Abstract
Objective
Exercise therapy as part of cardiac rehabilitation is one of the most effective treatments for patients with chronic heart failure (HF). The anaerobic threshold (AT) determined by an exhaled gas analysis during cardiopulmonary exercise testing (CPX) is used to prescribe the appropriate level of exercise therapy. However, CPX using an exhaled gas analysis is not widely performed because of its cost, complexity, and the need for skilled staff. Therefore, a simpler and inexpensive method for determining AT without respiratory gas measurements is required in patients with HF. The present study elucidated the relationship between the AT determined by the CPX ventilatory method (CPX-AT) and the AT determined by cardiac acoustic biomarkers (CABs), which are measured by acoustic cardiography (CAB-AT), in HF patients.
Methods
Patients underwent symptom-limited ramp CPX twice using a cycle ergometer. The ATs determined from the exhaled gas analysis were identified by three independent physicians. CABs, including the first heart sound (S1) and the second heart sound intensities (peak-to-peak amplitudes), electromechanical activation time (EMAT) defined as the time interval from the Q wave onset on electrocardiography to S1, heart rate, and other parameters, were collected during CPX.
Patients
Forty patients with HF were included in this study.
Results
A significant correlation (R=0.70; p＜0.001) was found between CPX-AT and CAB-AT, using the double product of S1 intensity and heart rate. CAB-AT using S1 intensity also showed a significant correlation with CPX-AT (R=0.71; p＜0.001).
Conclusion
The present study suggests a possible new method for determining AT without respiratory gas measurements in patients with HF.
Keywords: acoustic cardiography, anaerobic threshold, cardiac acoustic biomarkers, heart failure, cardiopulmonary exercise testing
Introduction
Exercise therapy as part of cardiac rehabilitation aimed at improving symptoms and improving tolerance is one of the most effective treatments for patients with chronic heart failure (HF). In fact, exercise is recommended as Class I and Level A evidence in the “Guidelines on Rehabilitation in Cardiovascular Diseases (revised 2012)” (1,2). The anaerobic threshold (AT), determined by an exhaled gas analysis during cardiopulmonary exercise testing (CPX), is used to prescribe a suitable level of exercise therapy with appropriate exercise intensity.
Although it has been established that the AT is useful for determining the intensity of cardiac rehabilitation (1), CPX with an exhaled gas analysis has not been widely implemented. One reason for this is that because the equipment is expensive, it is difficult to carry out at all institutions, which sometimes makes it difficult to determine the appropriate exercise level. Furthermore, determining AT using the CPX ventilatory method requires a certain degree of expertise and skills (3). In addition, the equipment used for an exhaled gas analysis, including gas masks, is complex. Therefore, a simpler and inexpensive method for determining AT without respiratory gas measurements is required in patients with HF.
Tanaka et al. (4) previously reported the relationship between the AT determined by CPX and that determined by the double product of the heart sound amplitude and heart rate during exercise, and they found that the correlation was relatively high (R=0.88). However, their study only evaluated healthy subjects.
Recently, the new technology of acoustic cardiography, which consists of a simultaneously recorded electrocardiogram and phonocardiogram that can be measured using an ambulatory device, has become available (5). It provides quantitative cardiopulmonary function information regarding the combination of systolic and diastolic time intervals and extra-cardiohemic vibrations (e.g., third and fourth heart sounds) as cardiac acoustic biomarkers (CABs). CABs can reportedly be used to determine cardiac contractility in patients with HF (5-9) and left ventricular dysfunction (10).
Against this background, the present study was conducted to elucidate the relationship between the AT determined by CABs (CAB-AT) and the AT determined by the CPX ventilatory method (CPX-AT) in patients with HF.
Materials and Methods
Study design and subjects
The AUDICOR-guided Cardiac Rehabilitation for Heart Failure patients (ACARE-HF) study was a single-center, observational study that involved outpatients with chronic HF who were scheduled to participate in cardiac rehabilitation between August 2017 and September 2018 at Nihon University Hospital (Tokyo, Japan). All subjects underwent CPX twice: before and after the cardiac rehabilitation period. The inclusion criteria were as follows: (1) chronic HF; (2) age ≥40 years old; (3) scheduled cardiac rehabilitation and CPX; and (4) voluntary written informed consent. The exclusion criteria were (1) acute HF and (2) patients judged by the investigator to be inappropriate. Baseline characteristics, including the age, sex, vital signs, medical history, laboratory data, and echocardiographic parameters, were obtained.
