Author + information
- Received September 3, 1997
- Revision received December 5, 1997
- Accepted December 17, 1997
- Published online March 15, 1998.
- Cristina Opasich, MDAB,*,
- Gian Domenico Pinna, MSB,
- Marco Bobbio, MDB,
- Massimo SistiB,
- Brunella Demichelis, MDB,
- Oreste Febo, MDAB,
- Giovanni Forni, MDAB,
- Roberto Riccardi, MDAB,
- P.Giorgio Riccardi, MDAB,
- Soccorso Capomolla, MDAB,
- Franco Cobelli, MDAB and
- Luigi Tavazzi, MD, FESC, FACCAB
- ↵*Dr. Cristina Opasich, Centro Medico di Montescano, Via per Montescano, 27040 Montescano (Pavia), Italy.
Objectives. This study sought to 1) assess the short-, medium- and long-term prognostic power of peak oxygen consumption (V̇o2) in patients with heart failure; 2) verify the consistency of a nonmeasurable anaerobic threshold (AT) as a criterion of nonapplicability of peak V̇o2; 3) develop simple rules for the efficient use of peak V̇o2in individualized prognostic stratification and clinical decision making.
Background. Peak V̇o2, when AT is identified, is among the indicators for heart transplant eligibility. However, in clinical practice the application of defined peak V̇o2cutoff values to all patients could be inappropriate and misleading.
Methods. Six hundred fifty-three patients consecutively considered for eligibility for heart transplantation were followed up. Outcomes (cardiac death and urgent transplantation) were determined when all survivors had a minimum of 6 months of follow-up.
Results. Contraindication to the exercise test identified very high risk patients. The relatively small sample of women did not allow inferences to be drawn. In men, peak V̇o2stratified into three levels (≤10, 10 to 18 and >18 ml/kg per min) identified groups at high, medium and low risk, respectively. The prognostic power of peak V̇o2≤10 ml/kg per min was maintained even when the AT was not detected. In patients in New York Heart Association functional class III or IV, peak V̇o2did not have prognostic power. In patients in functional class I or II, peak V̇o2stratification was prognostically valuable, but less so at 6 than at 12 or 24 months. Age did not influence peak V̇o2prognostic stratification.
Conclusions. A contraindication to exercise testing should be considered a priority for listing patients for heart transplantation. Only in less symptomatic male patients does a peak V̇o2≤10 ml/kg per min identify short-, medium- and long-term high risk groups. A peak V̇o2>18 ml/kg per min implies good prognosis with medical therapy.
In patients with chronic heart failure (HF), peak oxygen consumption (V̇o2), defined as the oxygen uptake at the maximal level of tolerated exercise, is a descriptive indicator with both prognostic power and decisional implications. A continuous variable, peak V̇o2has been categorized by many investigators [1–11]for practical purposes, and the prognostic power of the various V̇o2cutoff points (10, 12, 14, 16 and 18 ml/kg per min) has been repeatedly confirmed in published reports [1–11]. In a study considered the cornerstone of the validation of the prognostic power of peak V̇o2, Mancini et al. found that patients with a peak V̇o2≤10 ml/kg per min had the worst prognosis. The other cutoff values proposed for risk stratification were >10 to 14, >14 to 18 and >18 ml/kg per min. As a decisional indicator, peak V̇o2appears in the first statement of the 24th Bethesda Conference and in several other reports and statements as an indicator for heart transplant eligibility [2, 13–19].
Peak V̇o2should be considered only after the anaerobic threshold (AT) is detected, thus guaranteeing nearly maximal exercise performance. Mancini et al. , for instance, included in their survival analysis only patients whose AT could be identified for at least one of two exercise tests. However, in daily clinical practice, AT goes undetected in a large proportion of patients with a low peak V̇o2.
Despite such drawbacks, published studies have brought about both increasing use of cardiopulmonary exercise (CPX) testing in clinical practice and a simplistic application of defined peak V̇o2cutoff values to all patients with chronic HF. Intuitively, such a generalized application holds the risk of inappropriateness or wide approximation.
