Abnormal Heart Rate Recovery Immediately After Cardiopulmonary Exercise Testing in Heart Failure Patients

Abnormal Heart Rate Recovery Immediately After Cardiopulmonary Exercise Testing in Heart Failure Patients Tuba BILSEL, 1 MD, Sait TERZI, 1 MD, Tamer AKBULUT, 1 MD, Nurten SAYAR, 1 MD, Gultekin HOBIKOGLU, 1 MD, and Kemal YESILCIMEN, 1 MD SUMMARY An attenuated heart rate recovery (HRR) immediately after exercise has been shown to be predictive of mortality. It is not known whether HRR predicts mortality when measured in patients with heart failure. The present study was undertaken to evaluate the ability of HRR to predict mortality in patients with heart failure. We studied 84 NYHA class II or III chronic congestive heart failure patients who had a left ventricular ejection fraction 40%. All patients underwent symptom limited cardiopulmonary exercise testing. The value for the HRR was defined as the difference in heart rate between peak exercise and one-minute later; a value 18 beats per minute was considered abnormal. The patients were divided into 2 groups according to the value of HRR. Those with abnormal HRR were assigned to group I and those with normal HRR were assigned to group II. The 2 groups were compared with each other regarding baseline characteristics and exercise capacity assessed by peak VO 2. There were 26 patients (31%) in group I and 58 patients (69%) in group II. Group II patients had better performance on treadmill exercise testing than group I patients. They had greater exercise duration (7.5 ± 3.8 minutes versus 5 ± 3.5 minutes, P = 0.006), better heart-rate reserve (79 ± 25% versus 63 ± 27%, P = 0.01), and higher values of maximal heart-rate (141 ± 18 beats/min versus 132 ± 17 beats/min, P = 0.04). Group II patients also had higher peak VO 2 values (16.8 ± 4.4 ml/kg/min versus 14.4 ± 3.6 ml/ kg/min, P = 0.01). When we separated the groups according to beta-blocker usage, betablockers had no prominent effect on HRR. In the follow-up period (mean 14.1 ± 6.1 months), the presence of abnormal HRR and lower peak VO 2 ( 14 ml/kg/min) were the only significant predictors of mortality in our patient population (adjusted hazard ratio [HR] 5.2, 95% CI, 1.3 to 24, P = 0.03 and adjusted HR 13, 95% CI, 2.1 to 25.6, P = 0.005, respectively). It seems that the attenuated HRR value one minute after peak exercise appears to be a reliable index of the severity of exercise intolerance in heart failure patients and this study supports the value of HRR as a prognostic marker among heart failure patients referred for cardiopulmonary exercise testing for prediction of prognosis. (Int Heart J 2006; 47: 431-440) Key words: Heart-rate recovery, Heart failure, Cardiopulmonary exercise testing From the 1 Cardiology Clinic, Dr. Siyami Ersek Thoracic & Cardiovascular Surgery Training and Research Hospital, Istanbul, Turkey. Address for correspondence: Tuba Bilsel, MD, Acibadem Cad. Bilsel Apt. No: 128/9, 34718, Kadikoy, Istanbul, Turkey. Received for publication September 20, 2005. Revised and accepted March 23, 2006. 431

432 BILSEL, ET AL Int Heart J May 2006 THE fall in heart rate immediately after exercise is considered to be a function of reactivation of the parasympathetic nervous system, which is thought to be primarily important during the first minute after exercise. A delayed decrease in the heart rate during the first minute after treadmill testing has been shown to be an important predictor of overall mortality in healthy subjects, the presence or absence of coronary artery disease, and with preserved left ventricular function. 1-5) Abnormal heart rate recovery (HRR) is usually defined as a heart rate that declines 12 beats/min in the first minute after exercise for protocols that use a postexercise cool-down or 18 beats/min in the first minute after exercise for protocols that stop exercise abruptly. 1,6,7) Although some studies have demonstrated that HRR is attenuated in patients with left ventricular systolic dysfunction, 8,9) it is not known whether HRR predicts mortality in heart failure patients. Peak oxygen consumption (VO 2 ) during cardiopulmonary exercise testing (CPET) is a powerful independent prognostic marker to risk stratify patients with heart failure and an objective measure of functional capacity. 