Sex Differences in Neuromuscular Fatigue and Changes in Cost of Running after Mountain Trail Races of Various Distances

Abstract

Introduction 

Women have been shown to experience less neuromuscular fatigue than men in knee extensors (KE) and less peripheral fatigue in plantar flexors (PF) after ultratrail running, but it is unknown if these differences exist for shorter trail running races and whether this may impact running economy. The purpose of this study was to characterize sex differences in fatigability over a range of running distances and to examine possible differences in the postrace alteration of the cost of running (Cr).

Methods 

Eighteen pairs of men and women were matched by performance after completing different races ranging from 40 to 171 km, divided into SHORT versus LONG races (<60 and >100 km, respectively). Neuromuscular function and Cr were tested before and after each race. Neuromuscular function was evaluated on both KE and PF with voluntary and evoked contractions using electrical nerve (KE and PF) and transcranial magnetic (KE) stimulation. Oxygen uptake, respiratory exchange ratio, and ventilation were measured on a treadmill and used to calculate Cr.

Results 

Compared with men, women displayed a smaller decrease in maximal strength in KE (−36% vs −27%, respectively, P < 0.01), independent of race distance. In SHORT only, women displayed less peripheral fatigue in PF compared with men (Δ peak twitch: −10% vs −24%, respectively, P < 0.05). Cr increased similarly in men and women.

Conclusions 

Women experience less neuromuscular fatigue than men after both “classic” and “extreme” prolonged running exercises but this does not impact the degradation of the energy Cr.

In recent years, participation in trail and ultratrail (defined by the International Trail Running Association as trail races with a distance >80 km) running has increased considerably, with reports of numerous events having to limit registration numbers due to overdemand. However, participation in trail and ultratrail events remains considerably more popular in men than women. For instance, at the UTMB® 2019, one of the most popular ultratrail events in the world, only 257 of 2543 starters, that is, 10%, were women. Despite the low participation rates for women, ultratrail running is one of the rare disciplines in which women have outperformed men in some events (¹).

The increasing enthusiasm for trail running has led scientists and researchers to take a particular interest in the physiological and neuromuscular (NM) consequences of prolonged trail running events. Neuromuscular fatigue, also known as performance fatigability, can be defined as the progressive change that occurs in the central nervous system and/or muscles due to exercise, resulting in a force output that is less than anticipated for a given voluntary contraction or stimulation (²). It has been reported from isometric studies in isolated muscle groups that women are less fatigable than men, particularly at submaximal intensities (³). Although part of the sex difference in fatigability has been attributed to greater strength and consequent higher blood flow occlusion in men, previous studies have demonstrated that the sex difference in fatigability during submaximal isometric contractions persists when men and women were matched for strength (^4,5). Furthermore, Ansdell et al. (⁶) assessed sex differences in fatigability relative to the critical power and showed that women are less fatigable during intermittent isometric contractions above critical torque, and experience less muscle deoxygenation at intensities both above and below critical torque. Thus, it was suggested that the lower fatigability was mediated by lower muscle deoxygenation, likely due, at least in part, to a higher proportional area of type-I fibers (⁷), greater muscle capillarization (⁷) and greater vasodilatory capacity in women compared with men (⁸). Although the studies mentioned above [see Hunter (³) for a review] bring interesting insights to the mechanisms responsible for sex differences in fatigability after single-joint tasks, whether these findings are transferable to whole-body exercise, such as running, is unclear.

It is well known that trail and ultratrail running induces alterations to both central and peripheral components of NM function (^9–12). The reduction in maximal strength increases with exercise duration until a duration corresponding to ⁓100 miles (^13,14). Based on the lower level of fatigability in women during isometric tasks, as well as the out-performance of men by women in recent ultratrail events, one may be tempted to believe that women experience less NM fatigue during prolonged exercises. Surprisingly, this question remains understudied. After 2 h of running on a treadmill at gas exchange threshold, Glace et al. (¹⁵) reported a significant decrease in isokinetic maximal voluntary knee flexor and knee extensor (KE) forces in men only when measured at low angular velocities (60°·s⁻¹). On the contrary, Boccia et al. (¹⁶) reported a similar decrease between sexes in maximal KE force after a half-marathon performed in race conditions by amateur runners. Over a much longer distance, our group (¹²) assessed NM fatigue in women and men matched by relative level of performance after a 110-km ultratrail-running race. We observed that women showed (i) less peripheral fatigue in the plantar flexors (PF) and (ii) a lower decrease in maximal force loss in KE compared with their male counterparts. A recent review compiled physiological outcomes that can confer women an advantage over longer distances (¹⁷), such as greater distribution of type I fiber and better substrate efficiency (higher rate of lipid oxidation and lower carbohydrate utilization); however, a systematic comparison between men and women over various running distances has yet to be performed. In addition, whether or not this difference is because of NM fatigue resistance versus the competition intention of the participants (i.e., the effort put into the race) is not known. Indeed, this latter factor may influence the race-induced fatigue but has never been considered when comparing men and women. Sex differences, such as greater sporting motivation and competitiveness in men, could be expected (¹⁸), although the physiological impact of these psychological characteristics is still unknown.

In addition to NM fatigue, prolonged running may induce alterations in energy cost of running (Cr) (¹⁹). Interestingly, after a 2-h treadmill run, Glace et al. (¹⁵) reported significant maximal knee strength loss (see above) and an increase in Cr in men only. Additional motor unit recruitment to compensate for muscle fatigue could partly explain greater oxygen uptake (V˙O₂) demand (^19,20). As such, it is possible that attenuated muscle fatigue in women could reduce the compensatory increase in motor unit recruitment and thus V˙O₂ demand. However, it should be noted that conflicting evidence exists surrounding the effects of muscle fatigue on changes in V˙O₂, with some studies reporting no interaction between these variables (^21–24). In addition to Glace et al. (¹⁵) only two other studies have to the best of our knowledge examined sex difference on Cr change with fatigue and observed similar changes in men and women: one after a 1-h run at marathon pace (²⁵) and another one after a 5-km run at 80% to 85% of maximal oxygen uptake (V˙O_2max) (²⁶). However, these experiments were conducted on a treadmill, that is, not in ecologically valid conditions. Furthermore, because an effect of distance on sex differences in fatigability may exist, a comparison of the change in Cr between sexes for various distances is warranted.