The present study was conceived with the aim of reconsidering and elucidating major issues regarding the use of peak V̇o2as a routine prognostic and decisional indicator in chronic HF. We intended to 1) verify the clinical applicability of the stratification suggested by Mancini et al. in a large group of patients with moderate to severe chronic HF, consecutively considered for eligibility for heart transplantation; 2) verify the need for exclusion from peak V̇o2prognostic stratification of those patients without a determinable AT; 3) compare the short- (6 months), medium- (12 months) and long-term (24 months) prognostic power of peak V̇o2; and 4) develop simple rules, based on some common clinical indicators, for using peak V̇o2in individualized prognostic stratification and clinical decision making.
All patients with chronic HF and a left ventricular ejection fraction <40%, admitted for assessment or reassessment of indications for heart transplantation at the Heart Failure Unit of Montescano Medical Center or the Cardiological Department of Molinette Hospital of Turin from March 1992 to October 1995, were enrolled in the study.
On admission, all patients were given individualized therapy, which generally included angiotensin-converting enzyme inhibitors, vasodilators (nitroprusside, isosorbide dinitrate, hydralazine), digoxin, diuretic drugs, inotropic drugs (dopamine, dobutamine, enoximone) when needed and, in the absence of contraindications, antiacoagulants (international normalized ratio 2 to 3) or antiplatelet agents. At the time of recruitment, beta-adrenergic blocking agents were not yet considered routine therapy.
After therapeutic optimization, when patients were clinically stable, a functional evaluation was performed that included two-dimensional echocardiography; right heart catheterization, performed with a Swan-Ganz catheter for thermodilution, introduced through the internal jugular vein (Seldinger technique); and CPX testing, carried out on a treadmill with a modified Naughton protocol and with simultaneous monitoring of respiratory gases by a CAD/NET System 2001 Medical Graphics analyzer. The AT was defined as one or more of the following: 1) the point at which the ventilatory equivalent for oxygen (V̇e/V̇o2) was minimal, followed by a progressive increase; or 2) the point after which the respiratory gas exchange ratio exceeded the rest respiratory gas exchange ratio; or 3) the point after which a nonlinear increase in minute ventilation occurred relative to carbon dioxide consumption. The peak V̇o2cutoff values used were those proposed by Mancini et al. : ≤10, >10 to 14, >14 to 18 and >18 ml/kg per min.
Contraindications to CPX testing were symptomatic congestion at rest and the reappearance of rest symptoms after weaning from vasodilator or inotropic infusion treatment. Admission to the heart transplantation waiting list followed the current guidelines ; absolute and relative contraindications were considered according to standard practice .
All patients were closely followed up by means of clinical examination and other tests (echocardiography, CPX). All surviving patients had a minimum of 6 months of follow-up.
1.2 Cardiac Events
Sudden death (if unexpected, occurring in or out of the hospital within 1 h after the onset of, or change in, symptoms or during sleep in a patient who was symptomatically stable during the 24 h before death); progressive HF-related death; other cardiac-related death, including acute myocardial infarction and pulmonary embolism; and urgent transplantation (in Status I patients) were considered cardiac events.
1.3 Prognostic Variables
To individualize the use of the prognostic and decisional power of peak V̇o2, three descriptive clinical indicators commonly used in clinical practice were considered: gender, age and New York Heart Association functional class.
1.4 Statistical Analysis
All patients capable of performing the CPX test were initially subclassified by gender. Within the male group, survival analysis was performed separately for subjects in whom the AT was detected and in those in whom it was not. This analysis was carried out by stratifying patients according to the cutoff points of Mancini et al. and estimating corresponding survival functions by the Kaplan-Meier method. Data from patients who survived until the end of the follow-up period, died of noncardiac-related causes or underwent transplantation (except for Status I patients) were treated as “censored” observations. Survival curves were compared by the log-rank test. Besides allowing verification of the robustness of the criteria of Mancini et al. in a large group of patients, this analysis also allowed the prognostic stratification power of peak V̇o2to be tested in patients without a detected AT. The result was the identification of a subgroup of patients, with or without a detected AT, for whom peak V̇o2had a proven prognostic value. To refine this identification process on the basis of the chosen clinical indicators, these patients were then subclassified by age (≤55 vs. >55 years), and survival analysis was again performed to identify a further subgroup, if existing, in which peak V̇o2had a clear prognostic value. Finally, the same procedure was repeated after classifying patients by functional class (I to II vs. III to IV). The overall procedural tree is summarized in Fig. 1. In the final subgroup, identification of the relative risk of each peak V̇o2cutoff was estimated by Cox proportional hazards analysis (peak V̇o2cutoff was used as the explanatory variable in the model), and the survival experience was further examined over the first 6 months, 1 year and 2 years by the Kaplan-Meier method.
