Thirty endurance-trained male participants, regularly involved in competitive running, cycling, and/or triathlon events (4–5 times per week: ~10–15 h) in the past 5 years, and volunteered to take part in this study. Exclusion criteria were as: (i) smoking or other chronic issues; (ii) using medication or dietary supplements in the past 3 months, or during the duration of the study; and (iii) being exposed to altitude (≥2,000 m) in the past 3 months. The Ethics Committee of the University of Trás-os-Montes and Alto Douro approved this study (Reference 14A/CE/2017) and all participants provided written informed consent after being informed of all experimental procedures and possible risks associated with the experiments. All procedures complied with the ethics code of the Declaration of Helsinki.

This randomized, single-blind, placebo-controlled, and independent group study was carried out in a laboratory at sea level (Porto’s Exercise Medical Center, Portugal), which allows controlling for humidity, temperature, altitude, and % O₂ availability. This chamber reduced the fraction of O₂ in the inspired air by diluting it with extra nitrogen (normobaric hypoxic system: b-Cat®). The oxygenation level in the chamber was controlled by a specific O₂ sensor placed inside the chamber and remained stable in each training session/test. All participants had previously been involved with research in our laboratory and were familiar with the equipment and testing protocols implemented, thus familiarization sessions were not required. Consequently, V ˙ O₂max values were already known, and participants were randomly assigned to one of three experimental groups (HNO, HPL, and CON) with similar distributions for V ˙ O₂max, age, body mass, and height (55.1 ± 5.4, 57.3 ± 8.3, and 51.9 ± 7.9 ml·kg⁻¹·min⁻¹; 36.1 ± 4.6, 37.4 ± 5.7, and 34.9 ± 8.4 years old; 71.0 ± 7.7, 70.2 ± 7.3, and 73.2 ± 9.5 kg; and 175.7 ± 7.1, 173.3 ± 6.8, and 175.3 ± 6.9 cm, for HNO, HPL, and CON groups, respectively). Participants performed 12 high-intensity interval-training (HIIT) sessions during a 4-week period (three sessions/week, 48 h in-between each session), in one of three experimental groups: (i) HNO: high-intensity exercise training sessions in normobaric hypoxia (F_iO₂ = ~13%, ~3,000 m) with NO₃⁻ supplement; (ii) HPL: high-intensity exercise training sessions in normobaric hypoxia (F_iO₂ = ~13%, ~3,000 m) with placebo supplement, and (iii) CON: high-intensity exercise training sessions in normoxia (F_iO₂ = 20.9%) with placebo supplement.

Participants were instructed to maintain their normal daily activity throughout the 4-week period, and only an extra training session peer week was allowed (except for week 4). Also, they were advised to maintain their normal diet although they received a list of foods rich in NO₃⁻ to be avoided (e.g., leafy green vegetables, beetroot, and processed meats) to isolate supplemented NO₃⁻ as a cause of any potential beneficial effect. Moreover, participants were asked to abstain from the use of any chewing gum and antibacterial mouthwash products, as these have been shown to inhibit the conversion of NO₃⁻ to NO₂⁻ by bacteria in the oral cavity. A food diary was used to ensure adherence to the intervention.

In the week before (pre-intervention) and after (post-intervention) the 4-week exposure, participants performed an incremental, continuous test to exhaustion (INC), an exercise transition from rest to severe intensity to exhaustion (T_lim), and a 3 min all-out test (3AOT), in a cycle ergometer (Lode Excalibur Sport, Groningen, Netherlands) in normobaric hypoxia (F_iO₂ = ~13%, ~3,000 m).

From the 12 HIIT sessions, in each week, two sessions of short aerobic intervals (HIT; first- and third-week’s sessions) and one session of repeated-sprint training (RST; second week’s session) were performed on the cycle ergometer. A warm-up of 10 min at 100 W + 4 min at 50% of power associated with V ˙ O₂max (p V ˙ O₂max) and a cool-down period of 15 min of active recovery at 100 W were completed in each session. The main sets for HIT were as follows: two sets of 6 × 1 min at 90%∆ (90% between V ˙ O₂max and the first ventilatory threshold), with 1 min active recovery between repetitions and 3 min between sets, both at 50% of p V ˙ O₂max (short intervals). The main sets were for RST: four sets of 6 × 10 s all-out, with 20 s active recovery and 3 min between sets, both at 50% of p V ˙ O₂max (repeated-sprint training). The number of repetitions was increased from 6 (first and second weeks) to 7 (third and fourth weeks) in both sessions. The training intensities used were relative to specific p V ˙ O₂max (assessed in hypoxia for HNO and HPL, and in normoxia for CON group). For the total duration of all training sessions (ranging from 52 to 57 min), participants always remained inside the chamber. For the oxygenation level in the environmental chamber to remain stable, no one was allowed in or out of the chamber during the training sessions.