10-12) The goal of this study was to evaluate the importance of HRR in predicting mortality in patients with heart failure who had undergone CPET. METHODS Study population: The study population consisted of 84 patients (71 men, 13 women) referred for CPET. Inclusion criteria were stable congestive heart failure (CHF) symptoms (> 3 months), EF 40%, duration of CHF symptoms > 1 year, and exercise not limited by joint pain, peripheral vascular disease, or chest pain. Patients who had anemia (hemoglobin < 10 g/dl), atrial fibrillation, an implanted pacemaker, congenital or severe organic valvular heart disease, chronic obstructive pulmonary disease, exercise induced angina and/or ischemic ECG changes were excluded. The functional class of each patient was classified according to NYHA. Cardioactive medications were classified as beta-blockers, diuretics (furosemide, hydrochlorothiazide), spironolactone, digoxin, and angiotensinconverting enzyme (ACE) inhibitors. Hypertension was defined as systolic blood pressure of 140 mmhg and a diastolic blood pressure of 90 mmhg at rest and/or the presence of history of hypertension. Usage of hypoglycemic medications and/or fasting blood glucose 126 mg/dl established the diagnosis of diabetes mellitus. A history of coronary artery disease was considered present if diagnosed by prior angiography, if there was documented hospitalization for myocardial infarction or unstable angina. Patients were considered smokers if they regularly smoked cigarettes within the past year.

Vol 47 No 3 ABNORMAL HEART RATE RECOVERY IN HEART FAILURE 433 Left ventricular systolic function was analyzed by contrast ventriculography or transthoracic echocardiography. Cardiopulmonary exercise testing: Symptom-limited CPET with respiratory gas exchange analysis was performed in all patients without stopping their present medications. A one-minute step treadmill exercise testing protocol (with a constant treadmill speed of 1.7 mph at 0 slope and the slope increased 2 every minute) was used in all patients. All exercise tests were limited by dyspnea or fatigue. Blood pressure was recorded at baseline, and every 3 minutes during exercise and at the end of the exercise period. Heart rate was measured at rest, during each minute of exercise, at maximum exercise, and at 1 minute into recovery. Levels of peak VO 2 were determined breath-by-breath using a Cortex Biophysics Metalyzer 3B CPET system. Peak VO 2 was defined as the highest VO 2 level observed during exercise. Chronotropic response was assessed on the basis of the proportion of the heart rate reserve used as peak exercise [(peak heart rate - resting heart rate) / (220 - age-resting heart rate)]; 2) a value of 0.80 was considered chronotropic incompetence. After achieving a peak workload, the treadmill was stopped and patients were allowed to sit down without a cool down period. The value for the HRR was defined as the difference in heart rate between peak exercise and one minute later; a value 18 beats per minute was considered abnormal. According to the value of HRR, patients were divided into 2 groups. Patients with an abnormal HRR were assigned to group I and patients with a normal HRR were included in group II. The 2 groups were compared to each other regarding baseline characteristics and exercise capacity that has been assessed by peak VO 2. The primary end point was all-cause mortality, as determined by inspection of hospital files or by phone calls. Median follow-up time was 14.1 ± 6.1 months (6 to 27 months). Statistical analysis: Continuous variables are presented as the mean ± SD. Differences between groups were estimated using the chi-square test or Student's t- test as appropriate. The association of HRR with mortality was tested by Kaplan- Meier survival analysis and the Cox proportional hazard model. Multivariate proportional hazard regression analysis was performed to determine which variables significantly predict mortality and adjusted hazard ratios were calculated. Statistical analysis was performed using SPSS for Windows v.11.0 software. A P 0.05 was considered significant.

