Peak Oxygen Consumption as a Predictor of Death in Patients With Heart Failure Receiving ;-Blockers
http://www.100md.com
《循环学杂志》
the Department of Cardiovascular Medicine, The Cleveland Clinic Foundation, Cleveland, Ohio.
Abstract
Background— Peak oxygen uptake (peak O2) is a strong predictor of mortality and is commonly used in the evaluation of patients for cardiac transplantation. ;-Blockers reduce mortality in patients with heart failure, without influencing peak O2, raising the possibility that peak O2 is no longer suitable as an indicator of prognosis in these patients.
Methods and Results— We analyzed prospectively gathered data on 2105 patients referred for cardiopulmonary testing for all-cause mortality and for occurrence of death or transplantation. Patients receiving ;-blockers were younger, more likely to have coronary disease, and had a greater mean ejection fraction but had a similar peak O2. There were 555 deaths (26%) and 194 (9%) transplants during a median follow-up of 3.5 years. Peak O2 was a predictor of mortality irrespective of ;-blocker use; a decrease of 1 mL · kg–1 · min–1 resulted in an adjusted hazard ratio (HR) of 1.13 (95% CI 1.09 to 1.17, P<0.0001) in patients not receiving ;-blockers and 1.27 (95% CI 1.18 to 1.36, P<0.0001) in patients receiving ;-blockers. Similar findings were noted when considering death or transplantation as an end point. ;-Blocker use was associated with better outcomes until peak O2 values became very low (10 mL · kg–1 · min–1), at which level survival rates were equally poor.
Conclusion— Peak O2 is a determinant of survival in patients in heart failure even in the setting of ;-blockade. Because of improved survival in patients treated with ;-blockers, the cut point value of 14 mg · kg–1 · min–1 for referral for cardiac transplantation in these patients requires reevaluation, and a lower cut point may be more appropriate.
Key Words: heart failure ; exercise ; adrenergic beta-antagonists ; prognosis ; ventricular dysfunction, left
Introduction
Eligibility for cardiac transplantation relies heavily on the measurement of peak oxygen uptake (peak O2) during cardiopulmonary exercise testing. A cutoff of 14 mL · kg–1 · min–1 has been shown to predict an increased 1-year mortality1 and has been adopted as a threshold value for consideration for cardiac transplantation.2 These data were accumulated in the late 1980s and did not include patients receiving ;-blocker therapy. Later follow-up studies performed in the 1990s did include patients taking ;-blockers.3 ;-Blockers have been uniformly associated with a one-third reduction in mortality in patients with heart failure, without consistently changing peak O2.4–6 A recent study of 127 patients with heart failure concluded that peak O2 did not predict survival in the setting of ;-blocker therapy.7
Because accurate selection of cardiac transplant recipients remains such a challenging issue and in light of the increasing gap between donor supply and demand, we sought to reassess the prognostic significance of peak O2 in patients with heart failure according to whether they were taking regular ;-blocker therapy.
Methods
This was an observational prospective cohort study of consecutive patients referred to The Cleveland Clinic Foundation with left ventricular systolic ejection fraction 35% who were referred for metabolic treadmill exercise testing between January 1995 and December 2002. Exclusion criteria included age <20 years, absence of US Social Security number, congenital or primary valvular heart disease, end-stage renal disease, or history of previous cardiac transplantation. In patients with >1 metabolic exercise test, only the initial study was included in the analysis. Because sotalol has ;-blocking as well as class III antiarrhythmic effects, patients receiving this medication (n=18) were excluded. Data were prospectively recorded on a customized computer database, which was approved by the Institutional Review Board of The Cleveland Clinic Foundation. The requirement for obtaining informed consent was waived.
Clinical and Exercise Data
Before each metabolic stress test, a structured interview and chart review yielded prospectively obtained data on demographics, left ventricular ejection fraction, medications, etiology of heart failure, and various other clinical parameters as defined previously.8 We performed symptom-limited metabolic stress testing using the Naughton protocol and recorded on a MedGraphics cardiopulmonary system. Data were prospectively collected during each stage of exercise on symptoms, rhythm, and blood pressure. Measurements of oxygen consumption (O2), CO2 production (CO2), heart rate, minute ventilation (E), tidal volume (VT), and respiratory rate were made after steady state at rest and after every 30 seconds during exercise and recovery. The ventilatory response to exercise was defined as the value of E/CO2 at peak exercise.8 Anaerobic threshold was determined by the V-slope method9 or by inspecting ventilatory equivalents.10
Because of the inclusion of patients with atrial fibrillation and patients actively receiving ;-blocker therapy, chronotropic response (previously documented to have prognostic significance in exercise testing11) was not included in the analyses. In addition, because of the inclusion of patients with permanent pacemakers, heart rate recovery (also known to be of prognostic importance12) also was not considered. The Duke treadmill score13 was not included for the above reasons and because of the prevalence of abnormal resting ECGs in this population.
End Points
The primary end point was all-cause mortality, which was determined with reference to the Social Security Death Index.14,15 We have previously shown that this method has a sensitivity of 97% for detecting death in Cleveland Clinic Foundation exercise laboratory patients.11 Cross-referencing our unified transplant database identified patients later undergoing orthotopic heart transplantation. Thus, transplant-free survival (or time free of death or transplantation) was considered as a secondary end point.
