Changes in Running Performance After Four Weeks of Interval Hypoxic Training in Australian Footballers

A Single-Blind Placebo-Controlled Study

Abstract

McLean, BD, Tofari, PJ, Gore, CJ, and Kemp, JG. Changes in running performance after four weeks of interval hypoxic training in Australian footballers: A single-blind placebo-controlled study. J Strength Cond Res 29(11): 3206–3215, 2015—There is a paucity of data examining the impact of high-intensity interval hypoxic training (IHT) on intermittent running performance. This study assessed the effects of IHT on 17 amateur Australian Footballers, who completed 8 interval treadmill running sessions (IHT [F_IO₂ = 15.1%] or PLACEBO) over 4 weeks, in addition to normoxic football (2 per week) and resistance (2 per week) training sessions. To match relative training intensity, absolute IHT intensity reduced by 6% of normoxic vV[Combining Dot Above]O₂peak compared with PLACEBO. Before and after the intervention, performance was assessed by Yo-Yo intermittent recovery test level 2 (Yo-Yo IR2) and a self-paced team sport running protocol. Standardized effect size statistics were calculated using Cohen's d to compare between the interventions. Compared with PLACEBO, IHT subjects experienced (a) smaller improvements in Yo-Yo IR2 performance (Cohen's d = −0.42 [−0.82 to −0.02; 90% confidence interval]); (b) similar increases in high-intensity running distance during the team sport protocol (d = 0.17 [−0.50 to 0.84]); and (c) greater improvements in total distance (d = 0.72 [0.33–1.10]) and distance covered during low-intensity activity (d = 0.59 [−0.07 to 1.11]) during the team sport protocol. The lower absolute training intensity of IHT may explain the smaller improvements in Yo-Yo IR2 performance in the hypoxic group. Conversely, the data from the self-paced protocol suggest that IHT may positively influence pacing strategies in team sport athletes. In conclusion, IHT alters pacing strategies in team sport athletes (i.e., increased distance covered during low-intensity activity). However, IHT leads to smaller improvements in externally paced high-intensity intermittent running performance (i.e., Yo-Yo IR2), which may be related to a reduced absolute training intensity during IHT sessions.

Introduction

In recent years, the development of techniques that create a normobaric hypoxic environment has led to interest in performance benefits that may be induced by intermittent training bouts in hypoxic conditions, while living in a normoxic environment (Live Low, Train High [LLTH]). Much of the current literature examining the performance effects after LLTH interventions relates to endurance performance (²⁵). However, many of the physiological adaptations proposed following LLTH protocols (e.g., upregulation of monocarboxylate transporter-4 and carbonic anhydrase III, increased perfusion of fast twitch muscle fibers; for a review of mechanisms, see Faiss et al. (¹³)) are related to anaerobic metabolism (¹⁴), suggesting that this type of training may be beneficial for performance with a high anaerobic demand.

A number of studies have found improvements in high-intensity intermittent exercise after LLTH interventions in both endurance (¹⁴) and team sport athletes (¹⁶). However, improvements in high-intensity exercise performance are not universal (^26,29–31), and it seems that factors relating to the specificity and intensity of hypoxic training, as well as the volume and intensity of associated normoxic exercise (during LLTH interventions), are important in mediating performance outcomes (²⁵). Indeed, a wide range of training intensities exist within the LLTH literature (^25,28), ranging from continuous exercise (∼30–90 minutes) at ∼50% maximal aerobic work rate (¹¹) to repeated maximal sprints lasting 6–10 seconds (^14,16). The differences in these training intensities are proposed as key factors in determining the impact of the LLTH intervention (²⁷).

