom 40 (26%) were women. Mean LVEF was 27±11%. We divided the patients into a low BMI group (<18.5 kg/m2; n=33) and a normal BMI group (≥18.5 kg/m2; n=119). Clinical characteristics of both groups are summarized in Table 1. There were no significant differences in age; prevalence of old myocardial infarction; albumin, CRP, and BNP; or LVEF between the 2 groups. The low BMI group, however, had a lower prevalence of cardiovascular risk factors such as hypertension (33% vs. 65%; P=0.001) and dyslipidemia (36% vs. 67%; P=0.001) and lower hemoglobin (12.0±1.5 g/dL vs. 13.0±1.7 g/dL; P=0.002) than the normal BMI group. As expected, the low BMI group had significantly less skeletal muscle mass (20±4 kg vs. 24±5 kg; P=0.003), adipose tissue mass (8±2 kg vs. 15±4 kg; P=0.001), and knee extensor strength (22±6 kg vs. 29±11 kg; P=0.001). After knee extensor strength was corrected by body weight, there was no difference between the 2 groups (P=0.649).
Subject selection. BMI, body mass index; HF, heart failure.
| All (n=152) |
Normal BMI (n=119) |
Low BMI (n=33) |
P-value | |
|---|---|---|---|---|
| Age (years) | 64±13 | 65±12 | 61±16 | 0.139 |
| Male | 112 (74) | 92 (77) | 20 (61) | 0.054 |
| BMI (kg/m2) | 20.7±2.6 | 21.7±1.8 | 17.1±1.1 | <0.001 |
| Old myocardial infarction | 56 (37) | 46 (39) | 10 (30) | 0.452 |
| Medical history | ||||
| Hypertension | 88 (58) | 77 (65) | 11 (33) | 0.001 |
| Dyslipidemia | 92 (61) | 80 (67) | 12 (36) | 0.001 |
| Diabetes mellitus | 43 (28) | 38 (32) | 5 (15) | 0.063 |
| BNP (pg/mL) | 227 (109–442) | 219 (107–415) | 273 (171–576) | 0.198 |
| Hemoglobin (g/dL) | 12.8±1.7 | 13.0±1.7 | 12.0±1.5 | 0.002 |
| Creatinine (mg/dL) | 1.2±0.6 | 1.2±0.6 | 1.1±0.5 | 0.624 |
| Total cholesterol (mg/dL) | 165±36 | 165±37 | 167±31 | 0.789 |
| Albumin (g/dL) | 4.1±0.3 | 4.1±0.3 | 4.0±0.4 | 0.138 |
| CRP (mg/dL) | 0.09 (0.03–0.69) | 0.09 (0.03–0.26) | 0.09 (0.05–0.22) | 0.909 |
| LVEF (%) | 27±11 | 28±11 | 24±11 | 0.085 |
| Body composition parameters (n=62) | ||||
| Skeletal muscle mass (kg) | 23±5 | 24±5 | 20±4 | 0.003 |
| Adipose tissue mass (kg) | 14±5 | 15±4 | 8±2 | 0.001 |
| Knee extensor muscle strength (kg) | 27±10 | 29±11 | 22±6 | 0.001 |
| Knee extensor muscle strength/BW | 0.50±0.13 | 0.50±0.14 | 0.48±0.10 | 0.649 |
| Hand grip strength (kg) | 30±9 | 31±9 | 27±7 | 0.052 |
| Hand grip strength/BW | 0.53±0.12 | 0.52±0.13 | 0.57±0.11 | 0.142 |
| Medication at baseline | ||||
| β-blocker | 134 (88) | 105 (88) | 29 (88) | 0.950 |
| ACEI or ARB | 113 (74) | 94 (79) | 19 (58) | 0.013 |
| Diuretic | 119 (78) | 92 (77) | 27 (82) | 0.570 |
| CPX data | ||||
| Peak respiratory exchange ratio | 1.27±0.13 | 1.26±0.11 | 1.32±0.16 | 0.011 |
| Resting heart rate (beats/min) | 71±13 | 71±13 | 72±9 | 0.636 |
| Resting SBP (mmHg) | 108±18 | 110±17 | 100±19 | 0.005 |
| Resting DBP (mmHg) | 67±12 | 67±11 | 65±12 | 0.318 |
| Peak heart rate (beats/min) | 121±29 | 122±30 | 118±26 | 0.501 |
| Peak SBP (mmHg) | 138±29 | 141±27 | 126±32 | 0.010 |
| Peak work rate (W) | 85±30 | 89±30 | 69±25 | <0.001 |
| Peak V̇O2 (mL/min) | 955±348 | 1,016±354 | 732±212 | <0.001 |
| Peak V̇O2/BW (mL/min/kg) | 17.2±4.5 | 17.5±4.7 | 15.9±3.7 | 0.080 |
| Peak V̇O2 % predict (%) | 64±14 | 65±14 | 59±14 | 0.026 |
| V̇E/V̇CO2 slope (%) | 34.3±6.9 | 33.4±6.5 | 37.6±7.4 | 0.002 |