Statistical Analyses
Patients were first divided into 2 groups, based on whether they listed ;-blocker therapy among their regular medications. The Wilcoxon rank sum test was used for comparisons of continuous variables and the 2 test was used to test for comparisons of categorical variables.
Kaplan-Meier curves were constructed and nonparsimonious Cox proportional hazards modeling was performed to analyze the association of peak O2, in addition to a variety of other prospectively recorded parameters and survival or transplant-free survival. All of the variables listed in Table 1 were considered as potential confounders. For the primary end point of all-cause mortality, cardiac transplantation was considered as a time-dependent covariate irrespective of how long after stress testing the operation occurred. The proportional hazards assumption was confirmed by calculating Schoenfeld residuals. Possible nonlinear associations between the logarithm of hazard and outcome were tested with restricted cubic splines.16 We prospectively tested for an interaction between ;-blocker use and peak O2. Relative strengths of association were described based on Wald 2 statistics.
Model discrimination was tested by calculating a C-index for censored data,17 which corresponds to the area under an ROC curve; this index describes the possibility that for a randomly chosen pair of patients in which one had an event and the other did not but was followed for at least as long (or had an event after longer follow-up), the model correctly found a higher risk in the patient who had an event (or had the event earlier in follow-up). Model calibration was tested by grouping patients into quintiles of predicted risk and comparing actual 5-year Kaplan-Meier rates with predicted rates. For discrimination and calibration, 200 bootstrap resamplings were performed and the resulting 200 models used for calculating C-indexes and predicted risk.
In a supplementary analysis, we considered percent-predicted peak O2 based on age and sex as a predictor of survival or transplant-free survival.18
Data assembly and basic statistical comparisons were performed with SAS software (version 9.1, SAS Institute). Survival curves and Cox proportional hazards analyses were performed with Harrell’s Design and Hmisc libraries19 of the S-plus 6.2 software package (Insightful, Inc). A probability value of <0.05 was considered significant.
Results
Peak oxygen consumption was similar in both groups, but the E/CO2 ratio was slightly higher in patients receiving ;-blockers. The peak respiratory exchange ratio was slightly higher among patients not taking ;-blockers (1.13 versus 1.10, P<0.0001).
During a median follow-up of 3.5 years (25th and 75th percentiles 2.0 and 5.4 years) among survivors, 555 (26%) patients died and 194 (9%) underwent cardiac transplantation. Long-term survival was significantly better in patients taking ;-blockers, with an unadjusted mortality HR of 0.60 (95% CI 0.49 to 0.73, P<0.0001). The survival advantage was maintained when adjusted for the characteristics listed in Table 1 (adjusted HR 0.68, 95% CI 0.55 to 0.85, P=0.0006) and considering cardiac transplantation as a time-dependent covariate. Similarly, patients taking ;-blockers had a lower risk of experiencing death or cardiac transplantation (adjusted HR 0.63, 95% CI 0.52 to 0.76, P<0.0001).
Survival curves of patients subdivided into 4 groups on and off ;-blockers, with peak O2>14 mL · kg–1 · min–1 and 14 mL · kg–1 · min–1 are shown in Figures 1A and 2A. For both patients taking and not taking ;-blockers, a lower peak O2 was associated with a lower survival or transplant-free survival. When we repeated all of these analyses based on age- and sex-predicted peak O2, we noted essentially identical results. A lower percent-predicted peak O2 was associated with worse survival (Figure 1B) or transplant-free survival irrespective of ;-blocker use (Figure 2B).
The strongest predictor of death of peak O2 (Wald 2=76), whereas other strong predictors with a Wald 2 value >10 were men (2=21) and use of thiazide diuretics (2=13); predictors of better survival include use of ;-blockade (2=21) and undergoing cardiac transplantation (2=51). The interaction of ;-blockade and peak O2 was also strongly predictive of death (2=11).
When we repeated these regression analyses using percent-predicted peak O2 instead of absolute peak O2, the overall results were not changed materially.
Discussion
In this large cohort of 2105 patients with severe left ventricular dysfunction and symptomatic heart failure, peak O2 was a predictor of increased mortality or of a secondary end point of death or transplant, even in the so-called ;-blocker era. This remained true even after accounting for multiple possible confounding variables, including left ventricular ejection fraction, age, implantable cardioverter-defibrillators, and other concomitant medications. The relationship between peak O2 and prognosis was actually stronger in patients receiving ;-blockers than in those not receiving them.
