Novel Scoring System Using Cardiopulmonary Exercise Testing Predicts Prognosis in Heart Failure Patients Receiving Guideline-Directed Medical Therapy

Abstract

Background: Among variables obtained from cardiopulmonary exercise testing (CPXT), peak oxygen consumption (PV̇O₂) and the minute ventilation vs. carbon dioxide output (V̇E vs. V̇CO₂) slope were established as predictors of death of patients with heart failure (HF) at the cutoff points of 14 ml·min^–1·kg^–1 and 34, respectively. However, a recent update of guideline-directed medical treatment (GDMT) might alter the implication of these variables.

Methods and Results: We enrolled 77 HF patients receiving GDMT who had undergone symptom-limited CPXT between 2006 and 2014. Among them, 29 patients were re-hospitalized for HF and there were 13 cardiac deaths during the 4-year study period. Cox regression analyses demonstrated that the V̇E vs. V̇CO₂ slope, peak heart rate, peak systolic blood pressure, and PV̇O₂ were significant predictors of both re-admission and cardiac death at each cutoff point calculated by receiver-operating characteristic analyses. A new scoring system was constructed using the following criteria: 1 point was assigned to a variable meeting the cutoff point for re-admission; 2 points were assigned to that for cardiac death. The total scores calculated as the summation of each point (range, 0–8 points) had significantly highest area under the curves compared with each CPXT variable (P<0.05), and significantly stratified both event-free rate into 3 groups (P<0.05).

Conclusions: A novel scoring system using 4 CPXT variables simultaneously predicted re-admission and cardiac death even in patients with HF receiving GDMT. (Circ J 2015; 79: 1068–1075)

Cardiopulmonary exercise testing (CPXT) is currently an established tool for predicting the prognosis of patients with heart failure (HF).^1,2 Traditionally, both peak oxygen consumption (PV̇O₂) <14 ml·min^–1·kg^–1 and minute ventilation vs. carbon dioxide output slope (V̇E vs. V̇CO₂ slope) >34 measured during CPXT have been widely used to predict survival and decide optimal timing of listing for cardiac replacement therapy (left ventricular assist device (LVAD) or heart transplantation (HTx)).^1,3–9

However, there are few reports discussing the prognostic impact of these CPXT variables in the modern era of β-blocker-incorporated guideline-directed medical treatment (GDMT).^10–12 Moreover, although the prognostic impact of these CPXT variables has been discussed independently, there are few studies constructing scoring systems to predict prognosis by coupling several CPXT variables especially, including recently reported parameters relating to chronotropic incompetence.^13–15 Because most CPXT studies have been performed in Europe or the United States where patients’ physique is larger, the results may not be simply adopted for Japanese with smaller physique.⁸ Therefore, we constructed a novel scoring system to predict prognosis consisting of several CPXT variables from among Japanese HF patients receiving GDMT, and we discuss the optimal timing of cardiac replacement therapy.^16–18

Methods

Patient Selection

Of consecutive patients who had been diagnosed as HF according to the Framingham Criteria¹⁹ and followed between November 2006 and July 2014, 291 in-hospital patients were referred to CPXT before the discharge. Among them, 77 patients who had been treated with GDMT consisting of β-blockers, angiotensin-converting enzyme inhibitors (ACEIs), and aldosterone antagonists,¹² were enrolled. In other words, all participants were treated with the 3 anti-HF medications. Doses of these drugs were titrated according to each patient’s hemodynamics. No patients had musculoskeletal abnormalities, apparent lung diseases, or a history of cerebral vascular disease that would affect their ability to exercise. No patients had significant valvular disease or acute coronary syndrome within the previous month. Patients who had atrial fibrillation were excluded. Written informed consent was given by all patients before enrollment and the study protocol was approved by the Ethics Committee of the Graduate School of Medicine, University of Tokyo [application number 779 (1)].

CPXT Protocol

Symptom-limited CPXT using a cycle ergometer with a ventilator expired gas analyzer (AE-300S; Minato Igaku, Osaka, Japan) was performed in accordance with the guideline from the American Heart Association (AHA).¹ All patients started at 20 W for a 4-min warm-up, and then underwent a 10-W/min ramp increment protocol. Continuous data of V̇O₂, V̇CO₂, and V̇E were obtained during the procedure on a breath-by-breath basis. The V̇E vs. V̇CO₂ slope was calculated by the algorism of least-squares linear regression model (y=mx+b, m=slope). Considering the Borg scale, which indicates the patient’s objective fatigue ranging from 6 to 20, >17 points was targeted to terminate the procedure.²⁰ In the symptom-limited CPXT, the expiratory change ratio was targeted at ≥1.10.