Participants were instructed to keep their cadence between 70 and 90 rpm in HIIT’s short intervals, but in the RST sessions, the cadence was not fixed (isokinetic operation mode of the ergometer) and therefore, encouragement was given to perform at their maximal effort. In addition to the three supervised high-intensity training sessions, participants could perform one extra training session (that occurred on weekend days, in every case) below the moderate intensity domain (~60 min) in weeks 1–3 at normoxic conditions. No extra training session was allowed in week 4 to eliminate acute physiological effects.

Supplements were ingested 2.5–3 h prior to each session. NO₃⁻ was administered in the form of beetroot juice, containing 400 mg of a powdered standardized beetroot extract (containing 2% of NO₃⁻, ~8.4 mmol, according to the manufacturer label) dissolved in 150 ml of water (Sabeet®, Sabinsa Corporation). An equivalent volume of black currant juice, with negligible NO₃⁻ content, was served as a control drink. Both supplements were similar in appearance and taste. Supplementation was extended until post-intervention period (ingested 2.5–3 h prior each performance test).

In INC test, which was always conducted first, both V ˙ O₂max and p V ˙ O₂max were assessed. This test consisted of a 2-min step duration with 30 W increments between steps with a self-selected cadence between 70 and 90 rpm. Time sustained was determined through the T_lim test and performed at 80%∆ (i.e., first ventilatory threshold—VT₁, plus 80% of the difference between VT₁ and V ˙ O₂max). The test ended when the cadence could no longer be maintained within 10 rpm of the preferred cadence for >5 s. Critical power (CP) and the curvature constant (W′) were both assessed during the 3AOT test as previously described (Sousa et al., 2021). In this test, during the last 5 s of the baseline period, subjects were asked to increase their cadence to approximately 110 rpm. The braking resistance for the pedals to move was set (using the linear mode of the ergometer) so that the subjects would attain the power output halfway between V ˙ O₂max and the VT₁ (i.e., VT₁ + 50%∆) on reaching their preferred cadence (linear factor = power/cadence squared).

In both T_lim and 3AOT tests (performed randomly), a standard 5 min warm-up exercise at 50% of p V ˙ O₂max, followed by a 5-min passive rest, was performed. In all three tests, a standard 3-min period of unloaded baseline pedaling (8 W) at each participant’s preferred cadence, preceded each test. Verbal encouragement was given to motivate the participants to perform their best effort in the INC and 3AOT tests protocols and for as long as possible during the T_lim one. All tests were performed at room temperature (18–20°C) and 40–50% relative humidity (standardized and at the same time of the day (±2 h) for each participant to minimize an impact of the circadian variation on performance).

During each test, respiratory and pulmonary gas-exchange variables were continuously measured in a breath-by-breath mode (K5, Cosmed, Italy). The gas analyzer was calibrated before each test with gases of known concentration (16% O₂ and 5% CO₂) and the turbine volume transducer calibrated with a 3-L syringe, according to the manufacturer’s instructions. V ˙ O₂max (measured as the average of breaths over the last 30 s) was defined in one of the following cases: a plateau in V ˙ O₂ despite an increase in power, [La-] ≥ 8 mmol·L⁻¹, respiratory exchange ratio ≥1.0, HR >90% of [220—age], or a volitional exhaustion (controlled visually and case-by-case). p V ˙ O₂max was estimated as the power corresponding to the first stage of the INC test that elicited V ˙ O₂max. If a plateau of less than 2.1 ml·min⁻¹·kg⁻¹ could not be observed, the p V ˙ O₂max was calculated as previously described (Kuipers et al., 1985). VT₁ and VT₂ were determined as previously suggested (Sousa et al., 2021). The percentage of power, V ˙ O₂ and HR relative to p V ˙ O₂max, V ˙ O₂max, and HRmax associated to both VT₁ and VT₂ (VT₁-W, VT₁-O₂, VT₁-HR, VT₂-W, VT₂-O₂, and VT₂-HR) were considered for analysis.