The threshold of 14 mL · kg–1 · min–1 for referral for cardiac transplantation was based on data accumulated before the advent of ;-blockade for heart failure.1,20,21 ;-Blocker therapy is not independently associated with an improvement in peak O2,22–24 and in our group, peak O2 was similar in patients receiving and not receiving ;-blockers. These observations, coupled with the dramatic improvement in survival seen with ;-blockers, perhaps accompanied by a change in the natural history of the disease, has prompted some sources to question the prognostic significance of peak O2 in the ;-blocker era.7
Peak oxygen consumption assessment was developed as a means to formally assess functional status in patients with heart failure.25 Data from the first Veterans Administration Heart Failure Trial demonstrated that peak oxygen consumption independently predicted mortality.21 A study of 116 patients being considered for cardiac transplantation in a University of Pennsylvania program found that in patients with a peak oxygen consumption of 14 mL · kg–1 · min–1, the freedom from death or urgent cardiac transplantation was only 48% at 1 year.1 Patients without significant comorbidities and with a peak oxygen consumption of >14 mL · kg–1 · min–1 had a 1-year survival of 94%. A consensus exists that an ejection fraction <20% and a peak oxygen consumption of <14 mL · kg–1 · min–1 should be present to warrant referral for cardiac transplantation.2 Notably, however, some argue that there is an overreliance placed on this single end point for prognosis.26
A retrospective study by Shakar et al7 reviewed 127 patients who were taking ;-blockers for at least 3 months and had undergone metabolic exercise testing. They divided the patients into 2 groups, 1 with peak O2 >14 mL · kg–1 · min–1 (n=109) and the other with peak O2 <14 mL · kg–1 · min–1 (n=18). At 30 months, the number of patients who reached the combined end point of death or transplantation was similar. They concluded from these data that the current peak O2 cutoff does not predict survival without transplantation of patients on chronic ;-blocker therapy.
In another study of 408 patients with heart failure and ejection fraction <45%, Zugeck and colleagues assessed outcome according to ;-blocker use.27 They used a combined clinical end point of progressive heart failure (requiring hospital admission), inotropes, intravenous diuretics, and mechanical support or cardiac death or both at 1 year. A peak O2 14 mL · kg–1 · min–1 in patients who were receiving ;-blockers was associated with the combined end point 23% to 26%, whereas for patients who did not receive ;-blockers, it was 35% to 64%.
A further study by Peterson et al of 540 patients with heart failure compared the outcomes from nontransplanted patients with data from the International Society for Heart and Lung Transplantation transplant database, according to peak O2.28 They concluded that patients taking ;-blockers with peak O2 12 mL · kg–1 · min–1 had greater 1- and 3-year survival (with transplantation a censored event) than did post-transplant patients. For patients not taking ;-blockers and peak O2 <14 mL · kg–1 · min–1, however, survival was worse than it was for transplant recipients.
Our study is the largest to date to look at this issue. In addition, we used mortality as a hard end point. Some limitations of our study are worth noting, however. Because our data are derived from a cohort seen at a referral center with a high cardiac transplant volume, there is a need to confirm our results elsewhere. Using the Social Security Death Index meant that we did not have data regarding the cause of death. Our group and others have addressed the issue of assessing cause of death in patients with cardiovascular disease. Attempting to classify cause of death is problematic, whereas all-cause mortality is an objective, clinically relevant, and unbiased end point.29–31 Because of the inclusion of patients with atrial fibrillation, pacemakers, and ;-blocker use, we could not incorporate chronotropic response or heart rate recovery into our survival models. We did not have data on ;-blocker dose (although we did have data on resting heart rate), nor did we have data regarding who prescribed ;-blockers and why. We did not have detailed data regarding symptoms such as paroxysmal nocturnal dyspnea. Our analyses were based on peak O2 as an absolute value. Although some have proposed using percent-predicted O2 based on age and sex, this value has not been shown to be more effective at predicting survival than the simple value.3 Furthermore, in our supplementary analyses, results were essentially equivalent when using the percent-predicted value.
Cardiac transplantation was a competing event, which we treated as a time-dependent covariate and also as a secondary end point; this seems a reasonable strategy because it is unlikely that patients were specifically chosen for transplant on the basis of ;-blocker use; however, peak O2 most likely was used in selecting patients for transplantation.
It is noteworthy that at low levels of peak O2, ;-blockade was no longer associated with a survival benefit (see Figure 3 at about 10 mL · kg–1 · min–1). This suggests that the improved survival observed with ;-blockade may be weaker among patients with severely impaired functional capacity, in which mortality rates are high irrespective of treatment.
Our findings confirm the prognostic importance of peak O2 in patients with heart failure in the ;-blocker era. Indeed, a low peak O2 may be an even more powerful predictor of mortality in patients receiving ;-blocker therapy than in those treated without them. Because of improved survival in patients treated with ;-blockers, the traditional cut point of <14 mg · kg–1 · min–1 for referral for cardiac transplantation in these patients requires reevaluation, and a lower cut point value may be more appropriate. In addition, these data confirm the powerful survival advantage imparted by ;-blocker therapy, but primarily among those patients with a functional capacity exceeding 10 mL · kg–1 · min–1. Clinicians caring for patients with heart failure should recognize the value of prescribing ;-blockers for patients with impaired systolic ventricular function and that even in the setting of ;-blockade cardiopulmonary stress testing is a valuable prognostic tool.
Acknowledgments
Dr Lauer and Ms Pothier receive support from the National Heart, Lung, and Blood Institute (Grant HL-66004); Dr O’Neill receives support from the Fulbright Commission.
References
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.
Costanzo MR, Augustine S, Bourge R, Bristow M, O’Connell JB, Driscoll D, Rose E. 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.
Myers J, Gullestad L, Vagelos R, Do D, Bellin D, Ross H, Fowler MB. Clinical, hemodynamic, and cardiopulmonary exercise test determinants of survival in patients referred for evaluation of heart failure. Ann Intern Med. 1998; 129: 286–293.
Packer M, Bristow MR, Cohn JN, Colucci WS, Fowler MB, Gilbert EM, Shusterman NH. The effect of carvedilol on morbidity and mortality in patients with chronic heart failure. U.S. Carvedilol Heart Failure Study Group. N Engl J Med. 1996; 334: 1349–1355.
