2018 Volume 82 Issue 4 Pages 1033-1040
Background: Skeletal muscle pump function may play a key role in maintaining cardiac output (CO), because of the lack of cardiac contractility reserve during incremental exercise in heart failure (HF) patients. We aimed to investigate the relationship between lower leg pump function and surrogate measures of CO during cardiopulmonary exercise testing (CPX) in HF patients.
Methods and Results: Consecutive cardiac patients referred for CPX had their lower leg ejection fraction (LgEF) measured using strain gauge plethysmography as a marker of skeletal muscle pump function. We analyzed 88 patients, including 65 HF patients and 23 control subjects. Unlike the control subjects, LgEF correlated with peak oxygen consumption (V̇O2) and peak oxygen (O2) pulse (peak V̇O2: r=0.280, P=0.024; peak O2 pulse: r=0.540, P<0.001) in HF patients. Significant relationships among LgEF, peak V̇O2, and peak O2 pulse were observed in HF patients with reduced EF (peak V̇O2: r=0.367, P=0.026; peak O2 pulse: r=0.658, P<0.001), whereas LgEF in HF patients with preserved EF showed a weak correlation only with peak O2 pulse (r=0.407, P=0.032). LgEF was selected as an independent determinant of peak V̇O2 (β=0.187, P=0.036) and peak O2 pulse (β=0.520, P<0.001) in HF patients.
Conclusions: Lower leg skeletal muscle function may contribute to exercise capacity through an indirect mechanism on cardiac function in HF.
Reduced exercise capacity is one of the major manifestations in heart failure (HF) patients, and is widely recognized as a powerful prognostic factor of death.1–3 Added to this, in HF the peak oxygen consumption (V̇O2) has been reported as directly related to peak cardiac output (CO), which is maintained by augmentation of cardiac contractility, venous return, and heart rate (HR) during exercise.4–6 Another key determinant for exercise capacity is skeletal muscle function. In HF, muscle mass is known to be reduced in accordance with the disease process, and a strong association between skeletal muscle strength and peak V̇O2 is evident.7–9 However, the effect of muscle function on cardiac pump function remains unknown.
During exercise in an upright position, skeletal muscle function can contribute to CO through a marked blood volume shift from the legs to the heart, known as venous return or cardiac preload.10 In particular, HF patients with impaired cardiac function, or even in normal aged subjects, muscle pump function is essential for maintaining CO, because cardiac contractility does not increase sufficiently or decrease with incremental exercise.11,12 Thus, theoretically, the lower leg muscle pump can indirectly increase CO by increasing the venous return and thereby the cardiac preload. Despite this awareness, the role and correlating factors of the skeletal muscle pump in HF have not been established.
This study, therefore, aimed to clarify the function of the skeletal muscle pump by analyzing the association between lower leg ejection fraction (LgEF) measured by strain gauge plethysmography (SPG) and surrogate measures of CO during cardiopulmonary exercise testing (CPX) in HF patients. Additionally, we aimed to identify the correlated factors of the skeletal muscle pump.
We consecutively enrolled clinically stable cardiac patients who visited cardiac rehabilitation and underwent CPX to evaluate the effect of cardiac rehabilitation on exercise tolerance at Nagoya Ekisaikai Hospital from April 2013 to March 2014. Patients with primary hemodynamically significant uncorrected valvular heart disease, chronic obstructive pulmonary disease, primary renal or hepatic disease, complex congenital heart disease, constrictive pericarditis, infiltrative, or restrictive, or hypertrophic cardiomyopathy, and duplicate cases were excluded from this study. We measured LgEF in all enrolled patients except those with venous disease as indicated by physical examination or past history, orthopedic problems, and leg circumference >41 cm (technical limitation of strain gauge size). All participants in this study had either a negative myocardial perfusion scan or a negative exercise electrocardiogram (ECG).
Written informed consent was given by each patient. The study protocol conformed to the ethical guidelines of the 1975 Declaration of Helsinki as reflected in a priori approval by the human (ethics) subjects committee of Ekisaikai Hospital (approved no. 2012-047) and Nagoya University (approval no. 12-513).
Study DesignThe study was prospectively designed as an observational controlled study. All patients underwent CPX after ultrasound cardiography (UCG) and laboratory tests on the same day, and the LgEF measurement was performed within 7 days before or after CPX. LgEF was measured by the same examiner who showed high test-retest reliability of LgEF measurement (intraclass correlation coefficients: 0.78) prior to the study. To avoid measurement bias, the LgEF examiner was unaware of the results of CPX for each subject. The supervisor who decided on termination of CPX was also blinded to the LgEF data of each patient.