Variables Evaluated

Demographic, echocardiographic, and laboratory data were acquired on the same day of CPXT as baseline variables. Doses of anti-HF agents were normalized to each equivalent (ie, the dose of β-blocker was shown as an equivalent dose of carvedilol,²¹ and that of ACEIs was shown as an approximately equivalent dose of enalapril²²).

Endpoints

Endpoints were defined as (1) re-admission for HF or (2) cardiac death during the 4-year follow-up. Cardiac death was defined as death from HF, sudden death, or requiring implantation of a LVAD. No patients underwent HTx during the study period.

Statistical Analysis

All analyses were calculated using PASW Statistics 18 (SPSS Inc, Chicago, IL, USA). Data are expressed as mean±SD unless otherwise specified. Chi-square test or Fisher’s exact test was used for analysis of categorical data as appropriate. Continuous variables were analyzed using the unpaired t-test or Mann-Whitney U test as appropriate. Cox proportional hazards regression model was adopted to calculate predictors of each endpoint among the CPXT variables. Receiver-operating characteristics (ROC) analysis was used to obtain cutoff levels of each CPXT predictor, and to compare the area under the curve (AUCs) of each predictor.²³ Pearson product-moment correlation coefficient was used to evaluate relationship between doses of anti-HF medication and each CPXT variable. Kaplan-Meier analysis with log-rank test was adopted to examine the impact of a novel scoring system in stratifying the prognosis. Multicollinearity for each variable significant in the univariate analyses was analyzed by calculating variance inflation factors. All hypothesis tests were 2-tailed, and P<0.05 was significant.

Results

Patients’ Backgrounds

A total of 77 patients with HF (44±14 years old, 71% male) who had undergone CPXT were enrolled. New York Heart Association symptoms were Class III (47 patients) or IV (30 patients). Etiology of HF was dilated cardiomyopathy (55 patients), ischemic cardiomyopathy (7 patients), dilated phase of hypertrophic cardiomyopathy (4 patients), and other (11 patients). All patients accomplished CPXT without any complication (maximal load, 75±35 W). The expiratory change ratio was 1.3±0.2 (range 1.1–1.6), and the Borg scale was 18±2 points (range 17–20).

During the 4-year study period, 29 patients were re-admitted for HF at 316±300 days (range 11–1,460 days) after the discharge. Patients who were re-admitted had higher plasma levels of B-type natriuretic peptide (BNP) with lower doses of anti-HF agents and higher doses of furosemide than those not re-admitted (Table 1A).

