Environmental stress in elite soccer is an important consideration for both practitioners and policy makers (Taylor and Rollo, 2014). Indeed, eight of the last 19 Fédération Internationale de Football Association (FIFA) World Cups were hosted by countries located at either low (500–2000 m) or moderate (2001–3000m) altitudes (e.g., 2010 FIFA World Cup, South Africa, 1200–1700 m; Bartsch et al., 2008; Billaut and Aughey, 2013). Specific to the Union of European Football Associations (UEFA) region, fixtures are often played above sea level (e.g., Molde, Norway, 1000 m) and/or in hot environments (e.g., Madrid, Spain, 30°C—Taylor and Rollo, 2014). In relation to heat-stress, temperatures often exceeded 30°C (Maximum: 35°C) in the 2014 FIFA World Cup hosted by Brazil (Nassis et al., 2015). Furthermore, combinations of both high temperature and altitude (hypoxia) can be experienced during elite soccer match-play (e.g., Saint-Etienne, France, 30°C; 1000 m).

Soccer match-play data indicates a decline in physical performance in both heat (Ekblom, 1986; Mohr et al., 2010, 2012; Özgünen et al., 2010) and hypoxia (Aughey et al., 2013; Nassis, 2013; Garvican et al., 2014; Buchheit et al., 2015) due to a complex interplay between peripheral, central and perceptual mechanisms (Nybo and Secher, 2004; Billaut and Aughey, 2013; Goodall et al., 2014; Nybo et al., 2014). However, the combined permutations of heat and hypoxia during match-play have not been investigated, although logically their combination would likely exacerbate physical performance decrements. At 43°C (Mohr et al., 2012), total distance (−7%), and high-speed distance (−26%) covered are reduced, with these changes being attributed to a multitude of proposed mechanisms including increasing body temperatures (Nybo et al., 2014). Furthermore, alterations in tactical behavior (e.g., reduced pressing of the ball) has meant that sprint distance covered is unchanged and peak sprint speed is enhanced during heat-situated soccer match-play (Özgünen et al., 2010; Mohr et al., 2012; Taylor and Rollo, 2014; Flouris and Schlader, 2015). Soccer match-play at low altitudes [1200—(Nassis, 2013); 1600 m—(Garvican et al., 2014) above sea level] leads to a decline in total distance (3.1%) and high-speed distance (15%) covered as recovery from high-speed intermittent activity is prolonged, due to the onset of exercise-induced-arterial-hypoxemia caused by a reduction in partial pressure of oxygen within the atmosphere (Billaut and Aughey, 2013). However, sprint performance is enhanced in hypoxia due to improved aerodynamics and flight time of an athlete through the air (Levine et al., 2008), highlighting that different components of soccer performance (e.g., sprint performance) are likely to respond differently within heat and/or hypoxia (Mohr et al., 2012).

Soccer match-play data, including key physical performance measures (e.g., high-speed distance covered), shows high match-to-match variation due to a plethora of match factors, such as tactics, score, opposition, etc. (Gregson et al., 2010). This variability in key physical performance measures may be exacerbated in both heat (Mohr et al., 2010, 2012; Özgünen et al., 2010; Nassis et al., 2015) and hypoxia (Aughey et al., 2013; Nassis, 2013; Garvican et al., 2014; Buchheit et al., 2015) resulting in an altered “pacing strategy” and exercise intensity (Taylor and Rollo, 2014). Recently, Gregson et al. (2010) suggested that to obtain meaningful inferences from a soccer match-play research design, a minimum sample size of 80 players would be required. Consequently, it appears that the majority of match-play based studies examining environmental influences on soccer performance are underpowered (<25 participants; Özgünen et al., 2010; Mohr et al., 2012; Aughey et al., 2013; Garvican et al., 2014; Buchheit et al., 2015), compared to the sample size (n = 80) proposed by Gregson et al. (2010). Only two studies have utilized an appropriate sample size (>n = 80) to assess the performance decrements associated with soccer match-play in hypoxia (Nassis, 2013) and heat (Nassis et al., 2015) during the 2010 and 2014 FIFA World Cup's, respectively. In particular, Nassis et al. (2015) revealed that in hot environments, players preserved key physical performance measures (e.g., peak sprint speed) that are associated with the match outcome (Faude et al., 2012), by reducing the number of sprints and high-speed efforts performed during a match. However, irrespective of the environment, players are likely to modulate their physical performance to avoid an earlier onset of fatigue during a tournament (Dellal et al., 2013), making environmental-mediated-inferences difficult to ascertain from the international tournaments data (Nassis, 2013; Nassis et al., 2015).

