Functional reserve and sex differences during exercise to exhaustion revealed by post‐exercise ischaemia and repeated supramaximal exercise

Females have lower fatigability than males during single limb isometric and dynamic contractions, but whether sex‐differences exist during high‐intensity whole‐body exercise remains unknown. This study shows that males and females respond similarly to repeated supramaximal whole‐body exercise, and that at task failure a large functional reserve remains in both sexes. Using post‐exercise ischaemia with repeated exercise, we have shown that this functional reserve depends on the glycolytic component of substrate‐level phosphorylation and is almost identical in both sexes. Metaboreflex activation during post‐exercise ischaemia and the O2 debt per kg of active lean mass are also similar in males and females after supramaximal exercise. Females have a greater capacity to extract oxygen during repeated supramaximal exercise and reach lower PETCO2 , experiencing a larger drop in brain oxygenation than males, without apparent negative repercussion on performance. Females had no faster recovery of performance after accounting for sex differences in lean mass.


Introduction
During whole-body high-intensity exercise to exhaustion, it is likely that task failure occurs due to a higher energy demand than energy supply. The supply of energy may be curtailed by a reduction in the rate of ATP resynthesis due to partial blockade of the critical reactions in the metabolic energy pathways or by exhaustion of critical substrates. We define 'functional reserve' as the capacity to produce power at the same level or higher than reached at exhaustion. This functional reserve has been demonstrated in males by the capacity to perform sprint exercise at the end of an incremental exercise test to exhaustion followed by 10-60 s ischaemia, which impedes metabolic recovery Gelabert-Rebato et al. 2018, 2019a as well as after 3-5 s recovery without occlusion (Coelho et al. 2015). It remains unknown, however, whether this functional reserve is also present in females after an exercise bout supposedly utilizing the totality of the anaerobic capacity (Medbo et al. 1988;Medbo & Tabata, 1989). Moreover, the physiological factors determining the nature and magnitude of the functional reserve and whether sex differences exist have not been established.
Classically, during high-intensity exercise, i.e. in the severe exercise-intensity domain (Poole & Jones, 2012), task failure has been attributed to the accumulation of metabolites (AMP, ADP, lactate, H + and inorganic phosphate (P i )), particularly during exercise with marked recruitment of substrate-level phosphorylation (PCr and glycolysis) (Cady et al. 1989;Fitts, 1994;Allen et al. 2008;Black et al. 2017), and the associated production of reactive oxygen and nitrogen species (RONS) (Westerblad & Allen, 2011). At task failure in the severe intensity domain, the level of metabolite accumulation and peripheral fatigue is similar despite marked differences in relative intensity (Thomas et al. 2016). However, a study in humans indicates that task failure is not due to the accumulation of energy metabolites or blockade of glycolysis , in agreement with previous studies showing that the capacity to generate force is rapidly recovered despite a high concentration of lactate and H + (Sahlin & Ren, 1989).
Whether males and females fatigue similarly during whole-body exercise in the severe exercise-intensity domain remains unknown. Previous studies have reported lower fatigability in females during isometric contractions at low (20% of the maximal voluntary contraction (MVC)) but not at high (80% of MVC) intensities (Yoon et al. 2007;Hunter, 2014Hunter, , 2016aAnsdell et al. 2017). During single limb dynamic contractions, females fatigue less than males depending on the contraction velocity and the muscle groups recruited (Hunter, 2016b). When present, the sex differences are also observed during the first 45 min of recovery after task failure (Ansdell et al. 2019). At task failure during isometric intermittent quadriceps muscle contractions in the severe exercise-intensity domain, neuromuscular function is similar in both sexes (Ansdell et al. 2019). Consequently, it has been suggested that the lower fatigability and faster recovery of females should be due to less metabolic derangement (Senefeld et al. 2018;Ansdell et al. 2019). The latter could be due to the fact that females have a lower anaerobic capacity than males (Green et al. 1984). In support, near-infrared spectroscopy (NIRS) measurements at task failure and during recovery show lower muscle deoxygenation in females than males (Ansdell et al. 2019). Whether females have better muscle oxygenation and lower metabolic muscle impairment at task failure during whole-body exercise in the severe-intensity domain remains unknown (Ansdell et al. 2019). Recent work indicates that females have superior mitochondrial respiration than males (Cardinale et al. 2018), which may facilitate O 2 extraction. There is a paucity of data relating to sex differences in recovery (Ansdell et al. 2019) and no previous study has determined whether there is a sex-difference in whole-body performance recovery between males and females, after task failure in the severe exercise-intensity domain.
Input from type III and IV muscle afferents may also influence the rate of perceived exertion and contribute to task failure by a central mechanism (Sidhu et al. 2014(Sidhu et al. , 2017Hureau et al. 2019), although their exact role during whole-body exercise is poorly understood (Marcora, 2010;Torres-Peralta et al. 2016). Muscle afferent input during exercise may be higher in males than females, as suggested by the enhanced sympathetic and pressor response to metaboreflex activation in males compared to females (Jarvis et al. 2011;Welch et al. 2018;Joshi & Edgell, 2019;Samora et al. 2019), as well as the activation of chemoreflex responses secondary to metaboreflex activation (Edgell & Stickland, 2014;Wan et al. 2020). Metaboreflex activation may also directly modulate medullary respiratory centre output via relay through the nucleus tractus solitarii (Sander et al. 2010), resulting in increased pulmonary ventilation (Lam et al. 2019). The latter could contribute to an earlier task failure in males than females. However, most data have been obtained in males lying supine by eliciting the metaboreflex in the forearm muscles (Jarvis et al. 2011;Joshi & Edgell, 2019;Samora et al. 2019). Thus, if males have a higher metaboreflex sensitivity than females, the application of ischaemia after exercise should result in a higher heart rate (HR) and ventilatory responses during ischaemia in males than females.
Therefore, the main aim of this investigation was to determine whether males and females have a functional reserve at exhaustion following supramaximal exercise at 120% ofV O 2 max , and to ascertain what mechanisms may explain potential between-sex differences in task failure and functional reserve.
For these purposes, we applied a new experimental protocol which allows assessment of substrate-level phosphorylation for ATP re-synthesis non-invasively by the application of ischaemia at the end of exercise to impede the metabolic recovery (Harris et al. 1975;Morales-Alamo et al. 2015). Since muscle fatigue is task-specific (Gandevia, 2001), it is crucial to test the existence of a functional reserve by using the same pattern of movement. This can be accomplished by setting the ergometer in hyperbolic mode at a constant supramaximal power, as for example 120% of maximal oxygen consumption (V O 2 max ), such that the subjects will only be able to resume the exercise if they have a functional reserve in substrate-level phosphorylation at exhaustion sufficient to reach the power output corresponding to the supramaximal load.
We hypothesized that (1) females would have a lower anaerobic capacity than males, even when normalized to the lean mass of the lower extremities; (2) following high-intensity exercise to exhaustion, females would have a lower functional reserve than males; (3) during repeated fatiguing high-intensity exercise, females would recover from fatigue faster than males; (4) during repeated fatiguing high-intensity exercise females would achieve greater O 2 extraction than males; and (5) post-exercise ischaemia would reveal higher metaboreflex-induced heart rate and ventilatory responses in males than females.

