Cardiopulmonary exercise testing and risk stratification in heart failure with reduced, midrange or preserved ejection fraction: When nomenclature may not match with pathophysiology

Lately, the use of cardiopulmonary exercise testing (CPET) in cardiological practice has impressively grown, basically because of its high reliance on defining stable and reproducible prognostic-oriented measures.¹

In cardiac patients, most of the gas exchange variables measured during maximal exercise out of peak oxygen consumption (VO₂) provide a thorough prediction of outcome with a few of them that are surged as cornerstones in the clinical and prognostic assessment of heart failure with reduced ejection fraction (HFrEF), such as the rate of increase in ventilation to carbon dioxide production (VE/VCO₂ slope),² and the presence of exercise oscillatory ventilation (EOV) pattern.³

In order to improve the performance of these indicators, single metrics have been incorporated in statistical models to generate multi-parametric and composite scores^4,5 or collected in simple quantitative analyses by color-coded tables of risk in a simple ready-to-use format to be used in daily routine clinical practice.⁶

Validation of CPET-based prognostic algorithms has been carried on just in HFrEF and what is actually striking is that any application of those criteria to heart failure (HF) with different left ventricular ejection fraction (LVEF) levels, (i.e. preserved (p) EF and midrange (mr) EF) is not warranted, revealing how little we know about these two HF phenotypes and especially HFmrEF.⁷

Accordingly, open questions that physicians and exercise physiologists have quite often to face are how to interpret a similar CPET response in different categories of LVEF and whether LVEF is the exact categorical determinant that matches with specific exercise-gas-exchange phenotypes. Remarkably, an impaired cardiac output reserve during physical challenge is typical of any form of HF^8–10 and a reduced O₂ delivery is the main driver for an impaired performance. Nonetheless, studies performed in HFrEF and HFpEF, analyzing the putative respective role of Fick principle determinants have shown that additional hemodynamic contributory factors to exercise limitation in HF are a delayed and an inefficient O₂ diffusion from capillaries to mitochondria with a quite preserved O₂ peripheral extraction.^11,12

An isolated report on small numbers has pointed out the relevance of an unfavorable exercise-gas-exchange response by CPET analysis as tracking measure of disease severity,¹³ irrespective of reduced or preserved LVEF.

Under this aspect, LVEF at rest has been repeatedly found not to correlate with peak VO₂,¹⁴ it does not reflect cardiac contractility¹⁵ and does not predict, per se, stroke volume changes during exercise.¹⁶ Moreover, the link between LVEF and peripheral mechanisms implicated in O₂ diffusion and extraction remains elusive.^9,11 It has, however, to be acknowledged that LVEF, when reduced, is a strong predictive factor and although dependent on preload and afterload, is a sensitive indicator of LV eccentric geometry which definitively affects cardiac filling and output.¹⁷

Therefore, whether CPET data may provide a prediction of outcome in the three LVEF categories helping to overcome the differences in the clinical condition, comorbidity and responsiveness to therapy represent an additional intriguing unmet question.

In the December issue of the European Journal of Preventive Cardiology, Sato et al.¹⁸ report a single center prognostic analysis of CPET-derived variables performed in HF patients categorized according to the European Society of Cardiology Guideline’s LVEF classification. Specifically, they tested the prognostic ability of main CPET-derived variables, such as peak VO₂, VE/VCO₂ slope, oxygen uptake efficiency slope (OUES) and EOV, in a population of 1190 HF patients with HFrEF (41.9%) HFpEF (36.8%) and HFmrEF (21.3%). A similar rate of adverse cardiac-related events was observed in HFmrEF and HFpEF, whereas the highest rate of events occurred in HFrEF.

Peak VO₂ was the common independent predictor throughout the three groups and, in addition to it, there were the OUES for both HFrEF and HFmrEF and EOV for HFpEF.

There is an interesting amount of feedback from the present study that, even if single center and a retrospective one, provides sufficient numbers and evidence for drawing some conclusions.

This is actually the second study available extending clinical information based on exercise gas-exchange variables, to HFmrEF entity, showing that exercise performance may help to further characterize patients pertaining to this narrow LVEF range.

CPET results point to an intermediate level of performance, ventilation efficiency and metabolic pattern in this understudied category. At variance with the previous study by Nadruz et al.,¹⁹ the VE/VCO₂ slope did not emerge as prognostic, a finding that may be explained by a likely low rate of patients with combined pre- and post-capillary pulmonary hypertension.²⁰ Nonetheless, the main impactful message is that peak VO₂ has a full prognostic applicability throughout all LVEF groups, implying that the entire physiological processes contributing to O₂ chain utilization remain the mainstay pathways to be targeted and improved in all forms of cardiac failure.

Overall, findings by Sato et al.¹⁸ are challenging because, although we are assisting at a widespread use of classifying HF based on three levels of LVEF, they raise the question whether this nomenclature may truly impact precision in risk stratification as assessed by CPET.

Despite these merits, the study fails to focus on how much LVEF measures obtained by echo were reproducible, considering the range of variability determined by echocardiography, which questions the appropriateness of measures in the narrow window of HFmrEF.²¹

If the current evidence is confirmed by other studies, we will definitively need to address whether in HF syndromes there is a need to build up different prognostic toolkits based on LVEF categorization, or rather, we should have to rely on the pathophysiology behind exercise-gas-exchange phenotypes as a source of prognostic co-relates irrespective of LVEF, thus prioritizing its proven implications in grading the disease’s severity. Both approaches appear arbitrary at this stage but present observations seem to encourage us to point to the exercise-cardiopulmonary phenotype as the more trustful indication in the process of clinical decision-making and outcome prediction in HF.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: MG is supported by the Monzino Foundation Grant, Milano, Italy.

References