The aim of this study was to further characterize sex differences in NM fatigue and Cr changes by examining various running distances from 40 to 171 km. It was hypothesized that (i) women would exhibit less NM fatigue, particularly its peripheral component, and be better able to preserve their Cr than their male counterparts; and (ii) sex differences would increase with race distance.

METHODS

Participants

Seventy-five experienced trail runners were included in the study after medical examination (Table 1). Specific qualifying race criteria need to be met to register for the races of the study. To participate in the study, applicants have to meet inclusion criteria: men and women of older than 18 yr, free from muscular, bone or joint injuries, and free from neurologic disease. Participants were excluded if (i) they were taking neuroactive substances that can alter corticospinal excitability; (ii) they have contraindication to experimental procedures, including transcranial magnetic stimulation; and (iii) they were currently participating in a structured exercise program. Four participants dropped out of the study, 17 runners did not complete the race, and three runners were excluded from analysis because they did not complete the race in a competitive manner, because of completing the race by accompanying another race participant (Fig. 1). Thus, 56 participants participated in the posttest session of this study. A subgroup of 36 finishers was further separated into two groups of 18 men and 18 women matched by relative performance to the first male and the first female of their specific race (4, 5, 4, 1, and 4 pairs came from the De Martigny-Combe à Chamonix, Orsières-Champex-Chamonix, Courmayeur-Champex-Chamonix, Sur les Traces des Ducs de Savoie, and Ultra Trail du Mont-Blanc, respectively), respectively (158% ± 11% vs 158% ± 9%, respectively; P = 0.897). All pairs of male and female participants completed the same race. The male and female winners of each race had an International Trail Running Association (ITRA) performance index that was, on average, 95% of the world best ITRA performance index in their respective race category (more details are shown in the supplemental content; see Table, Supplemental Digital Content, best male and female levels and participation rate at the different races of the UTMB event, https://links.lww.com/MSS/C342). This shows that, on average, the relative level of the female winner is equivalent to the level of the male winner for each race despite a lower rate of participation in women. The study was performed according to the Declaration of Helsinki and was approved by the ethics committee (Comité de Protection des Personnes Ouest VI) and was registered at ClinicalTrials.gov (NCT04025138). All participants gave their written informed consent before their participation. The present study was part of a larger study investigating the effect of trail and ultratrail racing on different physiological and biomechanical responses in men and women.

TABLE 1 - Race details and participant characteristics.

Race	Group	Distance (km)	D+ (m)	ITRA KM-Effort	ITRA Category	Male Winning Time	Participants (n)	Age (yrs)	Height (cm)	Weight (kg)	V˙O2max (mL·kg−1·min−1)
MCC®	SHORT	40	2300	63	S	3:40:46	F: 7	39 ± 12	169 ± 6	62 ± 6	52.5 ± 4.0
							M: 7	34 ± 8	179 ± 6	73 ± 10	60.8 ± 14.7
OCC®		56	3500	91	M	5:19:24	F: 5	36 ± 8	168 ± 6	60 ± 5	54.5 ± 5.5
							M: 7	35 ± 7	180 ± 11	76 ± 15	64.7 ± 9.3
CCC®	LONG	101	6100	162	XL	10:28:49	F: 5	33 ± 6	163 ± 4	57 ± 8	57.9 ± 8.2
							M: 13	37 ± 10	180 ± 3	74 ± 6	61.5 ± 6.4
TDS®		145	9100	236	XXL	18:03:06	F: 2	47 ± 4	162 ± 10	54 ± 11	54.8 ± 5.7
							M: 4	43 ± 8	182 ± 5	75 ± 17	54.0 ± 7.6
UTMB®		171	10,300	274	XXL	20:19:07	F: 7	35 ± 3	164 ± 7	57 ± 6	53.3 ± 4.3
							M: 18	39 ± 7	177 ± 6	72 ± 9	61.5 ± 7.0

Values are presented as mean ± SD.

D+, positive elevation; F, females; M, males; MCC, De Martigny-Combe à Chamonix; OCC, Orsières-Champex-Chamonix; CCC, Courmayeur-Champex-Chamonix; TDS, Sur les Traces des Ducs de Savoie; UTMB, Ultra Trail du Mont-Blanc.

FIGURE 1: Inclusion table. M, males. F, females. *Not tested in POST because of inability to perform the test due to extreme fatigue or specific injury. **Excluded from analysis because not meeting the criteria (see detail in the text).

Design

Participants performed three testing sessions in total, comprising one familiarization session. One month before the race, participants visited the laboratory for a familiarization session. The other two testing sessions were performed before (PRE) and after (POST) a trail running race to assess NM function, Cr, blood parameters, and participants’ sensations in an isolated room close to the finish line (Fig. 2). The PRE session was completed at least 24 h and less than 128 h before the race, whereas the POST session was completed as soon as possible after the race. The delay between the end of the race and KE NM evaluation was 36 ± 14 min, 44 ± 13 min for PF, and 80 ± 19 min for Cr. The testing sessions were performed in a laboratory installed in a building located around 500 m from the finishing line at an altitude on 1035 m. After finishing their race, runners were offered food and drink and were allowed to lie down if necessary, for medical reasons.

FIGURE 2: Description of a testing session to assess the NM function (panel A) and energy Cr (panel B).

Familiarization Session

The familiarization session comprised a medical examination, familiarization with the NM testing protocol and a maximal running test to determine V˙O_2max. The NM familiarization consisted of sustained submaximal and maximal isometric contractions (MVC) on both KE and PF muscles. Participants were then familiarized with peripheral electrical nerve stimulation on both femoral and tibial nerves. During KE contractions, participants were also familiarized with transcranial magnetic stimulations (TMS). The maximal running test consisted of incremental running to exhaustion on a treadmill set with a 12% slope. Participants started at 5 to 6 km·h⁻¹, depending on their fitness, and the speed increased by 0.5 km·h⁻¹·min⁻¹ until exhaustion.

Trail Running Race

Participants completed various races across 5 d at the Ultra-Trail du Mont-Blanc® (the race characteristics are detailed in Table 1). In addition to sex, participants were further subdivided into two groups by distance of race completed: SHORT (<60 km) versus LONG (>100 km). The event took place at the end of August 2019 under summer temperatures, ranging from 11°C (at night) to 31°C in Chamonix throughout the duration of the event, with a weather very similar between days (data derived from https://www.timeanddate.com/weather). All races were mountain trail or ultratrail races, mostly composed of trail sections with a range of technical difficulties and gradients.