Because of the small sample size of the women who performed the CPX test, survival analysis in this subgroup was not carried out.
Results are expressed as mean value ± SD. A probability value <0.05 was considered statistically significant. Statistical analysis was performed with the SAS/STAT version 6.10 statistical package (SAS Institute Inc.).
The clinical characteristics of the 653 patients enrolled in the study are shown in Table 1. The average follow-up period was 498.8 ± 381 days (median 375). No patient was lost to follow-up. Fourteen patients died of noncardiac causes. Fifty-six patients (43 men) could not perform the CPX test because of clinical cardiac contraindications (20 of 24 patients in functional class IV were in this group). A cardiac event (cardiac death or heart transplantation in Status I patients) occurred in 77% of these patients, and in 70% (30 patients) within 100 days after the evaluation.
2.1.1 Female Patients
Ninety-two of 105 female patients performed the CPX test (55 in functional class I or II, 37 in functional class III). Twenty-six patients of these 92 patients died of cardiac causes and 1 of a noncardiac cause. Twenty-five patients underwent heart transplantation, of whom 12 were in Status I. In all, 38 events (cardiac death and heart transplantation in Status I patients) occurred (Table 2).
During the exercise test, the AT was detected in only 26 patients and was not detected in 66. Because of the small number of patients in both subgroups, survival analysis was not carried out.
2.1.2 Male Patients
In the male group, 505 of 548 patients performed the CPX test. Among these 505 patients, 126 (25%) died of cardiac causes and 11 (2.2%) of noncardiac causes; 93 (18.4%) underwent heart transplantation (24 were Status I). In all, 150 events occurred (Table 2).
During the exercise test, the AT was detected in 324 patients (64%). The survival curves for these patients according to the stratification of Mancini et al. are shown in Fig. 2. A peak V̇o2≤10 ml/kg per min identified patients with the worst prognosis (13 [59%] of 22 with a cardiac event), whereas a peak V̇o2>18 ml/kg per min identified patients with the best prognosis (14 [15%] of 96 with a cardiac event). No differences were found between survival of patients in the two intermediate groups (32 of 98 with a peak V̇o2>10 to 14 ml/kg per min and 35 of 108 with a peak V̇o2>14 to 18 ml/kg per min had a cardiac event, p = 0.66). We also repeated the analysis using a narrower stratification (10 to 12.5, 12.5 to 15, 15 to 17.5, 17.5 to 20, >20 ml/kg per min), but survival curves of patients with a peak V̇o2in the range 10 to 17.5 ml/kg per min again did not differ.
Because of the similarity of the survival curves of patients with a peak V̇o210 to 14 and 14 to 18 ml/kg per min, we considered these patients as a single group. The cardiac event rates of the three resulting groups (patients with a peak V̇o2≤10, 10 to 18, >18 ml/kg per min) were 62%, 33% and 15%, respectively. Death due to heart failure and Status I at the time of transplantation accounted for 62%, 59% and 50% of all cardiac events in the three groups, respectively, showing a homogeneous stratification power of peak V̇o2when only these end points were considered.