Ethical approval
The study was approved by the Ethical Committee of the University of Las Palmas de Gran Canaria (Ref.: CEIH-2017-13) and performed in accordance with the standards set by the latest revision of the Declaration of Helsinki, except for registration in a database. All subjects signed a written informed consent before the start of the study.

Subjects
Eighteen males and eighteen females, all healthy and physically active, agreed to participate in this investigation (age = 24.6 ± 4.4 years, height = 170.5 ± 8.1 cm, body mass = 63.8 ± 9.3 kg, body fat = 22.3 ± 5.5%, n = 36, Table 1). The inclusion criteria for participation in this investigation were: age 18-35 years; no chronic diseases or recent surgery; non-smoking; normal resting electrocardiogram; body mass index above 18 and below 30; no medical contraindications to exercise; and no history of disease requiring medical treatments lasting longer than 15 days during the preceding 6 months. Subjects were requested to avoid strenuous exercise 48 h before the laboratory test and not to drink carbonated, caffeinated and alcohol-containing beverages during the 24 h preceding all tests. During the study period, subjects were also requested to abstain from the consumption of drugs, medications and any dietary supplements. Sex and gender of the participants were defined based on self-report during participant recruitment. All participants reported cis-gender, and thereafter the terms males and females were applied in the study analysis and reporting (Heidari et al. 2016). All females were eumenorrhoeic, and five were taking oral contraceptives. Females not taking contraceptives were tested randomly in different phases of the menstrual cycle (Mattu et al. 2020), given the large number of tests performed. This approach is based on the similar sprint and high-intensity responses observed in different phases of the menstrual cycle (Botcazou et al. 2006;Bushman et al. 2006;Shaharudin et al. 2011).

Study overview
The study included the following consecutive phases: (1) recruitment, familiarization and pre-testing; (2) assessment ofV O 2 max and theV O 2 -intensity relationship; and (3) supramaximal exercise tests to determine the functional reserve at exhaustion (Fig. 1).

Recruitment, familiarization and pre-testing
Subjects agreeing to participate underwent a body composition examination and several familiarization sessions. Body composition was assessed by dual-energy X-ray absorptiometry (Lunar iDXA, GE Healthcare, Milwaukee, WI, USA) as described elsewhere (Calbet et al. 1998). Subjects were familiarized with the experimental procedures by performing an incremental exercise to exhaustion and sprint exercise (30 s Wingate all-out test), which were not used in subsequent analyses. After that, subjects reported to the laboratory to complete different experimental tests on separate days.

Assessment ofV O 2 max andV O 2 -intensity relationshiṗ
V O 2 max , maximal heart rate (HR max ) and maximal power output (W max ) were determined in normoxia (F IO 2 : 0.21, P IO 2 : 144 mmHg) with an incremental exercise test to exhaustion with verification (Poole & Jones, 2017). For this test, subjects reported to the laboratory at least 4 h after the last ingestion of food. The test started with 3 min at 20 W, followed by 15 W and 20 W increases every 3 min in females and males, respectively, until the respiratory exchange ratio (RER) was ≥1.00. After that, the load was increased by 10 W and 15 W every minute in females and males, respectively, until exhaustion. The highest intensity attained in the test was taken as the W max of the incremental exercise. At exhaustion, the ergometer was unloaded, and slow pedalling (30-40 rpm) continued for 3 min. At the third minute of active recovery, the verification test was initiated at W max + 5 W during 1 min, followed by a 4 and 5 W increase (females and males, respectively) every 20 s until exhaustion.
Oxygen uptake (V O 2 ) during all exercise tests was measured by open-circuit indirect calorimetry with a metabolic cart (Vyntus, Jaeger-CareFusion, Höchberg, Germany) operated in breath-by-breath mode. The gas analysers were calibrated immediately before each test using room air (20.93% O 2 and 0.04% CO 2 ) and high-grade certified gases provided by the manufacturer containing 16% O 2 and 5% CO 2 . The volume flow sensor was calibrated at low (0.2 l s −1 ) and high (2 l s −1 ) rates immediately before each test. The validity of this metabolic cart was established by a butane combustion test   In each of them, after a standardized warm-up, subjects performed three bouts of supramaximal constant intensity exercise at 120% ofV O 2 max until exhaustion, interspersed either with 20 s of recovery with application of immediate post-exercise ischaemia at exhaustion (ischaemic session) or with 20 s of recovery with free circulation (free circulation session), in random order. At the start of the 2nd and 3rd bouts in the ischaemic recovery sessions, the cuffs were deflated instantaneously, to allow for restoration of the circulation during the subsequent bout. The cuffs located around the two thighs were instantaneously inflated at 300 mmHg during the sessions with ischaemic recovery to elicit total occlusion of the circulation of both lower extremities and impede metabolic recovery. [Colour figure can be viewed at wileyonlinelibrary.com] J Physiol 599.16 were analysed breath-by-breath and averaged every 20 s during the incremental exercise tests. The highest 20-s averagedV O 2 recorded during the whole incremental test (i.e. including the verification phase) was taken as thė V O 2 max (Martin-Rincon et al. 2019). HR was recorded continuously with a sampling frequency of 1 s during all exercise tests via short-range radiotelemetry (RS400 and RS800, Polar Electro, Woodbury, NY, USA).