Table 1. HF Patients’ Backgrounds Stratified by Re-Admission or Cardiac Death

A. Re-admission for HF	Re-admission (+) (n=29)	Re-admission (−) (n=48)	P value
Demographic parameters
Age, years	41±17	46±13	0.141
Male, n (%)	17 (59)	38 (79)	0.068
BMI	20.9±3.4	22.5±3.9	0.071
NYHA
III, n (%)	10 (34)	37 (77)	<0.001†
IV, n (%)	19 (66)	11 (23)	–
Ischemic etiology, n (%)	3 (10)	4 (8)	0.532
CPXT parameters
V̇E vs. V̇CO2 slope	45.3±13.4	30.5±10.6	<0.001*
Rest HR, beats/min	80±15	77±14	0.432
Peak HR, beats/min	109±21	128±20	<0.001*
Rest sBP, mmHg	84±14	89±17	0.223
Peak sBP, mmHg	113±26	140±30	<0.001*
PV̇O2, ml·min−1·kg−1	13.3±5.1	17.2±4.7	0.001*
Echocardiographic parameters
LVDd, mm	67±15	64±10	0.427
LVEF, %	28±15	33±14	0.137
Laboratory parameters
Hemoglobin, mg/dl	13.2±1.7	13.6±1.7	0.417
Serum sodium, mEq/L	135±4	136±3	0.755
Serum potassium, mEq/L	4.3±0.4	4.3±0.5	0.910
Serum creatinine, mg/dl	1.1±0.7	1.0±0.4	0.215
Plasma BNP, pg/ml	571±383	196±316	<0.001*
Medications
β-blocker, mg/day	12.3±9.8	20.1±16.3	0.011*
ACEI, mg/day	2.9±1.5	4.1±3.2	0.032*
Aldosterone antagonist, mg/day	40±22	34±11	0.312
Furosemide, mg/day	43±32	22±26	0.002*
B. Cardiac death	Cardiac death (+) (n=13)	Cardiac death (−) (n=64)	P value
Demographic parameters
Age, years	42±13	44±15	0.635
Male, n (%)	8 (62)	47 (73)	0.387
BMI	21.2±3.8	22.1±3.8	0.456
NYHA
III, n (%)	3 (23)	44 (69)	0.002†
IV, n (%)	10 (77)	20 (31)	–
Ischemic etiology, n (%)	0 (0)	7 (11)	0.258
CPX parameters
V̇E vs. V̇CO2 slope	47.9±15.5	33.7±12.0	<0.001*
Rest HR, beats/min	79±18	78±13	0.775
Peak HR, beats/min	104±17	124±22	0.002*
Rest sBP, mmHg	81±11	88±17	0.174
Peak sBP, mmHg	104±21	135±30	0.001*
PV̇O2, ml·min−1·kg−1	11.5±3.1	16.6±5.1	0.001*
Echocardiographic parameters
LVDd, mm	73±19	64±10	0.015*
LVEF, %	23±16	33±14	0.016*
Laboratory parameters
Hemoglobin, mg/dl	13.2±1.4	13.5±1.8	0.566
Serum sodium, mEq/L	133±4	134±3	0.901
Serum potassium, mEq/L	4.4±0.5	4.3±0.4	0.579
Serum creatinine, mg/dl	1.2±0.9	1.0±0.4	0.119
Plasma BNP, pg/ml	667±404	270±349	<0.001*
Medications
β-blocker, mg/day	13.3±12.6	18.0±15.0	0.295
ACEI, mg/day	3.1±2.1	3.7±2.8	0.432
Aldosterone antagonist, mg/day	44±23	33±15	0.105
Furosemide, mg/day	51±35	26±27	0.006*

*P<0.05 by unpaired t-test or Mann-Whitney U test as appropriate. ^†P<0.05 by Chi-square test of Fisher’s exact test as appropriate. ACEI, angiotensin-converting enzyme II inhibitor; BMI, body mass index; BNP, B-type natriuretic peptide; HF, heart failure; HR, heart rate; LVDd, left ventricular diastolic diameter; LVEF, left ventricular ejection fraction; NYHA, New York Heart Association; PV̇O₂, peak oxygen consumption; sBP, systolic blood pressure; V̇E vs. V̇CO₂ slope, minute ventilation vs. carbon dioxide output.

During the 4-year study period, there were 13 cardiac deaths (9 patients received LVAD; 2 died of ventricular tachyarrhythmia; 2 died of HF) at 404±390 days (range 71–1,114 days) after discharge. The patients who died had a more remodeled LV, as well as high plasma levels of BNP, and had received higher doses of furosemide than those who survived (Table 1B).

Cox Regression Analyses

Univariate Cox regression analyses demonstrated that the V̇E vs. V̇CO₂ slope, peak HR, peak systolic blood pressure (sBP), and PV̇O₂ were significant predictors of re-admission (P<0.05 for all). Considering the cutoff levels calculated by ROC analyses, V̇E vs. V̇CO₂ slope ≥37.2 (hazard ratio 6.648), (2) peak HR ≤120 beats/min (hazard ratio 4.442), peak sBP ≤129 mmHg (hazard ratio 5.904), and PV̇O₂ ≤14.4 ml·min^–1·kg^–1 (hazard ratio 4.185) were significant categorical predictors of re-admission in the univariate Cox regression analyses (Table 2A).

