Do Agreements between HRV and Gas Exchange Thresholds Still Hold under Hypoxic Conditions?

Purpose: Heart rate (HR) variability thresholds (HRVT) based on detrended fluctuation analysis alpha 1 (DFA a1) generally show reasonable alignment of thresholds estimations based on gas exchange responses under normoxic conditions. This study examined whether acute hypoxia would affect the agreement between HRVTs and the gas exchange equivalents during incremental cycling. Methods: Twelve participants (5 females) completed an incremental ramp test in normobaric hypoxia (FIO2 ≈ 13.5%) and normoxia. Gas exchange and ventilatory responses alongside a high sampling rate electrocardiogram for DFA a1 computation were used to determine thresholds. Comparisons were made between the oxygen consumption (V̇O2) and HR at the gas exchange threshold (GET) and respiratory compensation point (RCP) with the responses at the first and second HRVTs (HRVT1 and HRVT2 respectively). Results: Mean V̇O2 and HR values were not statistically different for GET:HRVT1 (normoxia:1.74±0.41 vs 1.74±0.48 L·min-1,133±18 vs 133±16 bpm; hypoxia:1.47±0.21 vs 1.45±0.37 L·min-1, 135±14 vs 133±15 bpm) and RCP:HRVT2 (normoxia:2.38±0.55 vs 2.37±0.48 L·min-1, 158±13 vs 158±14 bpm, hypoxia:2.07±0.32 vs 1.90±0.43 L·min-1 and 156±13 vs 152±15 bpm) in any condition. All normoxic comparisons passed equivalence testing but only GET:HRVT1 responses passed during hypoxia. Pearsons r correlation coefficients were 0.86 to 0.96 in normoxia and 0.58 to 0.79 in hypoxia. Bland Altman analysis indicated higher degrees of bias and limit of agreements (LOA) during hypoxic testing. Conclusions: Although the V̇O2 and HR at HRVTs retained alignment with GET/RCP in both normoxia and hypoxia, the degrees of correlation, and equivalence were weaker and the bias and LOA were larger in hypoxia. Therefore, whilst using HRVT alone for training boundary guidance in hypoxia is a potential option, further investigation including incorporating complimentary surrogate markers is recommended.