Recently, a number of authors have reported beneficial effects of repeated sprint training in hypoxia (RSH) on high-intensity intermittent exercise performance in both cyclists (¹⁴) and team sport athletes (^7,16). For example, Galvin et al. (¹⁶) reported greater improvements in Yo-Yo Intermittent Recovery level 1 (Yo-Yo IR1) performance after 4 weeks of repeated sprint training (10 × 6 seconds sprints with 30-second passive recovery) in hypoxia (F_IO₂ = 13.0%) in well-trained rugby players. With improvements in high-intensity intermittent exercise capacity likely important for team sport athletes, due to the intermittent nature of match demands (¹), RSH interventions seem promising for improvements in team sport–specific exercise capacity (^14,16). However, the implementation of running-based RSH is often not practically feasible. This is because, ideally, RSH requires hypoxic training rooms that are equipped with nonmotorized treadmills or the use of hypoxic rooms or tents (¹⁸) that have sufficient space to complete overground repeated sprint or agility training (⁷). In contrast, interval hypoxic training (IHT) programs that can be implemented on motorized treadmills may be more accessible for running-based team sport athletes. To our knowledge, there are no studies that have examined directly the effect of IHT on high-intensity intermittent exercise capacity (²⁵). The influence of training interventions on self-paced activity is also important for team sport athletes, due to the self-paced nature of these sports (¹²). Therefore, the aims of this study were to assess the effects of 4 weeks of IHT on (a) externally paced high-intensity intermittent running performance and (b) self-paced performance during a team sport–specific running protocol. We hypothesized that IHT would lead to greater improvements in high-intensity running performance and in distance covered during the self-paced, team sport protocol.

Methods

Experimental Approach to the Problem

To assess the impact of IHT on intermittent running performance, a group of Australian Football players completed a 4-week training program, during which subjects completed 2 IHT (or PLACEBO), 2 resistance training, and 2 football training sessions (see Figure 1 for details). During the intervention period, all training variables were matched between groups, except for the IHT/PLACEBO treadmill running sessions (see below for details). Subjects' running performance was measured before and after the 4-week intervention by (a) an incremental treadmill run to exhaustion, (b) Yo-Yo intermittent recovery test level 2 (Yo-Yo IR2), and (c) an intermittent team sport running protocol on a nonmotorized treadmill (see below for details). The combination of football, resistance, and running sessions was chosen in an attempt to replicate the typical preseason training routine of Australian Football players. All IHT, PLACEBO, and resistance training sessions were prescribed and supervised by the investigators, whereas football training sessions were prescribed by team coaches. All training was monitored using the session rating of perceived exertion (RPE) method (¹⁵), which calculates a total load (arbitrary units) by multiplying the session RPE (Borg's category ratio 10 scale (⁵)) by the session duration, which is a valid method to quantify training loads in Australian Footballers (³²). Additional measurements of training load and intensity were used during running sessions (i.e., running volumes [global positioning system [GPS] or treadmill data]; mean session heart rate [HR]; see below for details).

Figure 1: Timeline of data collection and training. TSR = team sport running protocol; Famil. = familiarization; V[Combining Dot Above]O₂peak = treadmill run to exhaustion for assessment of peak oxygen consumption per minute; Yo-Yo = Yo-Yo intermittent recovery test, level 2; IHT/PLACEBO + resistance = supervised training session including running (under hypoxic or placebo conditions) and resistance training; football = football training session, as prescribed by team coaches.

Subjects

Twenty-one amateur Australian Footballers (23.2 ± 2.8 years; age range 18–30 years; 184.1 ± 8.1 cm; 80.8 ± 8.6 kg) were recruited to participate in this randomized controlled trial. Subjects were required to be currently training for Australian Football at least 3 times per week. Subjects were randomly assigned to an IHT (n = 11) or PLACEBO (n = 10) training group. Four subjects (IHT: n = 2; PLACEBO: n = 2) withdrew from the study due to injury. All training and testing took place during the Australian Football preseason period in February and March, following a 4-week period of preparatory training. This preparatory training involved subjects completing 2 football/running and 1–4 resistance (depending on individual history) training sessions per week; this training was unsupervised and compliance was self-reported by the subjects. All subjects provided written informed consent to participate in this study, which was approved by the Human Research Ethics Committee at Australian Catholic University. No subjects were under the age of 18.