NM Testing

Neuromuscular function was assessed on both KE and PF muscles. The protocol was the same between PRE and POST sessions, apart from the inclusion of a standardized warm-up of 10 submaximal isometric contractions in the PRE session only. During the POST session, the order of KE and PF NM function tests depended on the availability of testing stations to minimize the delay before assessment.

The NM testing protocol (Fig. 2) consisted of one MVC without stimulation, followed by two MVC with paired-pulse (100 Hz) peripheral nerve stimulation (PNS) delivered at peak torque, and on relaxed muscle separated by 3 s (100- and 10-Hz paired pulse and single pulse; Db100, Db10, and Pt, respectively). During MVC, participants were instructed to contract as strongly as possible for ~4 s. A 30-s resting period separated the first two MVC. Electrical PNS were delivered on the femoral nerve and on the tibial nerve for KE and PF, respectively. Further details on torque recordings and PNS are provided below.

For KE, participants performed two series of three contractions with TMS delivered at the target torque level (100%, 75%, and 50% MVC). Peripheral nerve stimulation was also delivered after the TMS stimulation during the 50% contraction. Real-time visual feedback of the torque level was given to the participants so that they could maintain the desired level of contraction. Contractions were separated by 5 s and series by ~30 s at PRE and ~10 s at POST.

Torque and EMG recordings

Knee-extensor torque was measured during MVC and evoked contractions with an isometric knee dynamometer (ARS dynamometry, SP2, Ltd., Ljubljana, Slovenia). Participants were seated on the chair in an upright position with the hip and the right knee at 90° of flexion. The right leg was attached to the chair by a noncompliant strap just proximal to the malleoli of the ankle joint. Hips were securely strapped to maintain the position during contractions.

Plantarflexor torque was measured by an instrumented pedal (CS1060 300 Nm; FGP Sensors Les Clayes Sous Bois, France). Participants were seated in a custom-built chair with hip, knee, and ankle angles of 90°. The chest was strapped to the chair, and heel and forefoot were securely attached to the pedal with noncompliant straps to avoid displacement of the foot during MVC.

EMG activity of KE (vastus lateralis [VL]) and PF (gastrocnemius medialis [GM] and soleus [SOL], respectively) was recorded using pairs of self-adhesive surface electrodes (Meditrace 100; Covidien, Mansfield, MA) with a 10-mm recording diameter. The electrodes were placed in bipolar configuration and spaced by a 30-mm interelectrode distance. A reference electrode was placed on the right patella and the right medial malleolus for KE and PF, respectively. Before placing the electrodes, the skin was prepared to obtain low impedance (<5 kΩ) by shaving, gently abrading the skin, and cleaning it with alcohol. The electrode placement was drawn on the skin using a permanent marker toensure the same placement between PRE and POST sessions. EMG data were recorded with PowerLab system (16/30-ML880/P; ADInstruments, Bella Vista, Australia) with a sampling frequency of 2000 Hz. The EMG signal was amplified with octal bioamplifier (Octal Bioamp, ML138; ADInstruments) with a bandwidth frequency ranging from 5 to 500 Hz (common mode rejection ratio, 85 dB; gain, 500), transmitted to the computer, and analyzed with LabChart 8 software (ADInstruments).

Electrical nerve stimulation

Single electrical stimuli of 1 ms duration and 400 V maximal output voltage were delivered via constant-current stimulator (DS7A or DS7R; Digitimer, Welwyn Garden City, Hertfordshire, United Kingdom) to both the right femoral and the tibial nerves. For the femoral nerve, stimulations were sent via a 30-mm-diameter surface cathode manually pressed into the femoral triangle (Meditrace 100) and a 10 × 5 cm self-adhesive stimulation electrode (Medicompex SA, Ecublens, Switzerland) located in the gluteal fold. Stimulations were delivered to the tibial nerve via a bipolar bar stimulating electrode with 30-mm anode-cathode spacing (Bipolar Felt Pad Stimulating Electrode part no. E.SB020/4 mm; Digitimer) placed in the popliteal fossa and parallel to the nerve.

For both KE and PF, the optimal intensity was determined by delivering single stimuli incrementally to the relaxed muscles until maximal M-wave (M_max) and the torque response plateaued. A stimulation intensity of 130% of the intensity that produced the maximal torque response, and M-wave amplitude was used to ensure supramaximality of the twitch responses. For both KE and PF, the stimulation intensity was determined at the beginning of both the PRE and the POST sessions.

Transcranial magnetic stimulation

Single TMS pulses were manually delivered to obtain motor-evoked potentials (MEP) and superimposed twitches (SIT) during isometric KE contractions. Transcranial magnetic stimulation was delivered to the left motor cortex using a magnetic stimulator (Magstim 200²; The Magstim Company Ltd., Whitland, United Kingdom) with a 110-mm concave double-cone coil (maximum output of 1.4 T) to induce a posteroanterior current. Participants wore a swim cap on which the optimal stimulation position was drawn to ensure a consistent coil position during the protocol. At the beginning of the protocol, the vertex was identified by drawing a line between the preauricular points and from nasion to inion. Six stimulation spots were drawn on the swim cap every centimeter from the vertex to 2 cm posterior along the nasal–inion line and 1 cm to the left of those three points. A single stimulation was delivered over each spot. The optimal coil position was determined during 20% MVC contractions and was chosen as the site that elicited the greatest SIT and MEP response. Optimal stimulus intensity was defined as the lowest stimulus intensity eliciting maximal MEP amplitude during short voluntary contractions at 20% MVC (²⁷). Two stimulations were delivered at each intensity. The same TMS intensity and coil position was used in POST. During the protocol, participants were asked to recontract as quickly as possible to the prestimulus torque level after TMS delivery.

Energy Cr

Following the NM assessments, participants were asked to run for two bouts of 4 min on a motorized treadmill (Pulsar 3p, h/p/cosmos, Munich, Germany) with a 1-min resting period in between 4 min at 0% incline (FLAT) and 4 min at 15% uphill incline (UH). The speed was set at 90% and 70% of the average speed sustained during the last minute of the maximal test performed during the familiarization for FLAT and UH, respectively. During this test, heart rate and breath-by-breath V˙O₂ were measured using a portable system (Metamax 3B; Cortex Biophysik, Leipzig, Germany).

