Isocapnic buffering: An inconvenient truth about cardiopulmonary exercise testing in heart failure

Cardiopulmonary exercise testing (CPET) is a proven contemporary diagnostic tool invaluable for evaluating heart failure severity and candidacy for mechanical circulatory support or heart transplantation.1,2 Support for these relationships is underpinned by heavy reliance on the “lowest hanging fruit” on the CPET tree or commonly reported as the “maximal or peak” oxygen uptake (⁠V·O2) measurement. Acceptance that multiple powerful CPET measurements exist should not be misinterpreted as tempered enthusiasm for peak V·O2 (⁠V·O2peak).1–3 Technological innovation, convenience, and simplicity have allowed CPET to withstand the test of time, whereas the unintended consequence of such modernization is the ease with which unique features of CPET can be neglected. The polymorphic nature of CPET is real, and understanding the “sum of the parts” as much as the “parts of the sum” is needed to narrate the complete V·O2 phenotype.2–5 Measurements involved in CPET should not be recognized as static or so straightforward in surreptitiously meeting pre-specified nomenclature norms.2,6,7 The quest for what makes measuring and understanding exercise capacity invaluable in the realm of heart failure has yet to completely embrace the notion that providing a layered phenotype of exercise V·O2 beyond “peak” and oversimplification of Fick’s Principle is necessary. The inconvenient truth is CPET must not be synonymous with a single variable or push of a button. Links shared between exercise capacity and the shifting imbalance between O2-dependent and O2-independent metabolism are known,8 yet continue to be overlooked. Skeletal muscle pathophysiology affecting O2-dependent metabolism is a recognized burden of heart failure. However, V·O2peak in itself is not intuitively useful for helping to provide a phenotype of (glycolytic) substrate level metabolic contributions to exercise capacity. How could meticulously providing a metabolic “domain” based phenotype of exercise V·O2 provide rich diagnostic information not extractable from what has become routine? While concepts of isocapnic buffering and the anaerobic threshold (AT) are not new,1,8,9 the specific V·O2 response rising, or lack thereof, between the AT and respiratory compensation point (RCP) during CPET has engendered little attention in heart failure. It is unclear whether this transient sector where oxidative phosphorylation and glycolysis impact the measurable V·O2 response is clinically benign. In this issue of European Journal of Preventive Cardiology, Carriere et al.10 have carried out an insightful clinical and exercise physiological study in which they tested for the first time whether the magnitude of increase in V·O2 within the isocapnic buffering domain predicts syndrome severity and prognosis in patients with heart failure and reduced ejection fraction (HFrEF). Across a 10 year period and multiple heart failure centers, 1995 well-phenotyped New York Heart Association class I–IV and American College of Cardiology/American Heart Association Class B or C patients performed an initial CPET (individualized upright cycle ramp-slope protocol) and thereafter were prospectively followed. The study end-point was the composite of cardiovascular mortality, urgent heart transplantation, and left ventricular assist device implantation. CPET data were retrospectively analyzed and the authors carefully identified whether both the AT and RCP were discernible, and, if so, the increase in V·O2 within this domain (ΔV·O2AT-RCP) was quantified. The AT and RCP were identified in 782 cases (∼40%), for whom the median ΔV·O2AT-RCP was 3.85 mL/kg per min (interquartile range (IQR), 2.69–5.46 mL/kg per min). As initially suspected by the authors, there was a detrimental effect on V·O2peak relating to ΔV·O2AT-RCP falling below (n = 390) the cohort median, approximately 4 mL/kg per min lower than V·O2peak of counterparts (n = 392). These differences were not explainable by variable CPET quality as both groups met peak effort criteria.1 Though failing to meet the median ΔV·O2AT-RCP was not predictive of the composite end-point at five years’ follow-up, importantly, demonstrating a low ΔV·O2AT-RCP was a significant univariate predictor of worse HFrEF severity and prognosis (median follow-up 3.15 years, IQR 1.51–5.61 years; hazard ratio 0.90; 95% confidence interval (CI) 0.82–0.99; p = 0.04). The univariate as opposed to multivariate nature of this association is misleading, with further commentary below. The additional finding of no stark impact on ventilatory efficiency as a consequence of a small ΔV·O2AT-RCP was perhaps most surprising. A modest-to-moderate severity ventilatory equivalent for carbon dioxide output (⁠V·E/V·CO2) slope was observed for all patients. Inhomogeneity between ΔV·O2AT-RCP and V·E/V·CO2 slope should not technically occur since, at least on a mathematical basis, any ΔV·O2AT-RCP superimposes the assumingly linear tail of the V·E/V·CO2 slope. Still, only the V·E/V·CO2 slope persisted in both univariate (hazard ratio 1.08; 95% CI 1.06–1.11; p < 0.001) and multivariate (hazard