Effect of metoprolol CR/XL in chronic heart failure: Metoprolol CR/XL Randomised Intervention Trial in Congestive Heart Failure (MERIT-HF). Lancet. 1999; 353: 2001–2007.
Packer M, Coats AJ, Fowler MB, Katus HA, Krum H, Mohacsi P, Rouleau JL, Tendera M, Castaigne A, Roecker EB, Schultz MK, DeMets DL. Effect of carvedilol on survival in severe chronic heart failure. N Engl J Med. 2001; 344: 1651–1658.
Shakar SF, Lowes BD, Lindenfeld J, Zolty R, Simon M, Robertson AD, Bristow MR, Wolfel EE. Peak oxygen consumption and outcome in heart failure patients chronically treated with beta-blockers. J Card Fail. 2004; 10: 15–20.
Robbins M, Francis G, Pashkow FJ, Snader CE, Hoercher K, Young JB, Lauer MS. Ventilatory and heart rate responses to exercise: better predictors of heart failure mortality than peak oxygen consumption. Circulation. 1999; 100: 2411–2417.
Beaver WL, Wasserman K, Whipp BJ. A new method for detecting anaerobic threshold by gas exchange. J Appl Physiol. 1986; 60: 2020–2027.
Wasserman K. Breathing during exercise. N Engl J Med. 1978; 298: 780–785.
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: 1392–1398.
Coronary artery surgery study (CASS): a randomized trial of coronary artery bypass surgery. Quality of life in patients randomly assigned to treatment groups. Circulation. 1983; 68: 951–960.
Mark DB, Hlatky MA, Harrell FE Jr, Lee KL, Califf RM, Pryor DB. Exercise treadmill score for predicting prognosis in coronary artery disease. Ann Intern Med. 1987; 106: 793–800.
Curb JD, Ford CE, Pressel S, Palmer M, Babcock C, Hawkins CM. Ascertainment of vital status through the National Death Index and the Social Security Administration. Am J Epidemiol. 1985; 121: 754–766.
Newman TB, Brown AN. Use of commercial record linkage software and vital statistics to identify patient deaths. J Am Med Inform Assoc. 1997; 4: 233–237.
Harrell FE Jr, Lee KL, Mark DB. Multivariable prognostic models: issues in developing models, evaluating assumptions and adequacy, and measuring and reducing errors. Stat Med. 1996; 15: 361–387.
Harrell FE Jr, Califf RM, Pryor DB, Lee KL, Rosati RA. Evaluating the yield of medical tests. JAMA. 1982; 247: 2543–2546.
Hansen JE, Sue DY, Wasserman K. Predicted values for clinical exercise testing. Am Rev Respir Dis. 1984; 129: S49–S55.
Harrell FE Jr. Regression Modeling Strategies with Applications to Linear Models, Logistic Regression, and Survival Analysis. New York, NY: Springer-Verlag; 2001.
Francis GS. Determinants of prognosis in patients with heart failure. J Heart Lung Transplant. 1994; 13: S113–S116.
Cohn JN, Johnson GR, Shabetai R, Loeb H, Tristani F, Rector T, Smith R, Fletcher R. 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–VI16.
Gilbert EM, Abraham WT, Olsen S, Hattler B, White M, Mealy P, Larrabee P, Bristow MR. Comparative hemodynamic, left ventricular functional, and antiadrenergic effects of chronic treatment with metoprolol versus carvedilol in the failing heart. Circulation. 1996; 94: 2817–2825.
Metra M, Giubbini R, Nodari S, Boldi E, Modena MG, Dei Cas L. Differential effects of beta-blockers in patients with heart failure: a prospective, randomized, double-blind comparison of the long-term effects of metoprolol versus carvedilol. Circulation. 2000; 102: 546–551.
Gullestad L, Manhenke C, Aarsland T, Skardal R, Fagertun H, Wikstrand J, Kjekshus J. Effect of metoprolol CR/XL on exercise tolerance in chronic heart failure—a substudy to the MERIT-HF trial. Eur J Heart Fail. 2001; 3: 463–468.
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.
Francis GS, Rector TS. Maximal exercise tolerance as a therapeutic end point in heart failure—are we relying on the right measure; Am J Cardiol. 1994; 73: 304–306.
Zugck C, Haunstetter A, Kruger C, Kell R, Schellberg D, Kubler W, Haass M. Impact of beta-blocker treatment on the prognostic value of currently used risk predictors in congestive heart failure. J Am Coll Cardiol. 2002; 39: 1615–1622.
Peterson LR, Schechtman KB, Ewald GA, Geltman EM, de las Fuentes L, Meyer T, Krekeler P, Moore ML, Rogers JG. Timing of cardiac transplantation in patients with heart failure receiving beta-adrenergic blockers. J Heart Lung Transplant. 2003; 22: 1141–1148.
Lauer MS, Blackstone EH, Young JB, Topol EJ. Cause of death in clinical research: time for a reassessment; J Am Coll Cardiol. 1999; 34: 618–620.
Gottlieb SS. Dead is dead—artificial definitions are no substitute. Lancet. 1997; 349: 662–663.
