Cardiopulmonary exercise testing (CPET) with minute ventilation (V˙E) and gas exchange measurement has proven to be useful in patients with chronic heart failure (CHF). It quantifies exercise tolerance and consequently is important to target for cardiac rehabilitation (15), the benefits of which seem to outweigh the risks (26). It has also shown particular value in selecting patients who would benefit from heart transplantation (18,21). Maximal oxygen uptake (V˙O2) reliably reflects cardiopulmonary capacity and is a good prognostic marker for these patients, although data are accumulating to support the use of the ventilatory efficiency index (V˙E–carbon dioxide product (V˙CO2) slope) (2) and the ventilatory anaerobic threshold (VAT) (13), either separately or together in addition.
In clinical practice, however, many patients perform a test with an RER < 1.0 and thus may be considered not to have performed a truly maximal exercise test; the reported value is therefore usually the V˙O2 that occurs at peak exercise (V˙O2peak). In a recent study performed by Ingle et al. (16), only 58% of the patients were able to reach an RER of ≥1.0, and the V˙O2peak and V˙E–V˙CO2 slope did not predict mortality in those who did not. In this respect, the oxygen uptake efficiency slope (OUES) is a promising index of exercise capacity (1) derived from the logarithmic relationship between V˙O2 and V˙E during incremental exercise (5) according to the following equation: V˙O2 (L·min−1) = OUES × log10V˙E (L·min−1) + k.
When V˙O2 is plotted on the y axis and log10V˙E is plotted on the x axis, OUES is the slope of the best fit line through the points. The constant is the y value when x = 0 (see Figure A, SDC 1, https://links.lww.com/MSS/A117, a representation of a subject’s relationship between V˙O2 and ventilation on different scales). The major advantage of OUES is that it is not dependent on maximal exercise. It also seems likely to be of prognostic value in heart failure (10), although it may not be superior to other markers (3). The aim of this study was to further investigate the relationship between OUES and V˙O2peak and other ventilatory indices in patients with heart failure to assess their interchangeability and improve the understanding of the determinants of OUES.
MATERIALS AND METHODS
We prospectively studied men who had been consecutively recruited for a study on the prevalence of sleep-disordered breathing in patients with CHF. These men were recruited from the cardiology clinic at the Royal Brompton Hospital between October 2002 and June 2003. All gave written informed consent, and the study was approved by the Royal Brompton and Harefield National Health Service Trust ethics committee.
Eligibility for participation and exclusion criteria were detailed in a prior publication (32). Briefly, male patients with mild stable CHF (New York Heart Association class II/III), secondary to left ventricular systolic dysfunction of at least 6 months’ duration, were eligible. Only those with idiopathic dilated or ischemic cardiomyopathy were included. “Stable” was defined as no change in symptoms or medication in the 4 wk before enrollment and no hospitalizations in the preceding 8 wk. Left ventricular systolic dysfunction was defined as a left ventricular ejection fraction < 45% as assessed by two-dimensional echocardiography. All patients were receiving optimal medical therapy and had no significant lung disease; the percent predicted forced expiratory volume in the first second/forced vital capacity ratio was always >70%. Subjects were instructed to take all prescribed medication before coming to the exercise laboratory, including β-blockers.
Exclusion criteria were unstable angina, primary valvular or congenital heart disease, a history of chronic respiratory or neurological disease, and the inability to perform an exercise test for any reason. Patients with biventricular pacing systems were excluded, as were patients taking hypnotic agents, theophylline, or opiates.
Sixty-three patients underwent exercise testing, and 54 data sets were eligible for analysis. Tests with poor signal quality or oscillatory breathing on exercise (n = 5) were discarded, as were those where submaximal effort was made, as evidenced by an RER at peak exercise < 1.1 (n = 2). Two further studies were discarded for other reasons (inability to pedal at a stable frequency, incorrect exercise protocol). Nineteen of the 54 retained patients performed a subsequent CPET between 3 and 9 months later. Subjects who completed both tests were not significantly different from those who performed only one test. The mean age of the subjects was 63.8 ± 8.9 yr at their first visit. Their mean body mass index was 28.9 ± 4.6 kg·m−2 (Table 1).
