Pick Your Threshold: A Comparison Among Different Methods of Anaerobic Threshold Evaluation in Heart Failure Prognostic Assessment

BACKGROUND: In clinical practice, anaerobic threshold (AT) is used to guide training and rehabilitation programs, to define risk of major thoracic or abdominal surgery, and to assess prognosis in heart failure (HF). AT of oxygen uptake (V)over dotO(2); (V)over dotO(2)AT) has been reported as an absolute value (V)over dotO(2)ATabs), as a percentage of predicted peak (V)over dotO(2) (V)over dotO(2) (V)over dotO(2)AT%peak_pred), or as a percentage of observed peak (V)over dotO(2) (V)over dotO(2)AT%peak_obs). A direct comparison of the prognostic power among these different ways to report AT is missing.RESEARCH QUESTION: What is the prognostic power of these different ways to report AT?STUDY DESIGN AND METHODS: In this observational cohort study, we screened data of 7,746 patients with HF with a history of reduced ejection fraction (< 40%) recruited between 1998 and 2020 and enrolled in the Metabolic Exercise Combined With Cardiac and Kidney Indexes register. All patients underwent a maximum cardiopulmonary exercise test, executed using a ramp protocol on an electronically braked cycle ergometer.RESULTS: This study considered 6,157 patients with HF with identified AT. Follow-up was median, 4.2 years (25th-75th percentiles, 1.9-5.0 years). Both (V)over dotO(2)ATabs (mean +/- SD, 823 +/- 305 mL/min) and (V)over dotO(2) AT%peak_pred (mean +/- SD, 39.6 +/- 13.9%), but not (V)over dotO(2) AT%peak_obs (mean +/- SD, 69.2 +/- 17.7%), well stratified the population regarding prognosis (composite end point: cardiovascular death, urgent heart transplant, or left ventricular assist device). Comparing area under the receiver operating characteristic curve (AUC) values, (V)over dotO(2)ATabs (0.680) and (V)over dotO(2)AT%peak_pred (0.688) performed similarly, whereas (V)over dotO(2)AT%peak_obs (0.538) was significantly weaker (P < .001). Moreover, the (V)over dotO(2) AT%peak_pred AUC value was the only one performing as well as the AUC based on peak (V)over dotO(2) (0.710), with an even a higher AUC (0.637 vs 0.618, respectively) in the group with severe HF (peak (V)over dotO(2) < 12 mL/min/kg). Finally, the combination of (V)over dotO(2)AT% peak_pred with peak (V)over dotO(2 )and (V)over dot per CO2 production shows the highest prognostic power.INTERPRETATION: In HF, (V)over dotO(2)AT%peak_pred is the best way to report (V)over dotO(2) at AT in relationship to prognosis, with a prognostic power comparable to that of peak (V)over dotO(2) and, remarkably, in patients with severe HF.