Lloyd-Jones DM, Martin DO, Larson MG, Levy D. Accuracy of death certificates for coding coronary heart disease as the cause of death. Ann Intern Med. 1998; 129: 1020–1026.(James O. O’Neill, MB, MRC)
Abstract
Background— Peak oxygen uptake (peak O2) is a strong predictor of mortality and is commonly used in the evaluation of patients for cardiac transplantation. ;-Blockers reduce mortality in patients with heart failure, without influencing peak O2, raising the possibility that peak O2 is no longer suitable as an indicator of prognosis in these patients.
Methods and Results— We analyzed prospectively gathered data on 2105 patients referred for cardiopulmonary testing for all-cause mortality and for occurrence of death or transplantation. Patients receiving ;-blockers were younger, more likely to have coronary disease, and had a greater mean ejection fraction but had a similar peak O2. There were 555 deaths (26%) and 194 (9%) transplants during a median follow-up of 3.5 years. Peak O2 was a predictor of mortality irrespective of ;-blocker use; a decrease of 1 mL · kg–1 · min–1 resulted in an adjusted hazard ratio (HR) of 1.13 (95% CI 1.09 to 1.17, P<0.0001) in patients not receiving ;-blockers and 1.27 (95% CI 1.18 to 1.36, P<0.0001) in patients receiving ;-blockers. Similar findings were noted when considering death or transplantation as an end point. ;-Blocker use was associated with better outcomes until peak O2 values became very low (10 mL · kg–1 · min–1), at which level survival rates were equally poor.
Conclusion— Peak O2 is a determinant of survival in patients in heart failure even in the setting of ;-blockade. Because of improved survival in patients treated with ;-blockers, the cut point value of 14 mg · kg–1 · min–1 for referral for cardiac transplantation in these patients requires reevaluation, and a lower cut point may be more appropriate.
Key Words: heart failure ; exercise ; adrenergic beta-antagonists ; prognosis ; ventricular dysfunction, left
Introduction
Eligibility for cardiac transplantation relies heavily on the measurement of peak oxygen uptake (peak O2) during cardiopulmonary exercise testing. A cutoff of 14 mL · kg–1 · min–1 has been shown to predict an increased 1-year mortality1 and has been adopted as a threshold value for consideration for cardiac transplantation.2 These data were accumulated in the late 1980s and did not include patients receiving ;-blocker therapy. Later follow-up studies performed in the 1990s did include patients taking ;-blockers.3 ;-Blockers have been uniformly associated with a one-third reduction in mortality in patients with heart failure, without consistently changing peak O2.4–6 A recent study of 127 patients with heart failure concluded that peak O2 did not predict survival in the setting of ;-blocker therapy.7
Because accurate selection of cardiac transplant recipients remains such a challenging issue and in light of the increasing gap between donor supply and demand, we sought to reassess the prognostic significance of peak O2 in patients with heart failure according to whether they were taking regular ;-blocker therapy.
Methods
This was an observational prospective cohort study of consecutive patients referred to The Cleveland Clinic Foundation with left ventricular systolic ejection fraction 35% who were referred for metabolic treadmill exercise testing between January 1995 and December 2002. Exclusion criteria included age <20 years, absence of US Social Security number, congenital or primary valvular heart disease, end-stage renal disease, or history of previous cardiac transplantation. In patients with >1 metabolic exercise test, only the initial study was included in the analysis. Because sotalol has ;-blocking as well as class III antiarrhythmic effects, patients receiving this medication (n=18) were excluded. Data were prospectively recorded on a customized computer database, which was approved by the Institutional Review Board of The Cleveland Clinic Foundation. The requirement for obtaining informed consent was waived.
Clinical and Exercise Data
Before each metabolic stress test, a structured interview and chart review yielded prospectively obtained data on demographics, left ventricular ejection fraction, medications, etiology of heart failure, and various other clinical parameters as defined previously.8 We performed symptom-limited metabolic stress testing using the Naughton protocol and recorded on a MedGraphics cardiopulmonary system. Data were prospectively collected during each stage of exercise on symptoms, rhythm, and blood pressure. Measurements of oxygen consumption (O2), CO2 production (CO2), heart rate, minute ventilation (E), tidal volume (VT), and respiratory rate were made after steady state at rest and after every 30 seconds during exercise and recovery. The ventilatory response to exercise was defined as the value of E/CO2 at peak exercise.8 Anaerobic threshold was determined by the V-slope method9 or by inspecting ventilatory equivalents.10
Because of the inclusion of patients with atrial fibrillation and patients actively receiving ;-blocker therapy, chronotropic response (previously documented to have prognostic significance in exercise testing11) was not included in the analyses. In addition, because of the inclusion of patients with permanent pacemakers, heart rate recovery (also known to be of prognostic importance12) also was not considered. The Duke treadmill score13 was not included for the above reasons and because of the prevalence of abnormal resting ECGs in this population.
End Points
The primary end point was all-cause mortality, which was determined with reference to the Social Security Death Index.14,15 We have previously shown that this method has a sensitivity of 97% for detecting death in Cleveland Clinic Foundation exercise laboratory patients.11 Cross-referencing our unified transplant database identified patients later undergoing orthotopic heart transplantation. Thus, transplant-free survival (or time free of death or transplantation) was considered as a secondary end point.
Statistical Analyses
Patients were first divided into 2 groups, based on whether they listed ;-blocker therapy among their regular medications. The Wilcoxon rank sum test was used for comparisons of continuous variables and the 2 test was used to test for comparisons of categorical variables.