A symptom-limited CPET was performed on a cycle ergometer (Ergometrics 800S®; Ergometrics, Bitz, Germany) in an air-conditioned laboratory, with the temperature set at 18°C–22°C. After at least 3 min of freewheeling, the resistance was increased by 10 or 20 W every minute to complete the test within 8–12 min. All subjects were encouraged to provide maximal effort until they could no longer pedal, unless there were clinical reasons to stop the test sooner.
Ventilatory gas analysis was performed using a metabolic cart (Oxycon Pro®; Erich Jaeger GmbH, Hoechberg, Germany) for 3 min of rest before the test, during the test, and for 5 min of recovery after pedaling. The flow meter and gas analyzers were calibrated for accuracy, linearity, and time response before each test according to the manufacturer’s specifications. A 12-lead ECG was continuously displayed. Blood pressure was measured every 2 min.
V˙E, V˙O2, and V˙CO2 were acquired breath by breath and averaged over nine breaths. A basic filter was applied to remove any data beyond three SD of the average of nine breaths around that point. The VAT was determined by two experienced independent reviewers, according to the initial recommendations of Wasserman et al. (34). This was the point at which the ventilatory equivalent for oxygen started to rise nonlinearly, whereas the ventilatory equivalent for carbon dioxide remained unchanged with the RER < 1.0. VAT was expressed in liters of oxygen per minute.
For each subject, OUES was calculated as follows: V˙O2 (L·min−1) = OUES × log10V˙E (L·min−1) + constant. OUES was calculated in two ways: using all data (OUES100) or only the data up to the point where the RER = 1 (OUESRER1). In both cases, the data gathered during the warm-up period of freewheeling were excluded. OUES was also normalized to body mass (OUES per kilogram).
In line with the current literature reporting its high prognostic value (13,27), the V˙E–V˙CO2 slope was calculated via least squares linear regression from nonaveraged V˙E and V˙CO2 values, including all data points from the beginning to the end of loaded exercise. It was also calculated at VAT, an intensity for which the factors affecting the relationship between V˙E and V˙CO2 are well identified (33).
Data were analyzed with STATISTICA (v. 5.5; StatSoft, Tulsa, OK). Unless specified, all continuous data are reported as mean values ± SD and pertain to the results of the 54 patients’ first visit. Statistical differences with a P value <0.05 were considered significant.
With regard to OUESRER1, one patient reached an RER of 1 in the first 2 min of loaded exercise, and his data are not included. VAT could not be determined in five participants, which is not unusual in this population (9), so only 49 data sets were used for the comparisons with VAT. OUES100 and OUESRER1 were compared using paired Student’s t-tests. In addition, the strength of the individual V˙O2–logV˙E relationships and the line intercepts were compared. Differences in the mean correlation coefficients at different exercise intensities were analyzed using the Fisher z transformation.
Least squares linear regression was used to fit a line to describe all the relationships between V˙O2peak, VAT, OUES100, and OUESRER1 in all subjects. Pearson correlation coefficients (two-tailed probabilities) were used to assess the significance of these relationships. Equations extracted from the linear regressions were used to calculate a predicted V˙O2peak. Repeated-measures ANOVAs assessed the differences between measured V˙O2peak and V˙O2peak predicted from VAT, OUES100, and OUESRER1.
Bland–Altman plots (7) were also used to assess the agreement between measured V˙O2peak and V˙O2peak predicted from OUES100, OUESRER1, and VAT to determine whether OUES from a restricted data set could be used to reliably predict V˙O2peak. We decided a priori that acceptable limits for interchangeability between V˙O2peak and other ventilatory parameters would be ±4 mL O2·min−1·kg−1. This will be discussed below. Pearson product moment correlations assessed the variation in test performance in the subgroup that attended twice.