Assessment of the functional reserve at exhaustion
The functional reserve at exhaustion was determined 1-2 weeks after the assessment ofV O 2 max . For this purpose, subjects reported to the laboratory on two occasions, 1 or 2 weeks apart, hereafter called ischaemic and free circulation recovery sessions (Fig. 1). Subjects were requested to ingest a similar light breakfast on the experimental days, which should have been ended at least 1 h prior to the scheduled time. One hour after their arrival to the laboratory, the protocol described in Fig. 1 was started. Each of the two sessions consisted of three bouts of constant-power exercise to exhaustion at 120% oḟ V O 2 max (hereafter referred as 120% CP) interspaced by 20 s recovery periods with (ischaemic session) or without (free circulation session) application of total occlusion of the circulation to the lower extremities. The intensity of the supramaximal exercise bouts was set at 120% ofV O 2 max , to facilitate the attainment of the maximal accumulated oxygen deficit during constant-power exercise on the cycle ergometer (Medbo & Tabata, 1989;Calbet et al. 1997).
In the ischaemic recovery session, the circulation of both lower extremities was instantaneously occluded at exhaustion after the first and the second 120% CP bouts, and the ischaemia maintained during the two 20 s recovery periods. Right at the restart of the second and third bouts, the cuffs were instantaneously released and the circulation wholly re-established. For the free circulation session, subjects performed unloaded pedalling at low cadence (∼20 rpm) during the two 20 s recovery periods, to minimize the risk of orthostatic post-exercise hypotension .
Before the exercise, while the subjects were resting supine, bilateral 10 cm-wide cuffs were placed around the thighs, as close as possible to the inguinal crease, and connected to a rapid cuff inflator (SCD10, Hokanson E20 AG101, Bellevue, WA, USA) as previously reported Torres-Peralta et al. 2016). Each session started with a warm-up consisting of 1 min of unloaded pedalling, followed by 2 min at 60 or 40 W, 3 min at 80 or 60 W, 1 min at 100 or 80 W, 1 min at 120 or 100 W and 1 min at 140 or 120 W for males and females, respectively. This was followed by 5 min of unloaded pedalling at low cadence (20-40 rpm). Then the subjects stopped pedalling, and the ergometer was set in hyperbolic mode at the load corresponding to their 120% ofV O 2 max to perform the three 120% CP bouts until exhaustion, interspaced by the corresponding 20 s recovery periods. During the 20 s recovery periods, at the 15th second subjects were given a 5 s reverse countdown and prompted to restart pedalling again at the same intensity (i.e. 120% ofV O 2 max ) until they were exhausted again. In the sessions with ischaemic recovery, the cuffs were instantaneously inflated at 300 mmHg to occlude completely the circulation of both lower extremities upon exhaustion. The cuffs were deflated instantaneously at the start of the second and third bouts and the circulation re-established during the subsequent exercise.
Cerebral and muscular oxygenation was assessed by NIRS (NIRO-200NX, Hamamatsu, Japan) employing spatially resolved spectroscopy to obtain the tissue oxygenation index (TOI) using a path-length factor of 5.92 (van der Zee et al. 1992). The first NIRS optode was placed on the right frontoparietal region at 3 cm from the midline and 2-3 cm above the supraorbital crest, to avoid the sagittal and frontal sinus areas . A second optode was placed in the lateral aspect of the thigh at middle length between the patella and the anterosuperior iliac crest, over the middle portion of the m. vastus lateralis. The vastus lateralis fractional extraction index (TOI O 2 extraction index) was obtained as TOI OBV − TOI MIN , where TOI OBV is the mean TOI value registered during exercise and TOI MIN is the minimal 1-s rolling average TOI value registered during ischaemia. The TOI is an indicator of absolute tissue O 2 saturation of the haemoglobin present in arteriolar, capillary and venular beds, combined with the O 2 saturation of myoglobin, and can be assessed using spatially resolved spectroscopy (Boushel et al. 2001;Quaresima & Ferrari, 2009). Assuming that blood is distributed similarly in the three vascular beds during supramaximal exercise, and that the TOI during ischaemia represents the maximal levels of deoxygenation reachable, a higher difference between the TOI observed during supramaximal exercise and the minimum reached during ischaemia represents the residual O 2 bound to the haemoglobin and myoglobin that has not been extracted. Recent studies carried out in our laboratory have shown that there is a linear relationship between the raw TOI value recorded by the NIRO-200NX, and the leg O 2 extraction and femoral vein O 2 content (see Appendix).
All exercise tests were performed on the same cycle ergometer (Lode Corival, Lode BV, Groningen, The Netherlands), which maintains the exercise intensity constant despite variations in pedalling rate. During all tests, subjects were requested to keep the pedalling rate at 80 rpm (±3 rpm). In all instances, exhaustion was defined by the incapacity of the subject to maintain a pedalling rate above 50 rpm during 5 s or by the sudden stop of pedalling. Strong verbal encouragement was provided throughout the entire exercise protocol and particularly approaching task failure. During the ischaemic recovery trials, subjects were requested to rate the level of pain in their thighs with a visual numerical rating scale from 0 to 10, with 10 being the highest muscle pain ever experienced during or after exercise. All subjects were accustomed to performing exercise to exhaustion and post-ischaemia all-out bouts of exercise, as well as to rate their level of pain with the same rating scales due to participation in other projects carried out in our laboratories using similar procedures. The ergometer seat and handlebar configuration were adjusted for comfort during the first visit and replicated in subsequent sessions. Exercise tests took place in an air-conditioned laboratory with an ambient temperature of ∼21°C, a relative humidity of 60-80%, and ∼735 mmHg atmospheric pressure.
The O 2 demand during the supramaximal exercise bouts was estimated from the linear relationship between the last minute averagedV O 2 of each load, from 20-40 W to the highest intensity with a RER < 1.00. The accumulated oxygen deficit (AOD, an estimate of the energy provided by substrate-level phosphorylation), representing the difference between O 2 demand and accumulated O 2 , was determined as previously reported (Calbet et al. 1997;Morales-Alamo et al. 2015). The contribution of anaerobic energy metabolism to the total energy yield was calculated as AOD × 100/O 2 demand. Since during the occlusions the myoglobin O 2 stores are depleted and PCr is not resynthesized (Harris et al. 1976;Blei et al. 1993;Quistorff et al. 1993;Morales-Alamo et al. 2015), the totality of the O 2 deficit measured during the subsequent bout corresponds to the energy supplied by the glycolytic component of substrate-level phosphorylation. This O 2 deficit was converted into moles of ATP assuming a volume of 22.4 litres per mole of oxygen (STPD), a muscle temperature of 38.5°C, and a phosphorus-to-O 2 ratio (P/O) of 2.5 (Hinkle et al. 1991). Lactate production was obtained knowing that 1.5 moles of ATP are produced per mole of lactate.
The oxygen debt incurred during the 20 s recovery by the lower extremities was calculated as the recoverẏ V O 2 with free circulation minus theV O 2 during ischaemic recovery. The obtained value was normalized to the lean mass of the lower extremities (LLM) for sex comparisons.