Blood Parameters

Peripheral venous blood samples were taken from an antecubital vein of the participants at PRE and POST sessions. Blood samples were collected in nonadditive collection tubes under sterile conditions. Tubes were then centrifuged for 10 min at 1000g and 4°C. A Cobas C501 integrated system (Roche, Basel, Switzerland) was used for simultaneous assay of C-reactive protein (CRP) and creatine phosphokinase (CPK) with reagents from the manufacturer.

Race Intensity

The average intensity at which the participants performed their race was calculated using the ratio of the mean speed achieved during their race relative to the maximal speed reached on the treadmill during the V˙O_2max test performed during the familiarization session. This parameter was calculated for both men and women and for both SHORT and LONG and was expressed in percentage of the speed reached during the V˙O_2max test.

Participants’ Sensations and Competitive Intention

In the PRE and POST races, participants were asked to report their fatigue, perceived pain (for both KE and PF), and level of gastrointestinal discomfort on a 10-cm visual analog scale. Each scale was anchored with the verbal descriptors “not at all” and “extremely.” Furthermore, at the end of the race, participants were asked their competitive intention going into the race by rating from 0 to 10, with 0 being performing the race as fast as possible (i.e., maximum effort) and 10 being for fun (i.e., minimal effort). The exact question provided to the participants was the following: How would you rate your state of mind between 0 and 10, between pleasure mode and competition mode? 0: Competition mode (I tried to do the best time possible). 10: Fun mode (my only goal was to finish the race).

Data Analysis

Voluntary and evoked torque

The maximal torque values were determined as the highest peak torque recorded from the MVC contractions (of three MVC for PF and of five MVC for KE). The ratio of the amplitude of the superimposed doublet to the resting doublet was then calculated to obtain the percentage of voluntary activation (VA) as follows:

${\mathrm{VA}}_{\mathrm{PNS}}=\left(1-\left[\frac{\text{superimposed}\mathrm{Db}100}{\text{resting}\mathrm{Db}100}\right]\right)\times 100$

The VA and amplitudes of Db100, Db10, and Pt were measured on the trial where the torque value was the highest when the superimposed doublet was delivered. The ratio of Db10 to Db100 (Db10:Db100) was calculated to evaluate the presence of low-frequency fatigue (²⁸).

EMG

M-wave peak-to-peak amplitude (Mmax) was analyzed from the single pulses elicited when the muscle was at rest. EMG root mean square (RMS) was calculated over a 500-ms period after the torque had reached a plateau and before the delivery of PNS during the best MVC trial. The RMS was then normalized to Mmax (RMS/Mmax).

TMS

Vastus lateralis peak-to-peak MEP amplitude (MEP_AMP) and MEP area (MEP_AREA) were obtained at each contraction level (MEP100, MEP75, and MEP50 corresponding to 100%, 75% and 50% MVC contraction, respectively) and used as an index of corticospinal excitability. Motor-evoked potential amplitude and area were then normalized to the amplitude and area, respectively, of the M-wave obtained during the 50% MVC contraction. Transcranial magnetic stimulation VA (VA_TMS) was measured by the twitch interpolation technique (²⁹). The estimated resting twitch was determined as the y-intercept of a linear regression of SIT amplitudes elicited by optimal TMS intensity and absolute voluntary force during the two series of the three contractions at 50%, 75%, and 100% MVC. Estimated resting twitch regression was considered not linear for two participants only (r < 0.9) who were then discarded from analyses. In all other participants, the regression was linear (r > 0.9) for at least one series at both PRE and POST sessions (³⁰). VA_TMS was then calculated with the following equation:

${\mathrm{VA}}_{\mathrm{TMS}}=\left(1-\left[\frac{{\mathrm{SIT}}_{\mathrm{MVC}}}{\text{estimated resting twitch}}\right]\right)\times 100$

The duration of the silent period, that is, as index of corticospinal inhibition, was visually determined and defined as the duration from the TMS stimulus to the return of continuous voluntary EMG (³¹). Both MEP (area and amplitude) and silent period were averaged from the two series of contractions.

Energy Cr

Oxygen uptake, carbon dioxide production and ventilation were measured during the Cr test during both level and uphill running (Cr_FLAT and Cr_UH, respectively). Cr was calculated from oxygen consumption using the energy equivalent of oxygen taking into account the respiratory exchange ratio (³²).

Participants’ sensations

For general fatigue, perceived pain (for both KE and PF) and level of gastrointestinal discomfort, the analysis was made on the PRE to POST change (measured in millimeters on the 100 mm scale).

Statistical Analysis

Statistics were performed the same way on two different groups: on pairs (i.e., performance matched) and on all participants (i.e., irrespective of performance; see Fig. 1 for more details). Statistical analyses were performed using Statistica software (Statsoft Inc., Tulsa, OK). Normality distribution was verified with a Shapiro–Wilk test and variance homogeneity using Levene’s test. A mixed-model ANOVA for time (PRE–POST) with sex (men–women) and distance (SHORT–LONG) as between-participant factors were used to assess PRE to POST alterations. A mixed-model ANOVA for time (PRE–POST) and voluntary contraction intensity (100%, 75% and 50% MVC) with sex (men–women) and distance (SHORT–LONG) as between-participants factors were used to evaluate changes in MEP (amplitude and area) and silent period. Because there were too few participants on LONG for Cr_UH (only two pairs), a two-way ANOVA (time–sex) was performed for Cr. Effect size is presented for significant findings as partial eta squared (η²_p). In the event of a significant time–sex–distance interaction, a time–sex mixed-model ANOVA was performed on both SHORT and LONG. Paired t-test was used to compare i) the relative performance of paired men and women and ii) the relative race intensity of paired men and women in both SHORT and LONG. An independent sample t-test was performed to compare the relative intensity of all men and women. When normality or homogeneity conditions were not met (for subject sensation parameters) a Mann–Whitney U test was performed on PRE to POST changes for men versus women in both SHORT and LONG. A Mann–Whitney U test was also used to compare the competitive intention of men and women in SHORT and in LONG. The POST concentrations of blood parameters (CRP and CPK) were assessed by independent sample t-test or Mann–Whitney U test (i.e., when assumption of normality or heterogeneity were not met) to compare men versus women in SHORT and LONG. The level of significance was set at P < 0.05.