Baseline characteristics of the patients are reported in Table 1. The subgroup that performed two exercise tests did not differ significantly from the other participants. The average test duration was 635 ± 218 s. The time to reach an RER of 1 was 446 ± 170 s, corresponding to an average of 65.1% ± 15.8% of the breath-by-breath data points.
There were no significant differences in the slope and constant of the relationships between V˙O2 and logV˙E depending on the amount of data incorporated into the calculation, but the r2 differed (Table 2).
OUES100, OUESRER1, and VAT were significantly correlated with measured V˙O2peak (Table 3).
The equations representing the best linear fit between OUES (either OUES100 or OUESRER1) and V˙O2peak for the whole group were the following:
The slopes and intercepts (df = 103, both t < 0.1, both P > 0.9) did not differ when either OUES100 or OUESRER1 was used to predict V˙O2peak (see Figure B, SDC 2, https://links.lww.com/MSS/A118, representation of the relationship between V˙O2peak and OUES in this study population). These equations were used to predict V˙O2peak on a case-by-case basis. ANOVA did not show any significant difference in the V˙O2peak predicted from OUES100, OUESRER1, or VAT and the measured V˙O2peak using values corrected or uncorrected for the body mass of the subject (df = 3, both F < 0.21, both P > 0.89). The differences between measured V˙O2peak and the values predicted by calculation from OUES100, OUESRER1, and VAT plotted against V˙O2peak are shown in Figure 1. It is observable on these Bland–Altman plots that the deviation between observed V˙O2peak and V˙O2peak predicted from OUES is independent on the level of fitness. This deviation can reach around 30% of the V˙O2peak whatever the submaximal index. Table 4 shows the limits of agreement and confidence intervals, which were moderately acceptable and in the same range for OUES100 and VAT. The agreement was lower when V˙O2peak was predicted from OUESRER1.
For the 19 subjects who undertook two cardiopulmonary exercise tests, the variation in OUESRER1 between the first and second tests expressed in percentage was not different from the VAT variation (−3.3% ± 17.2% and −7.5% ± 14.0%, respectively, P = 0.142). The variation in OUESRER1 between the first and second tests was significantly related to the variation in V˙O2peak between the tests (see Figure C, SDC 3, https://links.lww.com/MSS/A119, representation of the relationship between the change from the first to the second visit in V˙O2peak and in OUESRER1). Concomitant changes were observed for ΔOUESRER1 and ΔVAT, whereas ΔOUESRER1 seemed to be unrelated to ΔV˙E–V˙CO2 slope (Table 5). OUESRER1 was, however, significantly correlated with the V˙E–V˙CO2 slope when all subjects were considered (r = −0.444, P < 0.001; r = −0.557, P < 0.001 with V˙E–V˙CO2 slope calculated at VAT and including all data, respectively).
Since the original article by Baba et al. (5), OUES has been considered a potential method for predicting V˙O2peak from submaximal CPET. This assumption has relied mostly on appealing theoretical arguments, in particular, regarding putative independence from exercise duration, and on high correlation coefficients with V˙O2peak, which, however, were expected given that both indices are related to oxygen uptake.
Although there was a strikingly linear fit between oxygen uptake and the logarithm of V˙E, which left only small residuals, the last minutes of the test could have affected the OUES value, at least to some degree. Like others (14,28), we found that the OUES value calculated from the submaximal data set was slightly less than OUES100, although the mean is not significantly different. The values of V˙O2peak predicted from OUES100 and OUESRER1 did not differ, and the regression lines predicting V˙O2peak were not affected by the amount of data included, which is very compatible with previous observations. To validate OUES, most authors have assessed its linearity by comparing the OUES values calculated from different percentages of the total amount of data or equally, at different exercise intensities (8). If the use of OUESRER1 is to be developed, it will require more validation work because it is calculated over approximately 65% of the data set, which precludes the application of the findings usually based on 50% and 75% of the data set. Very interestingly, another approach was proposed recently. It consists in taking into consideration these changes at the end of test that occur in some patients and not in others. It was shown that the decrease in OUES was related to symptoms suggestive of CAD from a single-photon emission computed tomographic myocardial perfusion study (25).
