Cardiopulmonary exercise testing in advanced heart failure with reduced ejection fraction: Time for in-depth analysis for central and peripheral contributors to peak VO₂?

Traditionally, cardiopulmonary exercise testing (CPET) was used almost exclusively for the selection of heart transplantation (HTx) candidates suffering from heart failure (HF) with reduced ejection fraction (HFrEF).1 Now, peak oxygen consumption (VO2) is also used for risk stratification in a broader patient population including elderly patients with HF and mid-range or preserved ejection fraction,2 patients with other comorbidities such as obesity and anaemia,3 and left ventricular assist device recipients.4 Furthermore, CPET has been likewise established for follow-up evaluation of therapeutic interventions. The non-invasive measure of cardiac output (CO) with maximum exercise as described by Vignati et al. in this issue has the potential to add further value to CPET.5 Peak VO2 is the physiological equivalent of maximal oxygen uptake during exhaustive exercise and is usually associated with maximal CO. VO2 and CO are related by the Fick principle [VO2 = COċ(CaO2 – CvO2)], in which CaO2 is arterial oxygen content and CvO2 venous oxygen content. CaO2 is the product of haemoglobin concentration (Hb), oxygen saturation (SaO2) and a coefficient correcting for the ml of oxygen carried per gram of haemoglobin [CaO2 = [Hb]ċSaO2ċ1.34]. Arterial SaO2 and Hb do not vary largely in the healthy, underlining the pivotal role of CO for oxygen delivery to the periphery. However, exercise capacity in HF is not only limited by central factors such as CO and cardiac filling pressures but also by the periphery to the heart including the vasculature, endothelium and skeletal muscle.6 Invasive methods for a long time remained the only reliable CO measurement during exercise despite a significant effort to correlate echocardiographic indices with CPET parameters. More recently, studies have suggested that inert gas rebreathing (IGR) maneuvers performed during maximum exercise should allow for a non-invasive measure of CO.7 Vignati et al. in their current study investigated whether application of this emergent non-invasive measurement for CO affects CPET parameters in HF patients.5 Their test design compared in the same patient routine CPET with CPET plus IGR both performed within a short time interval. This comparison showed that IGR during maximum exercise does not affect anaerobic threshold or peak VO2 while peak workload was lower and the VE/VCO2 slope was higher with CPET plus IGR suggesting increased respiratory workload with IGR. The study participants were ambulatory chronic HFrEF patients on optimal medical treatment, who were predominantly men and middle-aged. The study population profile corresponds to the characteristics of advanced HF patients submitted to CPET in order to investigate the indication for HTx listing. This raises the question as to whether CO measurement during CPET provides added value to this group of HF patients. The 2006 guidelines on listing criteria for HTx indicated a peak VO2 cutoff of less than 14 ml O2*min–1* kg body weight for beta-blocker-intolerant HFrEF patients while it was less than 12 ml O2*min-1*kg body weight–1 for beta-blocker-tolerant HFrEF patients.1 In fact, a survival advantage with HTx was shown when HFrEF patients presented with peak VO2 below the respective cutoffs.1 The 10-year update in 2016 maintained these peak VO2 cutoffs despite the advent and broader application of cardiac resynchronisation therapy (CRT) in HF, which can increase left ventricular contractility. However, the effect of CRT treatment on peak VO2 was inhomogeneous with a neutral effect shown in the post hoc analysis of the COMPANION trial while a study in mildly symptomatic HF patients suggested an increase of peak VO2.1 HF drug medication has likewise changed significantly in the last few decades and a study pooling two HF patients’ cohorts with CPET performed between 1994 and 2008 showed a 21% increase of 2-year survival in HFrEF patients with peak VO2 less than 10 ml O2*min–1*kg body weight–1 when tested after 2000.8 The latter patients were more often treated with beta-blockers or mineralocorticoid receptor antagonists and the number of implanted devices was higher9 suggesting that nearly optimal therapy alters the predictive significance of peak VO2 less than 10 ml O2*min–1*kg body weight–1. In accordance, data stratified by peak VO2 and either the presence or absence of beta-blocker treatment show that the event rate is higher for those not taking beta-blockers at each stratum of peak VO2.1 However, we cannot exclude that patients on optimal treatment improved left ventricular function, as reported in the CHART-2 study in which the dynamic transition of HFrEF to either HF with mid-range ejection fraction or HF with preserved ejection fraction occurred in 18% and 22% of study participants within the first year after registration.10 In concordance, two serial CPET studies in HTx candidates have shown that even these patients can increase their peak VO2.1 Coming back to the Fick principle, there is an inverse relationship between peak VO2 and venous oxygen content, and this reciprocity becomes a pivotal determinant of peak VO2 if CaO2 and CO remain unchanged. In fact, CaO2 remains approximately unchanged at rest and with exercise, and in the healthy as well as the HF patient. Likewise, CO may not appreciably increase with exercise in more severe HF. This implies that for individuals with a given CaO2 and CO, peak VO2 is increased in those with the lowest CvO2. This concept brings us to the level of the capillary muscle and its delicate action transferring oxygen to the muscle mitochondrion. It is here that the CPET plus IGR measure can help to distinguish between the patient with peak VO2 increase due to favourable cardiac remodelling and the end-stage HFrEF patient who increases peak VO2 solely by increasing the peripheral extraction of oxygen but without concomitant CO augmentation.10 Unravelling of cardiac remodelling is the domain of transthoracic echocardiography or cardiac magnetic resonance imaging (MRI), but serial measurement of size and function must show considerable difference to achieve significance. Likewise, stress echocardiography may miss smaller increases in contractility as stroke volume increases proportionally less importantly with exercise (20–40%) than does heart rate (200–400%).6 For example, a relatively small increase in stroke volume such as 8 ml would increase exercise CO by 1 l/minute at a peak heart rate of 120 bpm. Echocardiography and cardiac MRI may therefore miss a minor improvement of left ventricular function while a subsequent increase in CO is reliably measurable by CPET plus IGR.7 This should identify the patient with favourable central remodelling and potential for further improvement, which may ultimately result in removal from the HTx list. On the other hand, those patients with increase of peak VO2 but an absence of appreciable change in CO should remain on the HTx list because of persisting severe cardiac dysfunction. Nevertheless, these patients merit particular scientific attention because comprehension of their favourable remodelling of the periphery may provide another therapeutic option for HF patients not candidates for HTx or assist device treatment. In summary, CPET plus IGR has the potential to become a valuable tool for exercise-based phenotyping in HFrEF. The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. The author(s) received no financial support for the research, authorship, and/or publication of this article.