During both T_lim and 3AOT tests, changes in local O₂ saturation (SmO₂), which represents the balance between O₂ delivery and extraction by the muscle, and in total hemoglobin (THb), an indicator of local blood volume, were assessed in the capillaries through a near-infrared spectroscopy (NIRS) monitor (Moxy monitor, Fortiori Design, Minnesota, United States). The Moxy monitor was positioned on the participant’s dominant leg vastus lateralis muscle, halfway between the greater trochanter and lateral epicondyle of the femur, perpendicular to the muscle fiber orientation. This muscle was chosen as it is part of the knee extensor group which is the primary contributor to force generation in the crank of the bicycle during down stroke of the pedal. Prior to placement, this area was trimmed with an electric razor and cleaned with alcohol swabs. Moxy was secured with a light shield and black athletic tape to block ambient near-infra red light from interfering with the detectors. Its exact position was recorded and replicated in all tests (pre-and post-intervention periods). As suggested, skinfold thickness at the site of the Moxy placement was measured using a skinfold caliper (Harpenden Ltd.) to ensure that the skinfold thickness was <1/2 the distance between the emitter and the detector (25 mm).

Capillary blood samples (5 μl) were collected from the earlobe for whole lactate concentration ([La⁻]) determination before exercise (at rest) and immediately at the end of each test, during the 1st, 3rd, 5th, and 7th minute of the recovery period until maximal ([La⁻]_max: for INC test) and peak values ([La⁻]_peak: T_lim and 3AOT tests) were reached (Lactate Pro2, Arkay, Inc., Kyoto, Japan).

As the gas analyzer did not measure directly F_iO₂, the corrected F_iO₂ (~13%) was inputted at the end of each test. V ˙ O₂ errant breaths (e.g., caused by swallowing, coughing, and signal interruptions) were omitted from the V ˙ O₂ analysis by including only those that were between V ˙ O₂ mean ± 4 SDs. After this process, the breath-by-breath data were used for V ˙ O₂ kinetic analysis in the T_lim test. The first 20 s of data after the onset of exercise (cardio-dynamic phase) was not considered for model analysis. To allow comparison of the V ˙ O₂ response, data were modeled using both a mono (Equation 1) and a double exponential approach (Equation 2) to isolate a possible V ˙ O₂ fast component response.

where VO₂(t) represents the relative VO₂ at the time t, A₀ is the VO₂at rest (ml·kg⁻¹·min⁻¹), and A₁ and A₂ (ml·kg⁻¹·min⁻¹) and τ₁ and τ₂ (s) are the amplitudes and time constants of the fast and slow VO₂ components, respectively. A₁ was used to determine the gain (G = A₁/power) of the primary component. To estimate VO₂ kinetics parameters, equations were fitted to the exercise data (GraphPad Prism, GraphPad Software, La Jolla, CA, United States) by minimizing the sum of the mean squares of the differences between modeled and measured VO₂ values.

Initially, the raw SmO₂ and [THb] data were treated using a smooth spline filter to reduce the noise created by movement and data presented every 2 s. Baseline SmO₂ (SmO_2base) and baseline [THb] were computed as a 30 s average while subjects performed the standard 3 min period of unloaded baseline pedaling (8 W) at their preferred cadence before the beginning of each test. Minimum SmO₂ (SmO_2min) was the lowest 6 s average reached during each test. Maximum SmO₂ (SmO_2max) and maximum [THb] were the highest 6 s average reached during each test with recovery phase included. Average SmO₂ from 30 to 120 s after the end of each test was used to assess recovery SmO₂ (SmO_2recovery). For each test, baseline SmO_2base and SmO_2min are expressed as % of SmO_2max (relative-SmO_2base and relative-SmO_2min, respectively). Change in SmO₂ (ΔSmO₂) and change in [THb] ([ΔTHb]) were calculated as the difference between relative-SmO_2min and relative-SmO_2base and, the difference between maximal and baseline [THb].