We analyzed the association between LgEF and CPX parameters, firstly in all subjects and then in subgroups based on left ventricular EF (LVEF) and prior history of HF hospitalization or B-type natriuretic peptide (BNP) as follows.
HF with reduced EF (HFrEF): LVEF <45%; HF with preserved EF (HFpEF): LVEF ≥45% and prior history of HF hospitalization with evidence of pulmonary congestion or BNP ≥100 pg/ml; controls: LVEF ≥45% and no prior history of HF hospitalization and BNP <100 pg/ml.13 We combined HFrEF and HFpEF as HF (Figure 1).
Classification of the patients in this study. BNP, B-type natriuretic peptide; HF, heart failure; HFrEF, HF with reduced EF; HFpEF, HF with preserved EF; LVEF, left ventricular ejection fraction.
CPX was performed using a cycle ergometer with breath-by-breath respiratory gas measurements using a computerized metabolic cart (AE-300S, Minato Medical Science, Osaka, Japan). After a 4-min rest on the ergometer, exercise began with a 3-min warm-up at 20 watts and 50 repetitions/min followed by 10 watts ramp loading every minute. ECG, HR, V̇O2, and carbon dioxide production (V̇CO2) were continuously monitored throughout the study, and blood pressure (BP) was measured every minute during exercise. The test was terminated when the patient showed (1) maximal volitional fatigue, (2) V̇O2 levelling off, or (3) excessive BP (systolic BP >250 mmHg or diastolic BP >115 mmHg). Peak oxygen (O2) pulse was calculated to divide peak V̇O2 by peak HR to estimate peak stroke volume during exercise.14
LgEF MeasurementLgEF was measured using SPG (EC6 Strain Gauge, Hokanson, Bellevue, WA, USA) as a function of the skeletal muscle pump (Figure 2). The adopted measurement protocol was proposed by Nicolaides et al.15 Before measurement, patients rested supine for 30 min with the right leg elevated more than 30° to gain adequate venous shrinkage, and the strain gauge was attached to the site of greatest circumference of right lower leg. Next, the patients were asked to stand up with a slightly flexed right knee and weight on the left leg. After an increase in leg volume was observed because of venous filling, blood volume changes were monitored in increments of 0.1%. If the volume did not change for more than 10 s, it was defined as the plateau level. When a plateau was reached, the patient was asked to do one calf raise movement with the weight on the right leg and return to the initial position. This one calf raise movement was repeated three times, and venous volume (VV) and expelled volume (EV) were determined. The LgEF of the lower extremity was derived from the formula LgEF=(EV/VV)*100. The average of 3 results was calculated and used for LgEF analysis. EV does not accurately reflect the actual venous return from the whole lower leg because EV is known to be affected by body size. LgEF has been recommended for use as a measure of lower skeletal muscle function,15 so we adopted the parameters of LgEF for the main analysis.
Diagrammatic representation of lower leg ejection fraction (LgEF) measurement. (A) Patient rested supine with the right leg elevated for 30 min and strain gauge was attached to the right lower leg. (B) Patient stood up with slightly flexed right knee and the weight on the left leg. (C) After plateau was reached, the patient did a tiptoe movement with the weight on the right leg and returned to the initial position. LgEF is derived from the formula LgEF=(EV/VV)*100. The average of 3 results was calculated and used for LgEF analysis. EV, expelled volume; VV, venous volume.
Echocardiography was performed with an Aplio XV (Toshiba, Tokyo, Japan) before CPX on the same day. LV end-diastolic dimension (LVDd) and left atrial dimension (LAd) were measured from standard M-mode, and LVEF was calculated using the biplane Simpson method. Peak early transmitral velocity (E) was obtained from transmitral flow, and peak early diastolic mitral annular velocity (E’) was also measured with pulsed-wave Doppler. E/E’ was calculated to estimate the LV filling pressure.
Laboratory ExaminationHemoglobin, serum creatinine, serum albumin, and BNP levels were also measured before CPX on the same day.
