A comprehensive review focusing specifically on altitude training for team-sports suggests that repeat-sprint training performed in hypoxia (RSH) may be a promising avenue for improving match-related performance (Faiss et al., 2013a). Indeed, a number of studies across a range of sports (e.g. football, rugby, lacrosse) have reported performance benefits following a RSH protocol (Brocherie et al., 2015; Faiss et al., 2013b; Galvin et al., 2013; Hamlin et al., 2017; Kasai et al., 2015), and a recent meta-analysis found that RSH induces greater improvements in mean repeated-sprint performance during sea-level repeated sprinting than the same training performed in normoxia (Brocherie et al., 2017). Conversely, select studies have not observed performance benefits following RSH (Goods et al., 2015; Montero and Lundby, 2017). Nevertheless, given the encouraging findings reported in some of the literature and the growing awareness of RSH as a training method (Girard et al., 2017), it is likely that RSH will be increasingly utilised within the team-sport setting.

To date, performance factors (e.g., power output, repeat-sprint ability) have been the main variable of interest in the RSH literature. However, relatively little is known regarding the effect of RSH on inflammatory markers, despite the fact that studies of the hypoxia signalling pathway consistently indicate that hypoxia induces inflammation (see review by Eltzschig and Carmeliet, 2011). This may be an important consideration for team-sport athletes given that inflammatory status has the potential to influence recovery and subsequent training performance (Davis et al., 2007). Indeed, athletes are already more susceptible to illness and infection during periods of intense training or competition (Gleeson, 2007), and there is recent evidence to suggest that a RSH program may alter immune functioning (Born et al., 2016). While select studies have investigated the acute inflammatory response following exercise in hypoxia (e.g., Badenhorst et al., 2014; Govus et al., 2014; Sumi et al., 2017), to the current authors’ knowledge only one previous study has measured an inflammatory biomarker (interleukin-6; IL-6) after the completion of a RSH protocol (Goods et al., 2016). Goods and colleagues (2016) had ten trained male team-sport athletes perform a repeat-sprint session in both simulated altitude (3000 m) and at sea level. This study reported increases in IL-6 concentrations in response to the sprint protocol in both conditions, but a large effect size (d = 0.80) indicated a trend for higher IL-6 concentrations one-hour post-exercise in the hypoxic environment. However, the authors noted that the IL-6 response elicited by the protocol was moderate, likely due to the use of cycling as the exercise modality (Nieman et al., 1998). Indeed, a reduced IL-6 response has been observed when comparing cycling exercise to running (Nieman et al., 1998), which is more specific to most team sports. Moreover, this previous study is limited by its sole use of IL-6 as a marker of inflammation; it is well understood that IL-6 has both pro-inflammatory and anti-inflammatory properties, with elevated post-exercise IL-6 concentrations playing a role in inhibiting pro-inflammatory (i.e., tumour necrosis factor-α; TNF-α) cytokines and facilitating anti-inflammatory cytokine (i.e., IL-1ra, IL-10) production (Petersen and Pedersen, 2005). Thus, the results obtained by Goods et al (2016) could be indicative of either an enhanced inflammatory status following RSH or a positive adaptive response to the training stimulus.

The present study aimed to assess the acute inflammatory response to one RSH session performed on a non-motorised treadmill when compared to the same exercise performed at sea level. Specifically, TNFα, IL-1β, IL-1ra, IL-6, IL-8, and IL-10 were measured before and after the performance of a team-sport specific RSH session. It was hypothesised based on the (albeit limited) existing literature that the RSH session would illicit a greater inflammatory response when compared to the same session performed in normoxic conditions.

Eleven amateur team-sport athletes (seven Australian rules footballers, three soccer players and a touch football player) volunteered to participate in the present study and provided their written informed consent. The Griffith University Human Research Ethics Committee approved all procedures used in the study (AHS/72/14/HREC). All participants had competed for a minimum of two consecutive years in their respective sport immediately prior to participation in the study, trained for a mean (± SD) of 152 ± 35 min·week^-1, and competed in their sport for a mean (± SD) of 108 ± 17 min·week^-1. Familiarisation commenced 3-4 weeks after each participant’s last competitive match for the season (i.e., at the beginning of their off-season). The physical characteristics of the participants are presented in Table 1.