Procedures

Blinding Procedure

To minimize placebo and nocebo effects, all subjects were informed that they were training at a simulated altitude of 3,000 m at the beginning of the study. During IHT and PLACEBO training sessions, ducted air conditioning was continuously running for both groups to control temperature and create airflow through the training room. Previous investigations have recommended simulated training altitudes of 2,000–3,000 m during LLTH interventions, so as not to significantly blunt absolute training quality and to maintain any additional stimulus produced by the hypoxic environment (²⁰). Therefore, during IHT sessions only, target oxygen concentration was set to 14.8% (simulated altitude of 3,000 m) and was continuously monitored using 2 wall-mounted oxygen sensors (SmarTox-I; The Canary Company, Lane Cove, NSW, Australia). At the conclusion of the posttesting, subjects were informed that 1 group actually trained “near sea-level” and at this time were asked if they believed they were training near sea-level or at “altitude”; 44% IHT and 50% PLACEBO subjects believed that they were training at altitude. All investigators were blinded to the allocation of groups, except for the principal investigator, who set the oxygen concentration of the training room and calculated individual training speeds based on percentage of the peak velocity achieved during a test of peak oxygen consumption (vV[Combining Dot Above]O₂peak; Table 1).

Table 1: Training distance, sRPE, and training HRR.*†

Hypoxic or Placebo Training

During the 4-week training block, subjects completed 8 IHT or PLACEBO training sessions. All training sessions were ∼30 minutes of duration and consisted of 30-second to 3-minute intervals of running on a motorized treadmill (Cybex 750T, Cybex, Medway, MA, USA). All subjects trained in a room (dimensions: 11.8 × 12.4 m) connected to 2 hypoxic generators (Altitude Training Systems, Lidcombe, NSW, Australia). Training intensity was set between 90 and 110% of normoxic vV[Combining Dot Above]O₂peak (see below for calculation) for normoxic subjects. Based on previous work (⁸) and pilot testing with the current training protocols, training intensity for the IHT group was reduced by 6% normoxic compared with the PLACEBO group; that is, training intensities between 84 and 104% normoxic vV[Combining Dot Above]O₂peak were used for IHT, to match relative training intensity between groups. During pilot testing, this percentage reduction in training intensity (6%) was determined as providing the highest training intensity that could be completed in hypoxia (14.8% O₂) when prescribing intervals between 90 and 110% of vV[Combining Dot Above]O₂peak, without changes in interval or rest times. In addition to session RPE measurements, training intensity during the final 2 weeks of the intervention was monitored using a team-based HR monitoring system (Polar Team; Pursuit Performance, Adelaide, Australia). Mean session HR was then calculated using commercially available software (Polar Precision Performance 4.03; Pursuit Performance, Adelaide, Australia), and a percentage of heart rate reserve (HRR (²³)) was calculated using the following formula: (HR_train − HR_rest)/(HR_max − HR_rest), where HR_train = average HR during training, HR_rest = resting HR and HR_max = maximal HR. Maximal HR was recorded as the peak value achieved during pre- and post-incremental treadmill test runs to exhaustion or Yo-Yo IR2 tests. Because of logistical challenges, full data sets for training HR were only available for 12 subjects (IHT: n = 8; PLACEBO: n = 4). If in a training session, subjects voluntarily dismounted the treadmill during an interval (due to fatigue), they were provided a 10-second rest before recommencing the interval to completion.

Resistance Training

During the intervention period, all subjects participated in 2 resistance training sessions per week supervised by the researchers. Resistance training sessions were ∼60 minutes of duration and comprised 4–5 compound exercises of the upper limbs (e.g., barbell bench press, barbell bench pull, dumbbell press, dumbbell row) and 2–3 compound exercises of the lower limbs (e.g., leg press, dumbbell step-ups). These programs followed a linear progression over the 4 weeks from moderate volume/moderate intensity (e.g., 4 sets, 10 repetitions each exercise) to low-moderate volume/moderate-high-intensity (e.g., 4 sets, 6 repetitions each exercise).