Table 2. Cox Regression Analyses for Re-Admission or Cardiac Death of HF Patients According to CPXT Variables

	P value	HR	95% CI	VIF
A. Re-admission for HF
Continuous parameters
V̇E vs. V̇CO2 slope	<0.001*	1.053	1.031–1.075
Rest HR, beats/min	0.331	1.014	0.986–1.042
Peak HR, beats/min	<0.001*	0.969	0.952–0.986
Rest sBP, mmHg	0.086	0.974	0.947–1.004
Peak sBP, mmHg	<0.001*	0.967	0.951–0.983
PV̇O2, ml·min−1·kg−1	0.001*	0.839	0.754–0.933
Categorical parameters
V̇E vs. V̇CO2 slope >37.2	<0.001*	6.648	3.345–17.49	1.737
Peak HR ≤120 beats/min	0.001*	4.442	1.894–10.42	1.045
Peak sBP ≤129 mmHg	<0.001*	5.904	2.378–14.66	1.051
PV̇O2 ≤14.4 ml·min−1·kg−1	0.001*	4.185	1.850–9.463	1.722
B. Cardiac death
Continuous parameters
V̇E vs. V̇CO2 slope	0.001*	1.061	1.026–1.098
Rest HR, beats/min	0.927	1.002	0.964–1.041
Peak HR, beats/min	0.005*	0.963	0.937–0.998
Rest sBP, mmHg	0.108	0.957	0.910–1.006
Peak sBP, mmHg	0.002*	0.962	0.938–0.986
PV̇O2, ml·min−1·kg−1	0.002*	0.714	0.579–0.881
Categorical parameters
V̇E vs. V̇CO2 slope >44.3	0.003*	5.500	1.797–16.83	1.680
Peak HR ≤115 beats/min	0.011*	5.340	1.469–19.41	1.154
Peak sBP ≤105 mmHg	<0.001*	7.231	2.680–25.28	1.312
PV̇O2 ≤11.8 ml·min−1·kg−1	0.006*	4.693	1.574–13.99	1.724

CPXT, cardiopulmonary exercise testing. Other abbreviations as in Table 1.

The same 4 variables were also significant predictors of cardiac death in the univariate Cox regression analyses (P<0.05 for all). Considering the results of ROC analyses performed in the same manner, V̇E vs. V̇CO₂ slope ≥44.3 (hazard ratio 5.500), peak HR ≤115 beats/min (hazard ratio 5.340), peak sBP ≤105 mmHg (hazard ratio 7.231), and PV̇O₂ ≤11.8 ml·min^–1·kg^–1 (hazard ratio 4.693) were significant categorical predictors of cardiac death (Table 2B). We could not construct significant multivariate models to predict re-admission rate or cardiac death.

Constructing the Prognostic Scoring System

Considering that the predictors of both re-admission and cardiac death were similar and the variation in hazard ratio was small, a scoring system was constructed by assigning 1 point when each variable met the re-admission-related cutoff level, and assigning 2 points when each variable met the cardiac death-related cutoff level (Table 3).

Table 3. Calculation of Novel Scoring System Using 4 CPXT Parameters

CPXT parameter	Points
	0	1	2
(1) V̇E vs. V̇CO2 slope	<37.2	37.2–44.3	>44.3
(2) Peak HR, beats/min	>120	120–115	<115
(3) Peak sBP, mmHg	>129	129–105	<105
(4) PV̇O2, ml·min−1·kg−1	>14.4	14.4–11.8	<11.8

Abbreviations as in Tables 1,2.

ROC analyses demonstrated that the calculated scores had the highest AUC among other CPXT variables (0.924 for re-admission, and 0.902 for cardiac death, P<0.05 for all) (Figure 1).

Figure 1.

Receiver-operating characteristics (ROC) analyses of the new scoring system, together with other CPXT variables for re-admission (A) or cardiac death (B). Area under the curve (AUC) of the new scoring system compared with other CPXT variables for the prognostic impact of re-admission (C) or cardiac death (D). *P<0.05 by ROC analysis. CI, confidence interval; CPXT, cardiopulmonary exercise testing.

The re-admission-free rate was stratified by the scoring system into 3 groups: low (0–2 points, 88%), intermediate (3–5 points, 40%), and high risk (6–8 points, 7%) during the 4-year study period (P<0.001) (Figure 2A). By using the same scoring system, the cardiac death-free rate was stratified into low (98%), intermediate (66%), and high-risk groups (48%) during 4-year study period (P<0.001) (Figure 2B).

Figure 2.

Kaplan-Meier curves of re-admission-free (A) and cardiac death-free rate (B) stratified by the new scoring system during 4-year study period. *P<0.005 by log-rank test. HF, heart failure.

Anti-HF Medication and CPXT Variables (Table 4)

Higher doses of β-blocker were associated with lower V̇E vs. V̇CO₂ slope and lower sBP at rest (P<0.05 for both). Other CPXT variables did not correlate with doses of ACEI or aldosterone antagonist.