| Low-intensity resistance training | 21 (14) | 13 (11) | 8 (24) | 0.050 |
| CR attendance during hospitalization (days) | 8±7 | 7±5 | 10±11 | 0.039 |
| CR attendance over 3 months (days) | 19±9 | 18±9 | 21±10 | 0.169 |
| Duration of hospital stay (days) | 48±83 | 38±52 | 84±146 | 0.095 |
Data given as mean±SD, median (IQR) or n (%). ACEI, angiotensin-converting enzyme inhibitor; ARB, angiotensin receptor blocker; BMI, body mass index; BNP, brain natriuretic peptide; BW, body weight; CPX, cardiopulmonary exercise; CR, cardiac rehabilitation; CRP, C-reactive protein; DBP, diastolic blood pressure; LVEF, left ventricular ejection fraction; SBP, systolic blood pressure; V̇CO2, carbon dioxide production; V̇E, minute ventilation; V̇O2, oxygen uptake.
Notably, the low BMI group had lower systolic blood pressure at rest (100±19 mmHg vs. 110±17 mmHg; P=0.005) and a lower frequency of angiotensin-converting enzyme inhibitor (ACEI) or angiotensin receptor blocker (ARB) use (58% vs. 79%; P=0.013) than the normal BMI group. The low BMI group attended more CR sessions during hospitalization (10±11 days vs. 7±5 days; P=0.039) and underwent LRT (24% vs. 11%; P=0.050) more frequently than the normal BMI group.
Exercise CapacityPeak respiratory exchange ratio was sufficiently high (average ≥1.2) both before and after CR in both groups. Table 1 lists CPX data for the low and normal BMI groups before CR. Peak work rate (69±25 W vs. 89±30 W; P<0.001) and peak V̇O2 (732±212 mL/min vs. 955±348 mL/min; P<0.001) were lower in the low BMI group than in the normal BMI group. Peak V̇O2 % predict was also significantly lower in the low BMI group than in the normal BMI group (59±14% vs. 65±14%; P=0.026).
Post-CR ChangesSkeletal muscle and fat mass increased more in the low BMI group than in the normal BMI group, but this difference was not significant (%∆skeletal muscle mass: 4.6±1.4% vs. 2.4±0.8%; P=0.181; %∆fat mass: 9.9±20.9% vs. 1.8±14.8%; P=0.109). Similarly, knee extensor and hand grip strength improved more in the low BMI group than in the normal BMI group, but this difference was not significant (%∆knee extensor strength: 22.1±4.3% vs. 15.4±1.7%; P=0.110; %∆hand grip strength: 12.9±4.5% vs. 7.1±1.7%; P=0.151). Figure 2 shows the change in exercise capacity after the CR program in the 2 groups. Patients with low BMI had significantly higher %∆peak V̇O2 (17.1±2.8% vs. 7.8±1.5%; P<0.001) and %∆peak work rate (20.2±3.5% vs. 8.7±1.8%; P<0.001) than the normal BMI group. A higher percentage of patients in the low BMI group had a positive change in peak V̇O2 (%∆peak V̇O2 ≥0%) than in the normal BMI group (91% vs. 70%, P=0.010). Furthermore, the change in peak V̇O2 was negatively correlated with BMI at baseline (r=−0.214, P=0.008). Of the patients with low BMI, the group with LRT had a greater improvement in peak V̇O2 than the group without LRT, but there were no significant difference, probably due to the small sample size (24.2±18.3% vs. 14.9±17.9%; P=0.208).
Change in exercise capacity after the cardiac rehabilitation program in heart failure patients according to body mass index (BMI) status: (A) percent change in peak oxygen uptake (%∆peak V̇O2); (B) percent change in work rate (%∆peak work rate); and (C) percentage of patients with a positive change in peak V̇O2.