A total of 10 participants per group were considered for a type I error of 5%, a power of 80%, with statistical significance, and an average to be detected with probability population effect size of 0.5 (as it indicates a moderate to large difference; G*Power software version 3.1.9.2). A priori type of power analysis was used considering the student’s t-test as the statistical test performed. Shapiro–Wilk test confirmed the data normality and Levene’s test the variance homogeneity. Data are presented as mean ± SD. A two-way (group × time) ANOVA test was used to test main and interaction effects for the studied variables in performance (post-hoc comparisons when appropriate: Bonferroni test). Student’s t-tests were used to test: (i) performance differences at moderate altitude when high-intensity exercise training sessions were performed at moderate altitude, compared to the condition where they were performed at normoxia (HNO + HPL vs. CON) and (ii) performance differences at moderate altitude when chronic dietary NO₃⁻ supplementation was used together with high-intensity exercise training sessions performed at hypoxia, compared to the condition where placebo condition was used (HNO vs. HPL). Repeated measures T-Tests were used to test both pre- vs. post-performance differences in-between each intervention group. Pearson’s correlation coefficient (r) was employed to analyze the bivariate correlations between the different variables. Magnitude of the changes using effect sizes (ES) = (post-test mean − pretest mean)/(pretest standard deviation) was obtained for each test, and threshold values considered were “trivial” (<0.2), “small” (0.2–0.6), “moderate” (0.6–1.2), “large” (1.2–2.0), and “very large” (>2.0). Magnitudes of standardized effects (Cohen’s d calculated as the difference in means divided by the pooled SD) were determined against the following criteria: small, 0.2–0.5; moderate, 0.5–0.8, and large, >0.8. All statistical procedures were conducted with SPSS 24.0 and the significance level was set at 5%.

Figure 1 shows the effects of 4 weeks of training in the INC test’ related parameters for all experimental groups. Except for VT₁-W (time effect: p = 0.04) which was higher in hypoxia compared to normoxia (p = 0.03, d = 0.62), no other differences were found, inclusively between chronic dietary NO₃⁻ supplementation vs. placebo condition. From all the INC test-related parameters, only VT₁-W, V ˙ O₂max (ES: 1.14 and 0.65, respectively, in HNO group), VT₂-W (ES: 0.72 in HPL group), and V ˙ O₂max (0.77 in CON group) percentage differences had a moderate effect, with the remaining one being small (0.2–0.6) and trivial (<0.2). From pre- to post-intervention period, VT₁-W was significantly higher (p = 0.006, d = 0.48) and [Lac]_max significantly lower (p = 0.009, d = 0.46) in HNO group, and V ˙ O₂max significantly higher (p = 0.037, d = 0.94, r = 0.7) in CON group. No significant differences were found for HPL group.

Figure 2 shows the effects of 4 weeks of training in the T_lim and 3AOT-related parameters for all experimental groups. No differences were found between hypoxia vs. normoxia intervention and between chronic dietary NO₃⁻ supplementation vs. placebo condition in any of the tests. Only time sustained (ES: 0.74, in HPL) and W′ (ES: 0.67 and 0.74 in HNO and CON groups, respectively) differences had a moderate effect, with the remaining one being small (0.2–0.6) and trivial (<0.2). From pre- to post-intervention period, W′ was significantly lower (p = 0.04, d = 0.74, r = 0.86) in CON group, and no differences were found for both HNO and HPL groups.

Table 1 shows the mean values for the V ˙ O₂ kinetics’ parameters obtained during the T_lim test. In the pre-intervention period, V ˙ O₂ kinetics’ analysis showed that from the 30 participants, only nine individual responses (30%) were better modeled using a double exponential approach (A₂ = 13.03 ± 3.4 ml·kg⁻¹·min⁻¹: mean ± SD) with an overall R² = 0.85 ± 0.06. In the post-intervention period, only eight individual responses (~27%) were better modeled using a double exponential approach (A₂ = 12.66 ± 4.1 ml.kg⁻¹.min⁻¹: mean ± SD) with an overall R² = 0.85 ± 0.09. For comparison purposes, only common V ˙ O₂ kinetics’ parameters were presented in Table 1.

No differences were found between hypoxia vs. normoxia intervention and between chronic dietary NO₃⁻ supplementation vs. placebo condition. V ˙ O₂ kinetics’ parameters differences were all trivial (ES < 0.2) for every experimental group. From pre- to post-intervention period, no differences were found for any experimental group.