The data on OUES in patients with CHF are not extensive enough to support its use in routine clinical practice. Some studies, however, found its accuracy to be satisfactory, in particular, when the statistical analysis focused only on correlation coefficients and regression analysis. These studies thus underlined its usefulness as an index of cardiorespiratory functional reserve (6,14,19). In our study, OUESRER1 was also strongly related to peak oxygen consumption, which corroborates the existing literature. However, this interesting concept hardly converts into a convincing clinical perspective. Some have concluded that OUES cannot accurately predict V˙O2peak in patients with CAD (28), overweight adolescents (12), or young healthy males (24). This is an example to confirm that the strength of the correlation between OUES and V˙O2peak, although necessary to characterize their relationship, is not sufficient to conclude about their interchangeability. Bland–Altman analyses are designed to overcome some of the limitations of the correlation approach for the purpose of agreement assessment (7). In the present study, on the basis of these analyses, we demonstrated that the OUESRER1 could not accurately predict V˙O2peak in men with mild stable CHF.
The agreement analysis between measured V˙O2peak and the V˙O2peak predicted from OUES was based on the examination of the initial linear relationships over the whole group. Because the strength of the relationships was weakened when body mass was introduced and the interest of OUES is its submaximal nature, we decided to focus the analysis on OUES calculated from a test finishing with an RER of 1.0 and V˙O2peak, neither of which is related to body mass.
The correlation coefficients we report between OUES, whatever the calculation mode, and V˙O2peak were more than 0.8. This gave a determination coefficient more than 60%, as currently reported in children with heart failure (5) and obesity (19), healthy children (20), adults with heart failure (6,22), and the healthy elderly (14).
Because of the remaining variability, the prediction of V˙O2peak was not accurate. The lower and upper limits of agreement were −0.57 and +0.57 L O2·min−1, respectively, similar to those for VAT and to previous observations in healthy subjects (24). This is proportionally even less acceptable in patients, however, because of the tiny ranges between very poor prognosis with peak oxygen uptake <10 mL O2·min−1·kg−1 (23), poor prognosis with <14 mL O2·min−1·kg−1 (18), and good prognosis with >18 mL O2·min−1·kg−1 (23). We used an a priori limit of ±4 mL O2·min−1·kg−1 for a 90-kg person, corresponding to ±0.36 L O2·min−1. Moreover, the maximal oxygen uptake variation related to biological and technological variability is likely to reach ±5.6% in trained subjects (17). Unfortunately, such information is not available for patients with heart failure. Thus, on the basis of the limits of agreement, OUES would not be better than VAT. The latter, however, is not identifiable in all patients (9), although it is of independent interest in the interpretation of exercise tests (34).
We investigated the sensitivity and specificity of OUES concerning changes in other ventilatory indices over time. This is important because these data are likely to contribute to our understanding of the factors that underlie this parameter.
We demonstrated a certain specificity of OUESRER1 to represent V˙O2peak. OUES varied from one occasion to another only when V˙O2peak varied and otherwise remained unchanged because the intercept of ΔV˙O2peak–ΔOUESRER1 was close to 0. This is very compatible with the previous observation that OUES values obtained on two separate occasions with an interval of no more than 7 d were highly correlated (4).