Statistical AnalysisContinuous variables are expressed as mean±standard deviation (SD) or median with lower and upper quartiles as appropriate. Data were compared between group differences using analysis of variance followed by Tukey-Kramer honest significant difference tests for normally distributed data and Kruskal-Wallis test with posthoc Steel-Dwass tests for skewed variables. Categorical variables were analyzed using chi-square test adjusted using the Bonferroni correction.
The association between LgEF and peak V̇O2 or peak O2 pulse was analyzed by Pearson’s correlation coefficient. Univariate and multivariate linear regression analyses were performed to assess predictors of peak V̇O2, peak O2 pulse, and LgEF. When collinearity of r≥0.50 or r≤−0.50 was not found between parameters, the parameters were selected as appropriate.
All statistical analyses were performed using JMP 10.0 software (SAS Institute, Cary, NC, USA) and P<0.05 was considered statistically significant.
Of the 104 subjects who were consecutively enrolled, we censored 12 duplicate cases and 4 cases of primary hemodynamically significant uncorrected valvular disease cases, so the data for 88 patients were analyzed in this study. Of these, 37 patients were classified as HFrEF, 28 as HFpEF, and 23 as controls (Figure 1). The baseline characteristics are listed in the Table 1. BNP was higher in HFrEF than in HFpEF patients and higher in HFpEF patients than in controls. Among the CPX parameters, all groups demonstrated an average peak RER >1.15, indicating maximum or near maximum exercise stress across the 3 groups. HR and systolic BP responses (baseline and peak) were not significantly different between groups except for peak systolic BP (lower in HFrEF and HFpEF patients when compared with controls). Peak V̇O2 was reduced in both HFrEF (15.6±4.1 mL/min/kg) and HFpEF (16.6±3.9 mL/min/kg) patients when compared with controls (19.8±4.6 mL/min/kg). Average values of LgEF in the HFrEF, HFpEF, and control groups were 47.0±22.8%, 43.9±21.6% and 49.4±22.0%, respectively, and were not different between groups.
| HF (n=65) |
HFrEF (n=37) |
HFpEF (n=28) |
Controls (n=23) |
|
|---|---|---|---|---|
| Age (years) | 67 (61–75) | 65 (59–73) | 71 (63–77) | 66 (60–73) |
| Male, n (%) | 54 (83.1) | 34 (91.9) | 20 (71.4) | 16 (45.7) |
| BMI (kg/m2) | 23.6±3.9 | 23.7±4.3 | 23.4±3.3 | 23.4±3.3 |
| NYHA functional class I/II/III | 3/39/24 | 2/18/17 | 1/21/7 | |
| Ischemic etiology of HF (%) | 34 (52.3) | 18 (48.6) | 16 (57.1) | 17 (48.6) |
| Atrial fibrillation (%) | 16 (24.6) | 7 (18.9) | 9 (32.1) | 3 (8.6) |
| Diabetes mellitus (%) | 37 (56.9) | 24 (64.9)* | 13 (46.4) | 4 (11.4) |
| Hypertension (%) | 43 (66.2) | 21 (56.8) | 22 (78.6) | 12 (34.3) |
| Medications, n (%) | ||||
| β-blocker | 58 (89.2) | 34 (91.9)* | 24 (85.7) | 15 (42.9) |
| ACEI or ARB | 54 (83.1) | 32 (86.5) | 22 (78.6) | 17 (48.6) |
| Aldosterone antagonist | 34 (52.3) | 24 (64.9)* | 10 (35.7)† | 1 (2.9) |
| Diuretic | 43 (66.2) | 29 (78.4)* | 14 (50.0)† | 3 (8.6) |
| Laboratory data | ||||
