Hypoxia Does Not Impair Resistance Exercise Performance or Amplify Post-Exercise Fatigue

ABSTRACT Purpose: To determine whether performing resistance exercise in hypoxia acutely reduces performance and increases markers of fatigue, and whether these responses are exaggerated if exercising at high versus low work rates (i.e., exercising to failure or volume matched non-failure). Methods: Following a within-subject design, 20 men completed two trials in hypoxia (13% oxygen) and two in normoxia (21% oxygen). The first session for hypoxic and normoxic conditions comprised six sets of bench press and shoulder press to failure (high work rate), while subsequent sessions involved the same volume distributed over 12 sets (low work rate). Physical performance (concentric velocity) and perceptual responses were measured during exercise and for 72 hr post-exercise. Neuromuscular performance (bench throw velocity) was assessed pre- and post-session. Results: Hypoxia did not affect physical performance, neuromuscular performance, and perceptual recovery when exercising at high or low work rates. Higher work rate exercise caused greater acute decrements in physical performance and post-exercise neuromuscular performance and increased perceived exertion and muscle soreness (p ≤ 0.006), irrespective of hypoxia. Conclusions: Hypoxia does not impact on resistance exercise performance or increase markers of physical and perceptual fatigue. Higher exercise work rates may impair physical performance, and exaggerate fatigue compared to low work rate exercise, irrespective of environmental condition. Practitioners can prescribe hypoxic resistance exercise without compromising physical performance or inducing greater levels of fatigue. For athletes who are required to train with high frequency, decreasing exercise work rate may reduce post-exercise markers of fatigue for the same training volume.

Normobaric hypoxia (reduced inspired fraction of oxygen [F i O 2 ]) during resistance training can further increase muscle strength and hypertrophy compared to normoxia (Manimmanakorn et al., 2013;Nishimura et al., 2010).These adaptations likely occur due to a cascade of physiological events, prompted by an increased reliance on anaerobic metabolism and subsequent accumulation of metabolites (Scott, Slattery, & Dascombe, 2015), which may also cause heightened muscle fiber recruitment for a given submaximal workload (Scott et al., 2017).However, increased physiological stress associated with resistance exercise in hypoxia could also hinder physical performance (i.e., repetition velocity), exacerbate muscle damage, or extend recovery requirements.
Currently, research investigating the impact of hypoxia on resistance exercise performance, neuromuscular performance, and muscle recovery have separately examined submaximal resistance exercise (i.e., low work rate or non-failure exercise) or maximal-load circuit training (i.e., high work rate or exercising to failure).Importantly, changing the exercise structure when exercising in hypoxia may alter the physical performance and neuromuscular performance responses to training (Ramos-Campo, Rubio-Arias, Dufour, et al., 2017;Ramos-Campo, Rubio-Arias, Freitas, et al., 2017;Scott, Slattery, Sculley, et al., 2015;Scott et al., 2018).Therefore, this study aimed to investigate the influence of hypoxia during both high work rate and low work rate resistance exercise on concentric repetition velocity, post-exercise neuromuscular performance, and perceptual responses.Furthermore, we aimed to investigate whether hypoxia would extend the recovery requirements and magnitude of post-exercise muscle soreness compared to normoxia following both high and low work rate sessions.We hypothesized that high work rate resistance exercise performed in hypoxia would induce greater decrements in physical and post-exercise neuromuscular performance than in normoxia, whereas an additional hypoxic stimulus would have a limited impact on these responses during and after low work rate exercise.We further hypothesized that hypoxic exposure would increase perceived exertion and delay the recovery of muscle soreness associated with high, but not low, work rate sessions.