All participants performed a repeat-sprint running test (i.e., the RSR₄₄₄; Morrison et al. (2015)) on two occasions. The RSR₄₄₄ consisted of sixteen (four sets of four) 4-s sprints separated by 26 s (and 2 min 26 s between sets) of passive recovery in a standing position, performed in a commercial normobaric hypoxic chamber (Synergy Physical Conditioning Systems, Yatala, Australia). During one trial participants breathed air with an inspired oxygen fraction (FiO₂) of 0.209, and in the other trial a FiO₂ of 0.145 was used to simulate an altitude of 3000 m (i.e., hypoxia). This FiO₂ at sea level creates a PiO₂ of approximately 110 mmHg. The study followed a crossover design with five of the participants performing the RSR₄₄₄ in normoxia and then in hypoxia 7-10 d later at the same time of day, whereas the remaining participants (n = 6) performed the RSR₄₄₄ tests in the reverse order with the same number of days between trials. Participants were required to: 1) refrain from caffeine consumption on testing days; 2) refrain from alcohol consumption 24 h prior to testing; 3) refrain from performing any exercise of moderate intensity or greater for 72 h prior to testing, as well as refrain from performing any exercise between POST and 3h blood samples; 4) consume only food/beverages that they would ordinarily consume during the 24 h prior to testing; 5) keep a detailed food diary so that food/beverage consumed prior to trial 1 could be replicated for trial 2, and; 6) consume only water in the 4 h prior to testing. Water was consumed ad libitum between POST and 3h, and no food was consumed during this time. The hypoxic environment was created via the extraction of oxygen from air that was subsequently pumped into the chamber. Oxygen concentration was monitored using a gas detector (KB-501, Kingsby Electronics,) which utilises an electrochemical sensor. Relative humidity and temperature were maintained between 45-50% and 19-21ËšC, respectively. Tests were performed on a non-motorised treadmill (Curve 3.0, Woodway, Waukesha, Wisconsin, USA). While we acknowledge that a passive recovery does not replicate team-sport movement demands, the protocol was designed to allow the maintenance of speed, acceleration, and total work (i.e., distance), as well as the rotation of up to four athletes on one treadmill (thus improving team-sport training efficiency).

Distance (m) data were obtained via a tachometer mounted on the treadmill drum and recorded using Pacer Performance System software (Fitness Technology, Adelaide, Australia). Earlobe capillary blood lactate concentration ([La^-]) measurements were obtained using a handheld lactate analyser (Lactate Pro, Arkray Factory Inc., KDK Corporation, Shiga, Japan). Samples were taken upon arrival to the laboratory (baseline; BL) in normoxic conditions, as well as immediately prior to the start of the RSR₄₄₄ (PRE) and immediately after sets 1, 2, 3, and 4 in the environmental condition for that testing day. Measurements of arterial oxygen saturation were made PRE and immediately after sets 1, 2, 3, and 4 using a portable pulse oximeter (Octive Tech, 300CSE, Beijing Choice Electronic Technology Co., Ltd. Beijing, China) in the environmental condition.

Prior to the commencement of the RSR₄₄₄, venous blood was sampled from an antecubital vein (PRE) with the participant lying in a supine position with elbow extended, after lying supine for 6 min to account for postural shifts in plasma volume. Subsequent to this, participants were required to complete a standardised warm-up, identical to that described in detail elsewhere (Morrison et al., 2015), with the addition of two 4-s all-out sprints separated by 24 s of recovery, following dynamic stretching. Participants were provided with strong verbal encouragement during all sprints. No feedback was given regarding the speed or distance achieved during each sprint. Subsequent to completion of the final sprint, participants rested in a seated position for 3 min, relocated to the blood sampling room and laid supine for 6 min prior to blood being sampled again (POST). A third blood sample was taken in the same manner 3 h following the completion of the RSR₄₄₄ (3 h).