Statistics
To establish the sample size, we first determined the correlation between mean power output in a control sprint of 10 s and the same type of sprints performed following 10 or 60 s occlusions applied at exhaustion ; the corresponding values ranged between 0.2 and 0.5. In this previous study, the standard deviation of the total work performed after the occlusion represented 22% of the mean value in each sprint. Simulations were run to determine what would be the sample size required to detect a mean difference between sexes of 20% in lean muscle mass normalized power output after the occlusions. This difference was established based on known sex differences in mean power output during supramaximal exercise anḋ V O 2 max (20-30%, when expressed per kg of body mass) (Perez-Gomez et al. 2008;Loe et al. 2013). This resulted in a sample size of 24, i.e. 12 males and 12 females, assuming a correlation between repeated measurements of 0.5 (effect size: 0.5, at α = 0.05 and 1 − β = 0.80; G * Power v.3.1, F-tests, ANOVA for repeated measures between factors). Thirty-six volunteers were recruited to allow for potential dropouts. Since there were no dropouts, the study was powered to detect a 13-16% sex difference, depending on the correlation between repeated measures assumed. Variables were checked for normal distribution by using the Shapiro-Wilks test. A three-way repeated-measures ANOVA was used with two within-subjects factors: exercise bout (with three levels) and occlusion (with two levels: ischaemia and free circulation) and with sex as a between-subjects factor (with two levels: male and female). To test for potential sex differences, the sex contrast was assessed. Besides, the following interactions were evaluated: occlusion by sex (to determine whether males and females responded differently to the occlusions) and bout by sex (to find out whether the repetition of bouts elicited different responses in males and females). The cardiorespiratory responses during the recovery periods were also tested using three-way repeated-measures ANOVAs with two within-subjects factors: exercise bout (with two levels: first and second recovery) and occlusion (with two levels: ischaemia and free circulation), and with sex as a between-subjects factor (with two levels: male and female). Mauchly's test of sphericity was run before the ANOVA, and in the case of violation of the sphericity assumption, the degrees of freedom were adjusted according to the Huynh-Feldt test. When a significant main effect or interaction was observed, specific pairwise comparisons were carried out with the Fisher's least significant difference (LSD) post hoc test when appropriate. Besides, Student's paired t-test (two-tailed) was used to compare the responses obtained during exercise to resting conditions or between two different bouts at particular time-points. An unpaired two-tailed t-test was also used to compared males and females at specific points. The relationship between variables was determined using linear regression analysis. Values are reported as the mean ± standard deviation (SD), effect size (ES) and 95% confidence intervals (CI). P < 0.05 was considered significant. Statistical analysis was performed using SPSS v.15.0 for Windows (SPSS Inc., Chicago, IL, USA). J Physiol 599.16
The O 2 expended per unit of work in the first bout was similar in males and females (142.1 ± 13.7 and 144.5 ± 19.3 ml kJ −1 , respectively, P = 0.67, ES = −0.14 (CI: −0.80, 0.51), sex contrast, ANOVA) ( Table 2). The O 2 expended per unit of work was increased from 144.9 ± 18.7 ml kJ −1 in the first bout with free circulation recovery to 171.7 ± 16.7 and 172.5 ± 18.1 ml kJ −1 , in  Table 2). The mean HR was 5.7 beats min -1 lower during the second and third bouts preceded by ischaemia compared to the same bouts preceded by recovery with free circulation ( Table 2).

Sex differences in fatigability and recovery during repeated supramaximal exercise to exhaustion
No sex differences were observed in times to exhaustion (and work), in the second and third bouts (P = 34, ES = −0.32 (CI: −0.98, 0.34), sex contrast; P = 0.88, occlusion by sex by bout interaction) ( Table 2).