Recent reviews (Taylor and Rollo, 2014; Roelands et al., 2015) have recommended a solution to this “sample size issue” is to utilize an individualized, valid and reliable soccer-specific simulation to quantify environmentally-mediated performance decrements with greater experimental control. Aldous et al. (2014) demonstrated that the intermittent Soccer Performance Test (iSPT) is a valid, reliable and individualized (i.e., individualized speed thresholds) laboratory and non-motorized treadmill (NMT) based soccer-specific simulation; which can ascertain changes in soccer performance more robustly compared to match-play data with limited sample sizes. By utilizing iSPT, changes in soccer performance between the identified conditions (e.g., hot and/or hypoxic) can be determined in a controlled environment, minimizing match factors (Gregson et al., 2010) and the within game (Mohr et al., 2005, 2010) and tournament (Dellal et al., 2013) enforced pacing strategies (Nybo et al., 2014; Périard and Racinais, 2015; Roelands et al., 2015), unlike previous environmentally-situated match-play derived data (Mohr et al., 2012; Garvican et al., 2014).

Therefore, the aim of this study was to utilize the iSPT to reliably quantify soccer performance in hot (HOT), hypoxic (HYP), and hot-hypoxic (HH) environments (Aldous et al., 2014). The first experimental hypothesis was that physiological strain would be increased in HOT, HYP, and HH compared with CON, causing a significant reduction in physical performance in HOT, HYP, and HH. The second experimental hypothesis expected the hot and hypoxic environments to enhance sprint performance in HOT and HYP. Finally, the third experimental hypothesis was that in HH, physiological strain would be exacerbated compared with HOT and HYP causing a larger decline in physical performance.

The present study examined the changes in simulated soccer performance in HOT, HYP, and HH conditions compared with CON, by utilizing the recently validated iSPT (Aldous et al., 2014). The main finding revealed a marked decline in total distance, high-speed distance, and variable run distance covered during HOT, HYP, and HH conditions when compared to CON (Figure 2), supporting the first experimental hypothesis. A secondary finding was that peak sprint speed, was increased in HOT compared with CON, HYP, and HH and that sprint distance covered was unchanged in HOT and HYP, supporting the second experimental hypothesis (Figure 2 and Table 3). Furthermore, a greater decline in physical performance was seen in HH even though physiological changes in body mass and temperatures (Figure 3), HR, subjective measures (Figure 4) and S_aO₂ (Figure 4) were not exacerbated compared to HOT and HYP. This change in physical performance was likely due to alterations in Bla concentration and plasma volume which were only present in HH, supporting the third experimental hypothesis.

The data from this study reveals a 4% reduction in total distance and high-speed distance covered in both HOT and HYP compared with CON, which agrees with previous match-play studies in the heat (43°C; Mohr et al., 2012) and at low altitudes (1600 m; Garvican et al., 2014). The performance decrements for total distance, high-speed distance and variable run distance covered between halves (Figure 2) were greater in HOT (8–11%), HYP (10%), and HH (13–14%) compared to CON (7%). In contrast to our results, Mohr et al. (2012) reported the performance decrements between halves was greater in temperate (21°C) compared to hot (43°C) conditions during match-play. This increased performance decrement is indicative of an adaptive match-play-specific pacing strategy which is postulated to preserve technical skill execution (Mohr et al., 2012; Nassis, 2013; Nassis et al., 2015). The environmental stress may likely reduce the “willingness” of an athlete to perform physical exercise during match-play (Mohr et al., 2012; Aughey et al., 2013). The iSPT (Aldous et al., 2014) prevents adoption of these pacing strategies (i.e., match factors; Gregson et al., 2010) with the same exercise performed in each half due to the individualized and externally-controlled speed thresholds. Therefore, players cannot preserve their sprinting characteristics during iSPT by minimizing their high-speed activity as observed during soccer match play (Nassis et al., 2015).