The study protocol was approved by the ethics committee of Nihon University Hospital, and the study complied with the ethical principles of the Declaration of Helsinki. All participants provided their written informed consent to participate in the study prior to the study. This study was registered in the UMIN Clinical Trials Registry (identifier: UMIN000028335).
CPX procedures
The subjects underwent symptom-limited ramp CPX on a cycle ergometer (StrengthErgo8; MEE, Nagoya, Japan) using a stress electrocardiograph system (ML-9000; Fukuda Denshi, Tokyo, Japan), followed by 3 min of rest and 3 min of warm-up at 0, 10, and 20 Watts (W), a ramp protocol at 10 and 20 W/min, and 1.5 min of cool-down. Workloads for the warm-up and ramp protocols were determined according to the functional capacity reported by the subjects during the clinical assessment.
Pulmonary gas exchange measurement and determination of the AT by the CPX ventilatory method
Respiratory gas was obtained breath-by-breath (Cpex-1 software program, Ver3.3.0, Cpex-1; Inter Reha, Tokyo, Japan), and ventilatory and metabolic variables, including real-time applied workload (W) and pedaling speed (rpm), as well as oxygen consumption (VO2), carbon dioxide production (VCO2), minute ventilation (VE), heart rate (HR), ventilatory equivalent values (VE/VO2 and VE/VCO2), respiratory exchange ratio, and tidal volume were measured.
A visual analysis of ventilatory variables (VO2 and VCO2) was performed on a moving average graph of 8-10 breathing cycles. AT was identified by three independent experts (HT, RK, and KI) using the V-slope and trend methods. The final AT value (CPX-AT) was determined using a previously described method (11). First, the independently identified results of two experts (HT and RK) were compared, and the mean value was calculated and used only when the difference between the two AT workloads was ＜5 W. Results in which both experts (HT and RK) were unable to identify the AT were excluded from the evaluation. For data in which the difference was ≥5 W or in which only one expert was unable to identify the AT in the first step, the result from a third expert (KI) was compared with the other two results; if there was a pair of data whose difference was ＜5 W in this comparison, the mean value was calculated and used. Results were excluded from the evaluation if the differences were ≥5 W for any pair of experts (HT, RK, and KI).
Measurements of CABs and estimation of the AT by CABs
CABs were recorded constantly throughout CPX by a simple and light (38 g) approved medical device, the AUDICOR AM-RT (Inovise Medical, Portland, USA), with three electrodes for electrocardiography placed at the right upper, left upper, and left lower chest areas, and the other two electrodes with an accelerometer attached at the V3 and V4 positions of the chest wall (Fig. 1). The electrocardiogram data used for CAB calculation were obtained from the V4-RA lead, whereas the heart sound data were obtained at the V4 position.
Figure 1.
AUDICOR AM-RT device and definitions of cardiac acoustic biomarkers (CABs). AUDICOR device and measurement site introduction and definitions of CABs: the first heart sound (S1) and second heart sound (S2) intensities, peak-to-peak amplitudes of S1 and S2, respectively. EMAT: time interval from the Q wave onset to S1, EMATc: ratio of EMAT to R-R interval, QoS2: time interval from the Q wave onset to S2, LDPT: time interval from S2 to the Q wave onset
CABs were quantitatively calculated every 10 s to create a time series (Fig. 1) using the AUDICOR algorithm, which automatically detects heart sounds with the aid of machine learning technology based on acoustic models consisting of synchronously recorded heart sounds and electrocardiography. The main CABs focused on were as follows (Fig. 1): 1) the first heart sound (S1) and second heart sound (S2) intensities, defined as the peak-to-peak amplitudes of S1 and S2, respectively; 2) electromechanical activation time (EMAT), defined as the time interval from the Q wave onset on electrocardiography to S1; 3) EMATc, defined as the ratio of EMAT to the R-R interval; 4) the total electromechanical systolic interval (QoS2), defined as the time interval from the Q wave onset on electrocardiography to S2; 5) LDPT, defined as the time interval from the second heart sound to the Q wave onset; 6) %LDPT, defined as the ratio of LDPT to the R-R interval; and 7) the double product of S1 intensity and heart rate, as reported by Tanaka et al. (4).