434 BILSEL, ET AL RESULTS Int Heart J May 2006 The 84 patients had a mean ejection fraction of 31% and a mean NYHA class of 2.4. The etiology of left ventricular systolic dysfunction was ischemic in 54 (64.3%) patients and nonischemic in 30 (35.7%) patients. Medication included ACE inhibitors (94%), diuretics (70.2%), beta-blockers (53.6%), spironolactone (29.8%), and digoxin (22.6%). The mean HRR was approximately 24.6 ± 12.2 beats/min with a range from 3 to 65 beats/min. Twenty-six patients (31%) had an abnormal HRR value. The baseline characteristics of the patients according to the value of their HRR are presented in Table I. When the 2 groups were compared to each other, there were no marked differences between the groups regarding age (61.7 ± 10.2 years versus 61.2 ± 9.6 years, P = ), sex (male, 80% versus 86%, P = 0.5), presence of coronary artery disease (57% versus 67%, P = ), diabetes (23% versus 13%, P = 0.2), hypertension (46% versus 43%, P = ), and left ventricular ejection fraction (29 ± 0.4% versus 31 ± 5.6%, P = 0.2). However, there was a difference with respect to NYHA class (2.6 ± 0.4 versus 2.3 ± 0.4, P = 0.006). There were also no marked differences between the groups concerning the use of cardioactive medications. Exercise characteristics according to the HRR are summarized in Table II. Patients in group II had a significantly better performance during the exercise testing as compared to those in group I. In particular, they had greater exercise duration (7.5 ± 3.8 minutes versus 5 ± 3.5 minutes, P = 0.006), a better chrono- Table I. Baseline Characteristics of Heart Failure Patients According to the Value of HRR Population characteristic HRR 18 (n = 26) HRR > 18 (n = 58) P Age Sex, male (%) Coronary artery disease, (%) Smoking, (%) LVEF, (%) NYHA class Diabetes, (%) Hypertension, (%) Beta-blocker usage, (%) ACE-I usage, (%) Digoxin usage, (%) Spironolactone usage, (%) Diuretic usage, (%) Resting systolic BP, mmhg Resting diastolic BP, mmhg 61.7 ± 10.2 21 (80) 15 (57) 3 (11.5) 29.8 ± 5.3 2.6 ± 0.4 6 (23) 12 (46) 12 (46) 23 (88) 5 (19) 7 (26) 19 (73) 124 ± 19 74 ± 8.5 61.2 ± 9.6 50 (86) 39 (67) 4 (6) 31 ± 5.6 2.3 ± 0.4 8 (13) 25 (43) 33 (56) 56 (82) 14 (24) 18 (31) 40 (68) 124 ± 16 73 ± 9.3 0.5 0.4 0.2 0.006 0.2 0.1 0.5 LVEF indicates left ventricular ejection fraction; ACE-I, angiotensin converting enzyme inhibitor; and BP, blood pressure.

Vol 47 No 3 ABNORMAL HEART RATE RECOVERY IN HEART FAILURE 435 tropic index (79 ± 25% versus 63 ± 27%, P = 0.01), and higher values of maximal heart rate (141 ± 18 bpm versus 132 ± 17 bpm, P = 0.04). Chronotropic incompetence was also more common in group I than group II (88% versus 51%, respectively). Patients in group I had a significantly lower peak VO 2 level compared to group II (14.4 ± 3.6 versus 16.8 ± 4.4 ml/kg/min, P = 0.01). When the groups were divided regarding beta-blocker usage, beta-blockers had no prominent effect on HRR (Table III). Patients taking a beta-blocker had significantly decreased resting heart rate (79 ± 14.5 versus 98.6 ± 13.9, P = 0.01) and maximum heart rate (134.4 ± 19.6 versus 143 ± 15.4, P = 0.02) compared with patients not taking a beta-blocker. Table II. Exercise Characteristics of Heart Failure Patients According to the Value of HRR HRR 18 (n = 26) HRR > 18 (n = 58) P Resting HR (beats/min) Peak HR (beats/min) Exercise time, minutes Chronotropic index Chronotropic incompetence, (%) Peak systolic BP, mmhg Peak diastolic BP, mmhg Peak VO 2 (ml/kg/min) 93 ± 17.2 132 ± 17 5 ± 3.5 3 ± 0.27 23 (88) 142 ± 15 76 ± 11 14.4 ± 3.6 86.3 ± 16.8 141 ± 18 7.5 ± 3.8 9 ± 0.25 30 (51) 144 ± 16 77 ± 12 16.8 ± 4.4 0.09 0.04 0.006 0.01 0.01 0.5 0.01 HRR indicates heart rate recovery; HR, heart rate; and BP, blood pressure. Table III. Effect of Beta-Blocker Usage on Exercise Parameters Beta-blockers (-) (n = 39) Beta-blockers (+) (n = 45) P Age, years LVEF, (%) NYHA class Resting systolic BP, mmhg Resting diastolic BP, mmhg 61.8 ± 9.2 29.7 ± 5.5 2.4 ± 0.5 123.4 ± 16.7 74.6 ± 9.3 61.2 ± 1 31.9 ± 5.4 2.4 ± 0.5 125.2 ± 18.2 72.7 ± 8.8 0.07 0.9 Exercise parameters Resting HR, bpm Peak HR, bpm Exercise time, minutes Chronotropic index Chronotropic incompetence, (%) Peak systolic BP, mmhg Peak diastolic BP, mmhg HRR 18, (%) Peak VO 2 (ml/kg/min) 98.6 ± 13.9 143 ± 15.4 6.4 ± 4.0 ± 0.2 22/39 (56.4) 142 ± 15 76 ± 11 14/39 (35.8) 15.6 ± 4.1 79 ± 14.5 134.4 ± 19.6 7.0 ± 3.8 ± 0.2 31/45 (68.8) 144 ± 16 77 ± 12 12/45 (26.6) 16.6 ± 4.5 0.01 0.02 0.5 0.1 0.2 LVEF indicates left ventricular ejection fraction; BP, blood pressure; HR, heart rate; and HRR, heart rate recovery.