A participants “willingness” to perform high-speed exercise at a self-paced speed was measured during iSPT, via the variable run component, which is designed to quantify high-speed running without an external cue (Aldous et al., 2014). However, when these external cues are removed in the variable run, participants choose a lower running speed in HOT, HYP, and HH compared to CON, which might be indicative of the environment-mediated performance decrements observed in soccer match play (Mohr et al., 2012; Garvican et al., 2014). Furthermore, significant reductions in variable run distance covered in HOT, HYP, and HH both between halves (Figure 2) and in the final 15 min compared with CON (Table 3) were observed. Conversely to soccer match-play at 43°C (Mohr et al., 2012) the performance decrements (high-speed distance, sprint distance, and variable run distance) between the first and last 15 min block was increased in HOT, HYP, and HH when compared with CON (Table 3); likely due to iSPT controlling pacing and match factors (Aldous et al., 2014). This decline in variable run distance supports the notion that the individualized externally-controlled movement patterns employed by iSPT prevented participants adopting an altered pacing strategy. However, previous soccer match-play data has identified that soccer players can preserve key physical performance measures (e.g., sprint distance covered) in hot and hypoxic environments (Nassis, 2013; Nassis et al., 2015), yet decrements in high-speed and sprint distance covered still occur in the final 15 min of match-play (76–90 min) when compared to the first 15 min (0–15 min; (Mohr et al., 2010, 2012)). These performance impairments may influence the match outcome as a number of studies have revealed more goals are scored/conceded in the final 15 min (76–90 min) of match-play (Abt et al., 2002; Armatas et al., 2007). This phenomenon in goals scored/conceded is likely due to an inability to maintain repeated sprint exercise or discrete episodes of non-fatigued maximal physical performance [central to match outcome (Faude et al., 2012)], within the final 15 min of match-play (Gregson et al., 2010; Faude et al., 2012) as supported by the presented data (Figure 2).

A further finding from the present study was that sprint distance covered was unchanged in HYP and HOT, however, peak sprint speed was also improved in HOT compared with CON, showing synergy with previous match-play data (Mohr et al., 2012; Nassis, 2013; Nassis et al., 2015). In HOT, the increase in peak sprint speed could be explained by an increase in estimated T_mu which has been shown to improve muscle contractile properties (Racinais et al., 2004), leading to a higher power production and in turn a better sprint performance (Racinais et al., 2005b). However, improvements in sprint performance during soccer match-play in hot environments has been only shown to occur when T_re is below 39°C (Mohr et al., 2012). Therefore, this could explain the significant reduction in sprint distance covered in the last 15 min in HOT (Table 3). Furthermore, Nassis (2013) identified that elite soccer players in the 2010 FIFA World Cup were able to preserve their peak sprint speed across match-play at low altitudes due to the altered composition to the atmosphere (i.e., air being thinner) which improves the aerodynamics and flight time of an athlete through the air (Levine et al., 2008). However, a hypobaric chamber was not available during this study, so a hypoxicator mask was used to simulate a low altitude environment despite the larger energy cost required when these types of masks are worn (Coppel et al., 2015). The mask was worn in all four experimental conditions to control for this potential confounding factor. Previous research has identified single and repeated sprint performance is maintained at altitude due to a greater anaerobic energy release (Calbet et al., 2003; Morales-Alamo et al., 2012). This is due to several metabolic pathways being stimulated to supplement energy production when aerobic metabolism is not capable of matching aerobic ATP production to consumption, especially the splitting of phosphocreatine (PCr) and glycolysis (Calbet et al., 2003). However, this is likely to manifest itself as a greater and earlier onset of fatigue toward the end of prolonged high-speed exercise as an increase in muscle lactate accumulation would account for a reduction in aerobic ATP production (Balsom et al., 1994; Billaut and Smith, 2010). Therefore, this could explain the exacerbated decline in sprint distance covered during the final 15 min in HYP (Table 3).