Study design
This study used a single-blind crossover study design.Participants visited the laboratory on five separate occasions; one familiarization visit was followed by four experimental sessions.While this study was completed as part of a larger project investigating the acute physiological responses during hypoxic resistance exercise with different work rates (Walden et al., 2020), the primary exercise-related outcome measures (neuromuscular and perceptual strain post-exercise) of the current study do not overlap with previous analyses.During experimental sessions, participants undertook resistance exercise with both free-weight seated barbell shoulder press and bench press exercises.All sessions were conducted inside an environmental chamber connected to hypoxic generators Training Systems Hypoxic Inc.,Bickenbach,GER) and comprised two experimental sessions in normoxia (F i O 2 = 0.21) and two in normobaric hypoxia (F i O 2 = 0.13).Both sessions in the same environmental condition were performed consecutively, and the order of conditions was randomized and counterbalanced between participants.In each environmental condition, the initial session included sets performed to failure (six sets, three of each exercise, 1 min rest between sets), and the second session included submaximal sets (same loads as failure session but separated over 12 sets, 6 of each exercise, 1 min rest).The loads lifted for each set were reduced across the session based on pilot testing, to allow participants to complete ~10 repetitions in each set (failure session: set 1 = 100%, set 2 = 80%, and set 3 = 60% 10RM; non-failure session: sets 1-2 = 100%, sets 3-4 = 80%, and sets 5-6 = 60% 10RM) (Walden et al., 2020).This design enabled volume-matching between sessions within each environmental condition to compare high (i.e., exercise to failure) versus low (i.e., same volume as for failure session but separated over twice as many sets) work rate exercise.Exercise performance, neuromuscular performance, perceived exertion, and recovery scores were collected during each session, while perceived fatigue and muscle soreness ratings were obtained up to 72 hr post-sessions.

Participants
Twenty healthy, physically active young men (age: 23.2 ± 1.9 yr, height: 178.6 ± 8.0 cm, body mass: 77.9 ± 11.1 kg, bench press 10RM: 64.0 ± 14.9 kg, barbell shoulder press 10RM: 48.6 ± 10.1 kg) with a minimum two years of resistance training experience volunteered for this study.Participants were excluded if they did not demonstrate a proficient lifting technique, were taking medications/supplements that could have affected exercise performance, had a history of altitude sickness or altitude exposure (>2000 m) in previous six months, or had medical conditions that could have exacerbated exposure to hypoxia.They were also encouraged to avoid any strenuous exercise during the study period, and to replicate their dietary and fluid intake for 24 hr before all testing sessions.Participants were provided with information detailing the purpose and requirements of the research and provided signed informed consent.This study was approved by the Institutional Human Research Ethics Committee.

Testing procedures
Participants first completed the pre-screening questionnaire and signed informed consent, and then undertook a 10RM test for the seated barbell shoulder press and bench press exercises using established testing protocols (Reynolds et al., 2006).Five attempts were given for both the shoulder press followed by the bench press to gather their 10RM, with 3 min rest between attempts.Prior to the 10RM testing, a standardized warm-up was performed, comprising light intensity cycling for 5 min, followed by 5 min upper body dynamic stretching and mobility exercises.
Prior to the commencement of each session protocol, participants performed a standardized 10-min warm-up (Table 1).Following the warm-up, participants performed baseline assessments of neuromuscular performance, comprising 3 × 1 repetition of Smith machine bench throws at each of three loads (40%, 70%, and 100% 10RM), each separated by 1 min.Participants then entered the environmental chamber, where they rested for 10 min.Warm-up sets of 10 reps (50% 10RM) followed by 5 reps (75% 10RM) of the seated barbell shoulder press were performed with 60 s of rest between sets.Subsequently, participants were given 2 min of rest prior to performing the protocol sets for shoulder press (60 s of rest between sets).After resting for 5 min, the same warm-up routine and working sets for the bench press exercise.Once the exercise protocol was completed, participants were afforded 5 min seated rest before leaving the chamber, which totaled ~40 min in the chamber for each session.Once participants left the chamber, they immediately completed the bench throw tasks again and provided their post-exercise perceptual response measures.Participants were also required to complete a visual analog scale (VAS) for muscle soreness and a well-being questionnaire at 24, 48, and 72 hr following trials (McLean et al., 2010).