436 BILSEL, ET AL Int Heart J May 2006 Table IV. Risk of Death According to Univariate Analysis Variable Alive (n = 70) Dead (n = 14) Odds ratio P Age, years Ischemic etiology, (%) Diabetes, (%) Female sex, (%) Present smoker, (%) LVEF, % Chronotropic incompetence, (%) Peak VO 2 14 ml/kg/min 6 ± 10.1 46/70 (65.7) 11/70 (15.7) 9/70 (12.8) 5/70 (7.1) 31.3 ± 5.7 45/70 (62.2) 18/70 (25.7) 65.1 ± 6.9 8/14 (57.1) 3/14 (21.4) 4/14 (28.5) 2/14 (14.2) 29 ± 4.4 8/14 (57.1) 10/14 (71.4) 4 (0.28-1.93) 1.36 (0.43-4.26) 2.15 (0.80-5.92) 1.83 (0.50-6.60) 0 (0.29-2.04) 5 (1.72-14.53) 0.1 0.5 0.1 0.1 0.0001 Exercise time, minutes Abnormal HRR, (%) 7.18 ± 3.9 18/70 (25.7) 4.7 ± 3.4 8/14 (57.1) 2.97 (1.14-7.70) 0.03 0.02 LVEF indicates left ventricular ejection fraction and HRR, heart rate recovery. Table V. Results of Multivariable Analysis Predictor Hazard ratio (95% CI) P Older age ( 65 years) LVEF Female sex Diabetes Chronotropic incompetence Exercise time Abnormal HRR Peak VO 2 14 ml/kg/min 0.4 (0.1-1.7) 0.9 (0.8-1.05) 1.6 (0.4-2.6) (0.1-3.7) 0.2 (0.03-1.2) 1.2 (0.8-1.5) 5.2 (1.3-24) 13 (2.1-25.6) 0.2 0.4 0.09 0.2 0.03 0.005 LVEF indicates left ventricular ejection fraction and HRR, heart rate recovery. During the follow-up period (mean 14.1 ± 6.1 months), there were 14 deaths. As seen in the Figure, patients with heart failure had significantly decreased survival if they had attenuated HRR. Predictors of death in univariate analysis are shown in Table IV. Presence of lower peak VO 2, exercise time, and attenuated HRR were variables associated with death. The results of multivariate Cox proportional hazard regression analysis are presented in Table V. An abnormal HRR and the presence of lower peak VO 2 ( 14 ml/kg/min) remained independently predictive of death in our study population (adjusted hazard ratio, 5.2, 95% CI, 1.3 to 24, P = 0.03, adjusted HR, 13, 95% CI, 2.1 to 25.6, P = 0.005, respectively).