Similar to a previous study (4), it was assumed that CABs related to cardiac contractility flexed at the AT, and the breakpoint was calculated from the CAB time series using the following method after pre-processing the moving average of six points for each CAB time series (CAB-AT): 1) the CAB time series existing in the ramp phase was divided into two consecutive phases (phase 1 and phase 2) with the breakpoint candidate as the boundary; 2) the regression line of phase 1 and the regression line of phase 2 were found using the method of least squares; and 3) the point at which the sum of the residual sums of squares of each of the two regression lines obtained in 2) was the minimum was the breakpoint in that time series.
Selecting analyzable data
To ensure the quality of the evaluated data, measurement data that met any of the following criteria were rejected: (1) the AT was not detected or could not be determined in the CPX test ramp load section; (2) the ratio of data that could be analyzed in the CAB time series in the CPX test ramp load section was ＜70%; and (3) the physician judged the data inappropriate for an evaluation.
Statistical analyses
Data are presented as the mean±standard deviation (SD) for continuous variables and as ratios (%) for categorical variables. Correlations between CPX-AT and CAB-AT values were assessed using Pearson's correlation coefficients. Bland-Altman plots were used to assess the mean differences and limits of agreement. In sub-analyses, correlations were compared between CPX-AT and CAB-AT values by study population, especially between (1) device user and non-device user groups, (2) patients with and without atrial fibrillation (AF), and (3) HF patients with a left ventricular ejection fraction (LVEF) ＜50% and HF patients with an LVEF ≥50%.
The significance level was set at p＜0.05, and statistical analyses were performed using the RStudio software program, ver. 1.2.5033 (RStudio, Boston, USA).
Results
Forty patients were enrolled in this study. Three patients who did not complete the data collection were excluded, and 71 measurements were obtained from 37 subjects who completed the data collection. Three experts (HT, RK, and KI) evaluated 71 CPX-AT measurements, of whom 61 were included in the final analysis (Fig. 2). The study population is summarized in Table 1. The mean±SD age of the subjects was 71.8±9.0 years old, 75.7% of the subjects were men, the mean±SD body mass index (BMI) was 23.8±3.9 kg/m2, and 40.5% of subjects had an implantable pacing device. The most common cause of HF was ischemic heart disease (51.4%), and 48.6% of the subjects had hypertension. The CPX test values in the study before (Trial 1) and after (Trial 2) exercise therapy are summarized in Table 2. The mean±SD of VO2 Peak was 15.7±4.7 mL/min/kg and 16.2±4.3 mL/min/kg in Trials 1 and 2, respectively. In addition, the mean±SD of VO2/HR Peak showed a significant difference (p＜0.01) between Trials 1 and 2 at 9.5±3.9 and 10.4±4.2 mL/min/bpm, respectively. As individual patient CPX test values varied considerably between Trials 1 and 2, as shown in Table 2, it was decided that there were no issues with treating them as different data.
Figure 2.
Number of patients enrolled in the study and number of data used for evaluation. Forty patients (A) and three patients who did not complete the data collection were excluded. Seventy-one measurements (C) were thus obtained from 37 subjects (B) who completed data collection, and 48 pairs of analyzable CPX-AT and CAB-AT data were available (E).
Table 1.
Baseline Characteristics of Study Patients.
| Item | n=37 |
|---|---|
| Age, y | 71.8±9.0 |
| Male | 28 (75.7%) |
| BMI, kg/m2 | 23.8±3.9 |
| LVEF, % | 50.5±17.3 |
| BNP, pg/mLa | 215.2±224.5 |
| Etiology of heart failure | |
| Ischemic heart disease | 19 (51.4) |
| Valvular disease | 11 (29.7) |
| Cardiomyopathy | 10 (27.0) |
| Hypertensive heart disease | 10 (27.0) |
| Implantable devices | |
| Pacemaker | 6 (16.2) |
| CRTD | 5 (13.5) |
| ICD | 4 (10.8) |
| Comorbidity | |
| Hypertension | 18 (48.6) |
| Dyslipidemia | 16 (43.2) |
| Chronic kidney disease | 16 (43.2) |
| Atrial fibrillation | 13 (35.1) |
| Diabetes mellitus | 11 (29.7) |
| Medication | |
| Beta-blockers | 35 (94.6) |
| ARBs/ACEIs | 23 (62.2) |
| Statins | 10 (27.0) |
| Nitrates | 10 (27.0) |
Values are means±SD or n (%).