Despite similar decrements in physical performance in both HOT and HYP compared to CON, the physiological underpinning of such responses differ. Elevated T_re, T_sk, and estimated T_mu (Figure 2) in HOT and HH were seen from 15 min onwards compared with CON and HYP, showing parity with previous soccer match-play research (Mohr et al., 2010, 2012; Özgünen et al., 2010). In HOT, the absolute rise in T_sk and estimated T_mu predicted total and high-speed distance covered, with end TS predicting the decrement in total distance. As both T_sk and TS have a strong relationship (Sawka et al., 2012), thermal comfort is likely central to the physical performance decrements seen in HOT. Interventions should target these specific factors (T_sk, estimated T_mu and TS) in an attempt to maintain “temperate-like” match play soccer performance.

An increase in HR at HYP when compared with CON, shows synergy with previous soccer match-play data at 1600 m above sea level (Garvican et al., 2014). The rise in HR seen in HYP can be attributed to a hemodynamic response arising from a reduction in S_aO₂ which drives a compensatory increase in cardiac output (Mazzeo, 2008; Stembridge et al., 2015a,b). However, during high-speed exercise bouts at altitude a decrease in stroke volume can decrease O₂ delivery to the active muscles as it cannot match the muscle demand, manifesting as a decline to physical performance in HYP (Mazzeo, 2008). A reduction in S_aO₂ by ~8% compared to CON was also apparent by the end of iSPT in both HYP and HH which indicates the onset of exercise induced arterial hypoxemia had occurred causing a plethora of detrimental physiological responses (Billaut and Aughey, 2013), driving the exacerbated performance decrements seen in HYP and HH (Figure 2). Indeed, reduced phosphocreatine re-synthesis at altitude is due to sub-optimal re-oxygenation of the active skeletal muscle elongating the recovery time between high-speed exercise bouts (Garvican et al., 2014). Changes in high-speed running are important for maintaining match-play physical performance, due to its association with game defining moments (Gregson et al., 2010), possibly impacting upon the match result (Taylor and Rollo, 2014). Furthermore, the employed design cannot distinguish precisely between whether the changes in S_aO₂ were apparent due to exercise and/or environmentally-induced-arterial-hypoxemia, highlighting that future work should look to explore these complex phenomena within an appropriate design. Data by Billaut and Smith (2010) indicates that intermittent running based exercise can induce exercise-induced-arterial-hypoxemia in University level soccer players. Therefore, although the employed design cannot distinguish precisely between exercise and environmentally-mediated-arterial-hypoxemia future work should look to explore these complex phenomena within an appropriate design.

In HH, the largest performance decrement both between halves (Figure 2) and 15 min blocks was evident (Table 3). However, all changes in TS (Figure 4), body mass and temperature (Figure 3) were similar compared with HOT. Furthermore, all changes to both S_aO₂ (Figure 4) and HR were comparable with HYP. This is despite a greater decline in total distance and high-speed distance covered, as well as an additional reduction in sprint distance covered in HH which were not present in HOT and HYP (Figure 2). This exacerbated reduction to physical performance in HH may have been due to a significant increase in Bla concentration which may indicate a greater anaerobic energy release compared with CON, HOT, and HYP (Amann et al., 2006). Furthermore, a 5% reduction in plasma volume (Table 3) which coincided with a 2% change in body mass post-iSPT in HH may have meant that the participants finished iSPT in a hypo-hydrated state, (Cheuvront et al., 2003) causing an increase to the rate of heat storage and sweat output which in turn can impair prolonged high-speed activities in hot environments (Cheuvront and Kenefick, 2014). Additionally, HR was also increased during HH, showing parity with previous research in a hot and low altitude environment (30°C; 1900 m; Buono et al., 2012). This augmented HR response in HH likely stemmed from an impaired stroke volume and/or cardiac output, previously seen during prolonged exercise bouts in heat (González-Alonso et al., 2008) and hypoxia (Mazzeo, 2008). Thus, the exacerbated decline in performance was likely caused by a combination of both hot and hypoxic-mediated fatigue mechanisms. It is already acknowledged that both heat and hypoxia induce performance decrements via these mechanisms during soccer match-play likely influencing match outcome (Taylor and Rollo, 2014). Therefore, the number of game defining moments may be further decreased within HH.