Table 4. Relationship Between Medication Doses and CPXT Variables in HF Patients

	β-blocker, mg/day		ACEI, mg/day		Aldosterone antagonist, mg/day
	P value	R value	P value	R value	P value	R value
V̇E vs. V̇CO2 slope	0.007*	−0.307	0.087	−0.223	0.112	0.210
Rest HR, beats/min	0.797	0.030	0.193	−0.150	0.492	0.080
Peak HR, beats/min	0.203	0.147	0.839	−0.024	0.132	−0.173
Rest sBP, mmHg	0.001*	−0.374	0.769	−0.035	0.063	−0.217
Peak sBP, mmHg	0.605	−0.061	0.505	0.079	0.068	−0.321
PV̇O2, ml·min−1·kg−1	0.138	0.231	0.584	0.063	0.087	−0.290

*P<0.05 by Pearson product-moment correlation coefficient. Abbreviations as in Tables 1,2.

Discussion

We demonstrated that CPXT prognostic predictors, including PV̇O₂, V̇E vs. V̇CO₂ slope, as well as variables relating to chronotropic incompetence, were still applicable in the modern era of β-blocker-incorporated GDMT in Japanese HF patients. We also proposed cutoff values of CPXT parameters for the prediction of re-hospitalization for HF for the first time as far as we know. Moreover, we constructed a novel scoring system consisting of 4 CPXT parameters to predict both long-term re-admission and cardiac death. The score had the strongest prognostic effect among other CPXT predictors, and could significantly stratify the prognosis of patients, including both the re-admission rate and survival, into 3 groups during the 4-year study period.

CPXT Variables as Classical Prognostic Predictors Before the Era of GDMT

PV̇O₂ is the most famous predictor of mortality in patients with various conditions including mild to severe HF, or systolic or diastolic left ventricular dysfunction.^3–5 The AHA consensus in 1995 stated a cutoff level of 14 ml·min^–1·kg^–1 as the indicative threshold for HTx.²⁴

The V̇E vs. V̇CO₂ slope is a 2nd-generation prognostic predictor, recently reported to be superior to PV̇O₂ in prognostic impact at a threshold of 34.^6–8

CPXT Variables as Prognostic Predictors in the Modern Era of GDMT

Because the prognostic value of PV̇O₂ and the V̇E vs. V̇CO₂ slope was traditionally evaluated before the era of β-blocker therapy, the clinical utility in HF patients receiving β-blocker has been a matter of debate.²⁵ Recent studies found that PV̇O₂ maintained its prognostic value irrespective of β-blockade, although the optimal threshold might require a downward adjustment.^1,11,26 We found that PV̇O₂ still maintained its prognostic value but at a lower threshold, probably because of the somewhat decreased exercise tolerance from β-blocker treatment.

Little is known about the prognostic impact of the V̇E vs. V̇CO₂ slope under β-blocker treatment, except for the study by Arena et al.¹⁰ The reason why our threshold for cardiac death was higher (V̇E vs. V̇CO₂ slope, 44.3) might result from the significant negative correlation between the dose of β-blocker and the V̇E vs. V̇CO₂ slope, which was consistent with an early study demonstrating recovery of inappropriate hyperventilation during exercise while on β-blocker treatment.²⁷ Titration of β-blocker might improve not only prognosis but also daily quality of life.

Recently, chronotropic incompetence, which is defined as insufficient exercise-induced HR response as well as less-elevated blood pressure,²⁸ was demonstrated to be an independent prognostic marker of cardiovascular death irrespective of β-blockade.¹⁴ The mechanism was attributed to reduced myocardial sensitivity to sympathetic modulation, together with β-receptor downregulation, which indicated progression of HF.²⁹ The dose of β-blocker had no correlation in the present study with CPXT variables related to chronotropic incompetence, which was consistent with the study by Jorde et al.¹⁵ We demonstrated for the first time that both the peak HR and sBP remained significant prognostic markers even under GDMT.

