Peak oxygen uptake in heart failure: Look behind the number!

An impaired peak oxygen uptake (⁠V·O2peak) during a maximal cardiopulmonary exercise test (CPET) has been shown extensively to correlate with a poor prognosis in heart failure patients.1 The abovementioned relationship remains effective irrespective of the left ventricular ejection fraction (LVEF), a recent study by Sato et al. showing an unaltered V·O2peak prognostic power also in early defined heart failure with mid-range ejection fraction class.2 The advantage of adopting V·O2peak to evaluate and, possibly, to manage clinically this setting of patients is linked to its composite character, which enables it to account for most of the mechanisms underlying the heart failure pathophysiology. In fact, as per the well-known Fick Law, V·O2 represents arithmetically the resulting number of the product between cardiac output (CO; = stroke volume*heart rate) and artero-venous O2 difference (ΔavO2). Thus, a reduced V·O2peak, as a single variable, mirrors variably a reduced LVEF and/or a concomitant chronotropic incompetence as well as a huge number of conditions able to impact negatively on the O2 transport and delivery (Figure 1). Diagram showing cardiac output (CO) as a function of artero-venous difference (ΔavO2) with superimposed oxygen uptake (⁠V·O2) isopleths (Fick Law). Notably, the V·O2peak impairment could depend variably on the CO reduction (i.e. preserved, midrange or reduced ejection fraction heart failure) but also, with a greater degree of complexity, on the ΔavO2 modification due to possible concomitant comorbidities (i.e. lung diseases, anaemia, obesity, myopathies, etc.). Diagram showing cardiac output (CO) as a function of artero-venous difference (ΔavO2) with superimposed oxygen uptake (⁠V·O2) isopleths (Fick Law). Notably, the V·O2peak impairment could depend variably on the CO reduction (i.e. preserved, midrange or reduced ejection fraction heart failure) but also, with a greater degree of complexity, on the ΔavO2 modification due to possible concomitant comorbidities (i.e. lung diseases, anaemia, obesity, myopathies, etc.). The present issue of the journal publishes an interesting report by Van Iterson and colleagues,3 which investigates the impact on V·O2peak of two prevalent heart failure clinical features, namely increased body mass index (BMI; >30 kg/m2) and reduced haemoglobin (Hgb) content (<13/12 g/dL in men and women, respectively). The authors found that the combination of these two conditions led to the greatest V·O2peak impairment and, interestingly, they speculate that the ‘detrimental’ effect of a low Hgb content exceeds the possible ‘protective’ effect of a high BMI. Although some relevant limitations are correctly acknowledged (i.e. the adoption of a unique BMI cut-off value, lack of lean mass distribution and anaemic status characterization), the greatest value of the present study lies in its attempt to look behind an easy number and, even, to analyse possible synergies and/or interferences between single comorbidities with respect to the final V·O2peak value. Indeed, by analysing their data, we have a special opportunity to outline two actually debated heart failure comorbidities, namely obesity and anaemia, in terms of both CPET variables’ modifications and, briefly, possible prognostic significance. The relationship between obesity and exercise-induced gas exchange variables is complex and full of confounders.4,5 Generally, V·O2 for a given work rate (WR) is increased due to the high energy required to move the body mass. Thus, as a somewhat paradoxical consequence, in spite of a reduced WR, it is usual to observe absolute V·O2peak values that are normal in obese subjects, or even higher than in non-obese subjects. Noteworthy, the abovementioned phenomenon could be easily identified by a single Wasserman plot which displays the relationship between V·O2 and WR: the VO2/WR slope shows a normal value, but in obese patients it starts ‘upshifted’. It is well, in interpreting the V·O2peak value in obese subjects, to keep in mind that adipose tissue contributes poorly to the peak ΔavO2. Accordingly, it is always preferable to express V·O2peak also as the percentage of the predicted maximum related to both the actual and the ideal weight using the specific formulas (i.e. the Hansen/Wasserman equations).4 As per the VO2/WR slope, in obese subjects, the excessive metabolic requirement might impact on ventilation (VE) derived variables, too. Indeed, higher VE for a given WR can be observed, as well as a peculiar respiratory pattern characterized by low tidal volume associated with high respiratory rate, the latter finalized most likely to minimize the increased work of breathing due to the excess elastic load linked to the adipose tissues. Eventually, obesity presence could influence