RESULTS

Pair Comparisons

Maximal voluntary contraction, VA, and EMG RMS

There was no time–sex–distance interaction for MVC, VA_PNS, VA_TMS, or EMG RMS for either KE or PF (P ≥ 0.18; Table 2). A significant time–sex interaction was found for KE MVC (P < 0.01, F = 11.7, η²_p = 0.28). Knee extensor MVC change was −36% ± 17% versus −27% ± 15% for men and women, respectively (Fig. 3A). The time–sex interaction did not reach the statistical significance (P = 0.051, F = 4.2, η²_p = 0.14) for PF MVC (−34% ± 11% vs −30% ± 15% for men and women, respectively; Fig. 3B). VA_PNS decreased significantly for both KE and PF (P < 0.001, F = 25, η²_p = 0.45 and P < 0.001, F = 23.2, η²_p = 0.47 for KE and PF, respectively; Figures 3E–F) but no significant time–sex interaction for either muscle group was found. Similarly, the ANOVA displayed a significant decrease in VA_TMS (P < 0.001, F = 20.1, η²_p = 0.44) but no significant time–sex interaction (P = 0.170, F = 2.0, η²_p = 0.07; Figure 3G). As reported in Table 2, RMS/Mmax decreased for VL (P < 0.001, F = 16.3, η²_p = 0.35), GM (P < 0.001, F = 14.9, η²_p = 0.36) and SOL (P = 0.028, F = 5.4, η²_p = 0.17), independently of sex.

TABLE 2 - EMG RMS and M-wave data of paired participants.

		Females		Males
		PRE	POST	PRE	POST
VL Mmax (mV)	SHORT	12.4 ± 1.1	12.2 ± 2.3	14.6 ± 6.7	14.7 ± 6.1	*
	LONG	8.5 ± 4.9	10.4 ± 6.1	12.5 ± 5.5	15.9 ± 4.6
VL RMS (mV)	SHORT	0.58 ± 0.18	0.49 ± 0.22	0.60 ± 0.32	0.47 ± 0.18	**
	LONG	0.46 ± 0.29	0.31 ± 0.14	0.64 ± 0.44	0.48 ± 0.18
VL RMS/Mmax (%)	SHORT	4.6 ± 1.2	3.9 ± 1.2	4.3 ± 1.4	3.6 ± 1.6	****
	LONG	5.5 ± 2.3	4.0 ± 3.0	5.0 ± 1.3	3.1 ± 1.0
GM Mmax (mV)	SHORT	8.9 ± 3.6	9.5 ± 3.7	8.1 ± 2.9	8.1 ± 2.9
	LONG	8.8 ± 4.4	8.5 ± 2.7	13.2 ± 6.9	11.5 ± 5.5
GM RMS (mV)	SHORT	0.19 ± 0.10	0.16 ± 0.10	0.23 ± 0.15	0.15 ± 0.06	***
	LONG	0.19 ± 0.12	0.11 ± 0.05	0.23 ± 0.15	0.13 ± 0.08
GM RMS/Mmax (%)	SHORT	2.2 ± 1.0	1.9 ± 1.0	3.0 ± 1.5	1.9 ± 0.7	****
	LONG	2.7 ± 1.8	1.5 ± 0.8	2.0 ± 1.0	1.5 ± 1.0
SOL Mmax (mV)	SHORT	10.6 ± 2.7	9.4 ± 3.0	12.3 ± 3.3	11.8 ± 3.8	**
	LONG	10.8 ± 4.6	9.5 ± 3.7	8.4 ± 3.0	8.6 ± 3.0
SOL RMS (mV)	SHORT	0.22 ± 0.09	0.21 ± 0.09	0.30 ± 0.16	0.23 ± 0.11	***
	LONG	0.31 ± 0.17	0.20 ± 0.06	0.30 ± 0.09	0.23 ± 0.06
SOL RMS/Mmax (%)	SHORT	2.2 ± 1.0	2.4 ± 1.2	2.8 ± 2.1	2.2 ± 1.4	**
	LONG	2.8 ± 0.8	2.3 ± 0.6	3.8 ± 1.2	3.1 ± 1.9

Values are presented as mean ± SD.

*Significant main sex effect: P < 0.05.

**^,***^,****Significant main time (PRE–POST) effect: P < 0.05, P < 0.01, P < 0.001, respectively.

FIGURE 3: PRE to POST change in MVC (panel A for KE and panel B for PF), in Pt (panel C for KE and panel D for PF), in VA_PNS (panel E for KE and panel F for PF) and VA_TMS (panel G). Continuous lines represent the group mean values and the dashed lines represent individual participants. ††P < 0.01: Significant time–sex interaction. $$P < 0.01: Significant time–sex–distance interaction (see results section for details).

Resting twitch responses

No time–sex interaction was found for KE Pt (P = 0.202, F = 1.7, η²_p = 0.05; Figure 3C) and KE Db100 (−15% ± 9% vs −17% ± 9% for men and women, respectively; P = 0.123, F = 2.5, η²_p = 0.07). There was a significant time–sex–distance interaction in PF Pt (P = 0.006, F = 8.7, η²_p = 0.25, Figure 3D). Then, a significant time–sex interaction was observed in PF Pt for SHORT (P = 0.019, F = 7.0, η²_p = 0.33), but no such interaction was observed in LONG (P = 0.122, F = 2.8, η²_p = 0.19). Change in PF Pt on SHORT was −24% ± 14% versus −10% ± 9% for men and women, respectively. A significant triple time–sex–distance interaction was also found for PF Db100 (P < 0.05, F = 4.7, η²_p = 0.15) followed by a significant time–sex interaction in SHORT (P = 0.005, F = 10.8, η²_p = 0.44) but not in LONG (P = 0.836, F = 0.1, η²_p < 0.01). In SHORT, the PRE–POST change was −19% ± 10% for men and −4% ± 9% for women. The analyses on the Db10:Db100 revealed the presence of low-frequency fatigue on both KE and PF, independent of sex and distance (P < 0.01, F = 8.5, η²_p = 0.22 and P < 0.05, F = 7.1, η²_p = 0.21, respectively). The decrease in Db10/Db100 was −5% ± 19% versus −7% ± 13% in KE and −3% ± 9% versus −5% ± 6% in PF for men versus women, respectively.