Football Training

Two football training sessions per week (∼90 minutes duration) were prescribed by team coaches. All 17 subjects completed 7 football sessions throughout the 4-week intervention, totaling 119 football sessions, of which 92 were monitored using global positioning system (GPS) technology (Catapult MinimaxX [n = 11] or Catapult MinimaxX S4 [n = 6] units; Catapult Innovations, Melbourne, Australia). Total distance, as well as distance below (“low-intensity running”) and above (“high-intensity running”) 84% vV[Combining Dot Above]O₂peak, was calculated. The threshold of 84% vV[Combining Dot Above]O₂peak for low- and high-intensity running was chosen because all running completed during IHT or PLACEBO sessions was performed ≥84% vV[Combining Dot Above]O₂peak. Therefore, this threshold allowed for direct comparisons between the volume of high-intensity running during football training and during IHT (or PLA) sessions.

Peak Oxygen Consumption (V[Combining Dot Above]O₂peak)

During the second visit to the laboratory, subjects completed an incremental run to exhaustion on a treadmill (Pulsar 3p; HP Cosmos, Nussdorf-Traunstein, Germany) while monitored through open-circuit spirometry (TrueOne 2400; Parvo Medics, Utah, USA) for assessment of V[Combining Dot Above]O₂peak. A run to exhaustion test was used as described previously (⁸), which involved initial speed being set to 10 km·h⁻¹ with a grade of 0%. Thereafter, speed was increased by 1 km·h⁻¹ every minute until volitional exhaustion. If the last stage was not fully completed, the peak treadmill speed was calculated using the following formula: peak treadmill speed = Sf + (t/60 × 0.5), where Sf was the last completed speed in kilometers per hours and t is the time in seconds of the uncompleted stage (^8,22). After completing the run to exhaustion, subjects rested for ∼10 minutes before completing an initial familiarization of the 30-minute team sport running protocol on a nonmotorized treadmill.

Yo-Yo Intermittent Recovery Level 2 Test

At the beginning and the end of the 4-week intervention (Figure 1), all subjects completed the Yo-Yo IR2 test (³), as a measure of high-intensity intermittent running performance. Briefly, the Yo-Yo IR2 requires subjects to complete 2 × 20 m shuttles at increasing speeds (controlled by audio signals), interspersed with a 10-second active recovery period (³). Subjects continued to complete shuttles until volitional exhaustion, or until they were unable to maintain the required shuttle speed. Subjects were given 1 warning if they did not complete the shuttle in the required time, and the test was terminated if they did not complete any subsequent shuttles in the required time. The total distance covered by an individual at the termination of the test was recorded as the final result. The Yo-Yo IR test has shown high reliability in intermittent team sport athletes (²⁴).

Team Sport Running Protocol

Before completing the team sport running protocol, subjects underwent a standardized warm-up, which involved 3 minutes of submaximal treadmill running at a self-selected speed, followed by a sequence of dynamic stretches of the major muscle groups of the lower limbs, and a 3-minute portion of the team sport protocol, which included 1 submaximal sprint. Subjects then rested for ∼5 minutes and their body mass was obtained (PW-200KGL; A&D Weighing, Kensington, Australia) wearing shorts only, before commencing the 30-minute team sport running protocol.