Kaplan-Meier curves were constructed and nonparsimonious Cox proportional hazards modeling was performed to analyze the association of peak O2, in addition to a variety of other prospectively recorded parameters and survival or transplant-free survival. All of the variables listed in Table 1 were considered as potential confounders. For the primary end point of all-cause mortality, cardiac transplantation was considered as a time-dependent covariate irrespective of how long after stress testing the operation occurred. The proportional hazards assumption was confirmed by calculating Schoenfeld residuals. Possible nonlinear associations between the logarithm of hazard and outcome were tested with restricted cubic splines.16 We prospectively tested for an interaction between ;-blocker use and peak O2. Relative strengths of association were described based on Wald 2 statistics.
Model discrimination was tested by calculating a C-index for censored data,17 which corresponds to the area under an ROC curve; this index describes the possibility that for a randomly chosen pair of patients in which one had an event and the other did not but was followed for at least as long (or had an event after longer follow-up), the model correctly found a higher risk in the patient who had an event (or had the event earlier in follow-up). Model calibration was tested by grouping patients into quintiles of predicted risk and comparing actual 5-year Kaplan-Meier rates with predicted rates. For discrimination and calibration, 200 bootstrap resamplings were performed and the resulting 200 models used for calculating C-indexes and predicted risk.
In a supplementary analysis, we considered percent-predicted peak O2 based on age and sex as a predictor of survival or transplant-free survival.18
Data assembly and basic statistical comparisons were performed with SAS software (version 9.1, SAS Institute). Survival curves and Cox proportional hazards analyses were performed with Harrell’s Design and Hmisc libraries19 of the S-plus 6.2 software package (Insightful, Inc). A probability value of <0.05 was considered significant.
Results
Peak oxygen consumption was similar in both groups, but the E/CO2 ratio was slightly higher in patients receiving ;-blockers. The peak respiratory exchange ratio was slightly higher among patients not taking ;-blockers (1.13 versus 1.10, P<0.0001).
During a median follow-up of 3.5 years (25th and 75th percentiles 2.0 and 5.4 years) among survivors, 555 (26%) patients died and 194 (9%) underwent cardiac transplantation. Long-term survival was significantly better in patients taking ;-blockers, with an unadjusted mortality HR of 0.60 (95% CI 0.49 to 0.73, P<0.0001). The survival advantage was maintained when adjusted for the characteristics listed in Table 1 (adjusted HR 0.68, 95% CI 0.55 to 0.85, P=0.0006) and considering cardiac transplantation as a time-dependent covariate. Similarly, patients taking ;-blockers had a lower risk of experiencing death or cardiac transplantation (adjusted HR 0.63, 95% CI 0.52 to 0.76, P<0.0001).
Survival curves of patients subdivided into 4 groups on and off ;-blockers, with peak O2>14 mL · kg–1 · min–1 and 14 mL · kg–1 · min–1 are shown in Figures 1A and 2A. For both patients taking and not taking ;-blockers, a lower peak O2 was associated with a lower survival or transplant-free survival. When we repeated all of these analyses based on age- and sex-predicted peak O2, we noted essentially identical results. A lower percent-predicted peak O2 was associated with worse survival (Figure 1B) or transplant-free survival irrespective of ;-blocker use (Figure 2B).
The strongest predictor of death of peak O2 (Wald 2=76), whereas other strong predictors with a Wald 2 value >10 were men (2=21) and use of thiazide diuretics (2=13); predictors of better survival include use of ;-blockade (2=21) and undergoing cardiac transplantation (2=51). The interaction of ;-blockade and peak O2 was also strongly predictive of death (2=11).
When we repeated these regression analyses using percent-predicted peak O2 instead of absolute peak O2, the overall results were not changed materially.
Discussion
In this large cohort of 2105 patients with severe left ventricular dysfunction and symptomatic heart failure, peak O2 was a predictor of increased mortality or of a secondary end point of death or transplant, even in the so-called ;-blocker era. This remained true even after accounting for multiple possible confounding variables, including left ventricular ejection fraction, age, implantable cardioverter-defibrillators, and other concomitant medications. The relationship between peak O2 and prognosis was actually stronger in patients receiving ;-blockers than in those not receiving them.
The threshold of 14 mL · kg–1 · min–1 for referral for cardiac transplantation was based on data accumulated before the advent of ;-blockade for heart failure.1,20,21 ;-Blocker therapy is not independently associated with an improvement in peak O2,22–24 and in our group, peak O2 was similar in patients receiving and not receiving ;-blockers. These observations, coupled with the dramatic improvement in survival seen with ;-blockers, perhaps accompanied by a change in the natural history of the disease, has prompted some sources to question the prognostic significance of peak O2 in the ;-blocker era.7
Peak oxygen consumption assessment was developed as a means to formally assess functional status in patients with heart failure.25 Data from the first Veterans Administration Heart Failure Trial demonstrated that peak oxygen consumption independently predicted mortality.21 A study of 116 patients being considered for cardiac transplantation in a University of Pennsylvania program found that in patients with a peak oxygen consumption of 14 mL · kg–1 · min–1, the freedom from death or urgent cardiac transplantation was only 48% at 1 year.1 Patients without significant comorbidities and with a peak oxygen consumption of >14 mL · kg–1 · min–1 had a 1-year survival of 94%. A consensus exists that an ejection fraction <20% and a peak oxygen consumption of <14 mL · kg–1 · min–1 should be present to warrant referral for cardiac transplantation.2 Notably, however, some argue that there is an overreliance placed on this single end point for prognosis.26
A retrospective study by Shakar et al7 reviewed 127 patients who were taking ;-blockers for at least 3 months and had undergone metabolic exercise testing. They divided the patients into 2 groups, 1 with peak O2 >14 mL · kg–1 · min–1 (n=109) and the other with peak O2 <14 mL · kg–1 · min–1 (n=18). At 30 months, the number of patients who reached the combined end point of death or transplantation was similar. They concluded from these data that the current peak O2 cutoff does not predict survival without transplantation of patients on chronic ;-blocker therapy.