Measures
For all repetitions during each exercise set, concentric bar displacement and time between data points were recorded by a linear position transducer (GymAware, PowerTool; Kinetic Performance Technology, Canberra, Australia) sampling at 50 Hz.Mean concentric velocity values were used across each set to track physical performance.To thoroughly examine the effects of hypoxia and work rate on physical performance, each repetition was classified as either slow, moderate, or fast velocity.This was achieved by first determining the fastest and slowest repetition across all conditions (i.e., not separately for each condition) for each relative load and exercise, specific to each participant.All repetitions for each of the three loads and two exercises were then allocated into specific velocity categories: slow (lower 33% of the velocity range), moderate (middle 33%), and fast (upper 33%).The number of repetitions completed within each velocity category was subsequently used to compare physical performance between the experimental sessions.This approach was taken to ensure that the velocity measures were individualized for each participant and did not simply reflect how heavy the load was (i.e., slower velocities would be expected with heavier loads).
To quantify the impact of exercise on neuromuscular performance following each protocol, participants performed a progressive-load bench throw task (using 40% [light], 70% [moderate], 100% [heavy] of 10RM loads) before and immediately following each experimental protocol (Morán-Navarro et al., 2017).This was completed on a counterbalanced Smith machine (Body-Solid Smith machine GSSM350, ACT, Australia) and comprised three single-repetition sets of a bench throw at each of three increasing loads (a total of nine repetitions).The participants demonstrated reliable performance in this task (coefficient of variation <5%) for peak velocity at each load during the familiarization session.The highest peak velocity at each load was measured via a linear position transducer and used for subsequent pre-to postexercise analysis.
We obtained RPE scores immediately following each set via Borg CR10 scales (Borg et al., 1987).During the low work rate sessions, set-RPE was averaged for both sets at each load to allow comparisons between high and low work rate sessions.A modified Borg's CR10 scale was used to assess recovery between sets on a scale from 0 ("very poorly recovered") to 10 ("very well recovered"), which was obtained 15 s prior to the commencement of each working set (Laurent et al., 2011), and also averaged across sets at each load as described for set-RPE data.
Participants rated their physical fatigue and pectoralis major muscle soreness by marking a 100-mm VAS from 0 ("no fatigue/ soreness") to 100 ("maximum fatigue/soreness") (Scott et al., 2018).To anchor the VAS scores, participants performed three push-ups followed by recording each score via the numeric scale.Physical fatigue VAS was recorded prior to and at 0, 20, and 40 min post-exercise.Muscle soreness VAS was recorded prior to and at 24, 48, and 72 hr post-exercise.In addition, participants completed a well-being questionnaire (McLean et al., 2012), which quantifies exercise-related factors (fatigue and general muscle soreness) and lifestyle factors (sleep quality, stress, and mood) on a 5-point scale.Well-being was determined by summing the scores from the five categories, recorded prior to, and 24, 48, and 72 hr post-exercise.

Statistical analyses
All data were tested using a Shapiro-Wilk test and were normally distributed.The main analysis used a 3-way repeatedmeasures ANOVA (hypoxia × work rate × repetition velocity category, time or set) to identify differences between the factors we manipulated (i.e., hypoxia and work rate) across different measurement classifications or time points (i.e., repetition velocity category, time or set) for physical performance and perceptual measures.Bonferroni post hoc analyses was used if a significant main effect or interaction effect was observed.Additionally, for the bench throw task, absolute change in peak velocity for each load were analyzed between sessions using 2-way repeated-measures ANOVA (hypoxia × work rate).All statistical analyses were completed using SPSS statistical software version 25.0 (IBM Corp., Somers, NY, USA), with the level of statistical significance set at p < .05.Effect sizes for within-subject comparisons are presented as partial eta squared (η 2 ).Data are expressed as mean ± SD.

Physical performance
For both exercises, there was no 3-way (p ≥ .659),hypoxia × work rate (p ≥ .087),or hypoxia × velocity category (p ≥ .203)interaction.A significant work rate × velocity category interaction (p < .001,η 2 = 0.878 to 0.918; Figure 1) was observed.Post hoc analysis indicated that fewer fast repetitions were performed in high versus low work rate sessions (p < .001),while more slow and moderate velocity repetitions were observed in high versus low work rate sessions (p < .001).

Neuromuscular performance
There was no significant hypoxia × work rate interaction for any load of the bench-throw task (p ≥ .329),or main effects of hypoxia (p ≥ .252).A significant main effect of work rate was observed for absolute change in peak velocity (pre-to postexercise) for each load of the bench-throw task (p < .001,η 2 = 0.724 to 0.832) (Table 2).Post hoc analysis indicated that for all three loads, change in peak velocity was significantly larger for high work rate sessions than for the low work rate sessions (p < .001).