The sensitivity of OUES to changes in V˙O2peak has been partly assessed in different populations by the correlation coefficients between the two indices. The team that originally proposed OUES as an index of cardiorespiratory functional reserve (5) and others (24) reported high correlation coefficients for healthy subjects, whether from parametric or nonparametric tests, which indicated that OUES is able to discriminate fitter from less fit subjects. Our results suggested that OUESRER1 significantly tracked variation in the V˙O2peak (r = 0.717) better than VAT did, although the range of V˙O2peak change was narrow and the number of subjects was low. OUESRER1 also followed VAT variations (r = 0.796). Because OUES has been demonstrated to follow V˙O2peak changes induced by exercise training (r ranging from 0.64 to 0.77) (11,30) or orthotopic heart transplantation (29), its sensitivity to V˙O2peak is confirmed in this population with known impaired skeletal muscle metabolism and high ventilatory response to exercise. However, the r value with V˙O2peak means that more than half of the variation in V˙O2peak was explained by the same parameters that contributed to the OUESRER1 variation.
From the present data, it can be suggested that OUESRER1 and the V˙E–V˙CO2 slope have some common physiological determinants, given their significant although not strong negative correlation over the whole group, as observed previously (22). This should not be surprising because the latter is considered an index of dead space ventilation, which suggests its weight in the assessment of exercise intolerance (33). OUES represents the absolute rate of increase in V˙O2 per 10-fold increase in V˙E; it is thus likely to be influenced by the physiologic dead space/tidal volume ratio and the overall level of ventilation, which are themselves basically dependent on the structural integrity of the lungs, ventilation–perfusion matching, CO2 production, development of metabolic acidosis, and CO2 set point (33). The remaining variability not explained by common factors has not yet been elucidated. On an individual basis, changes in OUESRER1 seem to be determined by factors other than the determinants of changes in the V˙E–V˙CO2 slope. This discrepancy between individual and interindividual data also reported earlier (31) is quite surprising. It could be attributed to limitations in V˙E–V˙CO2 slope, although both calculation modes reach the same pattern. In a number of subjects that is limited but sufficient to alter the interpretation of the relationship, the V˙E–V˙CO2 slope showed wide variation from the first to the second occasion, but we are not certain that these changes were physiologically significant. A small difference in exercise duration in patients with a very poor cardioventilatory condition can result in more than a five-point increase in the V˙E–V˙CO2 slope indeed.
Further research is required before the relationship between V˙O2 and logV˙E can be used in routine CPET. It is important to accumulate data to assess the day-to-day reproducibility of OUES measurements on a different sort of patients. Hopefully, results would confirm satisfactory reproducibility observed in healthy people (4). A greater understanding of this relationship should also increase its usefulness in the management of patients with heart failure. Additional measurements, such as cardiac output and tissue oxygenation, will be important, as will studies under different environmental conditions, such as hypoxia, and in a variety of populations, such as both male and female patients with diastolic and systolic heart failure. These studies should aim to explain the variability of OUES that is not explained by the variability of V˙O2peak. This would shed greater light on the physiological meaning of OUES, as well as the other indices from that relationship such as the logV˙E value for which V˙O2 is null, theoretically representative of the level of ventilation from which oxygen uptake starts to rise.
In conclusion, OUES, when calculated from an exercise test that finishes with an RER of 1, relates reasonably well to V˙O2peak in men with heart failure. However, the limits of agreement identified were not sufficient to preclude the possibility that, in the some cases, a classically measured V˙O2peak might change treatment decisions. The measure was also responsive to change. Further research is needed to more fully determine its potential as a complementary source of information and, in particular, to identify the lowest RER threshold or percentage of predicted maximal HR from which information may be usefully drawn.
S.A.J. was funded by the European Respiratory Society.
The authors thank the support of Dr. Anita Simonds and Dr. Mary Morrell from the Sleep and Ventilation Department at the Royal Brompton Hospital.
This study was made possible by the recruitment and data collection of a project supported by a grant from the British Heart Foundation (grant P.G./2001042).
The authors declare no conflict of interest.
The results of the present study do not constitute endorsement by the American College of Sports Medicine.
1. Agostoni P. Cardiopulmonary exercise testing for heart failure patients: a hodgepodge of techniques, parameters and interpretations. In other words, the need for a time-break. Eur Heart J. 2006; 27: 633–4.