A previously reported team sport running protocol was then used to assess self-paced, sport-specific running performance (³³). Briefly, this protocol involves subjects completing 30 minutes of self-paced, intermittent running on a nonmotorized treadmill (Woodway Curve 3.0., Woodway, USA). The protocol uses visual and audible commands to direct a participant's movement category (i.e., “Stand Still,” “Walk,” “Jog,” “Run,” “Sprint”). Before commencing the protocol, subjects were asked to follow visual and audible commands (as above) and instructed that during “run” periods, they should be completing a “hard run, as if pushing to the next contest within a match,” and during “sprint” periods to “sprint maximally.” Standardized verbal encouragement was provided by the investigator during the sprint periods. No other encouragement, verbal or visual feedback was provided (i.e., subjects had no knowledge of speeds/distances or time elapsed). The team sport–specific running protocol was designed to achieve mean running velocities above the Australian Football game mean (∼125 m·min⁻¹ (³⁴)) with the goal of creating significant physiological stress representing a similar running intensity that may be experienced during a 30-minute “worst case scenario” period of an Australian Football game (^6,34). In the 2 weeks before the study, subjects underwent 2 familiarization sessions on the nonmotorized treadmill, where 20 minutes and 15 minutes of this protocol were completed, respectively. This protocol has previously been shown to be reliable after 1 familiarization session (³³). In this study, the primary outcome variable measured during the team sport protocol was total distance, which included distance covered during different movement categories (i.e., Stand Still, Walk, Jog, Run, Sprint).

Statistical Analyses

A contemporary analytical approach involving magnitude-based inferences (²¹) was used to detect small effects of practical importance. All data were log transformed for analysis to account for nonuniformity of error. The differences within and between groups for the changes from pre- to post-intervention were assessed with dependent and independent t-tests for unequal variance (²¹). The magnitudes of change ([INCREMENT]) were assessed in relation to the smallest worthwhile change (SWC), where a small effect size (0.2 × between-subjects SD at pre intervention for both groups pooled) was used for all variables. For all performance variables, baseline measurements were used as a covariate in the analysis of changes, to account for different responses that may result from initial exercise capacity. Changes in the IHT group ([INCREMENT]IHT) were termed greater, similar, or smaller than changes in the PLACEBO group ([INCREMENT]PLACEBO). These differences were assigned a qualitative descriptor according to the likelihood of the difference exceeding the SWC as follows: 0–49%, trivial; 50–74%, possible; 75–94%, likely; 95–99%, very likely; and >99%, almost certainly (²¹). Effects where the 90% confidence interval overlapped the positive and the negative thresholds simultaneously were deemed unclear. Standardized effect size statistics were calculated using Cohen's d to allow for comparisons across different running performance measures, which exhibit different levels of normal variability. Effects of 0.2, 0.5, and >0.8 were considered small, moderate and large, respectively (¹⁰).

Results

Training Load and Intensity

Total training distances are displayed in Figure 2A. There were unclear differences in overall or high-intensity (≥84% vV[Combining Dot Above]O₂peak) weekly distance covered between the 2 groups during football and during IHT or PLACEBO training sessions. Likewise, there were unclear differences in weekly IHT/PLACEBO, resistance, football, or overall session-RPE training load between groups (Figure 2B). The session RPE was similar between groups for each of the 8 treadmill based running sessions; that is, all differences between groups were unclear. Mean training HRR was likely higher in the PLACEBO group during sessions 3.1 (Cohen's d ± 90% confidence interval = 0.69 ± 0.91), 3.2 (1.93 ± 1.16), and 4.1 (0.67 ± 1.01), as well as possibly higher in the PLACEBO group during session 4.2 (d = 0.46 ± 0.62).

Figure 2: A) Weekly running volumes, and (B) session rating of perceived exertion (sRPE) training load, during the 4-week intervention period. Values are mean and error bars are SD. IHT = interval hypoxic training group, PLACEBO = placebo training group, vV[Combining Dot Above]O₂peak = peak velocity achieved during a V[Combining Dot Above]O₂peak test, AU = arbitrary units.