To further identify whether changes in some parameters are associated with improved exercise tolerance, we evaluated the correlation between changes in parameters and changes in peak V̇O2. Change in peak V̇O2 after the CR program was significantly associated with change in skeletal muscle mass (r=0.311; P=0.014), fat mass (r=0.273; P=0.032), knee extensor strength (r=0.477; P<0.001), and hand grip strength (r=0.292; P=0.007). There were no significant correlations between change in peak V̇O2 and change in total cholesterol (P=0.093), CRP (P=0.993), or albumin (P=0.217).
Predictors of a Positive Change in Peak V̇O2Regarding the characteristics of patients with a positive change (n=113) vs. a decrease (n=39) in peak V̇O2, only age and prevalence of low BMI were significantly different (Table 2). On univariate analysis to identify the predictors of a positive change in peak V̇O2 after CR, age and low BMI were significantly associated with a positive change in peak V̇O2. Using the forced inclusion model with covariates with P<0.1 on univariate analysis, on multiple logistic regression age (OR, 0.94; 95% CI: 0.91–0.98; P=0.004) and low BMI (OR, 3.97; 95% CI: 1.10–14.31; P=0.035) were independent predictors for a positive change in peak V̇O2 after CR (Table 3).
| Decrease in peak V̇O2 (n=39) |
Increase in peak V̇O2 (n=113) |
P-value | |
|---|---|---|---|
| Age (years) | 70±10 | 62±13 | 0.001 |
| Male | 27 (69) | 85 (75) | 0.464 |
| BMI (kg/m2) | 21.2±1.9 | 20.5±2.7 | 0.693 |
| Low BMI | 3 (7.7) | 30 (27) | 0.014 |
| Old myocardial infarction | 18 (46) | 38 (33) | 0.291 |
| Medical history | |||
| Hypertension | 24 (62) | 64 (57) | 0.593 |
| Dyslipidemia | 27 (69) | 65 (58) | 0.197 |
| Diabetes mellitus | 10 (27) | 33 (31) | 0.685 |
| BNP (pg/mL) | 226 (134–420) | 231 (102–444) | 0.694 |
| Hemoglobin (g/dL) | 12.6±1.5 | 12.8±1.7 | 0.314 |
| Creatinine (mg/dL) | 1.2±0.6 | 1.2±0.6 | 0.551 |
| Total cholesterol (mg/dL) | 163±39 | 166±35 | 0.617 |
| Albumin (g/dL) | 4.2±0.3 | 4.1±0.3 | 0.281 |
| CRP (mg/dL) | 0.08 (0.03–0.24) | 0.09 (0.03–0.24) | 0.517 |
| LVEF (%) | 29±12 | 27±11 | 0.385 |
| Body composition parameters (n=62) | |||
| Skeletal muscle mass (kg) | 22±4 | 24±5 | 0.078 |
| Adipose tissue mass (kg) | 14±4 | 13±5 | 0.673 |
| Knee extensor muscle strength (kg) | 27±9 | 28±10 | 0.683 |
| Knee extensor muscle strength/BW | 0.49±0.13 | 0.49±0.13 | 0.890 |
| Hand grip strength (kg) | 29±8 | 30±9 | 0.576 |
| Hand grip strength/BW | 0.51±0.12 | 0.54±0.13 | 0.425 |
| Medication at baseline | |||
| β-blocker | 31 (80) | 103 (91) | 0.052 |
| ACEI or ARB | 27 (69) | 86 (76) | 0.397 |
| Diuretic | 29 (74) | 90 (80) | 0.490 |
| CPX data | |||
| Peak respiratory exchange ratio | 1.25±0.11 | 1.28±0.12 | 0.258 |
| Resting heart rate (beats/min) | 71±15 | 72±12 | 0.928 |
| Resting SBP (mmHg) | 107±17 | 108±18 | 0.749 |
| Resting DBP (mmHg) | 66±13 | 67±11 | 0.562 |
| Peak heart rate (beats/min) | 116±30 | 113±29 | 0.183 |
| Peak SBP (mmHg) | 134±27 | 139±29 | 0.293 |
| Peak work rate (W) | 80±21 | 87±33 | 0.227 |
| Peak V̇O2 (mL/min) | 904±232 | 972±379 | 0.288 |
| Peak V̇O2/BW (mL/min/kg) | 16.6±3.2 | 17.3±4.9 | 0.415 |
| Peak V̇O2 % predict (%) | 66.8±12.9 | 62.4±14.6 | 0.096 |
| V̇E/V̇CO2 slope (%) | 35.3±6.7 | 33.9±7.0 | 0.299 |
| Low-intensity resistance training | 4 (10) | 17 (15) | 0.445 |
| CR attendance during hospitalization (days) | 6±6 | 8±7 | 0.153 |
| CR attendance over 3 months (days) | 18±10 | 19±7 | 0.605 |
| Duration of hospital stay (days) | 64±152 | 42±40 | 0.163 |
Data given as mean±SD, median (IQR) or n (%). Abbreviations as in Table 1.