ACEI: ACE inhibitor, ARB: angiotensin receptor blocker, BMI: body mass index, BNP: B-type natriuretic peptide, CRTD: cardiac resynchronization therapy defibrillator, ICD: implantable cardioverter defibrillator, LVEF: left ventricular ejection fraction, NT-proBNP: N-terminal pro-B-type natriuretic peptide
aBrain natriuretic peptide (BNP) levels were measured in 31 patients, whereas N-terminal pro-B-type natriuretic peptide (NT-pro BNP) levels were measured in 6 other patients (2,437.2±2,579.9 pg/mL).
Table 2.
CPX Results of Study Patients.
| CPX items | Trial 1 | Trial 2 | p |
|---|---|---|---|
| Heart Rate Rest, bpm | 69±11 | 66±11 | 0.0048 |
| Heart Rate Peak, bpm | 107±21 | 104±22 | 0.1034 |
| Blood pressur e systolic Rest, mmH g | 124±21 | 123±20 | 0.3193 |
| Blood pressure diastolic Rest, mmHg | 71±13 | 72±12 | 0.4478 |
| Blood pressure systolic Peak, mmHg | 159±29 | 157±32 | 0.2523 |
| Blood pressure diastolic Peak, mmHg | 75±14 | 73±15 | 0.0362 |
| VO2 AT, mL/min/kg | 11.4±2.4 | 11.6±2.0 | 0.2006 |
| VO2 Peak, mL/min/kg | 15.7±4.7 | 16.2±4.3 | 0.1460 |
| VE/VCO2 SLOPE | 30.8±5.4 | 31.4±9.4 | 0.3472 |
| R Peak | 1.3±0.1 | 1.3±0.1 | 0.4590 |
| Respiration Rate Peak | 34.1±7.4 | 32.9±5.4 | 0.0498 |
| VO2/HR Peak, mL/min/bpm | 9.5±3.9 | 10.4±4.2 | 0.0077 |
| Workload AT, Watts | 45±14 | 46±14 | 0.2839 |
| Workload Peak, Watts | 75±25 | 78±27 | 0.0145 |
| Heart Rate Recovery, bpm | 17±10 | 16±11 | 0.2918 |
Values are means±SD or p value between Trial 1 and Trial 2, in which Trial 1 is the exercise test before exercise therapy, and Trial 2 is the exercise test after exercise therapy.
Relationship between CPX-AT and CAB-AT
A representative case of a CAB time series from start to cool-down is presented in Fig. 3, and the trend of each CAB through CPX can be observed. The S1 and S2 intensities, EMAT, HR, and the double product of S1 intensity and heart rate increased, whereas the EMATc, QoS2, LDPT, and %LDPT decreased during the observation period.
Figure 3.
Representative cardiac acoustic biomarker (CAB) time series data. Changes in each CAB (blue solid line) from start to cool-down and the anaerobic threshold (blue dotted line).
A significant correlation (R=0.70; p＜0.001) was found between the CPX-AT and CAB-AT using the double product of the S1 intensity and heart rate, similar to the results of a previous study (4) (Fig. 4a, b). The CAB-AT was also obtained in 7 of the 10 measurements that were excluded from the final analysis because experts were unable to identify the CPX-AT. The CAB-AT using the S1 intensity also showed a significant correlation with the CPX-AT (R=0.71; p＜0.001) (Fig. 5). Detailed time series and estimated breakpoints using the algorithm of each data used in the final analysis can be found in the Supplementary Materials.
Figure 4.
Relationship between the anaerobic threshold determined by the CPX ventilatory method (CPX-AT) and the anaerobic threshold determined by cardiac acoustic biomarkers (CAB-AT) using the double-product of S1 intensity and heart rate. Correlation between the CPX-AT and CAB-AT (using the double product of S1 intensity and heart rate) (a) and a Bland-Altman plot to assess mean differences and limits of agreement (b).
Figure 5.