Running Performance

V[Combining Dot Above]O₂peak

There were unclear differences between IHT and PLACEBO in V[Combining Dot Above]O₂peak at baseline. Improvements in V[Combining Dot Above]O₂peak and running performance after the training intervention are summarized in Table 2. All subjects experienced a possible increase in V[Combining Dot Above]O₂peak (d = 0.26 ± 0.17), with unclear differences between groups. Time to exhaustion was possibly greater in IHT (617 ± 91 seconds) vs. PLACEBO (581 ± 64 seconds) at baseline. Time to exhaustion during the V[Combining Dot Above]O₂peak test almost certainly increased (d = 0.56 ± 0.14), with trivial differences between groups.

Table 2: Summary of running performance outcomes.*†

Yo-Yo Intermittent Recovery Test Level 2

There were unclear differences between IHT and PLACEBO in Yo-Yo IR2 performance at baseline. All subjects almost certainly experienced increases in Yo-Yo IR2 performance (d = 0.88 ± 0.23), but these increases were likely smaller (d = −0.42 ± 0.40) in the IHT group than in the PLACEBO group.

Team Sport Running Protocol

At baseline, walking, jogging, running, sprinting, and total distance covered during the team sport running protocol were all likely greater in IHT (total = 4,565 ± 496 m) vs. PLACEBO (total = 4,258 ± 254 m). Pooled data for both groups showed a likely increase in total distance covered during the team sport protocol (d = 0.47 ± 0.28), with these increases very likely greater in the IHT group than the PLACEBO group (d = 0.72 ± 0.39). For both groups combined, there was a very likely increase in sprint (d = 0.57 ± 0.31) and run (d = 0.63 ± 0.30) distance and a likely increase in jog (d = 0.42 ± 0.29) distance. IHT subjects experienced a likely increase in walking distance (d = 0.72 ± 0.35), while the corresponding changes where unclear in the PLACEBO group (d = −0.22 ± 0.41).

Discussion

To our knowledge, this single-blinded, randomized controlled trial is the first to assess the effects of IHT on high-intensity intermittent, and self-paced, exercise performance. The main findings are that 4 weeks of IHT in Australian Footballers resulted in (a) smaller improvements in externally paced high-intensity running (Yo-Yo IR2) performance compared with training in normoxia; (b) IHT and PLACEBO subjects exhibited similar increases in high-intensity running distance during a self-paced team sport protocol; and (c) IHT subjects exhibited greater improvements in the self-paced protocol for total distance and distance covered during low-intensity activity.

Our findings with respect to Yo-Yo IR2 performance are in contrast to the results of Galvin et al. (¹⁶) who found greater Yo-Yo IR1 improvements in their hypoxic group after 4 weeks of RSH training with well-trained rugby players. Several differences in study design may explain these opposing findings, perhaps most importantly the training intensity used in each of the studies (repeated 6-second sprints in Galvin et al. (¹⁶) vs. 30-second to 3-minute intervals in this study). Differences in the initial running capacity of the groups in each study may also play a part, with the hypoxic group in the study of Galvin et al. (¹⁶) starting with a poorer high-intensity running capacity than the normoxic group (Yo-Yo IR1 distance: 1,237 ± 265 and 1,374 ± 361 m, respectively). This suggests that the hypoxic group may have had greater scope to improve performance after their training intervention, an idea supported by the similar Yo-Yo IR1 results at completion of the intervention (Yo-Yo IR1 distance 1,621 ± 364 m and 1,594 ± 379 m for the hypoxic and normoxic groups, respectively). Gatterer et al. (¹⁷) reported similar increases in Yo-Yo IR1 performance between hypoxic and normoxic subjects, after 5 weeks of shuttle run sprint training in a group of elite youth soccer players. However, the hypoxic and normoxic groups in this study also had substantially different Yo-Yo IR1 results preintervention (hypoxic ∼300 m lower than normoxic), which may have influenced the magnitude of changes in performance after the training intervention (¹⁷). In this study, there were unclear differences in baseline Yo-Yo IR2 performance, and we also controlled for any differences by using these baseline data as a covariate in statistical analysis. The hypoxic subjects in this study ended with substantially lower Yo-Yo IR2 results than those in the placebo group (1,262 ± 351 m and 1,373 ± 182 m, respectively). These differences may be related to the reduced absolute training intensity (≈6% vV[Combining Dot Above]O₂peak) of the hypoxic training sessions, which made up the majority of the high-intensity running volume (9–10 km per week) during the intervention. Given that IHT subjects completed only 2–3 km per week of high-intensity running in normoxia (during football sessions), this may have limited some of the training adaptations stimulated by high-intensity (absolute) interval training. A greater portion of high-intensity training in normoxia may overcome this limitation in future interventions and is an important consideration for practitioners when implementing IHT protocols (²⁵).