MEP and silent period

A significant time–sex interaction was observed in MEP_AREA (P = 0.022, F = 5.9, η²_p = 0.17). The post hoc revealed that MEP_AREA increased in men (P = 0.002) and did not change in women (P = 0.478 and P = 0.967, respectively). However, the analyses did not reveal any time–sex interaction in MEP_AMP (P = 0.239, F = 1.4, η²_p = 0.05) or silent period (P = 0.340, F = 0.94, η²_p = 0.03). All values and statistics concerning MEP and silent period are presented in Table 3.

TABLE 3 - MEP and SP data of paired participants.

		Females		Males
		PRE	POST	PRE	POST
MEPAMP 100% MVC (%)	SHORT	33.2 ± 15.3	34.9 ± 12.1	31.5 ± 5.9	38.2 ± 8.9	*
	LONG	37.5 ± 13.3	46.1 ± 14.0	35.9 ± 14.5	47.3 ± 9.3	*
MEPAMP 75% MVC (%)	SHORT	43.5 ± 14.2	42.4 ± 14.1	43.0 ± 14.7	42.8 ± 14.7	*
	LONG	47.3 ± 15.4	50.0 ± 13.8	44.2 ± 12.3	53.8 ± 9.5	*
MEPAMP 50% MVC (%)	SHORT	43.0 ± 16.8	44.4 ± 18.8	40.7 ± 11.6	43.9 ± 17.7	*
	LONG	46.8 ± 15.1	51.1 ± 13.2	42.9 ± 13.2	53.4 ± 12.9	*
MEPAREA 100% MVC (%)	SHORT	33.4 ± 14.2	37.4 ± 14.8	34.4 ± 6.4	47.2 ± 15.8	*,**
	LONG	42.5 ± 18.7	48.0 ± 18.8	40.2 ± 13.8	56.6 ± 13.9	*,**
MEPAREA 75% MVC (%)	SHORT	49.1 ± 17.3	48.3 ± 15.5	49.7 ± 17.1	58.2 ± 26.2	*,**
	LONG	53.5 ± 19.5	55.6 ± 18.6	53.1 ± 15.2	65.0 ± 11.4	*,**
MEPAREA 50% MVC (%)	SHORT	55.0 ± 21.5	55.1 ± 20.0	54.2 ± 15.2	61.3 ± 21.5	*,**
	LONG	59.2 ± 20.3	60.0 ± 18.8	56.7 ± 14.0	68.7 ± 15.6	*,**
SP 100% MVC (ms)	SHORT	212 ± 86	232 ± 94	185 ± 77	191 ± 78
	LONG	210 ± 58	188 ± 44	202 ± 91	184 ± 87
SP 75% MVC (ms)	SHORT	209 ± 82	213 ± 79	179 ± 78	178 ± 78
	LONG	198 ± 64	197 ± 47	201 ± 86	178 ± 81
SP 50% MVC (ms)	SHORT	218 ± 76	216 ± 77	179 ± 74	185 ± 82
	LONG	206 ± 62	199 ± 45	203 ± 79	175 ± 76

Values are presented as mean ± SD.

MEP_AREA and MEP_AMP are normalized to M-wave and presented as the % of M-wave.

*Significant main time (PRE–POST) effect: P < 0.01.

**Significant time–sex interaction: P < 0.05.

MEP_AREA and MEP_AMP are normalized to M-wave and presented as the percentage of M-wave.

Energy Cr

A significant increase in both Cr_FLAT (P = 0.011, F = 7.7, η²_p = 0.26) and Cr_UH (P = 0.023, F = 6.5, η²_p = 0.32) conditions (+6% ± 10% and +4% ± 7%, respectively) was observed (Figs. 4A–B), but no sex differences were identified (P = 0.208, F = 01.7, η²_p = 0.07 and P = 0.704, F = 0.2, η²_p = 0.01 for Cr_FLAT and Cr_UH, respectively).

FIGURE 4: PRE to POST change in Cr on both FLAT (Cr_FLAT, panel A) and UH (Cr_UH, panel B) conditions. Continuous lines represent the group mean values and the dashed lines represent individual participants.

Blood parameters

The statistical analyses did not reveal any sex differences in either CRP (P = 0.860 and P = 0.115 for SHORT and LONG, respectively) or CPK (P = 0.171 and P = 0.916 for SHORT and LONG, respectively).

Race intensity

No significant differences were observed between men and women in both SHORT (60% ± 5% vs 61% ± 11% of the speed reached during the V˙O_2max test, respectively; P = 0.798) and LONG (44% ± 7% vs 47% ± 5% of the speed reached during the V˙O_2max test, respectively; P = 0.131).

Participants’ sensations and competitive intention

Men reported more general fatigue compared with women in SHORT (P = 0.027) but not in LONG (P = 0.353, Fig. 5A). No sex differences were observed in perceived KE pain, perceived PF pain or digestive system feeling for either LONG or SHORT (Figs. 5A–D). The competitive intention was significantly different between men and women in SHORT (3.6 ± 1.9 vs 6.2 ± 2.9, respectively; P = 0.042) but not in LONG (5.8 ± 1.6 vs 7.0 ± 2.6, respectively; P = 0.171; Fig. 5E).

FIGURE 5: Subject sensation (panels A, B, C, and D) and competitive intention (panel E) of the participants. †P < 0.05: Significantly different between males and females.

All Participants

The race performance of all participants relative to the best male and the best female finishers was not different between sexes (157% ± 13% for men vs 159% ± 9% for women, P = 0.46). Analyses of all participants for NM parameters were not different than the pairs analyses already presented so the results are not presented here.

The significant time–sex interaction did not reach the level of significance for Cr_FLAT (P = 0.108, F = 2.7, η²_p = 0.06) or for Cr_UH (P = 0.057, F = 3.9, η²_p = 0.10), the deterioration being +7% ± 12% versus +3% ± 8% in FLAT and +10% ± 7% versus +2% ± 6% in UH for men versus women, respectively.

DISCUSSION

The purpose of this study was to investigate whether distance has an effect on the magnitude of sex differences in NM fatigue and energy Cr after trail running races. Our results showed that women are less fatigable as evidenced by a lower decrease in KE maximal strength, independent of the distance. Women also demonstrated less peripheral fatigue of the PF muscles compared with men on short distances, this result being possibly because of sex differences in competitive intentions. Furthermore, a sex difference was displayed in the fatigue-induced change in corticospinal excitability, with an increase in MEP_AREA only in men. The NM fatigue sex differences did not statistically translate into energy Cr sex differences.