Relationship between the anaerobic threshold determined by the CPX ventilatory method (CPX-AT) and the anaerobic threshold determined by cardiac acoustic biomarkers (CAB-AT) using S1 intensity. Correlation between the CPX-AT and CAB-AT (using S1 intensity) (a) and a Bland-Altman plot to assess mean differences and limits of agreement (b).
The correlations between the CPX-AT and CAB-AT using other major CABs are shown in Table 3. Using the EMATc and S2 intensity, Pearson's correlation coefficient showed a significant correlation of ≥0.70.
Table 3.
Correlations between the CAB-ATs and the CPX-AT.
| CAB | Correlation coefficient R (n) |
|---|---|
| S1 intensity | 0.71* (48) |
| Double product of S1 intensity and HR | 0.70* (48) |
| S2 intensity | 0.71* (48) |
| EMAT | 0.49* (48) |
| LDPT | 0.53* (41) |
| QoS2 | 0.68* (47) |
| EMATc | 0.72* (48) |
| %LDPT | 0.58* (41) |
*p<0.001
S1/S2 intensity: peak-to-peak amplitudes of the first/second heart sound (S1/S2), EMAT: time interval from Q wave onset to S1, EMATc: ratio of EMAT to the R-R interval, QoS2: time interval from Q wave onset to S2, LDPT: time interval from S2 to Q wave onset
A sub-analysis based on background characteristics
Correlations between the CPX-AT and CAB-AT using the double product of S1 intensity and heart rate, especially between (1) heart device users and non-users, (2) patients with and without AF, and (3) groups categorized by the LVEF, are shown in Fig. 6. A significant correlation was found in device users and non-device users, with R-values of 0.61 and 0.79, respectively, and in patients with and without AF, with R-values of 0.75 and 0.72, respectively. Significant correlations were also observed in HF patients with an LVEF ＜50% and ≥50%, with R-values of 0.67 and 0.84, respectively.
Figure 6.
Correlation between the anaerobic threshold determined by the CPX ventilatory method (CPX-AT) and the anaerobic threshold determined by cardiac acoustic biomarkers (CAB-AT) using the double-product of S1 intensity and heart rate in sub-analyses. (a) Device users, (b) non-device users, (c) patients with atrial fibrillation (AF), (d) patients without AF, (e) heart failure (HF) patients with left ventricular ejection fraction (LVEF) <50%, and (f) HF patients with LVEF ≥50%.
Discussion
Results and mechanism of the S1 intensity breakpoint
The present study found a significant correlation between the CAB-AT using the double product of S1 intensity and heart rate and CPX-AT (R=0.70; p＜0.001). Tanaka et al. (4) reported a significant correlation between the double product of S1 intensity and HR and plasma adrenaline level in normal subjects, suggesting the causal effects of an abrupt increase in sympathetic nervous activity. The results of the present study suggest that the double product of S1 intensity and heart rate can also be used to determine the AT in patients with HF. In general, it is well known that the sympathetic nerves of HF patients are hyperactive at rest, but recently, it has been confirmed that their hypersensitivity is further increased by exercise (12). The breakpoint detected by the proposed algorithm can be explained using the same mechanism. Furthermore, Tanaka et al. (4) also reported a significant correlation between the double product of S1 intensity and heart rate and LV dP/dt max, which is the major determinant of myocardial oxygen demand and determines the speed and intensity of mitral valve movement. The result of the CAB-AT using only S1 intensity (R=0.71; p＜0.001) suggests that the same mechanism could also be applied to patients with HF and could be useful for determining the AT using only S1 intensity.
In the present study, 70% of the measurements in which experts were unable to identify the CPX-AT were successfully determined using CABs. This suggests that the proposed method can be used in patients who are difficult to evaluate using a gas analysis. Thus, further research using a larger amount of data should be conducted to clarify the efficacy of CAB-AT and the relationship between the CAB-AT and gas analysis findings.
Sub-analyses
Although it is well known that both S1 and heart rate are unstable in AF patients, a significant correlation was observed between these parameters in both AF and non-AF patients. Therefore, the AT can be determined using CABs, regardless of AF status.