Table 2 shows the mean values for the oxygen saturation’ parameters obtained during the T_lim test in both pre- and post-intervention periods. Except for SmO_2recovery (group main effect: p = 0.001, d = 0.4), which was significantly higher in hypoxia compared with normoxia (p = 0.04, d = 0.82), no other differences were found, inclusively between chronic dietary NO₃⁻ supplementation vs. placebo condition. The 4 weeks of training induced an increase in both SmO_2min (p = 0.04) and Relative-SmO_2min (p = 0.05) in HNO group (ES = 0.72 and 0.67, respectively). Moreover, CON group presented significantly higher SmO_2base mean values in the post-intervention period (p = 0.03, ES = 1.05). No differences were found for the HPL group.

Table 3 shows the mean values for the oxygen saturation’ parameters obtained during the 3AOT in both pre- and post-intervention periods. Except for SmO_2recovery (group main effect: p = 0.008, d = 0.2; interaction effect: p = 0.02, d = 0.2) which was significantly higher in hypoxia compared with normoxia (p = 0.03, d = 1.43), no other differences were found, inclusively between chronic dietary NO₃⁻ supplementation vs. placebo condition. In HPL group, the 4 weeks of training induced an increase in SmO_2base (p = 0.001, ES = 1.5), Relative-SmO_2base (p = 0.05, ES = 1.14), and ΔSmO₂ (p = 0.013, ES = 1.32). Moreover, CON group presented significantly higher SmO_2base (p = 0.04, ES = 0.79) and Relative-SmO_2base (p = 0.016, ES = 1.11) mean values in the post-intervention period compared to baseline. No differences were found for the HNO group.

Although some improvements in anaerobic performance have been reported (Daniels and Oldridge, 1970; Gore et al., 1998; Friedmann et al., 2005; studies “uncontrolled”), it is extensively reported that no additional benefit on the aerobic performance of endurance athletes occurs when conducting interval training in a hypoxic environment, compared with the same training performed in normoxia (see Faiss et al., 2013a for a review). To overcome these limitations, it was recently proposed a new hypoxic training method: RST in hypoxia (RSH; Faiss et al., 2013b). As the training stimulus is maximal, allowing one to maintain high fast twitch fibers recruitment, additional positive results on performance could be expected when hypoxic stimulus is added. The consensus is that RSH leads to superior (1–5%) repeated-sprint ability in normoxic conditions, while larger increase in V ˙ O₂max was not observed (Brocherie et al., 2017).

In a protocol similar to ours (eight RSH sessions during a 4-week period: 3*10 s all-out with 20 s and 5 min recovery between repetition and sets, respectively), Faiss et al. (2013b) showed that RSH (~at 3,000 m) delayed fatigue during a repeated-sprint test to exhaustion. However, endurance performance (during a 3AOT) was not increased from pre- to post-intervention period in either group, neither hypoxic stimulus evidenced larger improvements compared to the same training performed in normoxia (for both 3AOT and Wingate test). Also, a 5-week intervention of RSH (30 s sprints interspersed by 4.5 min recovery intervals, 3 weekly sessions, 4–6 sprints per session), performed at ~2,750 m altitude, did not alter endurance ( V ˙ O₂max test and 30-min simulated time trial) or sprinting (30 s sprint test) performance to a higher extent compared to the same training stimulus performed in normoxia (De Smet et al., 2016). Even when exposure time was extended to 7 weeks (two sessions of high-intensity at 100 or 90% of p V ˙ O₂max), Roels et al. (2005) did not report a greater increase in performance (submaximal cycling test, 10 min cycle time trial and incremental test to exhaustion).

In line with most of the previous studies, we did not find differences in performance between hypoxic and control groups, except for VT₁-W in INC test (p = 0.09, d = 0.62). This was significantly higher when subjects were exposed to interval training in hypoxia compared to normoxia, achieving a moderate effect (ES = 1.14). In fact, only HNO group evidenced significantly higher VT₁-W (d = 0.48) and lower [Lac]_max mean values (d = 0.46) from pre- to post-intervention period. Moreover, in both T_lim and 3AOT, local O₂ saturation in the recovery period (SmO_2recovery) was significantly higher in hypoxia compared with the control group (d = 0.82 and 1.43, respectively). Therefore, the participants who were exposed to the hypoxic stimulus during the 4-weeks intervention were able to recover O₂ saturation faster after T_lim and 3AOT ended. Of interest, the addition of a single RSH session per week did not bring additional benefits and collectively, the combination of high-intensity sessions in hypoxia used in the present study (two HIT sessions and one RSH session) did not bring an additional aerobic and anaerobic performance enhancement compared to the same program performed in normoxia. However, an innovative and unexpected output of the present study is that this combination was also non effective for preparing athletes for simulated altitude performance.