| Hemoglobin (g/dL) | 13.2 (11.5–14.3) | 13.3 (11.9–14.9) | 12.6 (10.3–13.8) | 13.7 (12.2–14.8) |
| Creatinine (mg/dL) | 1.1 (0.9–1.4) | 1.0 (0.8–1.3)* | 1.1 (0.9–1.7)† | 0.8 (0.7–0.9) |
| Albumin (g/dL) | 3.7±0.4 | 3.8±0.4 | 3.5±0.4†,‡ | 3.9±0.4 |
| BNP (pg/mL) | 162 (107–390) | 255 (125–529)* | 132 (62–219)†,‡ | 54 (41–68) |
| Echocardiography parameters | ||||
| LVEF (%) | 42.9 (28.0–54.4) | 30.7 (20.2–41.1)* | 57.0 (49.4–68.4)‡ | 64.7 (57.8–70.0) |
| LVDd (mm) | 55.0 (48.3–63.7) | 62.9 (55.2–70.6)* | 48.3 (46.4–54.0)‡ | 45.3 (41.5–51.2) |
| LAd (mm) | 42.4 (39.1–45.9) | 43.6 (40.6–48.0)* | 40.4 (33.7–43.5)‡ | 35.1 (32.4–40.1) |
| E/E’ | 13.9 (9.8–17.8) | 14.8 (10.7–18.1)* | 12.1 (9.2–17.7) | 9.6 (8.8–13.1) |
| CPX parameters | ||||
| Baseline HR (beats/min) | 74 (68–83) | 74 (65–85) | 75 (69–81) | 68 (64–79) |
| Peak HR (beats/min) | 124 (107–141) | 129 (106–147) | 118 (106–133) | 133 (120–145) |
| Baseline systolic BP (mmHg) | 123 (108–141) | 122 (104–142) | 126 (110–141) | 139 (120–150) |
| Peak systolic BP (mmHg) | 164 (140–194) | 153 (139–196)* | 168 (143–194)† | 197 (185–218) |
| Peak V̇O2 (mL/min/kg) | 16.0±4.0 | 15.6±4.1* | 16.6±3.9† | 19.8±4.6 |
| Peak O2 pulse (mL/beat) | 8.2±2.4 | 7.9±2.5 | 8.5±2.4 | 9.3±2.3 |
| Peak RER | 1.18 (1.14–1.25) | 1.20 (1.17–1.28) | 1.17 (1.11–1.23) | 1.21 (1.16–1.25) |
| LgEF (%) | 45.7±22.2 | 47.0±22.8 | 43.9±21.6 | 49.4±22.0 |
| VV (mL/100 mL) | 4.3±1.6 | 3.9±1.3 | 4.9±1.9 | 4.3±1.7 |
| EV (mL/100 mL) | 1.9±1.1 | 1.8±1.1 | 2.0±1.2 | 1.9±0.9 |
Data are presented as mean±standard deviation, median (with lower and upper quartiles), or numbers (with percentages), where appropriate. *P<0.05 between HFrEF and controls. †P<0.05 between HFpEF and controls. ‡P<0.05 between HFrEF and HFpEF. ACEI, angiotensin-converting enzyme inhibitor; ARB, angiotensin-receptor blocker; BMI, body mass index; BNP, B-type natriuretic peptide; BP, blood pressure; CPX, cardiopulmonary exercise testing; E/E’, ratio of early transmitral velocity to early diastolic mitral annular velocity; EV, expelled volume; HF, heart failure; HFrEF, heart failure with reduced ejection fraction; HFpEF, heart failure with preserved ejection fraction; HR, heart rate; LAd, left atrial dimension; LgEF, lower leg ejection fraction; LVEF, left ventricular ejection fraction; LVDd, left ventricular end-diastolic dimension; NYHA, New York Heart Association; O2, oxygen; RER, respiratory exchange ratio; V̇O2, oxygen consumption; VV, venous volume.
In the HF patients, LgEF significantly and positively correlated with peak V̇O2 (r=0.280, P=0.024) and peak O2 pulse (r=0.540, P<0.001), but these associations were not observed in the control subjects (peak V̇O2: r=0.040, P=0.858; peak O2 pulse: r=0.154, P=0.483). When analyzed by HF subgroups, a numerically stronger association was found between LgEF and peak O2 pulse (r=0.658, P<0.001) in HFrEF (Figure 3).
Relationship between LgEF and peak V̇O2, and LgEF and peak O2 pulse. Relationship in patient with HF (Upper), with HFrEF (Middle), and with HFpEF (Lower). HF, heart failure; LgEF, lower leg ejection fraction; O2, oxygen; V̇O2, oxygen consumption.