However, in the sub-analysis results using the LVEF, the correlation tended to be lower in patients with an LVEF ＜50% than in those with an LVEF ≥50%. Many patients with an LVEF ＜50% had implanted devices, including pacemakers, and our observation of electrocardiogram data while processing the CPX showed that there were several patients whose devices were pacing over the entire CPX. In the sub-analysis results of patients whose devices were pacing over the CPX (n=7), the R value was 0.47, whereas it was 0.72 in the others (n=41). This suggests that there might be some limitations in the estimation algorithm using CABs, in which the features directly depend on a pacemaker, such as S1 intensity or HR. Thus, the algorithm using CABs must be modified to reduce the impact of a pacemaker.
Results using other CABs (without using S1 intensity and the double-product of S1 intensity with HR)
The results using other CABs, especially S2 intensity and EMATc which represent the cardiac contractions, also showed significant correlations between the CPX-AT and CAB-AT. These CABs have common aspects that are influenced by the sympathetic nervous system; however, they also have aspects that result from different mechanisms. Specifically, S2 intensity has been reported to correlate with central blood pressure (13) and EMATc is an indicator of HF severity (7). Split S2 is known to occur in patients with atrial septal defects and bundle branch blocks, and there is concern that CAB values may be affected as well. Since four patients with right bundle branch block were also included in the present study, S2 intensity was used for CAB-AT after confirming that no specific trend, including split S2, had been observed during CPX depending on the presence of right bundle branch block. These results therefore suggest that the development of an AT estimation method using several CABs together will lead to further improvements in AT estimation accuracy.
CABs, such as LDPT, which represents diastolic duration, showed a monotonous decrease during exercise, as seen in Fig. 3. Although it was unsuitable as a method for determining the breakpoint in the present study, it could be said that the change in LDPT depends linearly on the exercise workload. Cinquegrana et al. (14) showed that LDPT could be used to determine the effectiveness of exercise therapy in patients with coronary artery disease, suggesting that there might be other useful methods for determining AT using this CAB.
Limitations
Several limitations associated with the present study warrant mention. First, this was a single-center study with a relatively small sample size. Second, patients with implantable devices were included, and significant correlations were found between CAB-AT and CPX-AT among both the device users and non-device users, with R-values of 0.61 and 0.79, respectively. Third, CPX-AT workloads could not be determined in 10 of the 71 measurements because there were several cases in which AT determination is difficult in HF patients as follows: 1) at the appearance of oscillatory ventilation in two measurements; 2) in two measurements, AT point were recognized not as a single point, but in a range of 5 W or more; and 3) in 6 measurements confirmed to have a discrepancy of ≥5 W among any of the three experts based on the results of an evaluation or when values were undetectable due to an extremely low AT. For these reasons, the accuracy of the CPX-AT in the present study might be inferior to that reported in a previous study of healthy subjects (4). Fourth, CAB data could not be analyzed in 13 cases because of weak heart-sound signals, cloth friction, or respiratory noise. Fifth, cases in which the absolute error between the CAB-AT and CPX-AT could not be ignored have been observed; therefore, further improvement of the algorithm to reduce the error and make it more appropriate for clinical practice is necessary.
Conclusion
Significant correlations in the AT determined by CPX and CABs calculated from electrocardiograms and phonocardiograms were observed. Therefore, the present study demonstrated a new method for determining the AT without respiratory gas measurements in patients with HF. Determination of the AT using CABs is a low-cost, non-invasive, and easily implementable method compared to the respiratory gas analysis method.
All participants provided their written informed consent to participate in the study prior to the study.
Author’s disclosure of potential Conflicts of Interest (COI).
Kiyoshi Iida: Advisory role, Asahi Kasei. Osamu Saito: Employment, Asahi Kasei.
Financial Support
This work was supported by Asahi Kasei Corporation.
Supplementary Material
AT points are shown as judged by the three experts, as well as the CAB-AT positions determined by the estimation algorithm. The background also shows the rest, warm-up, ramp, and cooldown sections in blue, yellow, red, and green, respectively.
AT points are shown as judged by the three experts, as well as the CAB-AT positions determined by the estimation algorithm. The background also shows the rest, warm-up, ramp, and cooldown sections in blue, yellow, red, and green, respectively.
Acknowledgments
Data management and basic statistical analyses were performed by the EPS (Tokyo, Japan). The authors would also like to thank Dr. Hiroshi Takaki (Takaki Cardiovascular Medicine Clinic, Hyogo, Japan) and Dr. Reon Kumasaka (Department of Cardiovascular Medicine, Saitama Sekishinkai Hospital, Saitama, Japan) for their help in determining the anaerobic thresholds in this study.