Figure 3. Performance and overall energy expenditure during three bouts of constant power exercise to exhaustion at 120% ofV O 2 max , interspaced by 20 s recovery periods with (dark colours) or without (light colours) ischaemia
A, work; B, accumulated oxygen; C, O 2 deficit; D, anaerobic energy yield as percentage of the whole energy expenditure. In A-C values are normalized to the first bout, of each experimental day, i.e. the first bout is taken as 100%. Black bars: mean responses for the last two bouts in females; white bars: mean responses for the last two bouts in males. Error bars: standard deviation. * P < 0.05 males compared to females; †P < 0.05 free circulation compared to occlusion; ‡P < 0.05 compared to the second bout. n = 18 for males and n = 18 for females. [Colour figure can be viewed at wileyonlinelibrary.com] the occlusion and free circulation, respectively, P = 0.002, ES = −0.60 (CI: −0.96, −0.24), occlusion main effect, ANOVA) (Fig. 4C).
Brain oxygenation was 4-6 units (TOI) lower in females than males during the recovery periods (P = 0.003, ES = 1.10 (CI: 0.33, 1.85), sex contrast) ( Table 3). Brain oxygenation was ∼1.5 units lower in the second than in the first recovery period (P < 0.001, ES = 1.40 (CI: 0.92, 1.86), bout main effect, ANOVA), and this response was similar during the experiments with and without occlusion, and in men and females (Table 3).

Discussion
The main finding of the present investigation is the demonstration that males and females respond similarly to repeated supramaximal whole-body exercise to exhaustion and that at task failure a large functional reserve remains in both sexes. Although males and females performed the exercise at an intensity which should have exhausted the anaerobic capacity (Medbo & Tabata, 1993;Gastin et al. 1995;Calbet et al. 1997), a similar functional reserve in substrate-level phosphorylation energy supply per kg of lower extremities lean mass was observed at task failure in both sexes. Using indirect methods, we have shown that this functional reserve depends on the glycolytic component of substrate-level phosphorylation.
Accounting for their lower anaerobic capacity and blood haemoglobin concentration, females performed slightly less during the first bout of exercise and similarly to males during the subsequent bouts, likely due to their greater capacity to extract O 2 during high-intensity exercise. Compared to males, females achieved lower P ETCO 2 and brain oxygenation during supramaximal exercise to exhaustion, without an apparent negative repercussion for performance.

The anaerobic capacity does not determine task failure during supramaximal exercise to exhaustion
In agreement with previous studies, we have found that males incurred higher maximal O 2 deficits than J Physiol 599.16 females (Medbo & Burgers, 1990;Weyand et al. 1993;Ramsbottom et al. 1997;Hill & Vingren, 2014), but as a novelty, here we show that this greater capacity of males persists after accounting for differences in muscle mass. Nevertheless, after the second and third bouts to exhaustion, the capacity to produce energy through substrate-level phosphorylation was similar in both sexes, after normalization to the LLM. Neither in males nor in females was task failure caused by exhaustion of anaerobic capacity. Otherwise, resumption of exercise would have been impossible after the occlusion because the intensity was supramaximal. Since the occlusion of the circulation at exhaustion impedes the recovery of PCr and muscle pH (Harris et al. 1976;Blei et al. 1993;Quistorff et al. 1993;Morales-Alamo et al. 2015), and the exercise intensity was always aboveV O 2 max , our subjects would not have been able to restart pedalling at the end of the first and second 20 s occlusions without such a functional reserve in substrate-level phosphorylation. Since previous work has demonstrated that at the moment of occlusion PCr levels are very low, and ADP and Cr concentration increased , mitochondrial respiration should be maximally stimulated (Sahlin & Harris, 2011;Calbet et al. 2020). Consequently, in less than 3-5 s the small amount of O 2 remaining bound to myoglobin (Richardson et al. 1995) and the O 2 present in the capillary beds (Harris et al. 1975;Blei et al. 1993) is consumed, as previously explained  (Parolin et al. 1999). If anything, the metabolic state should be even worse at the end of the second bout of exercise (Parolin et al. 1999) and further impaired if preceded by ischaemia . Moreover, as the exercise should have started with very low levels of PCr (Harris et al. 1976;Morales-Alamo et al. 2015), the only metabolic pathway that can significantly supplement the aerobic energy production during post-occlusion supramaximal exercise (120% aboveV O 2 max , in the present study) is the glycolytic component of substrate-level phosphorylation . This allows for an estimate of the amount of lactate produced during the second and third bouts, which was 3.47 and 2.94 mmol kg LLM −1 (i.e. 0.16 and 0.17 mmol kg −1 s −1 , respectively). These values represent ∼17-20% of the maximal glycolytic rates achieved by six males during a Wingate test (Parolin et al. 1999). Despite this unfavourable metabolic environment, all subjects were able to perform some exercise during the third bout (from 4 to 52 s), revealing that even after the second bout of exercise, and despite impeding muscle recovery with the occlusion, some functional reserve remains at exhaustion. Nevertheless, four males and five females exercised for less than 10 s during the third bout, indicating that some subjects used a large fraction of their functional reserve after the second occlusion.