the relationship between VE and carbon dioxide production (VE/VCO2 slope) by increasing its value and, albeit still unproven, by modifying its prognostic power. Obviously, the degree of the changes in the CPET variables depends mostly on the obesity class (i.e. almost null in mildly obese). Regrettably, Van Iterson’s paper’s construction leaves us to imagine most of them. Although only mentioned by the authors, a few words are needed also about the so-called ‘obesity paradox’, that is, the hypothesized obesity beneficial effect on heart failure prognosis. Growing evidence tend to suggest this phenomenon is mainly due to a patient’s selection bias, with the obese heart failure patients studied most likely being those presenting a less advanced stage of heart failure. Accordingly, in a recent sub-analysis of the large Metabolic Exercise Cardiac and Kidney Indexes (MECKI score) dataset, the protective BMI disappeared when variables such as age, gender, LVEF and V·O2peak (as percentage of maximum predicted) were included into the multivariate model.6 The present study by Van Iterson and colleagues falls clearly outside the prognostic implications but does, albeit indirectly, support the thesis by showing any difference in V·O2peak between Group 1 (normal weight) and Group 3 (obese), which were well matched for age, gender, LVEF and Hgb content. However, contextually, the authors found a significant synergy in reducing the V·O2peak value of obesity and anaemia and this datum might be of interest in the management of systolic heart failure patients. While the BMI per se seems to impact marginally on V·O2peak, conversely the Hgb content plays a pivotal and well-described role in determining its value by affecting primarily the ΔavO2.4,5 Considering arterial O2 saturation normal, it could be possible to calculate in theory the amount of V·O2peak loss due to anaemia. Practically, given that each Hgb gram binds 1.34 ml of O2 at 100% saturation and considering an expected mean O2 extraction rate of around 70% at peak exercise,7 each Hgb gram should provide 1 ml of O2. Thus, by multiplying 1 ml of O2 by the CO (expressed in dL), we could obtain the amount of V·O2peak reduction due to the loss of 1 g/dL of Hgb. Unfortunately, neither peak CO nor peak ΔavO2 could be obtained through a simple CPET, but it is of note that a previous study on nearly 1000 systolic heart failure patients demonstrated (albeit in a semi-quantitative manner) an average of 109 ml/min change in V·O2peak for each Hgb gram.8 Regardless of its close relationship with a reduced V·O2peak, a low Hgb content represents also a strong independent prognostic factor in systolic heart failure patients.9 Supporting its powerful prognostic role, the previously cited MECKI score, built and validated considering several CPET, clinical, laboratory and echocardiographic parameters, identified the Hgb as one of the six independent predictors of total and cardiovascular mortality, the others being V·O2peak, VE/VCO2 slope, LVEF, renal function and sodium.10 However, a recent sub-analysis of the MECKI score dataset, focusing on the possible effect of different Hgb classes on the other five predictors, showed that in severe anaemic heart failure patients (Hgb < 11 g/dL) the VE/VCO2 slope lost its prognostic power.11 The present study by Van Iterson and colleagues also supports this intriguing datum, those patients with low Hgb content being characterized by worse ventilatory efficiency. Although the underlying mechanisms are far from elucidated, it might be hypothesized that an extreme O2 delivery reduction to the glomus cells of the carotid bodies may lead to an adjunctive VE stimulation which would overwhelm the physiologically CO2 dependent drive. All thing considered, the paper by Van Iterson and colleagues should encourage more rigorous pathophysiological studies dealing with the mechanisms underlying the reduced V·O2peak in heart failure patients. In such a context, in those cases where an expert analysis of all the CPET variables is not sufficiently conclusive, the combination of CPET with other diagnostic tools is beginning to play a relevant role.12 Specifically, depending on the centre skills and clinical indication, stress-echocardiography,13 inert gas rebreathing14 and full invasive haemodynamic investigation combined with CPET15 represent an approach as fully integrated as possible to identify the ‘weakest link’. Intriguingly, what emerges from the present report, as well as from this brief overview, is that V·O2peak in heart failure represents just the starting point, deep comprehension of the ΔavO2 changes being the greatest contemporary challenge. The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. The author received no financial support for the research, authorship, and/or publication of this article.