Blood samples were collected using a 21-gauge needle, deposited into 10 mL EDTA tubes (BD Vacutainer Systems, Plymouth, UK), and stored on ice until being centrifuged at 3000 rev.min^-1 for 10 min to separate plasma. Plasma was divided into 500 μL aliquots and stored at -80°C until subsequent analysis. Plasma concentrations of IL-1β, IL-1ra, IL-6, IL-8, IL-10, and TNFα were determined using a custom 6-plex cytokine suspension array (Bio-Rad Laboratories Pty Ltd; Hercules, CA, USA) according to the manufacturer’s instructions. Assays were completed using a series of six unique bead sets and samples were diluted 1:3 in the sample diluent provided. Samples from individual participants were analysed in a single assay to avoid interassay variation. The Bio-Plex Suspension Array reader (Bio-Rad Laboratories Pty Ltd; Hercules, CA, USA) was calibrated prior to being used to identify bead subsets and measure fluorescence intensity. A standard curve of cytokine concentration as a function of fluorescence units was constructed using Bio-Plex Manager software (Bio-Rad Laboratories Pty Ltd; Hercules, CA, USA). Recorded fluorescence values were used to determine cytokine concentrations by extrapolation from the respective standard curve.

Statistical analyses were performed using the IBM SPSS Statistics 19 software package. As the data were normally distributed, fully-factorial ANOVA with repeated measures were used to determine the interaction between, or main effect of, condition and time for performance and physiological variables. Least squares difference pairwise comparisons were used to detect the specific site of any significant effect identified. The null hypothesis was rejected at p < 0.05.

There was no difference between conditions in sprint distance (interaction: F =1.62, p = 0.08; main effect of condition: F = 2.45, p = 0.15) across the entire RSR₄₄₄; however, sprint distance deteriorated from 24.92 m to 23.48 m across the 16 sprints (decrease of ~ 6%, F = 7.92, p < 0.01).

Cytokine data are presented in Table 2, with the exception of IL-6 which is presented in Figure 1. There was no interaction effect (F = 0.35, p = 0.71), or main effect of time (F = 0.59, p = 0.57) or condition (F = 0.21, p = 0.66) observed for TNFa, nor IL-1β (F = 0.89, p = 0.43 for interaction; F = 0.61, p = 0.56 for main effect of time; and F = 2.51, p = 0.14 for main effect of condition) or IL-1ra (F = 0.73, p = 0.49 for interaction; F = 2.73, p = 0.09 for main effect of time; and F = 3.57, p = 0.09 for main effect of condition). There was no interaction effect (F = 1.22, p=0.32) or main effect of condition on IL-10 (F = 3.43, p=0.09), but a main effect of time indicated a lower IL-10 at 3 h when compared to the concentrations observed POST (p = 0.037, CI 0.08 – 2.13. There was no difference in IL-10 concentrations between PRE and POST (p = 0.10, CI -1.89 – 0.19). Similarly, there was no interaction effect (F = 0.56, p = 0.58) or main effect of condition (F = 0.03, p = 0.87) observed for IL-6; however, a main effect of time showed that IL-6 increased from PRE to POST (p = 0.03, CI -1.12 – -0.08), and then decreased from POST to 3 h (p = 0.02, CI 0.14 – 1.40) to a value not different from PRE (p = 0.25, CI -0.12 – 0.41). There was also no interaction effect (F = 0.45, p = 0.64) or main effect of condition (F = 0.07, p = 0.79) for IL-8, but a main effect of time showed that it decreased from PRE to POST (p = 0.04, CI 0.13 – 3.26), and was further reduced at 3 h to a value lower than that observed at POST (p = 0.02, CI 0.43 – 1.40).

Lactate and SpO2 data are presented in Table 3.