ratio 1.09; 95% CI 1.02–1.16; p < 0.007) models as a significant predictor of the study endpoint. In lieu of such statistical disparity, conventional wisdom suggests ΔV·O2AT-RCP in fact relates with ventilatory equivalents and the study endpoint. By assuming that V·E, true to its expected nature, rises quasi-exponentially relative to V·O2 beyond the AT, and ramp-slope exercise forces V·E and V·CO2 to maintain linearity up to the RCP,8 together these features imply that demonstrating a large ΔV·O2AT-RCP is indeed overwhelmingly dependent on the integrative degree to which the smallest increases in V·E and O2 delivery are able to properly accommodate rapidly increasing oxidative metabolic demand.4,5 Certainly, for such associations to persist beyond the AT, the need for acid-base regulation at both lung and skeletal muscle is synonymous with the appearance of “isocapnic buffering.” For post-AT exercise, the buffering of lactic acid (Na+HCO3– + H+lactate− ↔ Na+lactate– + H2CO3 ↔ CO2 + H2O) appreciably contributes to CO2 in the expirate (Figure 1). Included in these bicarbonate buffer system interactions are the underappreciated roles played by the Haldane effect and alkaline Bohr effect. Full details of these mechanisms are beyond the scope of this commentary, but, in general, binding of O2 to hemoglobin (Hb) within the pulmonary circulation favors the dissociation of H+ and CO2 from Hb, thereby shifting the bicarbonate buffer equilibrium to the right and eliminating CO2 and H+ from the circulation. Alternatively, increased gross metabolism at skeletal muscle yields surges in CO2 and H+ production, facilitating the offloading of O2 from Hb and allowing the binding of CO2 and H+ for transport towards lungs. This is synonymous with shifting of the bicarbonate buffer equilibrium to the left. Thus, it should be clear that without influences from both the Haldane effect and alkaline Bohr effect, bicarbonate buffer system interactions cannot occur. Generalized schematic illustrating simplified bicarbonate buffer system interactions, including the impact of the Haldane effect and alkaline Bohr effect, associated with the isocapnic buffering period of cardiopulmonary exercise testing. Exercise performed below the anaerobic threshold can be expected to result in the majority of carbon dioxide (CO2) expired to be equivalent to the rate of oxidation of substrate. For exercise performed above the anaerobic threshold and up to the respiratory compensation point, buffering of lactic acid (Na+HCO3– + H+lactate– ↔ Na+lactate– + H2CO3 ↔ CO2 + H2O) greatly contributes to CO2 in the expirate. ATP: adenosine triphosphate; Hb: hemoglobin A clearer benefit of the coupling that must occur between V·E and V·CO2 included in the process of maintaining lung and skeletal muscle acid-base equilibrium is creation of the required O2 pressure driving gradient along the length of the O2 transport chain (Figure 1: atmosphere→lungs→ capillaries→heart→arteries→ skeletal muscle→capillaries→mitochondria; and vice-versa for CO2 transport). From a global systemic level this is well-visualized as distinct phases of the exercise ventilatory equivalent for V·O2, illustrating that sub-AT CO2 production is not that of glycolysis, but of oxidative phosphorylation. Thus, the fact that for the present study the V·E/V·CO2 slope was not egregiously severe in either group, whereas failure of end-tidal CO2 tension (PETCO2) to rise adequately leading up to the AT was strongly predictive of HFrEF severity and prognosis, suggests that patients demonstrating the smallest ΔV·O2AT-RCP and poorest exercise capacity were limited by sub-AT ventilatory inefficiency occurring secondary to ineffective central and peripheral hemodynamics. Ventilatory inefficiency within the heavy O2-dependent metabolic domain imputed to the prognostic power of low PETCO2 at the AT was also likely impacted by pulmonary pathology relating to impaired lung diffusion. While the present study did not report such data, impaired lung diffusion in heart failure is germane to alveoli, bounding membranes of the alveolar–capillary interstitium, pulmonary capillaries, or most likely a composite of each.11,12 Regardless of the exact nature in which the size of the ΔV·O2AT-RCP can be explained using CPET data alone, credit must go to the authors for this original and insightful effort demonstrating that the addition of ΔV·O2AT-RCP to V·O2peak and the V·E/V·CO2 slope adds real-world value to the interpretive and prognostic power of CPET in patients with HFrEF. Integrative physiological considerations also challenge that the initial study observation in which isocapnic buffering could not be defined in nearly 60% of patients is another key finding with major clinical implications. These patients likely demonstrated the severest heart failure, although follow-up testing is needed to confirm this hypothesis. In conclusion, results of the present study support the continued push for reporting of the unconstrained and layered CPET and V·O2 phenotype. “Isocapnic buffering” is for the better, but not worse, an inconvenient truth about CPET in patients with 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.