Based on a previous study that observed decreases in mean arterial pressure (MAP) after hypoxic exercise (Horiuchi et al., 2016), a sample size estimation for the primary analysis (MAP) indicated that 10 participants were needed to produce an 80% chance of obtaining statistical significance at the level of 0.05 (G Power 3.1) (Faul et al., 2007).

Ten healthy young men with a mean age of 21 ± 1 years, a height of 174 ± 5 cm, and a body mass of 66.8 ± 8.7 kg [mean ± standard deviation (SD)] were enrolled. The participants engaged in regular physical activity (1–2 h per day, 3–5 days per week). None of the participants were exposed to an altitude higher than 1,500 m within 6 months prior to the study, and had a history of cardiovascular or orthopedic disease. They were non-smokers. After detailed explanations regarding the study, including the procedures, possible risks, and benefits of participation, a written informed consent was obtained from each participant. This study was approved by the Ethical Committee of the University of Yamanashi, in Japan, and was performed in accordance with the guidelines of the Declaration of Helsinki (No. 201509).

The participants were asked to abstain from caffeinated beverages for 12 h, strenuous exercise and alcohol consumption for 24 h before each session. The room temperature was set at 24 ± 1°C, and external stimuli were minimized. All experiments were performed using a cycle ergometer (Powermax V, Combi Co. Ltd., Tokyo, Japan). On the first visit, they familiarized themselves with 1) leg cycling with maximal pedaling revolutions while wearing all the devices as well as 2) the BP measurement and blood sampling. On the second and third visits, participants performed main experiment. The experiment consisted of three consecutive protocols; 1) a 10-min sitting/resting period, 2) warm-up cycling (5-min at the pedaling load corresponded to a 2.0% of body weight at 80 revolutions per minute of pedaling frequency, including maximal pedaling for 3 s at 2 and 4 min), followed by 3) the five sets of the sprint interval leg cycling exercise in hypoxia. In the normobaric hypoxic recovery condition, participants inspired hypoxic gas vias tube attached on a face mask, from a tent, while in the normobaric normoxic recovery condition, after the five sets of the sprint interval leg cycling exercise in hypoxia, the tube was taken out from the tent immediately, then, participants could inspire normobaric normoxic gas (room air). The load during pedaling corresponded to 7.5% of the individual’s body weight at the first set, and it gradually reduced from 6.5 to 5.5%, 4.5, and 3.5% of body weight in the second to third, fourth, and fifth sets, respectively. During each five set, participants performed leg cycling for 30 s. The applied load was set according to a previous study (Kasai et al., 2015). In the preliminary test at our laboratory, we decided that decremental settings of the load was an optimal protocol to allow the participants to produce the maximal mean power despite fatigue. A 4-min recovery at 0 W (unloaded cycling at 60 revolutions per minute) was inserted between the sets. After the five sets of leg cycling exercise, they rested for 30-min sitting recovery period: in 1) normobaric normoxia (room air; equivalent to 400 m), and 2) normobaric hypoxia [fraction of inspired oxygen (FiO₂) = 0.145; equivalent to 3,000 m]. These two tests were performed in a random order with a 1-week wash out period as shown in Figure 1.

A commercial hypoxic generator (Hypoxico Everest Summit II: Will Co., Ltd., Tokyo, Japan) supplied hypoxic gas based on a previous study (Horiuchi et al., 2017). Briefly, hypoxic gas was supplied into a custom-made tent (approximately 4,000 L) via one tube connected to the gas generator, and hypoxic gas was mixed in the tent. Participants inspired the hypoxic gas via another tube from the tent. They wore a two-way, non-rebreathing valve (for inspired and expired gas, respectively) and face mask (Horiuchi et al., 2017). Participants were blinded to inspire normoxic or hypoxic gas during recovery. The inspired oxygen concentration was verified before and after each experiment (AE-310; Minato Medical Science, Osaka, Japan).

The last 2 min of a 10-min resting hypoxic exposure was used to measure hypoxic baseline values before exercise. Blood pressure (BP) and blood lactate concentration (La) were measured at baseline, and at 5, 10, 20, and 30 min during recovery. BP was measured using the oscillometric method (HEM-907, OMRON, Tokyo, Japan). Resting BP was measured twice, and the average of the two measurements was considered as the participant’s BP value. We also confirmed the difference in the systolic BP or diastolic BP was <5 mmHg compared to the values of one before measurement at rest (Horiuchi et al., 2018). During the recovery period, BP was measured at 5, 10, 20, and 30 min of the period. Fingertip blood samples (0.3 μL) were taken to measure La (Lactate Pro 2LT-1730; Arkray, Tokyo, Japan). Heart rate (HR) and peripheral arterial oxygen saturation (SpO₂) were continuously monitored throughout the experiment using a wireless HR monitor (RS 800CX, Polar Electro, Japan) and pulse oximeter (TM-256, A&D, Tokyo, Japan), respectively. The probe of pulse oximeter was attached on the left middle finger. Local tissue oxygenation profiles of the vastus lateralis muscle were also continuously measured at 1-s intervals throughout the experiment using NIRS device (BOML1TRW, Omegawave, Tokyo, Japan) (Horiuchi et al., 2017). This instrument uses three laser diodes (780, 810, and 830 nm) and calculates the relative levels of oxygenated and deoxygenated hemoglobin (Hb) in the tissue according to the modified Beer–Lambert law. The oxygen saturation of the skeletal muscle (StO₂) was calculated by dividing oxygenated Hb by the total Hb, and was expressed as percentage (Horiuchi et al., 2017). NIRS optodes were placed on the lower third of the vastus lateralis muscle. The probe holder contained one light source probe, and two detectors were placed 2.0 cm (detector 1) and 4 cm (detector 2) away from the source at the vastus lateralis. Hb concentrations received by detector 1 were subtracted from those received by detector 2. This procedure minimized the influence of skin blood flow. Since NIRS signals can reach half the depth of the distance between the probe and detector, it can traverse at a depth of 10–20 mm at the target muscle.