Effect of Trail Race on NM Fatigue

The strength losses found in the present study in SHORT (−23% and −25% for KE and PF, respectively) and LONG (−36% and −33% for KE and PF, respectively) fit with previous studies which assessed NM fatigue after running races of distances shorter (^9,33) and longer than 100 km (^10,12,34,35). However, most of these studies did not include women or did not compare sexes. For both men and women and for both KE and PF, torque reduction was accompanied by central (i.e., decrease in VA) and peripheral (i.e., decrease in Pt, Db100 and Db10:Db100) alterations, as previously reported after trail-running races (^{9–12,33–36}).

Sex Differences in NM Fatigue

Maximal voluntary contraction

Maximal torque decreases in LONG KE for both men and women are comparable with Temesi et al. (¹²) (110 km; 10 men and 10 women), that is, −38% versus −40% in men, −29% versus −33% in women, respectively. However, the decrease of maximal strength in PF in LONG was higher in the present study compared with Temesi et al. (¹²) (−39% vs −26% in men, −37% versus −31% in women, respectively). The decrease in KE MVC was greater in men than in women, in accordance with Temesi et al. (¹²). Despite the interaction not reaching the level of significance (P = 0.051), the results observed in PF seem consistent with the results in KE. Contrary to our hypothesis, the sex differences observed in MVC decrease were independent of distance. Indeed, sex differences in NM fatigue did not increase with distance, and the results for peripheral fatigue in fact showed the opposite (i.e., a sex difference was only shown in SHORT; see below). In addition to the results from our previous study on this topic (¹²) and the known physiological differences between men and women which could give women a greater advantage over longer distances (e.g., muscle fiber type, muscle capillarization, and vasodilatory capacity, lower carbohydrate metabolism), it must be acknowledged that this hypothesis was based on anecdotal evidence from observing women beat men in ultraendurance races, that is, no direct scientific evidence existed.

Peripheral fatigue

Despite a greater strength loss in men in KE, no sex differences were observed in peripheral parameters (Pt, Db100, Mmax and Db10:Db100) in this muscle group. This result is in agreement with the study of Temesi et al. (¹²). The present data showed greater PF peripheral fatigue (greater decrease in Pt and Db100) in men compared with women, but unexpectedly, it was found in SHORT distance only (−24% vs −10% of PF Pt decrease after SHORT in men and women, respectively). Because Temesi et al. (¹²) reported a sex difference for PF Pt after 110 km, we would have expected such observations also in runners who performed races longer than 100 km. The sex difference observed in peripheral fatigue in SHORT cannot be explained by low-frequency fatigue because no sex differences were observed in Db10:Db100 ratio. Based on data collected in rat soleus (³⁷), Temesi et al. (¹²) speculated that the maintenance of work output driven by a large amount of eccentric component during a mountain ultramarathon could induce a smaller decrease in evoked responses in women because of more compliant Achilles tendon properties (³⁸). Yet, it seemed that a similar amount of muscle damage were observed in men and women in both SHORT and LONG because no sex differences were observed in CRP and CPK, although these markers are only indirect indices of muscle damage. Alternatively, sex differences in fatigability have previously been associated with contractile mechanisms (³). Glycogen depletion is also an important contributor to impairments in contractile function (Pt) after endurance exercise (^39,40). Men have been found to use ⁓25% more muscle glycogen than women during moderate-intensity exercise (⁴¹) which is likely the intensity at which ultramarathons are performed. Thus, a greater level of glycogen depletion in men could also be a plausible explanation for the smaller twitch amplitude reduction in women than men.

Yet, one must be cautious before attributing an attenuated force decrease to better fatigue resistance. Field studies performed during competitions are relevant for that type of experiment because (i) asking subjects to exercise over 10 to 20 h in a laboratory is extremely challenging (although doable (⁴²)) and (ii) competitions minimize motivation issues, that is, it is assumed that participants complete the race as fast a time as possible. However, this may not be the case and for the first time, the present study presented data on the competitive intention of the participants. The objective of the questionnaire was to better understand the competitive intention of the participants during the race. It is a relevant, yet underinvestigated, question because this type of event brings together runners from a wide range of backgrounds and experience, and with different motivations for the race (e.g., ranking, best personal performance, finishing the race within the time limits, enjoyment). Despite men and women being matched by relative performance, the present results showed that men were more competitively oriented and reported more general fatigue compared with women on SHORT but not on LONG. It has been suggested that during ultraendurance exercise, runners have a security reserve set by the brain to prevent excessive fatigue levels (¹⁴). This security reserve described in the Flush model is highly influenced by motivation. Thus, it could be speculated that with greater motivation for the competition, men could have stretched the limit of their security reserve to a greater extent, involving greater decreases of force capacities, explaining the greater peripheral fatigue in SHORT. Interestingly, although difficult to explain, men and women performed their race with similar competitive intention in LONG. This could partly explain the lower sex differences in terms of peripheral fatigue in LONG compared with SHORT. However, this suggestion should be balanced against the fact that performing exercise at a submaximal effort for a given distance would result in a longer duration of exercise, and it is unclear what effect this trade-off would have on muscle function postrace. Furthermore, it should be noted that despite the lower competitive intention during the short races in women, the present data showed that the intensity of exercise relative to the speed at V˙O_2max was similar between sexes. Taken together, these findings suggest i) that for a given relative intensity of prolonged exercise, women exhibit attenuated impairments in NM function, and ii) women might be capable of performing prolonged exercise at a greater relative intensity than men.

Central fatigue and corticospinal excitability and inhibition

No sex differences were observed in VA on KE and PF muscles. These results are in accordance with Temesi et al. (¹²) Despite the decrease in VA in KE being approximately twice as large among men compared with women (−18% ± 19% vs −9% ± 11% in VA_TMS and −22% ± 20% vs −11% ± 17% in VA_PNS for men and women, respectively), there was no statistically significant difference, likely because of the large variability, the power of the statistical analysis and/or the delay to POST evaluation.