Perceptual response measures
Physical fatigue and perceived soreness responses are shown in Figure 2.For perceived physical fatigue, there was no 3-way (p = .509),hypoxia × work rate (p=0.655), or hypoxia × time (p = .311)interaction, but a significant interaction was observed for work rate × time (p < .001,η 2 = 0.598).Post hoc analyses confirmed that all post-exercise values were increased from pre-exercise (p < .001)and that all post-exercise values for high work rate sessions were significantly greater than for low work rate sessions (p < .001).For muscle soreness, there was no 3-way (p = .684),hypoxia × work rate (p = .434),or hypoxia × time (p = .157)interaction.A significant interaction effect for work rate and time was observed (p < .001,η 2 = 0.410), with significantly higher muscle soreness values for high than for low work rate sessions at all post-exercise time points (p ≤ .006).Post hoc analysis also indicated muscle soreness was significantly higher at 24 and 48 hr post-exercise compared to both pre-exercise and 72 hr post across both high and low work rates (p ≤ .004),except for between pre-exercise and 48 hr in low work rate sessions (p = .563).Additional perceptual responses to exercise are summarized in Table 3.For set-RPE across both exercises, there was no 3-way (p ≥ .628)or hypoxia × work rate (p ≥ .508)interaction.There was also no hypoxia × set interaction for the shoulder press (p = .689).A significant work rate × set interaction effect was observed for set-RPE in both exercises (p < .001,η 2 = 0.778 to 0.796).Post hoc analysis indicated that all set-RPE values were significantly higher for high work rate compared to low work rate sessions for both exercises (p < .001).A significant hypoxia × set interaction (p = .014,η 2 = 0.203) was observed for the bench press, with post hoc analysis indicating differences between all sets across both hypoxia and normoxia (p ≤ .003).For mean perceived recovery scores in both exercises, there was no 3-way (p ≥ .773)or hypoxia × work rate (p ≥ .189),or hypoxia × set (p ≥ .645)interaction.A significant interaction effect was observed for work rate and set (p < .001,η 2 = 0.727 to 0.730).For the shoulder press, post hoc analysis indicated no between-work rate differences for set 1 (p = .383),while lower ratings of recovery were observed for the high compared to low work rate sessions for sets 2 and 3 (p ≤ .002).For the bench press, lower recovery values were observed across all sets for high compared with low work rate sessions (p < .001).For wellbeing scores, there was no 3-way (p = .679)or hypoxia × work rate (p = .123),hypoxia × time (p = .820),or work rate × time (p = .244)interaction.The main effect was observed for both work rate (p = .002,η 2 = 0.397) and time (p < .001,η 2 = 0.423) without interaction.Post hoc analysis indicated significantly higher well-being scores for low work rate compared to high work rate sessions (p = .002).Additionally, well-being scores were lower at 24 hr post-exercise compared to pre-exercise (p = .016),48 hr (p = .009),and 72 hr (p < .001)post exercise.Well-being scores were also higher at 72 hr than 48 hr postexercise (p = .001).
The physiological responses to each exercise session from this study are described elsewhere (Walden et al., 2020).Importantly, the hypoxic dose was confirmed via assessment of blood oxygen saturation, which was significantly lower during sets of exercise in hypoxic (87.4-89.8%)compared with normoxic (97.3-98.2%)sessions.