2. Arena R, Myers J, Guazzi M. The clinical and research applications of aerobic capacity and ventilatory efficiency
in heart failure: an evidence-based review. Heart Fail Rev. 2008; 13: 245–69.
3. Arena R, Myers J, Hsu L, et al.. The minute ventilation/carbon dioxide production slope is prognostically superior to the oxygen uptake efficiency slope. J Card Fail. 2007; 13: 462–9.
4. Baba R, Kubo N, Morotome Y, Iwagaki S. Reproducibility of the oxygen uptake efficiency slope in normal healthy subjects. J Sports Med Phys Fitness. 1999; 39: 202–6.
5. Baba R, Nagashima M, Goto M, et al.. Oxygen uptake efficiency slope: a new index of cardiorespiratory functional reserve derived from the relation between oxygen uptake and minute ventilation during incremental exercise. J Am Coll Cardiol. 1996; 28: 1567–72.
6. Baba R, Tsuyuki K, Kimura Y, et al.. Oxygen uptake efficiency slope as a useful measure of cardiorespiratory functional reserve in adult cardiac patients. Eur J Appl Physiol Occup Physiol. 1999; 80: 397–401.
7. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet. 1986; 1: 307–10.
8. Bongers B, Hulzebos H, Blank A, van Brussel M, Takken T. The oxygen uptake efficiency slope in children with congenital heart disease: construct and group validity. Eur J Cardiovasc Prev Rehabil. 2011; 18( 3): 384–9.
9. Cohen-Solal A, Benessiano J, Himbert D, Paillole C, Gourgon R. Ventilatory threshold during exercise in mild to moderate chronic heart failure: determination, relation with lactate threshold and reproducibility. Int J Cardiol. 1991; 30: 321–7.
10. Davies LC, Wensel R, Georgiadou P, et al.. Enhanced prognostic value from cardiopulmonary exercise testing in chronic heart failure by non-linear analysis: oxygen uptake efficiency slope. Eur Heart J. 2006; 27: 684–90.
11. Defoor J, Schepers D, Reybrouck T, Fagard R, Vanhees L. Oxygen uptake efficiency slope in coronary artery disease: clinical use and response to training. Int J Sports Med. 2006; 27: 730–7.
12. Drinkard B, Roberts MD, Ranzenhofer LM, et al.. Oxygen-uptake efficiency slope as a determinant of fitness in overweight adolescents. Med Sci Sports Exerc. 2007; 39 (10): 1811–6.
13. Gitt AK, Wasserman K, Kilkowski C, et al.. Exercise anaerobic threshold and ventilatory efficiency
identify heart failure patients for high risk of early death. Circulation. 2002; 106: 3079–84.
14. Hollenberg M, Tager IB. Oxygen uptake efficiency slope: an index of exercise performance and cardiopulmonary reserve requiring only submaximal exercise. J Am Coll Cardiol. 2000; 36: 194–201.
15. Hunt SA, Baker DW, Chin MH, et al.. ACC/AHA guidelines for the evaluation and management of chronic heart failure in the adult: executive summary a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (committee to revise the 1995 guidelines for the evaluation and management of heart failure): developed in collaboration with the International Society for Heart and Lung Transplantation; endorsed by the Heart Failure Society of America. Circulation. 2001; 104 (24): 2996–3007.
16. Ingle L, Witte KK, Cleland JG, Clark AL. The prognostic value of cardiopulmonary exercise testing with a peak respiratory exchange ratio of <1.0 in patients with chronic heart failure. Int J Cardiol. 2008; 127 (1): 88–92.
17. Katch VL, Sady SS, Freedson P. Biological variability in maximum aerobic power. Med Sci Sports Exerc. 1982; 14 (1): 21–5.