Vol 47 No 3 ABNORMAL HEART RATE RECOVERY IN HEART FAILURE 437 Figure. Kaplan-Meier curves of survival according to abnormal heart rate recovery. Log-rank = 4.96, P = 0.02 DISCUSSION In normal subjects during progressive treadmill exercise, cardiac output is increased through an augmentation of heart rate and stroke volume. The increased heart rate response is mediated in part by withdrawal of vagal tone and increased sympathetic tone. The fall in heart rate immediately after exercise is considered to be a function of the reactivation of the parasympathetic nervous system which is thought to be primarily important during the first minute after exercise. Although some studies have demonstrated that HRR is attenuated in patients with left ventricular systolic dysfunction, 8,9) prognostic data based on HRR are unavailable for patients with CHF. It has been known for many years that patients with chronic CHF have autonomic dysfunction and neurohormonal activation. Since then, a variety of measurements of autonomic function have demonstrated that patients with CHF exhibit attenuated parasympathic and increased sympathetic activity. Imai, et al demonstrated that vagally-mediated HRR after exercise is accelerated in athletes. 8) In contrast to athletes, patients with CHF have decreased tonic vagal activity. 13,14) Blunted vagal reactivation in patients with CHF may be associated with

438 BILSEL, ET AL Int Heart J May 2006 end organ subsensitivities to vagal stimulation. Vatner, et al demonstrated a decrease in cardiac muscarinic receptor density in dogs with experimental heart failure. 15) The higher frequency of abnormal HRR value in heart failure patients is most probably related to the influence of autonomic nervous system dysfunction on the HRR value. In previous studies, it was shown that patients with an attenuated HRR were older, had a more adverse risk profile, and an impaired functional capacity. However, in our study, only higher NYHA class was found to be related with an attenuated HRR. Age and other cardiovascular risk factors were found not to be significantly related with HRR. This conflict might be explained by the low number of cases involved in the study. Chronotropic incompetence is very common in patients with chronic CHF. 16) Most of our patients (63%) also could not successfully reach an expected maximal heart rate during CPET. It was shown that, when exercise is submaximal, attenuated HRR is still predictive of death in a healthy population. 3) In the present study, when exercise capacity was submaximal, attenuated HRR was also found to be a strong predictor of death even in heart failure patients. There is a paucity of data on the predictive value of HRR in patients treated with beta-blockers because most studies excluded such patients. Shetler, et al failed to detect an impact of HRR on survival in patients receiving beta-blockers. 6) However, this observation was reported in patients with preserved left ventricular ejection fraction. Our study on the other hand supports the value of HRR regardless of whether or not the patient is taking a beta-blocker in patients with impaired ejection fractions. We also found that compared to patients with normal HRR, patients with an abnormal HRR had more frequent chronotropic incompetence. A greater increase in the heart rate during exercise and a better chronotropic response during testing in patients with normal HRR have been reported in other studies as well. 2,11,17-19) The impact of a low peak exercise heart rate induced by beta adrenergic blockade has been reported by Desai, et al in healthy subjects and patients with coronary artery disease with preserved left ventricular function. 17) They reported that betablockers significantly attenuated HRR after exercise. Moreover, the degree of impairment in HRR mediated by beta-blockers was proportional to the chronotropic incompetence. However, in another study, Racine, et al reported that longterm beta-blocker therapy did not change HRR in patients with heart failure. 9) In our study group half of the patients were on beta-blockers, and there were no differences between the groups regarding the frequency of beta-blocker usage. Although our study population was small, we also may conclude that beta-blockers have no prominent effect on HRR in heart failure patients. Peak VO 2 measurements in patients with compensated CHF are useful in