Discussion
This study is the first to investigate whether manipulating the work rate of resistance exercise in hypoxia alters physical performance, neuromuscular performance, and perceptual responses compared with normoxic exercise.Our main findings indicate 1) hypoxia had no effect on within-session physical performance, post-exercise neuromuscular performance, and perceptual responses, irrespective of the exercise work rate, 2) high work rate resistance exercise resulted in more repetitions in the slow velocity category and greater decrements in post-exercise neuromuscular performance, compared with low work rate exercise, and 3) regardless of hypoxia, low compared to high work rate resistance exercise caused reduced levels of perceived exertion and soreness, in conjunction with better overall recovery scores up to 48 hr post-exercise.Practically, these findings suggest that hypoxia can be added to resistance training without exacerbating fatigue.Despite both high and low work rate protocols being matched for volume load, sessional exercise work rate has a larger impact than normobaric hypoxia (F i O 2 = 0.13) on physical performance, post-exercise fatigue, and perceptual responses.
In agreement with previous research (Scott, Slattery, Sculley, et al., 2015;Scott et al., 2018), our results demonstrate that the addition of moderate normobaric hypoxia does not compromise physical performance during high or low work rate resistance exercise.Contrastingly, Ramos-Campo et al. have reported significantly lower force and power output during circuit training in hypoxia (F i O 2 = 0.13) compared to exercising in normoxia (Ramos-Campo, Rubio-Arias, Dufour, et al., 2017) or less-severe hypoxia (F i O 2 = 0.16) (Ramos-Campo, Rubio-Arias, Freitas, et al., 2017).The disparity between our findings and Ramos-Campo, Rubio-Arias, Dufour, et al. (2017) and Ramos-Campo, Rubio-Arias, Freitas, et al. ( 2017) could be explained by differences in exercise methods used.Ramos-Campo, Rubio-Arias, Dufour, et al. ( 2017) and Ramos-Campo, Rubio-Arias, Freitas, et al. ( 2017) implemented maximal load full-body circuit training with very brief rest (<35 s between exercises) that caused an increased accumulation of metabolic by-products (Ramos-Campo, Rubio-Arias, Dufour, et al., 2017), which could impair excitation-contraction coupling (Green et al., 2011;Kelleher et al., 2010) and reduce contractile velocity (Green et al., 2011).However, the more traditional upper-body multijoint exercise modality prescribed in the current study may have allowed sufficient rest time between sets (60 s) for metabolic by-products to clear into systemic circulation (Krustrup et al., 2006;Walden et al., 2020).Hence, this could explain why no differences in physical performance between hypoxia and normoxia were observed here, regardless of work rate.
In line with acute exercise performance results, hypoxia had no impact on post-exercise measures of neuromuscular function.These findings also confer with previous research, which reported no differences in post-exercise neuromuscular performance for up to 48 hr after submaximal resistance exercise in moderate hypoxia (F i O 2 = 0.16) compared with normoxia (Scott et al., 2018).These results suggest that the addition of hypoxia during resistance exercise does not exacerbate post-exercise fatigue, possibly for similar reasons to those explained above regarding physical performance.Nevertheless, we observed that high work rate resistance exercise caused larger magnitudes of neuromuscular fatigue up to 48 hr when compared with low work rate exercise.Performing resistance exercise at a higher work rate has caused larger velocity losses post-exercise (Pareja-Blanco et al., 2017), which has also been associated with prolonged muscle recovery (Morán-Navarro et al., 2017).Lower work rate resistance exercise may be preferential for athletes who wish to avoid prolonged decrements in neuromuscular function, regardless of whether training in hypoxia or not.This may be particularly true for those who are required to train twice daily and/or on consecutive days, provided that the stimulus is still sufficient to elicit beneficial training responses.
Irrespective of the environmental condition, high work rate exercise caused greater perceived exertion and muscle soreness than volume-matched low work rate exercise.These findings contrast with Ramos-Campo, Rubio-Arias, Freitas, et al., (2017) who demonstrated that hypoxic circuit training exaggerates RPE scores than normoxia.These divergent results in perceived exertion could be explained by previously highlighted differences in energy demands and physiological responses associated with circuit training compared to traditional resistance training (Alcaraz et al., 2008;Ramos-Campo, Rubio-Arias, Freitas, et al., 2017;Skidmore et al., 2012).Circuit training is associated with a greater oxygen demand when compared to traditional resistance exercise (Alcaraz et al., 2008;Camacho et al., 2018;Skidmore et al., 2012); therefore, perceived exertion during high-load circuit training in hypoxia is likely to be greater due to reduced oxygen availability and increased cardiovascular stress (Camacho et al., 2018;Scott, Slattery, & Dascombe, 2015).
Although this study provides the first direct comparison of different resistance exercise structures performed in hypoxia versus normoxia, some limitations should be acknowledged.We implemented a standardized hypoxic stimulus (i.e., same absolute hypoxic condition; F i O 2 = 0.13), as opposed to eliciting a consistent decrease in arterial oxygen saturation, for all participants.Future research could investigate hypoxic resistance exercise by individualizing the hypoxic dose; however, this approach is challenging for practitioners to implement (Soo et al., 2020), and most real-world applications of hypoxic training prescribe the stimulus as we have here.In addition, the findings of this research are applicable to the cohort studied (healthy young men) and the size of our recruited sample, and it is possible that dissimilar results would be observed for a larger sample of clinical or untrained populations exercising in hypoxia.