The AT was not identified in 48 of the male patients (69%) with a peak V̇o2≤10 ml/kg per min, 70 (42%) with a peak V̇o2>10 to 14 ml/kg per min, 40 (27%) with a peak V̇o2>14 to 18 ml/kg per min and 23 (19%) with a peak V̇o2>18 ml/kg per min. There was no significant difference between the survival curves for patients with a peak V̇o2>10 to 14, >14 to 18 and >18 ml/kg per min (Fig. 3) for a cardiac event rate of 29%, 23% and 22% of 70, 40 and 23 subjects in each group, respectively. However, the survival of patients with a peak V̇o2≤10 ml/kg per min (22 [46%] of 48 patients with a cardiac event) was markedly worse (p = 0.007), indicating a high predictive power of this peak V̇o2level. When the survival curve of this subgroup of patients was compared with that of the corresponding subgroup with the same peak V̇o2but a measurable AT, substantial overlap of the two subgroups was seen (p = 0.52). For this reason, all 324 patients with a detectable AT and 48 patients with an undetected AT but with a peak V̇o2≤10 ml/kg per min were considered as a single group (372 patients in all) in further analyses.
In this group of patients the survival of older (n = 158) and younger (n = 214) subjects within the same peak V̇o2cutoff value was very similar (p > 0.5 for all comparisons) (Fig. 4). The incidence of cardiac events was also comparable (Table 3). For this reason, in the further analysis we did not consider age groups separately.
When survival according to functional class at the time of evaluation (235 patients in functional class I or II vs. 137 patients in functional class III or IV) (Table 3Fig. 5) was analyzed, different results were found depending on the value of this clinical indicator. Peak V̇o2cutoff values discriminated well between low, medium and high risk patients in functional class I or II, with mortality rates of 12%, 28% and 45%, respectively (p = 0.0005). The risk ratio was 2.5 (95% confidence interval [CI] 1.2 to 5) between peak V̇o210 to 18 and <10 ml/kg per min and 1.9 (95% CI 1.0 to 3.9) between peak V̇o2>18 and 10 to 18 ml/kg per min. In contrast, for patients in functional class III or IV, all peak V̇o2curves intersected each other, and the log-rank test was not significant (p = 0.15).
In the identified subgroup of patients in whom peak V̇o2was a significant prognostic indicator (235 men in functional class I or II), we verified the prognostic power of this marker for short-, medium- and long-term outcome. The percent of event-free subjects at 6 months and 1 and 2 years is shown in Table 4. It can be seen that a peak V̇o2≤10 ml/kg per min has a prognostic power that persists over time, with a mortality rate of 18% within 6 months, 36% within 1 year and 45% within 2 years. A peak V̇o210 to 18 ml/kg per min identified patients with a poorer medium-term outcome than those with a peak V̇o2>18 ml/kg per min, who had the best prognosis up to 2 years.
Peak V̇o2has become an integral part of prognostic stratification and evaluation for heart transplantation in patients with chronic HF [2, 13–19]. However, validation of its clinical applicability in a large population is still lacking, and the simplistic (now widespread) use of its cutoff values in all patients with moderate to severe chronic HF as a clinical tool in decision making for heart transplantation might be inappropriate.
In the present study, we verified the prognostic value of peak V̇o2stratification, originally proposed by Mancini et al. , in a large group of patients with moderate to severe chronic HF consecutively considered for candidacy for heart transplantation. We sought to exploit the information provided by a few simple clinical indicators to apply peak V̇o2cutoff values more efficiently in the individual patient.
3.1 Contraindication to CPX Testing
The present study verified that the presence of a clinical contraindication to performing the CPX test, despite optimized medical treatment, is the simplest and strongest negative prognostic indicator. This result confirms those previously demonstrated after acute myocardial infarction , after lung cancer resection and in advanced HF .
By identifying patients with a high probability of an unfavorable short-term outcome, the inability to perform an exercise test is also a decisional indicator in the process of eligibility for heart transplantation or ventricular assistance device implantation. Accordingly, this indicator should be added to the selection criteria for listing candidates with advanced HF and for timing heart transplantation.
3.2 Prognostic Power of the Peak V̇o2With and Without Detection of the AT
In patients able to perform an exercise test, the peak V̇o2cutoff points proposed by Mancini et al. were partially confirmed in our large cohort. A peak V̇o2≤10 ml/kg per min identifies high risk patients, and a peak V̇o2>18 ml/kg per min identifies low risk patients; however, all other values in between these cutoff values define a gray area of medium risk patients, without any further possible stratification. This result confirms the observations of Kao et al. in 178 patients with the same characteristics as ours. Moreover, our results do not confirm the prognostic and decisional value of the “magic number” of 14 ml/kg per min of peak V̇o2previously proposed ; for 10 to 18 ml/kg per min, such a cutoff point does not discriminate patients at different risk.