| Univariate analysis | Multivariate analysis | |||||
|---|---|---|---|---|---|---|
| OR | 95% CI | P-value | OR | 95% CI | P-value | |
| Age (per 1-year increment) | 0.94 | 0.91–0.98 | 0.002 | 0.94 | 0.91–0.98 | 0.004 |
| Male | 1.35 | 0.59–2.98 | 0.465 | |||
| Old myocardial infarction | 0.61 | 0.29–1.28 | 0.188 | |||
| Hypertension | 0.82 | 0.39–1.72 | 0.593 | |||
| Dyslipidemia | 0.60 | 0.27–1.30 | 0.199 | |||
| Diabetes mellitus | 1.19 | 0.53–2.83 | 0.680 | |||
| Low BMI | 4.34 | 1.24–15.13 | 0.021 | 3.97 | 1.10–14.31 | 0.035 |
| Hemoglobin (per 1-g/dL increment) | 1.12 | 0.90–1.41 | 0.313 | |||
| BNP (per 1-pg/mL increment) | 1.00 | 0.99–1.00 | 0.691 | |||
| Total cholesterol (per 1-mg/dL increment) | 1.00 | 0.99–1.01 | 0.614 | |||
| Albumin (per 1-g/dL increment) | 0.46 | 0.12–1.86 | 0.279 | |||
| CRP (per 1-mg/dL increment) | 1.08 | 0.70–1.67 | 0.720 | |||
| LVEF (per 1% increment) | 0.99 | 0.96–1.02 | 0.384 | |||
| Skeletal muscle mass (per 1-kg increment) | 1.10 | 0.97–1.25 | 0.140 | |||
| Knee extensor muscle strength (per 1-kg increment) | 1.01 | 0.98–1.05 | 0.696 | |||
| Knee extensor muscle strength/BW (per 1-unit increment) | 1.26 | 0.08–19.10 | 0.889 | |||
| Peak V̇O2 (per 1-mL/min increment) | 1.00 | 1.00–1.00 | 0.287 | |||
| Peak V̇O2/BW (per 1-mL/min/kg increment) | 1.04 | 0.95–1.13 | 0.413 | |||
| Peak V̇O2 % predict (per 1% increment) | 0.98 | 0.94–1.00 | 0.098 | 0.99 | 0.96–1.01 | 0.301 |
| Low-intensity resistance training | 1.56 | 0.50–4.92 | 0.458 | |||
| CR attendance in 3 months (per 1-session increment ) | 1.01 | 0.97–1.05 | 0.602 | |||
| CR attendance in the hospital (per 1-session increment ) | 1.05 | 0.98–1.12 | 0.160 | |||
| Duration of hospital stay (per 1-day increment) | 0.99 | 0.99–1.01 | 0.212 | |||
Abbreviations as in Table 1.
The present study had 2 major findings: (1) a significantly greater proportion of HF patients with low BMI have a positive change in peak V̇O2 after CR and average %∆peak V̇O2 is significantly greater in patients with low vs. normal BMI; and (2) low BMI is an independent predictor of a positive change in peak V̇O2 after CR.
Exercise Capacity of Low-BMI HF PatientsPatients with HF have exercise intolerance due to dyspnea and fatigue, and exercise capacity is the best predictor of survival in patients with HF.21 HF is a syndrome of circulatory failure secondary to ventricular dysfunction, followed by a variety of neurohumoral, peripheral circulatory, respiratory and skeletal muscle adaptations.22 In most patients with HF, skeletal muscle dysfunction rather than cardiac dysfunction limits exercise capacity.23 Skeletal muscle impairment in HF patients is caused by not only physical deconditioning with excessive rest but also HF-induced systemic changes, such as inflammation, catabolism, and high levels of catecholamines and renin activity.24
In the present study, HF patients with low BMI had lower exercise capacity than patients with normal BMI. The low BMI group had less skeletal muscle mass and lower knee extensor strength, suggesting that skeletal muscle impairment might be strongly related to low exercise capacity in patients with low BMI. Regarding factors related to HF severity, inflammation, and catabolism, the 2 groups had similar LVEF and levels of BNP, hemoglobin, CRP, cholesterol, and albumin. Patients with low BMI had a similar severity of HF and had similar HF-induced systemic changes as patients with normal BMI, so we speculate that the greater skeletal muscle impairment in patients with low BMI is attributable to more sereve physical deconditioning. Patients with low BMI tended to stay in the hospital longer than patients with normal BMI (84 days vs. 38 days; P=0.095). Physical deconditioning could have played an important role in skeletal muscle impairment in patients with low BMI.