Novel Scoring System as a Strong Prognostic Predictor

Based on the results of our Cox regression analyses, we constructed a novel scoring system to predict prognosis using 4 simple CPXT variables: (1) V̇E vs. V̇CO₂ slope, (2) peak HR, (3) peak sBP, and (4) PV̇O₂. These variables were all routinely measured during general CPXT and are easy to obtain with high reproducibility. We did not assign different weights to each variable because of the low variability of the hazard ratios. The scoring system had the highest AUC among other CPXT variables, and would be useful especially in Japanese patients, because the results of European studies are not easily extrapolated to subjects with smaller physique. Not only cardiac death but also re-admission could be predicted by the one simple scoring system.

When to Consider Cardiac Replacement Therapy?

Now is the era of cardiac replacement therapy with HTx or LVAD treatment, and clinicians need to decide the optimal timing of such interventions. In the present study, 4 patients died without cardiac replacement therapy, and 9 had treatment. The International Society of Heart and Lung Transplantation recently stated that PV̇O₂ 12 ml·min^–1·kg^–1 was the threshold for considering HTx under β-blocker treatment,³⁰ considering the recently accumulated evidence in the era of β-blockade.^11,26 In contrast, PV̇O₂ 14 ml·min^–1·kg^–1 is practically considered as the definitive threshold for HTx in Japan, irrespective of background medical treatment, despite little evidence from Japanese studies.

Because most Japanese patients cannot receive HTx without LVAD support, owing to the greater than 2-year waiting period after listing,³¹ the indication of HTx is in reality equivalent to that of implantable LVAD treatment. The Japanese Registry for Mechanically Assisted Circulatory Support reported that the 2-year survival under EVAHEART support, most widely distributed device in Japan currently, was 87%,³² which was almost comparable to that of patients with PV̇O₂ <14 ml·min^–1·kg^–1 (76%) or of patients in the intermediate risk group (80%). The high-risk patients would be good candidates for implantable LVAD treatment after listing for HTx, considering their lower prognosis compared with those with the EVAHEART LVAD (2-year cardiac death-free survival, 58% vs. 87%). Indeed, the specificity of PV̇O₂ <14 ml·min^–1·kg^–1 for cardiac death was only 0.625, whereas that of the high-risk group was 0.891. At least the traditional simple criterion (ie, PV̇O₂ <14 ml·min^–1·kg^–1) would still be able to distinguish optimal candidates for cardiac replacement therapy in the modern era of GDMT.

In contrast, those in the low-risk group experienced neither cardiac death nor re-admission. Such patients can be managed by ambulatory practice, and titration of GDMT is not so difficult. Those with intermediate risk are not necessarily too sick to receive LVAD, but have higher probability of re-admission. Such patients may be assigned to the so-called “frequent flyer” group,³³ and be good candidates for HTx listing without LVAD implantation.

Study Limitations

The most important limitation was the small number of subjects, because the present study was executed in a single center. The result should be confirmed in the larger, multicenter study. Mortality and re-hospitalization rates vary according to the clinical skills of the institute. Therefore, it should be noted that the study was conducted at an institute expert in the management of advanced HF, and the result may not be simply adopted in the other institutes having less experience in the treatment of advanced HF. We could not construct a multivariate model, probably because of the small sample size and mild multicollinearity among the CPXT variables. The present study was performed only in HF patients receiving GDMT who could tolerate exercise. The novel scoring system cannot simply be adopted for patients who are too ill to receive CPXT or GDMT. A validation study using the novel scoring system in another population with similar backgrounds would be a future concern. We could not analyze predictors of mortality because only 4 patients died, owing to improvement of survival from GDMT. However, all 4 deceased patients were assigned to the high-risk group.

Conclusions

A novel scoring system using 4 CPXT variables predicts both re-admission and cardiac death even in HF patients receiving current GDMT.

Disclosures

None.