In another study of 408 patients with heart failure and ejection fraction <45%, Zugeck and colleagues assessed outcome according to ;-blocker use.27 They used a combined clinical end point of progressive heart failure (requiring hospital admission), inotropes, intravenous diuretics, and mechanical support or cardiac death or both at 1 year. A peak O2 14 mL · kg–1 · min–1 in patients who were receiving ;-blockers was associated with the combined end point 23% to 26%, whereas for patients who did not receive ;-blockers, it was 35% to 64%.
A further study by Peterson et al of 540 patients with heart failure compared the outcomes from nontransplanted patients with data from the International Society for Heart and Lung Transplantation transplant database, according to peak O2.28 They concluded that patients taking ;-blockers with peak O2 12 mL · kg–1 · min–1 had greater 1- and 3-year survival (with transplantation a censored event) than did post-transplant patients. For patients not taking ;-blockers and peak O2 <14 mL · kg–1 · min–1, however, survival was worse than it was for transplant recipients.
Our study is the largest to date to look at this issue. In addition, we used mortality as a hard end point. Some limitations of our study are worth noting, however. Because our data are derived from a cohort seen at a referral center with a high cardiac transplant volume, there is a need to confirm our results elsewhere. Using the Social Security Death Index meant that we did not have data regarding the cause of death. Our group and others have addressed the issue of assessing cause of death in patients with cardiovascular disease. Attempting to classify cause of death is problematic, whereas all-cause mortality is an objective, clinically relevant, and unbiased end point.29–31 Because of the inclusion of patients with atrial fibrillation, pacemakers, and ;-blocker use, we could not incorporate chronotropic response or heart rate recovery into our survival models. We did not have data on ;-blocker dose (although we did have data on resting heart rate), nor did we have data regarding who prescribed ;-blockers and why. We did not have detailed data regarding symptoms such as paroxysmal nocturnal dyspnea. Our analyses were based on peak O2 as an absolute value. Although some have proposed using percent-predicted O2 based on age and sex, this value has not been shown to be more effective at predicting survival than the simple value.3 Furthermore, in our supplementary analyses, results were essentially equivalent when using the percent-predicted value.
Cardiac transplantation was a competing event, which we treated as a time-dependent covariate and also as a secondary end point; this seems a reasonable strategy because it is unlikely that patients were specifically chosen for transplant on the basis of ;-blocker use; however, peak O2 most likely was used in selecting patients for transplantation.
It is noteworthy that at low levels of peak O2, ;-blockade was no longer associated with a survival benefit (see Figure 3 at about 10 mL · kg–1 · min–1). This suggests that the improved survival observed with ;-blockade may be weaker among patients with severely impaired functional capacity, in which mortality rates are high irrespective of treatment.
Our findings confirm the prognostic importance of peak O2 in patients with heart failure in the ;-blocker era. Indeed, a low peak O2 may be an even more powerful predictor of mortality in patients receiving ;-blocker therapy than in those treated without them. Because of improved survival in patients treated with ;-blockers, the traditional cut point of <14 mg · kg–1 · min–1 for referral for cardiac transplantation in these patients requires reevaluation, and a lower cut point value may be more appropriate. In addition, these data confirm the powerful survival advantage imparted by ;-blocker therapy, but primarily among those patients with a functional capacity exceeding 10 mL · kg–1 · min–1. Clinicians caring for patients with heart failure should recognize the value of prescribing ;-blockers for patients with impaired systolic ventricular function and that even in the setting of ;-blockade cardiopulmonary stress testing is a valuable prognostic tool.
Acknowledgments
Dr Lauer and Ms Pothier receive support from the National Heart, Lung, and Blood Institute (Grant HL-66004); Dr O’Neill receives support from the Fulbright Commission.
References
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.
Costanzo MR, Augustine S, Bourge R, Bristow M, O’Connell JB, Driscoll D, Rose E. 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.
Myers J, Gullestad L, Vagelos R, Do D, Bellin D, Ross H, Fowler MB. Clinical, hemodynamic, and cardiopulmonary exercise test determinants of survival in patients referred for evaluation of heart failure. Ann Intern Med. 1998; 129: 286–293.
Packer M, Bristow MR, Cohn JN, Colucci WS, Fowler MB, Gilbert EM, Shusterman NH. The effect of carvedilol on morbidity and mortality in patients with chronic heart failure. U.S. Carvedilol Heart Failure Study Group. N Engl J Med. 1996; 334: 1349–1355.