There were also no differences in V ˙ O₂ kinetics’ parameters assessed in T_lim. It was previously reported that τ₁ and A₂ are highly plastic and decrease substantially with endurance exercise training in previously untrained individuals and that these adaptations are remarkably rapid (7–14 days; Womack et al., 1995). The authors proposed that the initial speeding (faster V ˙ O₂ kinetics) might be related to faster muscle blood flow kinetics or increased capillary-to-fiber ratio. In our study, at least for the HNO group, SmO_2min was significantly lower in T_lim at post-compared to pre-intervention period. Although acute exposure to simulated altitude induces a slower V ˙ O₂ kinetics (Sousa et al., 2021), there are no studies that analyzed the chronic impact of hypoxic exposure on altitude’ V ˙ O₂ kinetics performance.

The lack of performance differences found in our study may be due to several factors. It has been purported that intermittent exposure and/or training under hypoxic conditions may enhance exercise sea-level performance through increasing circulating erythropoietin levels and hemoglobin mass, upregulation of hypoxia-inducible factor 1α (causing structural changes within the muscle fiber), and ventilatory adaptation (Chapman et al., 2013). However, intermittent training in hypoxia is quite likely to have a minimal effect on erythropoiesis since a large “hypoxic dose” is required for significantly “stimulating the erythropoietic pathway” to the point that it enhances post-altitude sea-level endurance performance [Levine and Stray-Gundersen, 2006; e.g., ≥3,000 m for at least 3 h/day for 1–3 weeks (Katayama et al., 2003)]. Intermittent training in hypoxia has more potential to induce muscle skeletal and ventilatory adaptations (Faiss et al., 2013a), being suggested that an exposure altitude ≥4,000 m for at least 1.5 h/day repeated for 5–6 days is required to stimulate ventilatory acclimatization and concomitant improvements in performance (Muza, 2007). Moreover, greater buffering capacity, lactic acid tolerance, and/or O₂ extraction in the working muscle can be expected as well (Brocherie et al., 2017).

Collectively, it seems that the simulated altitude used in our study (~3,000 m), as well as the exposure time (~3 h per week for 4 weeks), may have been insufficient to promote the total ergogenic effect of training in hypoxia, either through hematological changes (hemoglobin mass) or via ventilatory acclimatization. It is also possible that the high-intensity exercise stimulus, which necessarily induces a higher acidosis, has blunted the erythropoiesis, making HIIT less effective, as previously suggested (Millet et al., 2010).

It is well documented that the one-electron reduction of NO₂⁻ to NO is enhanced in conditions of hypoxia (Kelly et al., 2014). Therefore, it would be expected that lowering skeletal muscle oxygenation would augment the potential for dietary NO₃⁻ supplementation to improve physiological responses during exercise and performance at simulated altitude. However, we did not find evidence that chronic dietary NO₃⁻ supplementation potentiates simulated altitude performance’ responses when high-intensity training sessions were held at moderate altitude, as no differences were found between HNO and HPL groups in any of the variables studied. This latter corroborates previous reports where acute NO₃⁻ supplementation has shown to be ineffective in enhancing cycling and running’ performance in hypoxia (Arnold et al., 2015; Bourdillon et al., 2015). However, significant results were also shown for single-leg knee extension high-intensity (Vanhatalo et al., 2011) and cycling exercise (Masschelein et al., 2012; Kelly et al., 2014). Some factors may potentially contribute to the lack of agreement in these data, namely, the supplementation period (Wylie et al., 2016), dosage used (Wylie et al., 2013), and training status of the participants (Wilkerson et al., 2012).

