Participants' hypoxic chemosensitivity was assessed using the pure nitrogen breathing test (N 2 T) and the Richalet hypoxia sensitivity test (RT) ( 30,31 ). During each test, V E (Quark CPET; Cosmed) and SpO 2 (Wristox; Nonin Medical Inc) were continuously monitored.
Briefly, for the N 2 T, subjects were exposed to a series of 10 periods of 100% nitrogen (N 2 ; 1-8 consecutive breaths). The number of breaths of N 2 was varied and randomized to produce a wide range of arterial oxygen desaturations. Transitions from room air to N 2 breathing were accomplished using a three-way valve with the participants blinded to this maneuver. Consecutive N 2 exposures were separated by at least 2 min to allow respiratory parameters to return to baseline values. During each N 2 period, the nadir SpO 2 was plotted against the highest V E . The hypoxic ventilatory response (HVR) was determined as the slope of the relationship between SpO 2 and V E .
The RT was performed as previously described ( 31). Briefly, participants were exposed to the following sequence: 4-min rest in normoxia, 4-min rest in hypoxia (F I O 2 = 0.115), 4-min exercise (at 30% of normoxic V O 2max ) in hypoxia, and 4-min exercise at the same intensity in normoxia. HVR was calculated as the ratio between normoxia-to-hypoxia changes in ventilation over saturation normalized by body weight ( 31 ).
Briefly, for the N 2 T, subjects were exposed to a series of 10 periods of 100% nitrogen (N 2 ; 1-8 consecutive breaths). The number of breaths of N 2 was varied and randomized to produce a wide range of arterial oxygen desaturations. Transitions from room air to N 2 breathing were accomplished using a three-way valve with the participants blinded to this maneuver. Consecutive N 2 exposures were separated by at least 2 min to allow respiratory parameters to return to baseline values. During each N 2 period, the nadir SpO 2 was plotted against the highest V E . The hypoxic ventilatory response (HVR) was determined as the slope of the relationship between SpO 2 and V E .
The RT was performed as previously described ( 31). Briefly, participants were exposed to the following sequence: 4-min rest in normoxia, 4-min rest in hypoxia (F I O 2 = 0.115), 4-min exercise (at 30% of normoxic V O 2max ) in hypoxia, and 4-min exercise at the same intensity in normoxia. HVR was calculated as the ratio between normoxia-to-hypoxia changes in ventilation over saturation normalized by body weight ( 31 ).