Effects of post-exercise ischaemia on subsequent exercise
This research shows that muscle O 2 extraction capacity is increased similarly in males and females when the exercise is performed after ischaemia, as indicated by the lower TOI values observed during exercise after ischaemia. Following post-exercise ischaemia, time to exhaustion is reduced similarly in males and females due to a lower provision of energy by oxidative phosphorylation. The lowerV O 2 during exercise after ischaemia is partly compensated for by increasing the rate of substrate-level phosphorylation to 1.8-fold above that observed when the exercise is performed after recovery with free circulation, with similar responses in males and females. The latter demonstrates that task failure is not caused by the achievement of a limiting glycolytic rate. Moreover, the fact that the O 2 deficit was the same during the bouts performed after ischaemia and free circulation implies that fatigue occurs after a defined amount of energy has been produced through substrate-level phosphorylation, and that the mechanism causing task failure is likely linked to a direct or indirect effect of the metabolites and ion disturbances associated with substrate-level phosphorylation in critical subcellular compartments (Fitts, 1994;Allen et al. 2008).
Post-exercise ischaemia may have contributed to either facilitate or impair recovery by several mechanisms. Post-exercise ischaemia may have facilitated recovery, first, through an increase of muscle temperature (Coupland et al. 2001;Pedersen et al. 2003) caused by the elevated metabolic rate during post-exercise ischaemia , which combined with the occlusion of the circulation, would limit the transfer of heat (Westerblad et al. 1997;Pedersen et al. 2003); second, by the positive effects of lactate and H + in the recovery of sarcolemmal excitability (Pedersen et al. 2003;Allen & Westerblad, 2004;Copithorne et al. 2020); third, by reducing the rate of RONS production by mitochondrial respiration due to the lack of O 2 (Bruton et al. 2008); fourth, by facilitating the entry of P i into the sarcoplasmic reticulum, where it might precipitate with Ca 2+ (Fryer et al. 1995;Ferreira et al. 2021;Hinks et al. 2021); and fifth, by increasing the firing of group III/IV muscle afferents, facilitating motoneuron discharge, as recently suggested by neurophysiological experiments in humans (Brandner et al. 2015;Copithorne et al. 2020).
Alternatively, post-exercise ischaemia could have exerted a negative influence in the recovery of muscle contractile capacity by several mechanisms. First, it may have done so by increasing nicotinamide adenine dinucleotide phosphate (NAPDH) oxidase (NOX) and xanthine oxidase (XO) RONS production at the start of the exercise by an ischaemia-reperfusion mechanism, which may have elicited the production of OH· and H 2 O 2 through the Fenton reaction (Morales-Alamo & Calbet, 2014;Larsen et al. 2016;Calbet et al. 2020). This is supported by the fact that administration of polyphenolic compounds with potent direct (by quenching RONS) and indirect (by inhibiting XO and NOX RONS production) antioxidant effects enhances O 2 consumption and performance when subjects are requested to sprint maximally with normal blood flow after a short period of post-exhaustive exercise ischaemia (Gelabert-Rebato et al. 2018;Gelabert-Rebato et al. 2019b). Second, it may have exerted a negative influence by reducing the levels of muscle glycogen (Gejl et al. 2014;Jensen et al. 2020) due to the reliance on glycolysis to maintain cellular viability in the absence of O 2 (Morales-Alamo et al. 2015). However, since the reduction in muscle excitability is rather small (Senefeld et al. 2018;Copithorne et al. 2020) and all subjects were able to perform the second and the third exercise bouts despite the occlusions, our results suggest that task failure is elicited by a muscle mechanism that can be recovered very quickly (within seconds) despite the progressive accumulation of metabolites during ischaemia, and/or by a mechanism residing outside the skeletal muscles, i.e. in the central nervous system (Kayser, 2003;Marcora & Staiano, 2010;Calbet et al. 2015;Morales-Alamo et al. 2015;Torres-Peralta et al. 2016;Angius et al. 2018).

During a single bout of supramaximal exercise, females produce less ATP through substrate-level phosphorylation, but both sexes recruit anaerobic pathways similarly in subsequent bouts of supramaximal exercise
In the present investigation, females incurred a lower O 2 deficit during the first exercise bout and had a lower increase in RER during the exercise, both compatible with less activation of glycolysis during the first exercise bout. This interpretation agrees with the low RER values observed at exhaustion in MacArdle's disease patients despite remarkable hyperventilation. MacArdle's disease patients cannot rely on the glycolytic component of substrate level phosphorylation during high-intensity exercise (Hagberg, 1982).
Our results concur with the reported lower muscle lactate accumulation after repeated sprint exercise in females than males (Jacobs et al. 1983), as well as lower activation of glycogenolysis in females than males in type I fibres after repeated Wingate tests (Esbjornsson-Liljedahl et al. 2002). Compared to males, females have lower activities of some enzymes regulating the glycolytic rate, such as lactate dehydrogenase (Green et al. 1984;Simoneau & Bouchard, 1989;Esbjornsson Liljedahl et al. 1996), hexokinase (Green et al. 1984;Simoneau & Bouchard, 1989), phosphofructokinase (Green et al. 1984;Simoneau & Bouchard, 1989;Jaworowski et al. 2002), pyruvate kinase (Green et al. 1984) and phosphorylase (Green et al. 1984). Although differences in enzyme activities could account for a lower maximal glycolytic rate, they cannot explain why during the second and third bouts the contribution of glycolysis was the same in both sexes, as demonstrated by the almost identical O 2 deficits and calculated glycolytic rates observed in the second and third bouts performed after ischaemia.
Interestingly, the O 2 deficits were similar in the bouts with free circulation compared to the bouts performed after ischaemia. It should be considered that in the bouts performed after ischaemia, the O 2 stores were likely exhausted at the start of the exercise, meaning that all the deficit corresponds to the glycolytic component of substrate-level phosphorylation. In contrast, during recovery with free circulation some PCr resynthesis and replenishment of muscle O 2 stores should have occurred (Harris et al. 1975;Bangsbo et al. 1990). Therefore, the actual anaerobic energy production must have been 10-20% lower in the bouts preceded by free circulation recovery than calculated from the assessment of the O 2 deficit. The latter would match even more closely the actual O 2 deficits of the bouts preceded and not preceded by ischaemia.
Since performance was 42% greater after recovery with free circulation and the rate of utilization of the O 2 deficit 1.8-fold higher after exercise preceded by ischaemia, these results strongly suggest that task failure occurs after the utilization of a fixed amount of the anaerobic capacity. This is likely related to the accumulation of muscle metabolites altering sarcolemmal excitability, which is rapidly restored upon cessation of contractile activity and by the application of ischaemia (Copithorne et al. 2020) or the sequestration of P i into the sarcoendoplasmic reticulum (Ferreira et al. 2021). The recovery with free circulation should have allowed a higher recovery (yet incomplete) of ion perturbations, PCr and O 2 stores, permitting a slowing of the glycolytic substrate-level phosphorylation rate and, therefore, prolonging the exercise bout.
To recapitulate, our data and previous research support that females have a lower anaerobic capacity than men and this may be the main reason why they produce a smaller amount of energy through anaerobic pathways than men in the first exercise bout. In the subsequent exercise bout, the male advantage disappears likely due to the partial inhibition of glycolysis elicited by the accumulation of metabolites (Gaitanos et al. 1993), with a greater influence in a muscle with more type II fibres, which are more abundant in males than females. A lower capacity to utilize PCr in females than males is unlikely since no sex differences have been reported for PCr utilization during sprint exercise (Esbjornsson-Liljedahl et al. 2002). J Physiol 599.16