The novel combination of externally and self-paced performance tests in this study highlights the importance of test specificity when assessing the influence of LLTH, and other training interventions. Yo-Yo intermittent recovery test level 2 is often used as a measure of team sport–specific running performance (³); however, this measure is externally paced and, therefore, does not detect changes in pacing strategies that may be adopted, which is important given the self-paced nature of team sport competition (¹). Within the current investigation, Yo-Yo IR2 results would suggest that the hypoxic intervention blunted some of the training-induced improvements in team sport high-intensity intermittent running performance. In direct contrast, the IHT group had greater increases in total distance covered in the team sport running protocol, with no differences between groups in the amount of high-intensity activity (i.e., running and sprinting) performed. Thus, IHT subjects covered extra distance during low-intensity periods, which may support the theory that team sport athletes regulate low-intensity activity throughout a match to maintain high-intensity activity (²). Although the reasons for the change in pacing strategies between the IHT and PLACEBO groups are unclear, these changes may be related to central regulation. Indeed, cortical voluntary activation has been shown to parallel changes in cerebral oxygen delivery (¹⁹), and hypoxic training interventions can lead to enhanced cerebral oxygen delivery during high-intensity intermittent exercise (¹⁶). Therefore, training in hypoxia presents an intervention that may influence central fatigue during exercise, which in turn could impact on self-pacing strategies. It should be acknowledged that cerebral oxygenation was not measured in this study and, therefore, our contention that our hypoxic intervention has altered oxygen delivery to the brain and influenced self-pacing is speculative. However, this is the first study to report changes in self-pacing during high-intensity intermittent exercise after an LLTH intervention.

Pooled data from both groups in this study show a small increase in V[Combining Dot Above]O₂peak after the training intervention. However, there were trivial differences between ΔIHT and ΔPLACEBO, which suggest that the hypoxic stimulus had little additional effect. Training in hypoxia is known to limit absolute exercise intensity (⁸), which invariably leads to a reduction in cardiovascular load (training HR and V[Combining Dot Above]O₂). This was evident in our study, where hypoxic subjects were prescribed velocities at 6% less vV[Combining Dot Above]O₂peak than normoxic subjects, which resulted in likely lower training heart rates (Table 1). Despite this apparent reduction in cardiovascular load, both groups experienced similar gains in V[Combining Dot Above]O₂peak. Therefore, it may be that a 4-week training intervention was too short to elicit differences between IHT and PLACEBO subjects in cardiovascular adaptations. Alternatively, reductions in central overload may have been offset by an increase in peripheral adaptation in hypoxic subjects; indeed, training in hypoxia has been shown to augment a range of peripheral mRNA transcripts related to glycolytic potential and mitochondrial biogenesis and metabolism (³⁵). Regardless of the underlying mechanism, both groups experienced similar increases in V[Combining Dot Above]O₂peak during our 4-week intervention, which suggests that any differences in performance are not related to changes in aerobic capacity.

The placebo and nocebo effects are always an important consideration in intervention studies, and logistical challenges often make these difficult to control in hypoxic training studies (^4,25). In this study, all subjects were informed that they were training in hypoxia, and training sessions were set-up to promote this belief. Indeed, all subjects trained in a hypoxic training room with air conditioning running (thereby creating similar air flow in both the hypoxic and normoxic conditions), and the alterations in absolute training intensity (of which the subjects were not informed) led to similar session ratings of perceived exertion between groups. These data indicated that we adequately controlled for placebo/nocebo effects given that, when asked, ∼50% of subjects from both groups believed that their training was performed in hypoxia.