In the univariate and multivariate linear regression analyses, we excluded controls because of their high peak V̇O2 level in which peak V̇O2 was limited other than by reduced cardiac function.14 The results of univariate and multivariate linear regression analyses for peak V̇O2 and peak O2 pulse are listed in Table 2.
| Univariate | Multivariate | ||||||
|---|---|---|---|---|---|---|---|
| R | P value | Unstandardized coefficients (95% CI) |
SE | Standardized coefficients (β) |
t statistic | P value | |
| Peak V̇O2* | |||||||
| Hemoglobin | 0.153 | 0.224 | |||||
| LVEF | 0.263 | 0.034 | 0.029 (−0.011 to 0.069) | 0.020 | 0.130 | 1.45 | 0.153 |
| LVDd | −0.307 | 0.013 | |||||
| LAd | −0.449 | <0.001 | −0.133 (−0.240 to −0.026) | 0.054 | −0.231 | −2.48 | 0.016 |
| E/E’ | −0.385 | 0.002 | −0.080 (−0.186 to 0.026) | 0.053 | −0.135 | −1.51 | 0.137 |
| Baseline HR | 0.100 | 0.430 | |||||
| Peak HR | 0.507 | <0.001 | 0.045 (0.018 to 0.071) | 0.013 | 0.301 | 3.37 | 0.001 |
| Baseline systolic BP | 0.310 | 0.014 | |||||
| Peak systolic BP | 0.563 | <0.001 | 0.037 (0.018 to 0.056) | 0.009 | 0.359 | 3.89 | <0.001 |
| Peak RER | 0.049 | 0.701 | |||||
| LgEF | 0.280 | 0.024 | 0.035 (0.002 to 0.067) | 0.016 | 0.187 | 2.15 | 0.036 |
| Peak O2 pulse** | |||||||
| Hemoglobin | 0.132 | 0.294 | |||||
| LVEF | 0.266 | 0.032 | 0.023 (−0.004 to 0.050) | 0.013 | 0.173 | 1.71 | 0.093 |
| LVDd | −0.101 | 0.423 | |||||
| LAd | −0.011 | 0.931 | |||||
| E/E’ | −0.266 | 0.033 | −0.041 (−0.116 to 0.033) | 0.037 | −0.118 | −1.12 | 0.268 |
| Baseline HR | −0.354 | 0.004 | |||||
| Peak HR | −0.176 | 0.160 | −0.029 (−0.048 to −0.010) | 0.009 | −0.331 | −3.12 | 0.003 |
| Baseline systolic BP | 0.018 | 0.889 | |||||
| Peak systolic BP | 0.275 | 0.028 | 0.013 (−0.001 to 0.026) | 0.007 | 0.208 | 1.90 | 0.063 |
| Peak RER | 0.020 | 0.872 | |||||
| LgEF | 0.540 | <0.001 | 0.057 (0.034 to 0.080) | 0.011 | 0.520 | 5.04 | <0.001 |
*R2=0.616. **R2=0.437. CI, confidence interval; HR, heart rate; SE, standard error. Other abbreviations as in Table 1.
When performing multivariate analyses by items with P<0.1 in the univariate analyses and peak HR, LgEF was extracted as an independent determinant of both peak V̇O2 (β=0.187, P=0.036) and peak O2 pulse (β=0.520, P<0.001) (Table 2). For the rest, LAd (β=−0.231, P=0.016), peak HR (β=0.301, P=0.001), and peak systolic BP (β=0.359, P<0.001) were selected as independent associated factors for peak V̇O2, and peak HR (β=−0.331, P=0.003) for peak O2 pulse.
Predictors of LgEF in HFIn the univariate analyses, all echocardiography parameters and the CPX parameters (peak HR, peak BP, peak respiratory exchange ratio) were not significantly associated with LgEF. When populated P<0.1 variables were out into the multivariate linear regression analysis, body mass index (BMI: β=0.335, P=0.006), albumin (β=0.237, P=0.047), and BNP (β=−0.222, P=0.037) were extracted as significant independent contributors to LgEF (Table 3).
| Univariate | Multivariate | ||||||
|---|---|---|---|---|---|---|---|
| R | P value | Unstandardized coefficients (95% CI) |
SE | Standardized coefficients (β) |
t statistic | P value | |
| LgEF* | |||||||
| Age | −0.410 | <0.001 | −0.308 (−0.770 to 0.154) | 0.231 | −0.159 | −1.34 | 0.187 |
| BMI | 0.520 | <0.001 | 1.920 (0.584 to 3.257) | 0.667 | 0.335 | 2.88 | 0.006 |
| Hemoglobin | 0.214 | 0.088 | −0.312 (−2.590 to 1.966) | 1.138 | −0.031 | −0.27 | 0.785 |
| Creatinine | −0.026 | 0.835 | |||||
| Albumin | 0.467 | <0.001 | 12.747 (0.182 to 25.312) | 6.275 | 0.237 | 2.03 | 0.047 |
| BNP | −0.362 | 0.003 | −0.018 (−0.035 to −0.001) | 0.008 | −0.222 | −2.14 | 0.037 |
| LVEF | 0.041 | 0.746 | |||||
| LVDd | 0.026 | 0.834 | |||||
| LAd | 0.046 | 0.714 | |||||
| E/E’ | −0.198 | 0.117 | |||||
| Baseline HR | −0.112 | 0.373 | |||||
| Peak HR | 0.161 | 0.200 | |||||
| Baseline systolic BP | −0.008 | 0.952 | |||||
| Peak systolic BP | 0.176 | 0.165 | |||||
| Peak RER | 0.140 | 0.264 | |||||
*R2=0.426. Abbreviations as in Tables 1,2.