Moderately active males and females have similaṙ V O 2 max values when expressed per kg of active lean mass
Males and females had a similarV O 2 max when normalized to the lean mass of the lower extremities, in agreement with our previous observation (Perez-Gomez et al. 2008). Consequently, during the exercise at 120% ofV O 2 max (first bout) the mean rate ofV O 2 per kg of muscle mass was almost identical, as it was also during the two last bouts of exercise with or without occlusion during the recovery periods. This indicates that the relative intensity was well matched between sexes and the metabolic demand elicited by the supramaximal exercise and the recoveries (with or without occlusion) on the subsequent exercise bouts was similar in both sexes. In both cases, the occlusion limited the capacity to utilize O 2 in the subsequent bout, potentially through partial inhibition of oxidative phosphorylation (Jubrias et al. 2003) or due to reduced O 2 delivery to the contracting muscles. The fact that during the second and third bouts of exercise with free circulation both males and females were able to reach mean metabolic rates close toV O 2 max indicates that there was minimal if any inhibition of oxidative phosphorylation. Since during whole-body exerciseV O 2 max is limited by O 2 delivery, the fact that during the second and third bout preceded by free circulation recoveryV O 2 was close to maximal can only be explained by an almost maximal O 2 conductance and mitochondrial utilization of O 2 in both sexes. Since females have lower haemoglobin concentration, this can only be explained by a larger O 2 extraction. This agrees with the recent observation of higher complex I, complex II and oxidative phosphorylation capacity (complex I + complex II) respiratory capacity (normalized to the amount of mitochondrial protein content) in females than males (Cardinale et al. 2018). The latter finding has the effect of lowering mitochondrial activation rate, lowering P 50 (higher mitochondrial O 2 affinity) in females compared to males (Cardinale et al. 2018) and enhancing O 2 extraction capacity (Cardinale et al. 2019), explaining the similarV O 2 observed in the present investigation. In agreement, our NIRS data also indicate greater O 2 extraction capacity in females.

Oxygen debt
Here we report the first assessment of post-exercise O 2 debt of the lower extremities after whole-body exercise to exhaustion using a novel methodological approach. We have demonstrated that the LLM accounts for just 24% of the total O 2 debt paid during the first 20 s of recovery, with the observed values being almost identical between sexes when normalized to the LLM. Our results agree well with the values obtained for quadriceps muscle after exhaustive single-leg knee-extension exercise (Bangsbo et al. 1990). The fact that there were no sex differences indicates a similar muscle metabolism after exhaustion in males and females. Only a minor fraction of the O 2 debt is expended in PCr, ATP and glycogen resynthesis (Bangsbo et al. 1990); the rest likely represents the energy expended by the ATPases involved in ion regulation, as previously discussed . Interestingly, the value of the O 2 debt incurred in the first 20 s of the exercise represented 24% of theV O 2 max , a metabolic rate slightly lower than the 32% ofV O 2 max measured for the vastus lateralis during 60 s ischaemia applied at exhaustion ). The small difference may relate to differences in the metabolic rates during recovery between different muscles of the lower extremities (Heinonen et al. 2012).