References

• 1.    Fletcher  GF,  Balady  GJ,  Amsterdam  EA,  Chaitman  B,  Eckel  R,  Fleg  J, et al. Exercise standards for testing and training: A statement for healthcare professionals from the American Heart Association. Circulation 2001; 104: 1694–1740.
• 2.    Ingle  L. Theoretical rationale and practical recommendations for cardiopulmonary exercise testing in patients with chronic heart failure. Heart Fail Rev 2007; 12: 12–22.
• 3.    Weber  KT,  Kinasewitz  GT,  Janicki  JS,  Fishman  AP. Oxygen utilization and ventilation during exercise in patients with chronic cardiac failure. Circulation 1982; 65: 1213–1223.
• 4.    Mancini  DM,  Eisen  H,  Kussmaul  W,  Mull  R,  Edmunds  LH Jr,  Wilson  JR. Value of peak exercise oxygen consumption for optimal timing of cardiac transplantation in ambulatory patients with heart failure. Circulation 1991; 83: 778–786.
• 5.    Myers  J,  Gullestad  L,  Vagelos  R,  Do  D,  Bellin  D,  Ross  H, et al. Clinical, hemodynamic, and cardiopulmonary exercise test determinants of survival in patients referred for evaluation of heart failure. Ann Intern Med 1998; 129: 286–293.
• 6.    Chua  TP,  Ponikowski  P,  Harrington  D,  Anker  SD,  Webb-Peploe  K,  Clark  AL, et al. Clinical correlates and prognostic significance of the ventilatory response to exercise in chronic heart failure. J Am Coll Cardiol 1997; 29: 1585–1590.
• 7.    Ponikowski  P,  Francis  DP,  Piepoli  MF,  Davies  LC,  Chua  TP,  Davos  CH, et al. Enhanced ventilatory response to exercise in patients with chronic heart failure and preserved exercise tolerance: Marker of abnormal cardiorespiratory reflex control and predictor of poor prognosis. Circulation 2001; 103: 967–972.
• 8.    Tsurugaya  H,  Adachi  H,  Kurabayashi  M,  Ohshima  S,  Taniguchi  K. Prognostic impact of ventilatory efficiency in heart disease patients with preserved exercise tolerance. Circ J 2006; 70: 1332–1336.
• 9.    Arena  R,  Myers  J,  Abella  J,  Peberdy  MA,  Bensimhon  D,  Chase  P, et al. Development of a ventilatory classification system in patients with heart failure. Circulation 2007; 115: 2410–2417.
• 10.    Arena  RA,  Guazzi  M,  Myers  J,  Abella  J. The prognostic value of ventilatory efficiency with beta-blocker therapy in heart failure. Med Sci Sports Exerc 2007; 39: 213–219.
• 11.    O’Neill  JO,  Young  JB,  Pothier  CE,  Lauer  MS. Peak oxygen consumption as a predictor of death in patients with heart failure receiving beta-blockers. Circulation 2005; 111: 2313–2318.
• 12.    Hunt  SA,  Abraham  WT,  Chin  MH,  Feldman  AM,  Francis  GS,  Ganiats  TG, et al. ACC/AHA 2005 Guideline Update for the Diagnosis and Management of Chronic Heart Failure in the Adult: A report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Writing Committee to Update the 2001 Guidelines for the Evaluation and Management of Heart Failure): Developed in collaboration with the American College of Chest Physicians and the International Society for Heart and Lung Transplantation: Endorsed by the Heart Rhythm Society. Circulation 2005; 112: e154–e235, doi:10.1161/CIRCULATIONAHA.105.167586.
• 13.    Myers  J,  Arena  R,  Dewey  F,  Bensimhon  D,  Abella  J,  Hsu  L, et al. A cardiopulmonary exercise testing score for predicting outcomes in patients with heart failure. Am Heart J 2008; 156: 1177–1183.
• 14.    Magri  D,  Corra  U,  Di Lenarda  A,  Cattadori  G,  Maruotti  A,  Iorio  A, et al. Cardiovascular mortality and chronotropic incompetence in systolic heart failure: The importance of a reappraisal of current cut-off criteria. Eur J Heart Fail 2014; 16: 201–209.
• 15.    Jorde  UP,  Vittorio  TJ,  Kasper  ME,  Arezzi  E,  Colombo  PC,  Goldsmith  RL, et al. Chronotropic incompetence, beta-blockers, and functional capacity in advanced congestive heart failure: Time to pace? Eur J Heart Fail 2008; 10: 96–101.
• 16.    Imamura  T,  Kinugawa  K,  Hatano  M,  Fujino  T,  Inaba  T,  Maki  H, et al. Low cardiac output stimulates vasopressin release in patients with stage d heart failure. Circ J 2014; 78: 2259–2267.