18. Mancini DM, Eisen H, Kussmaul W, Mull R, Edmunds LH Jr, Wilson JR. Value of peak exercise consumption for optimal timing of cardiac transplantation in ambulatory patients with heart failure. Circulation. 1991; 83: 778–86.
19. Marinov B, Kostianev S. Exercise performance and oxygen uptake efficiency slope in obese children performing standardized exercise. Acta Physiol Pharmacol Bulg. 2003; 27: 59–64.
20. Marinov B, Mandadzhieva S, Kostaniev S. Oxygen-uptake efficiency slope in healthy 7- to 18-year-old children. Pediatr Exerc Sci. 2007; 19 (2): 159–70.
21. Mudge GH, Goldstein S, Addonizio LJ, et al.. 24th Bethesda conference: cardiac transplantation. Task Force 3: recipient guidelines/ prioritization. J Am Coll Cardiol. 1993; 22: 21–31.
22. Myers J, Gujja P, Neelagaru S, Burkhoff D. Cardiac output and cardiopulmonary responses to exercise in heart failure: application of a new bio-reactance device. J Card Fail. 2007; 13: 629–36.
23. Opasich C, Pinna GD, Bobbio M, et al.. Peak exercise oxygen consumption in chronic heart failure: toward efficient use in the individual patient. J Am Coll Cardiol. 1998; 31: 766–75.
24. Pichon A, Jonville S, Denjean A. Evaluation of the interchangeability of V˙O2MAX
and oxygen uptake efficiency slope. Can J Appl Physiol. 2002; 27: 589–601.
25. Pinkstaff S, Peberdy MA, Kontos MC, Fabiato A, Finucane S, Arena R. Usefulness of decrease in oxygen uptake efficiency slope to identify myocardial perfusion defects in men undergoing myocardial ischemic evaluation. Am J Cardiol. 2010; 106 (11): 1534–9.
26. Selig SE, Hare DL. Evidence-based approach to exercise prescription in chronic heart failure. Br J Sports Med. 2007; 41 (7): 407–8.
27. Tabet JY, Beauvais F, Thabut G, Tartière JM, Logeart D, Cohen-Solal A. A critical appraisal of the prognostic value of the V˙E
slope in chronic heart failure. Eur J Cardiovasc Prev Rehabil. 2003; 10: 267–72.
28. Van Laethem C, Bartunek J, Goethals M, Nellens P, Andries E, Vanderheyden M. Oxygen uptake efficiency slope, a new submaximal parameter in evaluating exercise capacity in chronic heart failure patients. Am Heart J. 2005; 149: 175–80.
29. Van Laethem C, Goethals M, Verstreken S, et al.. Response of the oxygen uptake efficiency slope to orthotopic heart failure transplantation: lack of correlation with changes in central hemodynamic parameters and resting lung function. J Heart Lung Transplant. 2007; 26: 921–6.
30. Van Laethem C, Van de Veire N, De Backer G, et al.. Response of the oxygen uptake efficiency slope to exercise training in patients with chronic heart failure. Eur J Heart Fail. 2007; 9: 625–9.
31. Van Laethem C, Van de Veire N, De Sutter J, et al.. Prospective evaluation of the oxygen uptake efficiency slope as a submaximal predictor of peak oxygen uptake in aged patients with ischemic heart disease. Am Heart J. 2006; 152: 297.e9–15.
32. Vazir A, Hastings PC, Dayer M, et al.. A high prevalence of sleep disordered breathing in men with mild symptomatic chronic heart failure due to left ventricular systolic dysfunction. Eur J Heart Fail. 2007; 9 (3): 243–50.
33. Wasserman K. Principles of Exercise Testing and Interpretation: Including Pathophysiology and Clinical Applications. 3rd ed. Baltimore (MD): Lippincott Williams & Wilkins; 1999: 206.
34. Wasserman K, Whipp BJ, Koyal SN, Beaver WL. Anaerobic threshold and respiratory gas exchange during exercise. J Appl Physiol. 1973; 35: 236–43.