This study primarily aimed to investigate the inflammatory response after performing a repeat-sprint running protocol under both normoxic and hypoxic conditions. There was a time effect observed for certain cytokines (i.e., IL-6, IL-8, IL-10) which indicates a response to the exercise stimulus; however, hypoxia did not alter the pro- or anti-inflammatory cytokine response, which is in opposition to the proposed hypothesis. This finding suggests that performing RSH elicits a comparable inflammatory response to RSN. However, blood lactate was higher, and blood oxygen saturation was lower, in RSH when compared to RSN.

Although a handful of studies have investigated the inflammatory response to performing exercise under hypoxic conditions (e.g., Badenhorst et al., 2014; Govus et al., 2014; Sumi et al., 2017), only one previous study (Goods et al., 2016) has measured an inflammatory biomarker (i.e., IL-6) after completion of a RSH protocol. Goods and colleagues (2016) reported an increase in IL-6 concentrations both immediately and 60-min post exercise, with a trend for higher IL-6 in the hypoxic condition 60-min after completion of the exercise protocol. In line with these findings (Goods et al., 2016) and those reported by the previous interval training studies conducted in hypoxia (Badenhorst et al., 2014; Govus et al., 2014; Sumi et al., 2017), the current study observed an increase in IL-6 concentrations at the cessation of exercise when compared to baseline levels. However, unlike the prior RSH study (Goods et al., 2016), no difference was observed in IL-6 concentrations between the hypoxic and normoxic conditions at any time point. Interestingly, only the study by Goods et al (2016) reported a difference in IL-6 concentrations between the hypoxic and normoxic trials; the studies utilising longer intervals (3-4 min) all reported a comparable IL-6 response irrespective of condition (Badenhorst et al., 2014; Govus et al., 2014; Sumi et al., 2017). When measured 3-h post-exercise, IL-6 had returned to pre-exercise levels in the present study, which supports the findings of Govus et al (2014) who observed IL-6 concentrations comparable to baseline levels 3 h after completion of an interval training protocol in hypoxia. Conversely, Goods et al (2016) and Sumi et al (2017) reported elevated IL-6 concentrations (when compared to pre-exercise values) 60 min and 120 min post-exercise, respectively. This disparity is likely due to the different sampling times utilised; differences may have been observed in the present study had the blood been sampled at 60-min or 120-min post exercise as opposed to 3-h post-exercise. However, the fact that IL-6 concentrations returned to baseline levels within 3 h, and that comparable concentrations were observed for both conditions, suggests that performing RSH does not elicit an exacerbated IL-6 response. Interestingly, the magnitude of the increase in IL-6 in the present study was similar to that observed by Goods et al (2016), despite the present study utilising a running protocol which typically induces a greater inflammatory response when compared with cycling exercise (Nieman et al., 1998). Indeed, greater increases in IL-6 concentrations were reported after each of the longer interval-training studies (Badenhorst et al., 2014; Govus et al., 2014; Sumi et al., 2017), all of which were performed on a treadmill. Thus, the relatively modest increase in IL-6 concentrations (~75%) observed after the current protocol can most likely be attributed to the relatively short duration of the exercise bouts rather than the mode of exercise utilised.