The findings of our study were evidence that, in patients with HF, skeletal muscle pumping increases CO indirectly by increasing venous return and thereby cardiac preload during upright exercise. One of our main findings, a stronger correlation between LgEF and peak O2 pulse in HFrEF, suggests that skeletal muscle pumping performs a complementary role in augmenting stroke volume for patients with systolic cardiac dysfunction, and in turn CO through the Starling mechanism. This is the first observational study to demonstrate an underlying mechanism of the effect of skeletal muscle function on exercise-stressed hearts in HF patients, and to indicate the lower leg muscles as a new target for improving exercise capacity.
Marked blood volume shifts by leg muscle contraction, or muscle pump, from the lower extremity to the heart play a major role in increasing cardiac preload against gravity.10,15 LgEF, which has been used to assess the functionality of venous disease, is calculated from the calf volume changes induced by muscle contraction and this parameter has been assumed to indicate venous return derived from the entire leg (i.e., muscle pump function).15 In patients with HF, exercise capacity is considered to be strongly determined by peripheral abnormalities, and both muscle strength and muscle volume are reported to be strongly related to peak V̇O2.7–9 However, the detailed mechanism of skeletal muscle’s effect on peak V̇O2 or CO has not been elaborated. The findings of our study suggested that leg muscle strength may not have a direct causal effect on peak V̇O2, but has an effect through the skeletal muscle pump function, which underlies maintenance of cardiac preload. The findings of this study are new evidence of the mechanism that explains how skeletal muscle function affects exercise capacity through a central function in a HF population. Furthermore, a modest relationship between LgEF and calf circumference was also observed (r=0.571, P<0.001). This result also confirmed that LgEF might be the link between muscle mass and peak V̇O2 because calf circumference has strongly correlated with calf muscle mass assessed by magnetic resonance imaging.16
In our study, there was a significant correlation between the skeletal muscle pump function and exercise capacity in the HF group only. The control subjects were patients with cardiac disease, including old myocardial infarction and angina pectoris, but with almost normal LVEF and preserved exercise capacity. This implied that preserved cardiac contractility may not drive peripheral compensation to increase CO. HR augmentation during exercise is another determinant of CO during exercise. Indeed, the magnitude of the HR increase during CPX, or the chronotropic response, which is decreased by β-blocker usage and autonomic nervous system imbalance,17,18 was significantly lower in the present HF patients than in the control subjects (44 beats/min vs. 57 beats/min, P=0.009). This result suggested that cardiac contractility as well as HR increase was a primary contributor to peak V̇O2.
In HFrEF, the increase in cardiac contractility induced by load stress has been reported to be low.6,19 Therefore, it is necessary to compensate for this insufficient contractility reserve by other mechanisms to maintain V̇O2. Our results suggested that reduced contractility reserve was effectively substituted by cardiac preload augmentation, or skeletal muscle pump function, and the effect of muscle pump function on exercise capacity became stronger in HFrEF. On the other hand, in HFpEF, LgEF did not show a significant relation with peak V̇O2. It has been reported that the arteriovenous O2 difference during exercise is not driven in HFpEF in contrast to normal subjects or HFrEF.20 They also indicated that, in HFpEF, the HR and arteriovenous O2 difference rather than stroke volume mainly determined peak V̇O2, suggesting less effect of preload on peak V̇O2 in HFpEF. Thus, the mechanism of a central or peripheral function limiting exercise capacity would not be the same between HFrEF and HFpEF.
