Author(s): ACEITUNO PINTO, H., CANO-CAPPELLACCI, M., RAMOS-ÁLVAREZ, J.J., SERÓN, P., ROULIEZ-ANAYA, K., OPAZO-DÍAZ, E. , Institution: UNIVERSIDAD DE CHILE, Country: CHILE, Abstract-ID: 2346

INTRODUCTION:
Cardiorespiratory training in heart failure (HF) improves functional capacity, aerobic fitness, and quality of life, and may reduce hospitalizations. Individualized intensity prescription using cardiopulmonary exercise testing (CPET)-derived ventilatory thresholds (VTs) is recommended, but gas-analysis and expert visual interpretation limit feasibility. Muscle oxygen saturation (SmO₂) may provide practical breakpoints, yet its validity in HF is unclear. Therefore, this study aimed to examine the agreement between visually determined ventilatory thresholds and SmO₂ breakpoints in HF.
METHODS:
Twenty clinically stable adults with heart failure with reduced ejection fraction (HFrEF; left ventricular ejection fraction [LVEF] ≤40%) and New York Heart Association (NYHA) class II–III performed a symptom-limited incremental CPET on a cycle ergometer (4 min unloaded; ramp +6 W·min⁻¹ for NYHA II or +10 W·min⁻¹ for NYHA III) with breath-by-breath gas exchange and vastus lateralis SmO₂ (Moxy Monitor). Ventilatory threshold 1 (VT1) and ventilatory threshold 2 (VT2) were visually identified by two blind observers, in case of disagreement, third observer analyzed the data to resolve the disagreement. SmO₂ was plotted against oxygen uptake (VO₂) or workload (W) and analyzed in MATLAB using an automated four-knot segmented regression to identify two breakpoints (BP1, BP2). Thresholds were expressed as VO₂ (L·min⁻¹) and workload (W). Agreement was assessed using Bland–Altman analysis and intraclass correlation coefficients (ICC).
RESULTS:
Between VT1 and BP1 expressed as oxygen uptake, bland–Altman Analysis showed a mean bias of -0,02 L·min⁻¹ with 95% limits of agreement (LOA) from -0.39 L·min⁻¹ to 0.34 L·min⁻¹ and an ICC of 0.67 (0.35 - 0.86). Between VT1 and BP1 expressed workload, bland–Altman Analysis showed a mean bias of 23 W with 95% LOA from -18 to 63 W and an ICC of 0.13 (-0.33 - 0.53). Between VT2 and BP2 expressed as oxygen uptake, bland–Altman Analysis showed a mean bias of 0,04 L·min⁻¹ with 95% LOA from -0.35 L·min⁻¹ to 0.44 L·min⁻¹ and an ICC of 0.80 (0.56 - 0.92). Between VT2 and BP2 expressed workload, bland–Altman Analysis showed a mean bias of 8 W with 95% LOA from -42 to 58 W and an ICC of 0.28 (0 – 0.64).
CONCLUSION:
Agreement between VT1 and BP1 was low when expressed as VO₂ (L·min⁻¹) or workload (W), suggesting limited interchangeability between BP1 and the first ventilatory threshold in patients with HFrEF. In contrast, agreement between VT2 and BP2 was high when expressed as VO₂, supporting BP2 as a potential surrogate to estimate the second ventilatory threshold. However, when expressed as workload, the strength of agreement decreased, warranting cautious interpretation and further validation.