Vol 47 No 3 ABNORMAL HEART RATE RECOVERY IN HEART FAILURE 439 risk stratifying patients with CHF to determine the subsequent incidence of cardiac events. Several reports have suggested that a cut off point of 14 ml/kg/min for peak VO 2 be used as a criterion for listing patients for transplantation, since it appears that for patients who achieve values 14 ml/kg/min, the one-year mortality rate is greater than 30%. 11,20) In our study population, patients with an abnormal HRR had a lower peak VO 2 value than the patients with a normal HRR. Moreover, abnormal HRR was found to be as strong a predictor of death as a lower peak VO 2. This is the first and only study in the literature, to the best of our knowledge, in which HRR was evaluated together with peak VO 2 in CPET. The findings of the present study support using the value of HRR as a prognostic marker among heart failure patients referred for CPET for evaluation of their prognosis. This value is simple to calculate from data that are already contained in the results of standard exercise tests. REFERENCES 1. Cole CR, Blackstone EH, Pashkow FJ, Snader CE, Lauer MS. Heart-rate recovery immediately after exercise as a predictor of mortality. N Engl J Med 1999; 341: 1351-7. 2. Nishime EO, Cole CR, Blackstone EH, Pashkow FJ, Lauer MS. Heart rate recovery and treadmill exercise score as predictors of mortality in patients referred for exercise ECG. JAMA 2000; 284: 1392-8. 3. Cole CR, Foody JM, Blackstone EH, Lauer MS. Heart rate recovery after submaximal exercise testing as a predictor of mortality in a cardiovascularly healthy cohort. Ann Intern Med 2000; 132: 552-5. 4. Vivekananthan DP, Blackstone EH, Pothier CE, Lauer MS. Heart rate recovery after exercise is a predictor of mortality, independent of the angiographic severity of coronary disease. J Am Coll Cardiol 2003; 42: 831-8. 5. Diaz LA, Brunken RC, Blackstone EH, Snader CE, Lauer MS. Independent contribution of myocardial perfusion defects to exercise capacity and heart rate recovery for prediction of all-cause mortality in patients with known or suspected coronary heart disease. J Am Coll Cardiol 2001; 37: 1558-64. 6. Shetler K, Marcus R, Froelicher VF, et al. Heart rate recovery: validation and methodological issues. J Am Coll Cardiol 2001; 38: 1980-7. 7. Watanabe J, Thamilarasan M, Blackstone EH, Thomas JD, Lauer MS. Heart rate recovery immediately after treadmill exercise and left ventricular systolic dysfunction as a predictors of mortality: the case of stress echocardiography. Circulation 2001; 104: 1911-6. 8. Imai K, Sato H, Hori M, et al. Vagally mediated heart rate recovery after exercise is accelerated in athletes but blunted in patients with chronic heart failure. J Am Coll Cardiol 1994; 24: 1529-35. 9. Racine N, Blanchet M, Ducharme A, et al. Decreased heart rate recovery after exercise in patients with congestive heart failure: effect of β-blocker therapy. J Card Fail 2003; 9: 296-302. 10. Cohn JN, Johnson GR, Shabetai R, et al. Ejection fraction, peak exercise oxygen consumption cardiothoracic ratio, ventricular arrhythmias and plasma norepinephrine as determinants of prognosis in heart failure. The V- HeFT VA Cooperative Studies Group. Circulation 1993; 87: VI5-6. 11. Passantino A, Scrutinio D, Lagioia R, Forleo C, Ricci A, Rizzon P. Value of peak oxygen consumption during exercise for the prognostic stratification of patients with severe systolic dysfunction of the left ventricle. Cardiologia 1997; 42: 947-52. (Italian) 12. Metra M, Faggiano P, D'Aloia A, et al. Use of cardiopulmonary exercise testing with hemodynamic monitoring in the prognostic assessment of ambulatory patients with chronic heart failure. J Am Coll Cardiol 1999; 33: 943-50. 13. Eckberg DL, Drabinsky M, Braunwald E. Defective cardiac parasympathetic control in patients with heart disease. N Engl J Med 1971; 285: 877-83.

440 BILSEL, ET AL Int Heart J May 2006 14. Saul JP, Arai Y, Berger RD, Lilly LS, Colucci WS, Cohen RJ. Assessment of autonomic regulation in chronic congestive heart failure by heart rate spectral analysis. Am J Cardiol 1988; 61:1292-9. 15. Vatner DE, Lee DL, Schwarz KR, et al. Impaired cardiac muscarinic receptor function in dogs with heart failure. J Clin Invest 1988; 81: 1836-42. 16. Colucci WS, Ribeiro JP, Rocco MB, et al. Impaired chronotropic response to exercise in patients with congestive heart failure. Role of postsynaptic beta adrenergic desensitization. Circulation 1989; 80: 314-23. 17. Desai MY, De la Pena-Almaguer E, Mannting F. Abnormal heart rate recovery after exercise as a reflection of an abnormal chronotropic response. Am J Cardiol 2001; 87: 1164-9. 18. Fletcher GF, Balady G, Froelicher VF, Hartley LH, Haskell WL, Pollock ML. Exercise standards. A statement for healthcare professionals from the American Heart Association Writing Group. Circulation 1995; 91: 580-615. 19. Lauer MS, Francis GS, Okin PM, Pashkow FJ, Snader CE, Marwick TH. Impaired chronotropic response to exercise stress testing as a predictor of mortality. JAMA 1999; 281: 524-9. 20. 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: 778-86.