In our multiple linear regression analysis, LgEF, LAD, peak HR, and peak systolic BP were selected as independent variables for peak V̇O2. These results are consistent with previous publications that have proved the relationship between peak V̇O2, hemodynamic indexes, and echocardiographic findings.5,6,21,22 Our study adds evidence that LgEF and peak HR are also independent contributors to peak O2 pulse. The negative association of peak HR and peak O2 pulse suggests that increasing the HR would compensate for reduced stroke volume or that increasing HR would lead to reduced stroke volume through inadequate time for compromised diastolic filling. It is noteworthy that LgEF was a significant determinant in both peak V̇O2 and peak O2 pulse. Although we could not analyze the effect of LgEF based on HF phenotype because of our sample size, our model implied that a peripheral factor, LgEF, may contribute to the increase in stroke volume during exercise rather than augmentation of cardiac contractility.
Venous congestion caused by leg raising before LgEF measurement may be a considerable factor affecting LgEF, peak O2 pulse and peak V̇O2. Leg raising while in a supine position, theoretically, is likely to raise central venous pressure, which resists the augmentation of stroke volume during exercise. However, skeletal muscle pump function in this study was measured with subjects in an upright posture in which cardiac preload is remarkably reduced by hydrostatic stress, implying the effect of a possible increase in central venous pressure while supine may not need to be considered. Further study of the postural effect on augmentation of cardiac contractility in HF patients is needed.
By additional analysis, BMI and albumin were selected as significant contributing factors to LgEF. Because of technical limitations of LgEF measurement, we excluded obese patients. Therefore, this analysis indicated that reduced BMI or muscle volume and albumin could lead to a decline in muscle pump function. In addition, LgEF was negatively associated with BNP, suggesting the HF may affect skeletal muscles, known as cardiac cachexia. These observations could explain the reduced exercise capacity in cachectic or sarcopenic patients. Prevention and improvement of cachexia or sarcopenia are indispensable for maintaining exercise capacity or to ameliorate patients’ symptoms. It has been reported that muscle pump function increases with lower extremity exercise, and exercise training effectively improves exercise capacity through increasing CO, even in frail patients.23,24 In addition, it has become evident that there is an additive effect of amino acid supplements to exercise regimens to increase peak V̇O2 in the HF population.25 These reports suggest that physical training and nutritional care could improve skeletal muscle pump function and lead to increased exercise capacity by modifying cardiac preload in HF patients.
Study LimitationsThere are several to describe. Firstly, results were derived from a small patient population in a single center, and patients were ambulatory, stable, and able to perform exhaustive exercise testing, suggesting selection bias. Additionally, we had a limited number of female patients in the HFpEF group, which may limit generalization of our findings to the whole spectrum of patients with HF. We speculate that this may be caused by limited numbers of female patients who could undergo CPX because of the greater prevalence of frailty in female patients with HFpEF.26 Secondly, this study was not designed to clarify the causal effect of muscle pump function to exercise capacity, but to find an association between lower leg pump function and CPX parameters. Thirdly, our measurement of LgEF was determined from several calf muscle contractions in a standing position, which may be different to the activation during cycling in CPX. Nevertheless, we consider that LgEF is a novel parameter of venous return against hydrostatic stress during upright exercise. Finally, we did not measure possible confounding factors, such as circulating blood volume, blood shift from abdominal organs, comprehensive diastolic function, or right ventricular function. Nevertheless, the findings in this study are novel evidence of the effect of muscle pump function measured by LgEF using SPG on exercise capacity in patients with HF. Further study will be needed to establish the causal effect of muscle pump function on peak V̇O2.
The findings of this study suggested that the lower leg muscle pump function is a factor in generating exercise capacity in HF patients.
T.K., C.A., D.T. have no conflicts of interest. S.Y. has received lecture fees from Daiichi Sankyo, Fukuda Denshi, MSD, and Toa Eiyo and research grants from Epson Kenpokumiai, and Inter Reha outside the submitted work. T.O. has received research grants from Ono, Bayer, Tanabe Mitsubishi, and Daiichi Sankyo outside the submitted work. T.M. received lecture fees and unrestricted research grant for the Department of Cardiology, Nagoya University Graduate School of Medicine from Astellas, Daiichi Sankyo, Dainippon Sumitomo, Kowa, MSD, Tanabe Mitsubishi, Nippon Boehringer Ingelheim, Novartis, Otsuka, Pfizer Japan, Sanofi-aventis, Takeda and Teijin outside the submitted work.
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.
None.