The use of recreationally active male volunteers, rather than elite soccer players, is a limitation of this study; so any generalization of the results to such populations should be considered cautiously. However, our sample included participants with a V˙O2max>55 mL^.kg^−1.min⁻¹, demonstrating some parity with elite soccer (Tonnessen et al., 2013). The assessment of technical skills and multi-directional movements were unable to be quantified by iSPT (Aldous et al., 2014). Therefore, to assess these within a similarly valid soccer-specific simulation, match factors and protective adaptive pacing strategies must be controlled, in order to robustly assess whether technical skills would remain unchanged in line with previous match-play data (Mohr et al., 2012; Nassis, 2013; Nassis et al., 2015).

The data from this study can be utilized to ascertain the efficacy of any ergogenic intervention to offset the environmentally-induced-decrements. For example, pre- and/or half-time-cooling has been reported to have an ergogenic effect upon both aerobic (Duffield et al., 2010) and repeated-sprint performance in the heat (Castle et al., 2006). Dietary nitrate has also been shown to improve muscle oxygenation during sub-maximal and maximal exercise in acute severe hypoxia (Masschelein et al., 2012; Wylie et al., 2013; Thompson et al., 2015). Furthermore, key physical performance measures (e.g., high-speed distance and sprint distance covered) associated with the match outcome in soccer (Gregson et al., 2010; Faude et al., 2012) are impaired in hot, hypoxic and hot-hypoxic environments, potentially decreasing the number of game defining events during match-play (Taylor and Rollo, 2014). Therefore, the efficacy of these interventions may be important for practitioners and governing bodies to attenuate these decrements present for key physical performance measures during soccer match-play in hot and hypoxic environments.

In conclusion, the present study shows that during simulated soccer performance, total distance, high-speed distance and variable run distance covered are significantly impaired within hot (30°C), hypoxic (1000 m above sea level), and hot-hypoxic (30°C; 1000 m above sea level) conditions when compared to a normoxic-temperate environment. Furthermore, peak sprint speed, was increased in HOT compared with CON, HYP, and HH. However, sprint distance covered was unchanged in HOT and HYP and only decreased in HH compared with CON. It is also revealed that the reduction in soccer physical performance is exacerbated in HH, compared to HOT and HYP alone. The heat-induced-decrements in HOT stem from increasing body temperatures, TS and the 2% reduction in body mass. The hypoxic-induced-decrements in HYP were most likely initiated by a decrease in SaO2 and increase in HR. Similar changes in TS, body mass and temperatures were seen in HOT compared with HH, whilst similar changes in HR and SaO2 were evident in HH compared to HYP. Furthermore, both Bla and plasma volume change alterations were only seen in HH compared with CON, highlighting that both these measures may play a role in the exacerbated decrements seen in HH. However, a deductive design to assess whether simulated soccer performance would still decrease in HH if plasma volume was maintained is needed to understand the mechanistic cause of these findings. The aforementioned physiological changes seen in the present study may influence the decrements to physical performance seen in HOT, HYP, and HH. Therefore, a detrimental effect on the match outcome may be seen in soccer match-play in these environments, which would be important to practitioners within soccer.

Conceived and designed the experiments JA, BC, IA, BD, GA, and LT. Performed the experiments JA, BC, IA, BD, GA, and LT. Analyzed the data JA, BC, IA, BD, GA, and LT. Contributed reagents/materials/analysis tools JA, BC, IA, BD, GA, and LT. Wrote the paper JA, BC, IA, BD, GA, and LT.

This research was funded by the João Havelange Research Scholarship on behalf of the Fédération Internationale de Football Association (FIFA). No commercial or financial incentives were provided that can have caused any potential conflict of interest. The authors would like to thank the participants for their involvement in this study. There was no conflict of interest for any author in this study.