In the Mancini et al. study , peak V̇o2had prognostic power provided that the AT was reached. When we performed a survival analysis in the subgroup of our patients in whom the AT was not detected, we found that identification of the AT did not influence the prognostic power of a peak V̇o2≤10 ml/kg per min. The low level of symptom appearance is probably a valuable indicator of abnormal metabolic response to exercise and replaces the information of the low level of AT. This result has important practical implications because the AT is not detectable in the great majority of the patients with a peak V̇o2≤10 ml/kg per min [25–27](Table 1) and, if reached, the AT is hard to identify by noninvasive methods. Thus, the other “magic number,” peak V̇o210 ml/kg per min should be taken into account in the prognostic stratification of and decision making for patients with chronic HF, without the constraint of AT detection. In contrast, for all other patients without a detectable AT and a peak V̇o2>10 ml/kg per min, the peak V̇o2value does not provide prognostic information. In these patients, a repeat exercise test is advisable.
3.3 Prognostic Power of Peak V̇o2in the Individual Patient
To individualize the use of peak V̇o2as a prognostic and decisional marker, gender, age and clinical severity were considered in the survival analysis. The first variable used to subgroup patients was gender, in accordance with published suggestions [19, 28]of the limited prognostic power of peak V̇o2in women with HF and the poor informative results of exercise testing previously seen in women with coronary disease [29–32]. Although our group of female transplant candidates is, to our knowledge, the largest to be evaluated with exercise testing, the sample size achieved was considered inadequate for a reliable survival analysis to be performed. However, a major finding was that women, compared with men, comprise both a higher proportion of patients with low exercise capacity and good prognosis and a lower proportion with a detectable AT.
In male patients grouped according to age (≤55 vs. >55 years), no differences were seen in the prognostic power of peak V̇o2cutoff values. This result could be explained in part by the absence of very elderly subjects among our transplant candidates and in part by the observation that the strength of the prognostic power of peak V̇o2is concentrated at its lowest and highest values (≤10 and >18 ml/kg per min), which are far more influenced by severity of functional limitations than by age.
The prognostic power of peak V̇o2was confirmed only in patients in functional class I or II. Although the cardiac events rate was higher in patients in functional class III or IV, survival curves corresponding to the three peak V̇o2strata were not significantly different. Thus, peak V̇o2shows less prognostic sensitivity in those patients for whom decisional contributions are most needed.
3.4 Short-, Medium- and Long-Term Prognostic Power of Peak V̇o2
In summary, we found that peak V̇o2provides prognostic information in male patients with moderate to severe HF, who are between 30 and 70 years old and are in functional class I or II after optimal therapy. A peak V̇o2>18 ml/kg per min maintains its predictive power of a good outcome for at least 2 years. Thus, in these patients, repetition of the CPX test for prognostic reasons might not be necessary throughout this period. A peak V̇o2≤10 ml/kg per min shows a prognostic power that is strong in the short term and is maintained over time.
3.5 Toward the Efficient Use of Peak V̇o2in the Individual Patient
The prognostic and decisional (evaluation for heart transplantation) value of peak V̇o2is summarized in Fig. 6, which may also be viewed as a flowchart for the efficient use of CPX test results in individual patients. A clear contraindication to exercise testing is both a prognostic and a decisional indicator: The patient is at very high risk, and aggressive treatment (i.e., urgent heart transplantation, ventricular assistance) should be considered. When the CPX test is not contraindicated and the patient is a woman, the current view is that peak V̇o2has little value for prognostic stratification and decision making; however, larger studies are needed to confirm this current provisional policy.
For male patients, CPX test results should be taken into consideration in all subjects with a detectable AT and in those who reach a peak V̇o2≤10 ml/kg per min without an evident AT.
In patients in functional class III or IV, peak V̇o2is neither prognostic nor decisional. Our data are partially in contrast with the results of Stevenson et al. , who showed that among exercising patients in an advanced functional class, the threshold of 10 ml/kg per min was the only cutoff value that identified a higher risk group. Currently, most heart transplantation centers rely heavily on peak V̇o2evaluation to decide which patients with advanced HF should be considered for heart transplantation [12, 18]. We believe that this decision should rely more on other indicators of prognosis, such as contraindication to exercise testing and exhausted therapeutic alternatives.