• 17.    Imamura  T,  Kinugawa  K,  Hatano  M,  Fujino  T,  Inaba  T,  Maki  H, et al. Status 2 patients had poor prognosis without mechanical circulatory support. Circ J 2014; 78: 1396–1404.
• 18.    Imamura  T,  Kinugawa  K,  Kato  N,  Muraoka  H,  Fujino  T,  Inaba  T, et al. Late-onset right ventricular failure in patients with preoperative small left ventricle after implantation of continuous flow left ventricular assist device. Circ J 2014; 78: 625–633.
• 19.    McKee  PA,  Castelli  WP,  McNamara  PM,  Kannel  WB. The natural history of congestive heart failure: The Framingham study. N Engl J Med 1971; 285: 1441–1446.
• 20.    Borg  GA. Psychophysical bases of perceived exertion. Med Sci Sports Exerc 1982; 14: 377–381.
• 21.    Hori  M,  Nagai  R,  Izumi  T,  Matsuzaki  M. Efficacy and safety of bisoprolol fumarate compared with carvedilol in Japanese patients with chronic heart failure: Results of the randomized, controlled, double-blind, Multistep Administration of bisoprolol IN Chronic Heart Failure II (MAIN-CHF II) study. Heart Vessels 2014; 29: 238–247.
• 22.    McKelvie  RS,  Yusuf  S,  Pericak  D,  Avezum  A,  Burns  RJ,  Probstfield  J, et al. Comparison of candesartan, enalapril, and their combination in congestive heart failure: Randomized evaluation of strategies for left ventricular dysfunction (RESOLVD) pilot study: The RESOLVD Pilot Study Investigators. Circulation 1999; 100: 1056–1064.
• 23.    DeLong  ER,  DeLong  DM,  Clarke-Pearson  DL. Comparing the areas under two or more correlated receiver operating characteristic curves: A nonparametric approach. Biometrics 1988; 44: 837–845.
• 24.    Costanzo  MR,  Augustine  S,  Bourge  R,  Bristow  M,  O’Connell  JB,  Driscoll  D, et al. Selection and treatment of candidates for heart transplantation: A statement for health professionals from the Committee on Heart Failure and Cardiac Transplantation of the Council on Clinical Cardiology, American Heart Association. Circulation 1995; 92: 3593–3612.
• 25.    Pohwani  AL,  Murali  S,  Mathier  MM,  Tokarczyk  RL,  Kormos  RL,  McNamara  DM, et al. Impact of beta-blocker therapy on functional capacity criteria for heart transplant listing. J Heart Lung Transplant 2003; 22: 78–86.
• 26.    Kato  TS,  Collado  E,  Khawaja  T,  Kawano  Y,  Kim  M,  Farr  M, et al. Value of peak exercise oxygen consumption combined with B-type natriuretic peptide levels for optimal timing of cardiac transplantation. Circ Heart Fail 2013; 6: 6–14.
• 27.    Agostoni  P,  Guazzi  M,  Bussotti  M,  De Vita  S,  Palermo  P. Carvedilol reduces the inappropriate increase of ventilation during exercise in heart failure patients. Chest 2002; 122: 2062–2067.
• 28.    Brubaker  PH,  Kitzman  DW. Chronotropic incompetence: Causes, consequences, and management. Circulation 2011; 123: 1010–1020.
• 29.    Roche  F,  Pichot  V,  Da Costa  A,  Isaaz  K,  Costes  F,  Dall’Acqua  T, et al. Chronotropic incompetence response to exercise in congestive heart failure, relationship with the cardiac autonomic status. Clin Physiol 2001; 21: 335–342.
• 30.    Mehra  MR,  Kobashigawa  J,  Starling  R,  Russell  S,  Uber  PA,  Parameshwar  J, et al. Listing criteria for heart transplantation: International Society for Heart and Lung Transplantation guidelines for the care of cardiac transplant candidates – 2006. J Heart Lung Transplant 2006; 25: 1024–1042.
• 31.    Fukushima  N,  Ono  M,  Saito  S,  Saiki  Y,  Kubota  S,  Tanoue  Y, et al. Heart donation in Japan before and after revision of Japanese Transplantation Act. Transplant Proc 2014; 46: 2050–2053.
• 32.    Saito  S,  Yamazaki  K,  Nishinaka  T,  Ichihara  Y,  Ono  M,  Kyo  S, et al. Post-approval study of a highly pulsed, low-shear-rate, continuous-flow, left ventricular assist device, EVAHEART: A Japanese multicenter study using J-MACS. J Heart Lung Transplant 2014; 33: 599–608.
• 33.    Kinugawa  K. How to treat stage D heart failure?: When to implant left ventricular assist devices in the era of continuous flow pumps? Circ J 2011; 75: 2038–2045.