CR for Low-BMI HF PatientsExercise training in patients with HF is an accepted adjunct in an evidence-based management program. Exercise training in HF patients has produced improvements of 18% in peak V̇O2.25 CR also has beneficial effects on HF-induced skeletal myopathy, and is an effective exercise therapy for patients with low body weight.26–28 Those studies, however, did not focus on patients with low BMI.
The present study has shown that a greater proportion of patients with low BMI have a positive change in peak V̇O2 after CR, and that average %∆peak V̇O2 is significantly greater in patients with low vs. normal BMI (Figure 2). Moreover, low BMI was an independent predictor of a positive change in peak V̇O2 after CR (Table 3). We would infer that exercise intolerance due to physical deconditioning and/or malnutrition was more severe but the improvement in exercise capacity after CR was relatively greater in the low BMI group than in the normal BMI group. In the present study, change in peak V̇O2 was significantly correlated with change in muscle mass and strength, but not with change in factors related to inflammation or catabolism. This suggests that improvement in exercise capacity is more directly related to improvement in the physical function of skeletal muscles than to inflammation or catabolism. Moreover, the effect of CR in the low BMI group could be related to LRT, which has been shown to improve exercise capacity in patients with HF.29,30 LRT training was prescribed to lean patients based on physician discretion at the present institution, and patients with low BMI underwent LRT more frequently than patients with normal BMI. The use of LRT, however, was not a significant predictor of %∆peak V̇O2 on multivariate analysis.
While previous studies found that lower exercise tolerance is associated with a greater effect of CR,31,32 in the present study, baseline peak V̇O2 was not significantly different between patients who had a decrease vs. increase in peak V̇O2 (Table 2). One possible explanation for this discrepancy is that patients with an increase in peak V̇O2 were much younger than patients with a decrease in peak V̇O2 (62±13 years vs. 70±10 years; P=0.001).
In addition, the rate of medication use in the low BMI group was lower due to lower blood pressure than in the normal BMI group. As noted in a previous report,7 patients with low BMI in the present study were at risk for not receiving optimal drug therapy. Therefore, CR is more effective and necessary for HF patients with low BMI than those with normal BMI.
Clinical ImplicationsThe number of patients with sarcopenia or frailty has been rapidly increasing worldwide because of population aging. Sarcopenia and frailty are related to exercise capacity and prognosis in patients with HF,33,34 making them major health problems. In the present study, we did not evaluate the prevalence of sarcopenia or frailty, therefore it is difficult to discuss the direct relationship between the present results and sarcopenia or frailty. The low BMI group, however, had less skeletal muscle mass and lower muscle strength than the normal BMI group, which implied that the low BMI group had a higher prevalence of sarcopenia or frailty. Therefore, the present results suggest that CR can be effective for HF patients with sarcopenia or frailty; this effectiveness of CR is worth assessing in further studies.
Study LimitationsFirst, this was a small observational cohort study in a relatively short period. Second, this study was performed at a single cardiology hospital. Although the CR program used a guideline-directed standardized protocol, there might have remained a possibility of referral bias. Third, we did not examine nutrition and sarcopenia status, which can be related to exercise capacity in this population.6 Future studies should consider the relationship between these factors and exercise capacity in HF patients with low BMI. Fourth, because we studied patients who underwent CPX of those admitted with HF, they were relatively young and the prognosis was less poor compared with general HF patients.1 It remains uncertain whether the present results could apply to patients with advanced HF. Finally, we did not evaluate quality of life as an outcome. Quality of life measurements need to be included in future trials.
HF patients with low BMI are at greater risk of not being treated with optimal drug therapy, but had a greater beneficial effect from CR than patients with normal BMI. Notably, low BMI is an independent predictor of a positive change in peak V̇O2. Thus, CR should be strongly recommended for HF patients with low BMI.
The authors are grateful to Ayumi Otani and Keiko Katayama for technical support.
The authors declare no conflicts of interest.