A number of limitations must be considered when interpreting results from the current study. This study was conducted with a group of sub-elite amateur Australian Footballers and, therefore, their responses may differ from responses of highly trained professional football athletes. However, comparison with published data for professional Australian Footballers (⁹) shows that the athletes in the current study displayed similar Yo-Yo IR2 running performance at baseline as professional athletes who had completed a 3-week training camp in the heat (current study ≈1,013 m, professional ≈1,024 m). Despite this comparatively elite level of team sport running performance before the intervention, both groups experienced large (d = 0.88) increases in Yo-Yo IR2 performance after the training intervention, suggesting that elite players may also benefit from similar protocols.

Matching prescription of training intensity between hypoxic and normoxic groups is difficult when implementing submaximal training protocols (i.e., any training intervention that is not a maximal sprint protocol). Although it is well established that training in hypoxia will limit absolute exercise intensity (⁸), there is currently no consensus on the magnitude of such limitations (²⁵). In this study, we used the recommendations of Buchheit et al. (⁸) whereby interval training intensity is reduced by 6% vV[Combining Dot Above]O₂peak in hypoxia, in an attempt to match relative training intensity between groups. This method resulted in similar perceived exertion between groups but substantially higher training HR in the normoxic group, reflecting the greater absolute training intensity compared with the IHT group.

The prescription of training intensity based on vV[Combining Dot Above]O₂peak also has a number of limitations, which are apparent in the current study. During training sessions 1.1, 1.2, and 3.2, approximately, half of the subjects required ≥one 10-second rest period to complete the exercise protocol. Similar numbers of subjects required 10-second rest periods in both the IHT and PLACEBO groups, which suggest this prescription issue was not caused by the hypoxic stimulus. Although these results suggest that the training velocities prescribed were too high, ∼50% of subjects were still able to complete the stipulated protocol in these 3 sessions (i.e., sessions 1.1, 1.2, and 3.2). Therefore, we propose that the sole use of %vV[Combining Dot Above]O₂peak is insufficient for accurate prescription of high-intensity interval training in team sport athletes. Some combination of aerobic power measurement and anaerobic speed reserve/maximal speed may be more efficacious for training prescription in these athletes.

We recently highlighted the importance of sufficient volume and intensity of normoxic training during IHT interventions (²⁵) and, in the current study, we attempted to provide this training stimulus through “football” training sessions, as prescribed by team coaches. However, in retrospect, this resulted in only small amounts of high-intensity running (Figure 2A) in normoxia. Thus, the normoxic training for the IHT group may not have been of sufficient volume and intensity to maximize adaptations related to high-intensity (absolute) interval training. Future IHT interventions should consider the inclusion of structured high-intensity normoxic training sessions to optimally overload all physiological systems that are important in team sport running performance.

Practical Applications

The main findings of this study are that 4 weeks of IHT led to greater improvements in self-paced performance during a team sport running protocol, but smaller improvements in externally paced intermittent running performance (i.e., Yo-Yo IR2), when compared with matched training in normoxia, where an attempt was made to match relative training intensity. Thus, we propose that hypoxic training positively influences pacing strategies in team sport athletes and, therefore, improves movement strategies adopted in team sport competition. Differences in the volume of high-intensity training in normoxia may explain the smaller improvements in externally paced running performance in the hypoxic group, and we recommend that practitioners and researchers include greater volumes of targeted normoxic training in all groups undertaking hypoxic training.

Acknowledgments

The authors acknowledge the generous participation of all athletes involved in this research study. The authors acknowledge the contribution of Collingwood Football Club and their staff involved in making this project run smoothly. The authors specially thank Craig Stewart for his assistance with this project. This article has been read and approved by all the listed coauthors and the authors declare no potential conflicts of interest.