Limitations
The main limitation of this study resides in the indirect measurement of substrate-level phosphorylation energy turnover. For this purpose, we have assumed constant muscle efficiency during supramaximal exercise regardless of the level of fatigue (Medbo & Tabata, 1993;Jubrias et al. 2008). Although in vitro experiments indicate that the muscle efficiency may deteriorate with fatigue (Barclay, 1996), conclusive evidence is lacking in humans (Woledge, 1998). Nonetheless, the effect of severe fatigue on muscle efficiency is likely small during dynamic contractions in humans (Myburgh, 2004;Jubrias et al. 2008;Morales-Alamo et al. 2015). In our previous study, we applied 10 and 60 s occlusions after incremental exercise to exhaustion . Despite a greater level of metabolic disturbance after 60 than 10 s occlusions, theV O 2 per power output was lower after 60 s of ischemia than after 10 s, which is not compatible with a marked reduction of muscle efficiency during supramaximal exercise under severe fatigue conditions . Nevertheless, if muscle efficiency should deteriorate with fatigue, we could have underestimated the actual glycolytic rates. However, the latter would imply that the metabolic functional reserve is even greater than we have estimated.
Although the menstrual cycle phase was not controlled, most studies have shown no impact of the menstrual cycle on high-intensity exercise performance (Botcazou et al. 2006;Bushman et al. 2006) nor on oxygen deficit (Shaharudin et al. 2011).
Despite the high number of subjects included in the study, we cannot rule out a type 2 error in some sex comparisons. For example, males achieved 10 and 12% higher glycolytic rates during the second and third exercise bouts preceded by ischaemic recovery, respectively. Since we report the size effects, these can be used to define the sample size required to determine whether this sex difference is real, and at least 404 volunteers (202 males and 202 females) will be required. On the other hand, our conclusion regarding the lack of sex differences in most of the variables assessed is robust given the small magnitude of the differences observed. In summary, our data indicate that both males and females fatigue to a similar extent and recruit the aerobic and anaerobic energy pathways during repeated supramaximal exercise in a remarkably similar fashion when differences in lean mass (a surrogate of muscle mass in the present study) are considered. Males have higher anaerobic capacity than females, even after normalization to the lower extremities lean mass, but this advantage is only manifested during the first bout of supramaximal exercise. Through the utilization of post-exercise ischaemia, we have shown that both sexes have a functional reserve which depends on the glycolytic component of substrate-level phosphorylation. In the fatigued state, the amount of exercise that can be performed at supramaximal exercise intensities depends on the rate of utilization of the remaining functional reserve, which supplements the aerobic energy production. We have shown that after ischaemia, muscle O 2 extraction capacity is increased, but muscleV O 2 is reduced similarly in males and females. Metaboreflex activation during post-exercise ischaemia and the O 2 debt per kg of active lean mass are also similar in males and females after supramaximal exercise. Finally, P ETCO 2 and brain oxygenation are lower in females than males, without apparent negative repercussion on performance. Females had no faster recovery of performance after accounting for sex differences in lean mass.  Table 4. [Colour figure can be viewed at wileyonlinelibrary.com]  -2017-13). In this study, volunteers were catheterized in both femoral veins and equipped with NIRS optodes in both vastus lateralis before doing incremental exercise to exhaustion, followed immediately by ischaemia. Data obtained in the first 12 subjects (seven males and five females) were available to validate the O 2 extraction index here defined. The seven males (age = 22.0 ± 2.0 years, height = 180.1 ± 9.4 cm, body mass = 76.0 ± 6.6 kg, body fat = 19.6 ± 6.8%, andV O 2 max = 40.4 ± 6.8 ml kg −1 min −1 ) and five females (age = 22.6 ± 2.2 years, height = 161.1 ± 2.7 cm, body mass = 52.8 ± 7.3 kg, body fat = 26.4 ± 1.1%, andV O 2 max = 36.3 ± 5.3 ml kg −1 min −1 ) were healthy, physically active sports science students. Subjects were catheterized in both femoral veins as previously described . Also, a hand vein was catheterized and heated to obtain arterialized blood. After catheterization, NIRS optodes were carefully placed over both vastus lateralis, as described in Methods by the same research personnel. The same device used in the main experiments was employed in the validation study Hamamatsu,Japan). Arterial saturation was estimated by haemoglobin oxygen saturation (S pO 2 ) measured with a finger pulse oximeter (OEM III module, 4549-000, Plymouth, MN, USA). The exercise protocol J Physiol 599.16 consisted of an incremental exercise test to exhaustion starting at 20 W, which was increased by 15 and 20 W every 3 min in females and males, respectively, until the RER was >1.00. During this phase, 1 ml blood samples were withdrawn with heparinized Radiometer syringes (PICO50, Radiometer Medical ApS, Brønshøj, Denmark) simultaneously from both femoral veins, 30 s before each increase of the load. After completing the highest workload with an RER ≥ 1.00, the ergometer was unloaded while the subject continued pedalling at a slow cadence (∼30-40 rpm) for 2 min. At the end of the 2 min, the subjects started an incremental exercise test at the load reached in the previous phase, increased by 10 and 15 W every minute until exhaustion. Close to exhaustion, a final blood sample was obtained from both femoral veins and the heated hand vein. At task failure, the circulation of one leg, randomly assigned, was occluded instantaneously with a cuff connected to the same rapid cuff inflator described in Methods (Hokanson E20 AG101, Bellevue, WA, USA), using the same pressure, i.e. 300 mmHg.

Table 5. Relationship between leg O 2 extraction and TOI (tissue oxygenation index) O 2 extraction index computed as the difference between TOI observed (TOI OBV ) and the minimum TOI registered (TOI MIN ) during a 60 s ischaemia applied instantaneously at task failure, at the end of an incremental exercise to exhaustion on the cycle ergometer
The blood samples were immediately analysed to determine blood gases, electrolytes and haemoglobin concentration (ABL90, Radiometer), as previously reported . The arterial O 2 content (C aO 2 ) was calculated using S pO 2 . The concentration of O 2 in blood samples was calculated from the saturation and [Hb] (i.e. (1.34 × [Hb] × S O 2 ) + (0.03 × P O 2 )). The femoral vein O 2 extraction was calculated for each leg as (C aO 2 − C fvO 2 ) × 100/C aO 2 , where C fvO 2 represents the O 2 concentration in the corresponding femoral vein. The TOI data were averaged every 20 s around the sampling point. During the 1 min post-exercise ischaemia, the TOIs were averaged every 5 s, and the lowest 5 s average was used as the minimum TOI to calculate the extraction index. The relationship between O 2 extraction, TOI, and NIRS-derived extraction index, was determined by linear regression (SPSS v.15.0 for Windows; SPSS Inc., Chicago, IL, USA).
As depicted in Fig. 5, a linear relationship was observed during exercise between leg O 2 extraction and TOI values for each leg in all subjects. In Tables 4 and 5, the equation parameters defining the relationship between leg O 2 extraction and TOI (Table 4 and Fig. 5) and between leg O 2 extraction and the TOI-derived O 2 extraction index (i.e. TOI-derived O 2 extraction index = TOI OBV − TOI MIN ) are reported for all subjects and for each leg. The intercepts and slopes of the regression lines were compared between legs (two-tailed paired t-test) and between males and females (two-tailed unpaired t-test). No significant differences were observed either between legs or between sexes. Similar results were obtained when the O 2 content of the femoral vein was regressed against TOI values.