Effect of metoprolol CR/XL in chronic heart failure: Metoprolol CR/XL Randomised Intervention Trial in Congestive Heart Failure (MERIT-HF). Lancet. 1999; 353: 2001–2007.
Packer M, Coats AJ, Fowler MB, Katus HA, Krum H, Mohacsi P, Rouleau JL, Tendera M, Castaigne A, Roecker EB, Schultz MK, DeMets DL. Effect of carvedilol on survival in severe chronic heart failure. N Engl J Med. 2001; 344: 1651–1658.
Shakar SF, Lowes BD, Lindenfeld J, Zolty R, Simon M, Robertson AD, Bristow MR, Wolfel EE. Peak oxygen consumption and outcome in heart failure patients chronically treated with beta-blockers. J Card Fail. 2004; 10: 15–20.
Robbins M, Francis G, Pashkow FJ, Snader CE, Hoercher K, Young JB, Lauer MS. Ventilatory and heart rate responses to exercise: better predictors of heart failure mortality than peak oxygen consumption. Circulation. 1999; 100: 2411–2417.
Beaver WL, Wasserman K, Whipp BJ. A new method for detecting anaerobic threshold by gas exchange. J Appl Physiol. 1986; 60: 2020–2027.
Wasserman K. Breathing during exercise. N Engl J Med. 1978; 298: 780–785.
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: 1392–1398.
Coronary artery surgery study (CASS): a randomized trial of coronary artery bypass surgery. Quality of life in patients randomly assigned to treatment groups. Circulation. 1983; 68: 951–960.
Mark DB, Hlatky MA, Harrell FE Jr, Lee KL, Califf RM, Pryor DB. Exercise treadmill score for predicting prognosis in coronary artery disease. Ann Intern Med. 1987; 106: 793–800.
Curb JD, Ford CE, Pressel S, Palmer M, Babcock C, Hawkins CM. Ascertainment of vital status through the National Death Index and the Social Security Administration. Am J Epidemiol. 1985; 121: 754–766.
Newman TB, Brown AN. Use of commercial record linkage software and vital statistics to identify patient deaths. J Am Med Inform Assoc. 1997; 4: 233–237.
Harrell FE Jr, Lee KL, Mark DB. Multivariable prognostic models: issues in developing models, evaluating assumptions and adequacy, and measuring and reducing errors. Stat Med. 1996; 15: 361–387.
Harrell FE Jr, Califf RM, Pryor DB, Lee KL, Rosati RA. Evaluating the yield of medical tests. JAMA. 1982; 247: 2543–2546.
Hansen JE, Sue DY, Wasserman K. Predicted values for clinical exercise testing. Am Rev Respir Dis. 1984; 129: S49–S55.
Harrell FE Jr. Regression Modeling Strategies with Applications to Linear Models, Logistic Regression, and Survival Analysis. New York, NY: Springer-Verlag; 2001.
Francis GS. Determinants of prognosis in patients with heart failure. J Heart Lung Transplant. 1994; 13: S113–S116.
Cohn JN, Johnson GR, Shabetai R, Loeb H, Tristani F, Rector T, Smith R, Fletcher R. 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–VI16.
Gilbert EM, Abraham WT, Olsen S, Hattler B, White M, Mealy P, Larrabee P, Bristow MR. Comparative hemodynamic, left ventricular functional, and antiadrenergic effects of chronic treatment with metoprolol versus carvedilol in the failing heart. Circulation. 1996; 94: 2817–2825.
Metra M, Giubbini R, Nodari S, Boldi E, Modena MG, Dei Cas L. Differential effects of beta-blockers in patients with heart failure: a prospective, randomized, double-blind comparison of the long-term effects of metoprolol versus carvedilol. Circulation. 2000; 102: 546–551.
Gullestad L, Manhenke C, Aarsland T, Skardal R, Fagertun H, Wikstrand J, Kjekshus J. Effect of metoprolol CR/XL on exercise tolerance in chronic heart failure—a substudy to the MERIT-HF trial. Eur J Heart Fail. 2001; 3: 463–468.
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.
Francis GS, Rector TS. Maximal exercise tolerance as a therapeutic end point in heart failure—are we relying on the right measure; Am J Cardiol. 1994; 73: 304–306.
Zugck C, Haunstetter A, Kruger C, Kell R, Schellberg D, Kubler W, Haass M. Impact of beta-blocker treatment on the prognostic value of currently used risk predictors in congestive heart failure. J Am Coll Cardiol. 2002; 39: 1615–1622.
Peterson LR, Schechtman KB, Ewald GA, Geltman EM, de las Fuentes L, Meyer T, Krekeler P, Moore ML, Rogers JG. Timing of cardiac transplantation in patients with heart failure receiving beta-adrenergic blockers. J Heart Lung Transplant. 2003; 22: 1141–1148.
Lauer MS, Blackstone EH, Young JB, Topol EJ. Cause of death in clinical research: time for a reassessment; J Am Coll Cardiol. 1999; 34: 618–620.
Gottlieb SS. Dead is dead—artificial definitions are no substitute. Lancet. 1997; 349: 662–663.
Lloyd-Jones DM, Martin DO, Larson MG, Levy D. Accuracy of death certificates for coding coronary heart disease as the cause of death. Ann Intern Med. 1998; 129: 1020–1026.(James O. O’Neill, MB, MRC)