The previous studies investigating the inflammatory response to training under hypoxic conditions have used IL-6 as the sole inflammatory biomarker, which may be problematic when trying to apply the findings given IL-6 exhibits both pro-inflammatory and anti-inflammatory properties (Petersen and Pedersen, 2005). Thus, the current research sought to provide a more nuanced understanding of the inflammatory response to RSH through the inclusion of other key pro-inflammatory (IL-1β, IL-8, TNF-α) and anti-inflammatory (IL-1ra, and IL-10) cytokines. There was no apparent effect of the exercise protocol (during either condition) on the production of TNFa or IL-1β. This finding is perhaps unsurprising, as although TNFa and IL-1β are the first cytokines to be released during sepsis/infection, generally speaking they are not induced in response to an exercise stimulus alone (Pedersen et al., 1998; Petersen and Pedersen, 2005). Interestingly, no change was detected in IL-1ra concentrations across the course of the protocol, which could be considered unusual given that IL-6 was elevated and IL-1ra typically acts as a cytokine inhibitor (Petersen and Pedersen, 2005). Furthermore, there was no difference in the anti-inflammatory cytokine IL-10 between pre- and immediately post exercise; in studies investigating longer exercise durations (e.g., distance running), increased concentrations of IL-10 have typically been reported immediately post-exercise (Nieman et al., 2001; Ostrowski et al., 1999; Peake et al., 2005) in response to dramatic increases in IL-6 (Pedersen, 2000). These results may be explained, at least in part, by the relatively modest increase observed for IL-6 (IL-6 concentrations can be increased up to 100-fold after a marathon, for example; Pedersen, 2000) and the short duration of the exercise protocol, both of which play an important role in determining the extent of the post-exercise cytokine response (Flynn et al., 2007; Petersen and Pedersen, 2005). Finally, the chemokine IL-8 decreased at each time point throughout the protocol. Exercise does not typically increase the concentrations of circulatory IL-8 (Moldoveanu et al., 2001), although there is relatively strong evidence to suggest that IL-8 is released from working skeletal muscle during exercise and acts locally (Frydelund-Larsen et al., 2007). The reason behind the observed decrease in plasma IL-8 from pre- to immediately post-exercise, and again from immediately post-exercise to 3 h post-exercise, is not immediately clear, and thus requires further investigation. However, the addition of hypoxia did not differentially influence the IL-8 response, which was the primary focus of the present research. Indeed, the current findings suggest that the relative risk of a repeat-sprint training session inducing significant inflammation is low regardless of environment, given the protocol did not elicit significant inflammatory responses in either hypoxic or normoxic conditions.

As expected, blood oxygen saturation was lower in the hypoxic trial when compared with the normoxic trial (Goods et al., 2016; West et al., 1962), whereas blood lactate concentrations were higher. The findings for blood lactate corroborate the findings reported by previous studies performed in hypoxia utilising longer exercise intervals (Sumi et al., 2017) and continuous incremental exercise (Hughson et al., 1995), as well as another RSH study performed on a non-motorised treadmill (Goods et al., 2014). However, they differ from the findings reported by Goods and colleagues (2016) who observed no difference in blood lactate in hypoxia when compared to sea-level during a repeated-sprint cycling protocol. This is likely due to the fact that subjects in the Goods et al (2016) study performed less work in the hypoxic trial, whereas there was no effect of condition on sprint distance observed in the current study. It should also be noted that there is a ~10% (non-significant) difference in pre-test [La] that may reflect slight differences in warm-up intensity or pacing, which may have influenced the overall lactate kinetics throughout the protocol. Nevertheless, the magnitude of the difference in peak blood lactate in the present study was small (10.5 compared to 11.5 mmol/L) and well below the limits typically achieved during repeated-sprint training (Spencer et al., 2005). Thus, the observed differences between the hypoxic and normoxic condition are unlikely to be practically meaningful in terms of the athletes’ recovery.

The ability of athletes to maintain sprint distance despite the lower F_iO₂ during RSH in the present study may be explained, at least in part, by the increase in anaerobic energy release that has been suggested to compensate for a reduction in aerobic ATP production during sprinting in hypoxia (Calbet et al., 2003; Girard et al., 2017). Furthermore, in comparison to the Goods et al (2016) study, the present study utilised shorter (4 v 5 s) and fewer (16 v 27) sprints, a lower within-set work-to-rest ratio (1:6 v 1:4), and a passive recovery, all of which are likely to have contributed to the maintenance of sprint distance during RSH when compared with RSN.

There was no difference between environmental conditions for any of the pro- or anti-inflammatory cytokines measured in the present study, which indicates a comparable inflammatory response between RSN and RSH. Furthermore, while IL-6 increased in both trials immediately after the performance of the RSR₄₄₄ test, the increase was very small relative to what has been previously observed after intense endurance exercise, and blood samples collected 3 h post-exercise indicated comparable plasma IL-6 concentrations to those reported at baseline. This suggests that RSH elicits only a modest inflammatory response, and thus, inflammation should not have any negative impact on subsequent training and performance in team-sport athletes performing RSH.