With a protocol of 6 weeks in duration (five training sessions/week, 5 × 30 min/week at 4–6 mmol/L blood lactate), Puype et al. (2015) concluded that dietary NO₃⁻ supplementation did not enhance the effects of intermittent hypoxic training on endurance exercise performance at sea level. However, the authors pointed out that higher daily doses of NO₃⁻ (in contrast with the 0.07 mmol NO₃⁻/kg body weight used) and administered as a bolus 2–3 h (and not 2.5–2 h) prior to the training sessions could result in significant training adaptations. With a protocol similar to ours (5 weeks: 4–6*30-s sprints interspersed by 4.5 min recovery intervals; three sessions/week), and a NO₃⁻dosage of 6.45 mmol administered 3 h before each session, De Smet et al. (2016) showed that performance at sea level increased in all experimental groups (with and without NO₃⁻ supplementation) to a similar extent. The only exception was that NO₃⁻ supplementation combined with training sessions performed in hypoxia increased the proportion of type IIa fibers in muscle. Taken together, it would appear that even when chronic NO₃⁻ supplementation is combined with HIIT training sessions performed in hypoxia, no effect on performance at sea level in healthy and/or moderately trained participant is observed. Our results extend these findings by suggesting that dietary NO₃⁻ does not have the potential to improve training adaptation even at simulated altitude conditions. Therefore, given the lack of consistent outcomes, at least for well-trained endurance athletes, we suggest that chronic dietary NO₃⁻ supplementation cannot be currently recommended to improve performance or training adaptation in hypoxic conditions, even when combined with HIIT training sessions, as previously suggested (Mujika et al., 2019).

One important feature of the present study has to do with an O₂ “sparing-effect” of oral NO₃⁻ supplementation, which was initially demonstrated by Larsen et al. (2007). In addition to liberating bioactive NO, NO₂⁻ acts as both a potent vasodilator in hypoxia and as an alternative electron acceptor, replacing O₂ in respiration (Bourdillon et al., 2015). Therefore, it would be expected the O₂ “sparing effect” of NO₃⁻ supplementation to be reflected in muscle tissue oxygenation in hypoxic conditions. In fact, our results show that from pre- to post-intervention period only the HNO group increased SmO_2min (from ~3 to 6%), at least in T_lim test, with the remaining experimental groups not evidencing this difference in any of the tests performed. These findings were previously observed by Masschelein et al. (2012) during steady-state cycling where an improvement in muscle tissue oxygenation index (and reduced HHb) during hypoxic exercise after 6 days of oral NO₃⁻ supplementation occurred. In this present study, as performance was not significantly different between HNO and HPL groups, we attribute the greater improvement in SmO_2min to a better matching of O₂ delivering to the metabolic requirements of the working muscles in T_lim test.

In our assay, we did no measure total baseline hemoglobin mass (Hb_mass) or both plasma [NO₂⁻] and [NO₃⁻]. Although baseline Hb_mass influence on the potential increase of total Hb_mass after exposure to a Living High-Training Low design is unclear and remains debated (Millet et al., 2019), this may have influenced the results of the present study. In recent years, several studies have assessed the acute and chronic effect of NO₃⁻ supplementation and reported increases in plasma [NO₃⁻] (~500%) and [NO₂⁻] (~40–140%; see Jones et al., 2018 for more details). Participants were instructed to maintain their normal diet, but they chose their own meals and were responsible for keeping their diets consistent. These could have influenced the effects of NO₃⁻ supplementation on the outcome variables. Moreover, the well-trained status of the participants recruited, and consequently, their higher plasma NO₂⁻ and NO₃⁻ concentrations, may have played an influential role (Wilkerson et al., 2012). In support of this, data from Bescos et al. (2012) showed that 7 out of 13 trained endurance athletes were non-responders, with only small elevations in plasma NO₃⁻ concentration observed following a 3-day supplementation protocol. Moreover, well-trained athletes experience less severe localized hypoxia and acidosis in the muscle compared with untrained populations at normoxia conditions (Wilkerson et al., 2012). However, when exposed to hypoxic environments, decreasing pulse oxygen saturation (SpO₂) is larger in endurance-trained athletes (compared to less trained counterparts) due to the prevalence of exercise-induced hypoxemia caused by a higher cardiac output (Mollard et al., 2007). Collectively, this seems to suggest that the training status of some of our participants may have influenced any NO₃⁻ supplementation effects or at least, expressed them differently. The fact that only male participants were part of this study reduces the applicability of the results found since our results are relevant to 50% of the population at large. Considering these latter, an important future direction for research is to investigate the moderating effect of training status and gender on the response to NO₃⁻ supplementation in hypoxic conditions. Moreover, the inclusion of another experimental group (normoxia + supplementation) could have helped in the interpretation of the results obtained. Other sources of inter-individual variability are likely to be influential. Consequently, the study design would have been improved by better controlling these factors, particularly for the allocation of the participants in the different groups.