Practical implications
This study highlights that performing resistance exercise with the addition of hypoxia does not alter physical performance, neuromuscular performance, or perceived exerciserelated responses.Practitioners who wish to implement hypoxic resistance exercise into their training programs can do so without compromising the physical performance or inducing greater levels of fatigue, irrespective of work rate.Nevertheless, performing repetitions to failure was substantially more demanding than not, with increases in postexercise neuromuscular fatigue and muscle soreness for up to 72 hr compared with non-failure exercise.Therefore, when recovery is paramount and training volume is to be maintained (i.e., with high session frequencies or around competition times), prescribing training programs with lower work rate sessions may reduce post-exercise neuromuscular fatigue and muscle soreness, as long as the stimulus is not so weak that it does not elicit the desired training response.Importantly, exercise work rate may be more influential than total training volume of a session on these responses, as both high and low work rate sessions were matched for volume load in this research.

Disclosure statement
No potential conflict of interest was reported by the author(s).

Front
of the resistance band with each hand, keeping arms straight out in front of the chest then pull the band apart, keeping arms straight and parallel to the ground 10 × 2 sets 90° External rotation with resistance band Position arm to the side parallel to the ground with the elbow at 90°, while holding the resistance band, then externally rotate the shoulder and arm 10 reps × each arm Passovers with resistance band Hold either end of the resistance band with each hand keeping arms straight out in front of the chest then raise the band overhead while holding the band 10 × 2 sets Y-Press scapula with resistance band Hold either end of the resistance band with elbows bent and tucked into obliques (both hands in-line with shoulders) then press hands over-head until arms fully extended 10 × 2 sets Medicine ball chest throw While standing 3 m from a wall, press the medicine ball explosive off the chest into the wall 5 × 2 sets Practice bench throw task (30% of 10RM) Performed with 30 s rest between warm-up sets 3 × 2 sets Reps = repetitions, min = minutes, each arm = exercise performed using each arm individually, RB = resistance band, m = meter, 10RM = 10 repetition maximum.

Figure 1 .
Figure 1.Repetitions performed within each concentric velocity category across experimental sessions for the shoulder press (a) and bench press (b) exercises.a Significantly different to high work rate sessions within the same velocity category.HYP high = hypoxia and high work rate session, NORM high = normoxia and high work rate session, HYP low = hypoxia and low workrate session, NORM low = normoxia and low work rate session.

Figure 2 .
Figure 2. Perceived levels of physical fatigue (a) and muscle soreness (b) before and after experimental sessions measured using VAS. a Significantly different to high work rate sessions at same time point, b significantly different to all post-exercise assessments, c significantly different 24 hr assessment, d significantly different to 48 hr assessment.HYP high = hypoxia and high work rate session, NORM high = normoxia and high work rate session, HYP low = hypoxia and low workrate session, NORM low = normoxia and low work rate session.
to low work rate sessions.a Significantly different to 24-hr post-exercise.b Significantly different to 48-hr post-exercise.c Significantly different to other sets for same exercise in hypoxia and normoxia.NOTE: Set-RPE and perceived recovery scores for low work rate sessions were averaged across both sets using the same load to allow for comparisons with high work rate sessions.HYP high = hypoxia and high work rate session, NORM high = normoxia and high work rate session, HYP low = hypoxia and low work rate session, NORM low = normoxia and low work rate session, RPE = rating of perceived exertion.

Table 1 .
Standardized warm-up performed prior to experimental sessions.

Table 2 .
Peak concentric velocity (m • s −1 ) across the three loads of the progressive-load bench throw task at pre-exercise and post-exercise time points.Significantly different to absolute change in low work-rate sessions at same load.HYP high = hypoxia and high work-rate session, NORM high = normoxia and high work-rate session, HYP low = hypoxia and low work-rate session, NORM low = normoxia and low work-rate session, RM = repetition maximum. a

Table 3 .
Mean set-rating of perceived exertion (RPE), mean set-perceived recovery, and overall well-being scores recorded during and after experimental sessions.