References
- 1.Makita S, Yasu T, Akashi YJ, et al. ; the Japanese Circulation Society/the Japanese Association of Cardiac Rehabilitation Joint Working Group . JCS/JACR 2021 guideline on rehabilitation in patients with cardiovascular disease. Circ J 87: 155-235, 2023. [DOI] [PubMed] [Google Scholar]
- 2.Kamiya K, Sato Y, Takahashi T, et al. Multidisciplinary cardiac rehabilitation and long-term prognosis in patients with heart failure. Circ Heart Fail 13: e006798, 2020. [DOI] [PubMed] [Google Scholar]
- 3.Koike A. Estimation of anaerobic threshold, a classical index drawing renewed attention, by new methods without respiratory gas measurements. Circ J 77: 2694-2695, 2013. [DOI] [PubMed] [Google Scholar]
- 4.Tanaka H, Matsuda T, Tobina T, et al. Product of heart rate and first heart sound amplitude as an index of myocardial metabolic stress during graded exercise. Circ J 77: 2736-2741, 2013. [DOI] [PubMed] [Google Scholar]
- 5.Diller R, Zuber M, Arand P, Erne S, Erne P. Assessment of systolic and diastolic function in heart failure using ambulatory monitoring with acoustic cardiography. Ann Med 43: 403-411, 2011. [DOI] [PubMed] [Google Scholar]
- 6.Wen YN, Lee AP, Fang F, Jin CN, Yu CM. Beyond auscultation: acoustic cardiography in clinical practice. Int J Cardiol 172: 548-560, 2014. [DOI] [PubMed] [Google Scholar]
- 7.Wang S, Lam YY, Liu M, et al. Acoustic cardiography helps to identify heart failure and its phenotypes. Int J Cardiol 167: 681-696, 2013. [DOI] [PubMed] [Google Scholar]
- 8.Jamshidi P, Kobza R, Toggweiler S, Arand P, Zuber M, Erne P. Impact of preload changes on positive and negative left ventricular dP/dt and systolic intervals: preload changes on left ventricular function. Indian Heart J 64: 314-318, 2012. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Zuber M, Kipfer P, Jost CA. Systolic dysfunction: correlation of acoustic cardiography with Doppler echocardiography. Congest Heart Fail 12: 14-18, 2006. [DOI] [PubMed] [Google Scholar]
- 10.Maisel AS, Peacock WF, Shah KS, et al. Acoustic cardiography S3 detection use in problematic subgroups and B-type natriuretic peptide “gray zone”: secondary results from the Heart failure and Audicor technology for Rapid Diagnosis and Initial Treatment Multinational Investigation. Am J Emerg Med 29: 924-931, 2011. [DOI] [PubMed] [Google Scholar]
- 11.Myers J, Goldsmith RL, Keteyian SJ, et al. The ventilatory anaerobic threshold in heart failure: a multicenter evaluation of reliability. J Card Fail 16: 76-83, 2010. [DOI] [PubMed] [Google Scholar]
- 12.Notarius CF, Floras JS. Sympathetic neural responses in heart failure during exercise and after exercise training. Clin Sci (Lond) 135: 651-669, 2021. [DOI] [PubMed] [Google Scholar]
- 13.Otsubo K, Kiri H, Matsuda T, Obara S, Kiyonaga A, Tanaka H. In: Relationship between second cardiac sound amplitude and blood pressure during gradual exercise load in patients with illness. Paper presented at: Japanese Association of Cardiac Rehabilitation Annual Meeting; July 17, 2010; Kagoshima, Japan. [Google Scholar]
- 14.Cinquegrana G, Spinelli L, D'Aniello L, Landi M, D'Aniello MT, Meccariello P. Exercise training improves diastolic perfusion time in patients with coronary artery disease. Heart Dis 4: 13-17, 2002. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
AT points are shown as judged by the three experts, as well as the CAB-AT positions determined by the estimation algorithm. The background also shows the rest, warm-up, ramp, and cooldown sections in blue, yellow, red, and green, respectively.
AT points are shown as judged by the three experts, as well as the CAB-AT positions determined by the estimation algorithm. The background also shows the rest, warm-up, ramp, and cooldown sections in blue, yellow, red, and green, respectively.