The mean arterial pressure (MAP) was calculated using the following equation (Horiuchi et al., 2018): MAP = ( systoloc BP  [ SBP ] − distolic BP  [ DBP ] ) 3 + DBP

The values of HR and SpO₂ were adopted at baseline, and at 5, 10, 20, and 30 min of recovery for further analysis. These values were averaged during the last 1 min at the sitting baseline and during the 10 s immediately after each recovery period.

Baseline values of StO₂ were determined by the average value during the last 5 min prior to the warm-up exercise. The minimum StO₂ values were defined as the lowest StO₂ values during the five sets of the sprint interval leg cycling exercise. The hyperemic phase (i.e., the time from baseline values to the end of 30 min of recovery) was analyzed for the peak StO₂ (%) and the area under the hyperemic curve (AUC; % s). The peak StO₂ value was defined as the highest StO₂ value achieved during the hyperemic phase. The AUC was calculated as the total area under the reperfusion curve above the baseline values (Soares et al., 2018; Horiuchi and Okita, 2020). The mean power was calculated for each set of sprint cycling, and the mean values of five sets was identified as exercise performance.

Data are presented mean ± standard deviation. Statistical analyses were performed using the R programming language (RKWard, Version 0.7.2). A paired t-test was performed to compare the sprint interval leg cycling exercise performance and NIRS indices between the two conditions. A two-way (time × recovery condition) repeated measures analysis of variance (ANOVA) was used to compare time course changes in other physiological variables (SBP, DBP, MAP, HR, La, and SpO₂). Effect size was calculated as Cohen’s d and η², defined as a trivial (<0.2), small (0.2), moderate (0.5), or large (0.8) for paired t-test “Cohen’s d” (Hopkins et al., 2009) and as small (η ² = 0.01), medium (η ² = 0.06), and large (η ² = 0.14) for ANOVA “η ²” effects (Lakens, 2013). Statistical significance was set at p < 0.05.

First, only young men were enrolled in this study, which makes the results inapplicable to other populations, such as women and elite sport athletes. Second, the present study did not conduct normoxic exercise, followed by normoxic recovery as a control condition. However, our main aim was to compare between normoxic and hypoxic recovery after sprint interval leg cycling exercise in hypoxia. Although future studies are needed, our results may have following future outlooks. Finally, we evaluated only StO₂ among the indicators of NIRS technique. Venous occlusion technique has been suggested to estimate O₂ supply and consumption in working muscles during exercise (Homma et al., 1996). Similarly, a recent study demonstrated that combination of arterial and venous occlusion provides valid estimates of blood flow and oxygen consumption at rest and during exercise (Dennis et al., 2021). Although we must acknowledge that these techniques are useful and that we should have performed this technique to obtain more insight interpretations, we considered that further arterial occlusion may not be required after exhaustion. This is because, in the preliminary test, most of the participants complained of intolerable pain and nausea during arterial occlusion after exercise. Thus, we considered a safety for participants preferentially in the present study.

Sprint interval training in hypoxia has been considered to be a useful strategy to improve exercise both at sea level and high-altitude (Kon et al., 2015). Similarly, it was reported that repeated-sprint training, defined to consist of the repetition of short (<30 s) “all-out” sprints with incomplete recoveries (<60 s) (Girard et al., 2020), in hypoxia significantly improved cardiorespiratory fitness level (Camacho-Cardenosa et al., 2020). Additionally, even though under normoxic environments, a recent review summarized that sprint interval training improves glucose metabolism (Jiménez-Maldonado et al., 2020). Although these previous findings are beyond of the present study, studies including various combination of hypoxic (normoxic) exercise and recovery conditions may be required. As the present study is the first study to compare the effects of a normoxic versus hypoxic recovery on post-exercise hemodynamics, the present results may be informative for populations who perform an exercise at high-altitude, particularly from the viewpoint of safe implementation for conducting sprint interval training in hypoxia. For example, while the average magnitude of PEH in the present study was approximately 6 mmHg, regardless of different recovery conditions, previous studies reported PEH magnitudes of approximately 3 mmHg after 45 min of a high-intensity running exercise (85% of maximal HR) (Kleinnibbelink et al., 2020), approximately 6 mmHg after a 2-h leg cycling exercise (50% of peak oxygen uptake) (Horiuchi et al., 2016), and more than 10 mmHg after a resistance exercise (bilateral leg squat) (Horiuchi et al., 2018). These different magnitudes of PEH may be influenced by different study settings. Since these previous studies (Horiuchi et al., 2016; Horiuchi et al., 2018; Kleinnibbelink et al., 2020) and our study did not evaluate symptoms leading to presyncope, it is still uncertain whether sprint interval exercise in hypoxia increases the risk of presyncope. Nevertheless, the current findings suggest that the risks of presyncope may be similar with either normoxic or hypoxic recovery. Moreover, a recent study demonstrated that an acute exercise-induced reduction in BP was related to a greater chronic reduction in the BP-lowering effect of exercise training (Kleinnibbelink et al., 2020). Therefore, the greater magnitude of PEH in the present study, compared with the previous study (Kleinnibbelink et al., 2020), might indicate that regular sprint interval training in hypoxia could be a future strategy for antihypertensive therapy.