The present study showed a sex difference in the fatigue-induced change in MEP_AREA, with a postexercise increase in MEP_AREA in men only. These results are in contrast to that of Keller et al. (⁴³) and Hunter et al. (³⁰) who reported similar results between MEP_AMP and MEP_AREA and did not show any effect of sex on the fatigue-induced change in MEP_AREA after isometric exercise, but the fatiguing task differed considerably to the present study. Speculatively, the greater increase in MEP_AREA in men could be related to the greater strength loss in the KE in men. For example, during sustained MVC, MEP have been shown to increase, with this increase thought to occur because of an increase in cortical output to compensate for impairments in NM function occurring downstream of the motor cortex (^31,44). Surprisingly, the sex differences observed in MEP_AREA were not found in MEP_AMP, and the lack of difference in MEP_AMP between men and women is consistent with our previous study (¹²). Although the reasons why MEP_AREA and MEP_AMP behaved differently are unclear, this could indicate a reduction in the firing frequency of the multiple descending volleys elicited by TMS, causing an elongation of the MEP and thus an increase in its area (⁴⁵). Silent period was not differently altered in men and women, suggesting that corticospinal inhibition does not change with fatigue after ultramarathon races and that is true in either sexes (¹²).

Sex Differences in Energy Cr

It was hypothesized that women would be better able to preserve their Cr than their male counterparts based on changes observed after a 2-h treadmill run (¹⁵). Specifically, given that men have been shown to exhibit greater decrements in Pt after trail running (¹²), it was expected that men would be required to increase motor unit recruitment to compensate for greater impairments in contractile function relative to women, with a consequent increase in oxygen demand and Cr. Although we observed greater reductions in Pt in men in the PF after SHORT, no sex differences were observed in Cr_FLAT or Cr_UH. First, although the present results on NM alterations are consistent with existing literature (¹²) and seem to go in same direction across muscles (i.e., with women appearing less fatigued), differences are probably not strong enough to differently alter Cr between men and women. Second, as mentioned in the introduction, contrary to popular hypothesis, NM fatigue is possibly not as strongly related to oxygen uptake kinetics (^21–24) and therefore energy Cr. However, a type II error cannot be ruled out given that out of the 18 matched pairs who finished the race, only 13 were able to perform the Cr POST evaluations (7 pairs in SHORT and 6 in LONG). Furthermore, when running the analysis on all subjects (i.e., on 17 women and 26 men), the deterioration in Cr_UH in men was approximately five times larger than in women despite the difference not reaching statistical significance (P = 0.057). These results are not readily comparable with the existing literature because the few studies that have assessed fatigue-induced sex differences in Cr (^15,25,26) were performed over shorter durations on a treadmill without gradient. Nevertheless, the present findings are intriguing, and future research should further investigate potential sex differences in Cr after ultramarathons.

Limitations

In the present study, there was an imbalance between the number of male race versus female race finishers, impairing our ability to match pairs of men and women based on performance and decreasing our statistical power. It was difficult to recruit women in the study even though the percentage of women of the present sample (34%) was much larger than the rate of participation at these type of events (e.g., 10% on UTMB). The sample size was estimated to take into account an anticipated ⁓30% of dropout rate (calculated using data on the previous 3 yr), however, it is difficult to predict how many runners would be able to perform the POST sessions after such a demanding effort. Another limitation is that women were in different phases of their menstrual cycle, as assessed using a medical questionnaire before the race. Conflicting findings exist in the literature about the effects of menstrual cycle on fatigability (^46–49). It is worth mentioning that studies showing effects of the menstrual cycle on fatigability were conducted on fatiguing tasks using local, single-joint exercise (⁵⁰), whereas no effects have been found when considering whole body exercise (^51,52) as in the present study. The delay to POST evaluation is another limitation of this study despite measurements were done as soon as possible after the race. It should be noted that the time to postassessment was similar between men and women for both KE (37 ± 15 min vs 34 ± 13 min, respectively) and PF (44 ± 14 min vs 45 ± 12 min, respectively) NM assessment and for Cr evaluation (81 ± 19 min vs 77 ± 20 min, respectively). Although most of the metabolic perturbations occurring during the race likely recovered by the postassessment, such perturbations were likely to be minimal given the low intensities at which trail runs are performed. Indeed, slower recovery is associated with more prolonged exercises, as shown for instance by Kruger et al. (⁵³) in cycling, and the low frequency fatigue induced by muscle damage during trail running most likely even further delayed recovery. Although the delay before Cr assessment was similar between men and women, this delay could still have affected the results by changing their Cr as well as their substrate use because participants had some time to eat and digest after the race. Furthermore, given that carbohydrate intake was not controlled during and after the race, it is possible that differences in carbohydrate intake, or simply the “mouth rinsing” effect, could have impacted the degree of NM impairment (^54,55). Finally, the potential limitations behind our matching of participants based on performance level relative to the winner of the race warrants discussion. In the races of the UTMB, there is a considerable sex difference in participation, such that substantially more men than women compete. Consequently, matching performance relative to the winner might be associated with limitations if the relative standard of the male race is greater than the female races. However, the ITRA performance index of both male and female winners of each race was over 95% of the world’s best as the UTMB is one of the most renowned races in the world. Thus, despite the differences in participation, we are confident that our experimental approach is reliable to address sex differences in fatigability and Cr among men and women of a similar participation level. Moreover, our analysis examining the average speed at which the race was completed relative to the speed at V˙O_2max revealed that these participants were also not different in terms of the relative intensity of exercise throughout the races. Finally, important limitations surrounding the use of the competitive intention scale should be acknowledged. The goal of the scale was to understand the level of effort put into the race and any potential sex differences. Although the results interestingly revealed a lower competitive intention in women, which warrants further investigation and has potential implications for sex differences in NM function, it is important to note that the competition intention scale has not undergone a validation process. Thus, the results surrounding this scale should be interpreted with caution. Future studies, performed in collaboration with a psychologist, should reconsider this question.

CONCLUSIONS

The present study showed lower decrements in force in the KE and lower decrements in contractile function after SHORT in the PF in women compared with men after prolonged running exercises. These data are consistent with our previous study performed on a 110-km trail running race (¹²). The present study performed on distances ranging from 40 to 171 km allowed the conclusion that, contrary to our hypothesis, sex differences in fatigability do not increase with race distance. This study also brings a novel element on the competition intention of the participants. Although it likely does not explain all physiological outcomes, the effort extended during self-paced race events is an important consideration for future research when comparing men and women after such efforts.