Male patients 30 to 70 years old, in functional class I or II after optimal therapy and with a peak V̇o2≤10 ml/kg per min have a cardiac event probability higher than all other patients in functional class I or II and similar to that of patients in functional class III or IV. Hence, in these patients peak V̇o2is a powerful decisional indicator for transplantation (and priority), whereas the simple evaluation of symptoms could be misleading. These data are in agreement with those of Aaronson et al. , who showed that among ambulatory patients referred for heart transplantation, peak V̇o2alone gave a satisfactory prognostic discrimination.
A patient with a peak V̇o2>18 ml/kg per min, has a good long-term prognosis. In this case, peak V̇o2is also decisional, making any aggressive treatment or close follow-up unnecessary.
Finally, patients with a peak V̇o2in the range 10 to 18 ml/kg per min fall into a gray (and unfortunately very large) prognostic region, and peak V̇o2does not provide information on which decisions can be made. In such patients, further data should be collected to improve prognostic capability and to address decision making (e.g., serial cardiopulmonary exercise tests, other exercise/recovery indexes [34, 35]), evaluation of responses to other therapy, other diagnostic tests ).
The results of the present study suggest that the decision to list patients with advanced HF for heart transplantation should rely not only on peak V̇o2but on other indicators of prognosis as well. Contraindication to exercise testing appears to be a powerful short-term unfavorable prognostic indicator. Only among less symptomatic male patients able to exercise is peak V̇o2a valuable prognostic indicator for decision making in that it can identify low and high risk subjects and thus offer a practical operative contribution for nearly 20% of all patients with moderate to severe chronic HF. For transplantation, only a peak V̇o2≤10 ml/kg per min is a definite decisional indicator, with or without AT. These considerations, together with the lack of prognostic power of peak V̇o2in patients in a functional class III or IV, limit its role in clinical practice.
- anaerobic threshold
- confidence interval
- cardiopulmonary exercise
- heart failure
- oxygen consumption
- Received September 3, 1997.
- Revision received December 5, 1997.
- Accepted December 17, 1997.
- The American College of Cardiology
- ↵Cohn JN, Johnson GR, Shabetai R, et al., for the V-HeFT VA Cooperative Studies Group. Ejection fraction, peak oxygen consumption, cardiothoracic ratio, ventricular arrhythmias, and plasma norepinephrine as determinants of prognosis in heart failure. Circulation 1993;87 Suppl VI:VI-5–16.
- Mancini DM,
- Eisen H,
- Kussmaul W,
- Mull R,
- Edmunds LH,
- Wilson JR
- Pina IL
- Roul G,
- Moulichon ME,
- Bareiss P,
- et al.
- Madsen BK,
- Hansen JF,
- Stokholm KH,
- Brons J,
- Husum D,
- Mortensen L
- van den Broek S,
- van Veldhuisen D,
- de Graeff P,
- Landsman M,
- Hillege H,
- Lie H
- Rickenbacher P,
- Trindade P,
- Haywood G,
- et al.
- Haywood G,
- Rickenbacher P,
- Trindade P,
- et al.
- Hunt SA
- Stevenson LW,
- Steimle A,
- Fonarow G,
- et al.
- Singer A,
- Vagellos R,
- Wilson K,
- Nejedly M,
- Fowler M
- Costanzo MR,
- Augustine S,
- Bourge R,
- et al.
- Stevenson LW,
- Couper G,
- Natterson B,
- et al.
- Stelken A,
- Younis L,
- Jennison S,
- et al.
- Hachamovitch R,
- Berman DS,
- Kiat H,
- et al.
- Aaronson K,
- Scwartz J,
- Chen T,
- Wong K,
- Goin J,
- Mancini D
- DeGroote P,
- Millaire A,
- Decoulx E,
- Nugue O,
- Guimier P,
- Ducloux G
- Chomsky DB,
- Lang CC,
- Rayos GH,
- et al.