Custom foot orthoses improve performance, but do not modify the biomechanical manifestation of fatigue, during repeated treadmill sprints

The capacity to reiterate “all out” efforts with incomplete recoveries (i.e. repeated sprint ability) depends, not only, on metabolic but also neuro-mechanical factors (Girard et al. 2011a). Identifying the onset of fatigue through ground reaction forces (GRF) monitoring during repeated sprint exercise (RSE) may be helpful in providing early warnings of increased injury risk in many team sports. An important limitation of existing RSE literature is the lack of continuous measurements of running velocity and GRF including horizontal force production (Girard et al. 2011b, c). In a pioneer instrumented sprint treadmill study, Morin et al. (2011) provided a thorough description of changes in kinetics and kinematics during a multiple-set repeated sprint series. Along with the expected progressive slowing of running velocity, these authors observed both a significant decrease in the capability to produce total (resultant) GRF and a significant and even larger decrease in the ability to apply it with a forward orientation during acceleration.

Runners often use custom foot orthotics (CFOs) in order to rehabilitate from injuries and/or improve comfort. Additional benefits also concern run-induced fatigue reduction through a better preservation of the ability of the body to absorb plantar loading (Lucas-Cuevas et al. 2014a). However, findings have not been unanimous. For instance, wearing foot orthoses altered neuromuscular control during a sub-maximal, 1-h constant-velocity treadmill run and partly protected from the resulting fatigue-induced reductions in rapid force development characteristics of the plantar flexors (Kelly et al. 2011). Contrastingly, CFOs had no positive effect on running mechanical adjustments induced by a 12-min run at ~14.5 km h⁻¹ (Patzkowski et al. 2012; Lucas-Cuevas et al. 2014b). Identification of the biomechanical manifestation of fatigue during RSE with versus without CFOs is currently lacking.

Flexible shank-dependent CFOs are derived from a three-dimensional representation of the individual’s foot and are commonly made of ethyl-vinyl acetate (EVA), while the use of more durable thermoplastic polyurethane (TPU) foams is increasingly popular. By increasing tibial accelerations, peak eversion and tibial internal rotation parameters, this new TPU material is thought to promote energy return (Sinclair et al. 2014), in comparison with traditional EVA footwear midsoles, which may translate into more economical movements (i.e. lower oxygen uptake; Sinclair et al. 2016), thereby minimizing fatigue occurrence. In support, running shoes with softer and more resilient midsoles were found to improve running economy by ~1% on average during treadmill running at a speed slightly below the anaerobic threshold (Worobets et al. 2014). Contrasting results have also been observed regarding sub-maximal running at constant pace with softer shoes with reports of either increase (Bosco and Rusko 1983) or unchanged (Nigg et al. 2003) oxygen uptake values. No previous RSE investigation has investigated whether CFOs with different materials, but identical in construction, protect from the biomechanical manifestation of fatigue.

The primary purpose of offering CFOs choices with varying specifications is to promote comfort, reduce injury risk, improve performance or a combination of the three. In the absence of research studying the biomechanical effects of wearing CFOs during RSE, clinicians require evidence of favourable biomechanical or physiological adjustments to support their prescription. A better understanding of the effects of inserts made of different materials will help to determine if repeated sprint ability can be improved and how this is brought about biomechanically. We therefore determined the effect of CFOs manufactured from EVA and TPU materials, both compared to a control condition (CON; shoes only) during repeated sprints on alterations in running kinetics and kinematics, with special reference to horizontal force production. We hypothesised that wearing CFOs will improve repeated sprint ability by mitigating the effects of fatigue on stride pattern, while only subtle running mechanical differences may be observed between inserts.

Distance ran in 5 s decreased from the first to the last sprint (−3.9 ± 3.1%; P < 0.001), independent of condition (P = 0.682). Averaged distance covered values for the eight sprints were higher for both EVA (23.3 ± 1.4 m; P = 0.004) and TPU (23.1 ± 1.5 m; P = 0.018) than CON (22.7 ± 1.6 m) (Fig. 1). The eight sprints for both EVA and TPU produced lower heart rates (~4 bpm; P < 0.001) compared to CON, while oxygen uptake (P = 0.280) and ratings of perceived exertion (P = 0.680) values did not differ between conditions (Fig. 1).

Regardless of the footwear condition, mean horizontal forces (−11.1 ± 5.8%) and step frequency (−5.7 ± 4.8%) decreased from sprint 1 to sprint 8 (P < 0.001), while peak and mean vertical forces remained unchanged (P > 0.261) (Fig. 2). Contact time lengthened from the first to the last sprint (+9.6 ± 8.7%; P < 0.001) as a result of significantly longer braking (+11.5 ± 10.3%; P < 0.001) and propulsive (+4.2 ± 4.0%; P = 0.031) phases. Averaged contact time values for the eight sprints were also significantly shorter for EVA (−2.3 ± 2.8%; P = 0.010), but not TPU (−1.1 ± 3.3%; P = 0.432), compared to CON. Specifically, the duration of the propulsive phase was globally shorter for both EVA (−4.7 ± 4.6%; P = 0.002) and TPU (−4.0 ± 5.7%; P = 0.021) versus CON, while braking phase duration was similar (P = 0.919). In the horizontal direction, peak propulsive forces (−8.0 ± 6.7%; P < 0.001) decreased from sprint 1 to sprint 8, independently of conditions, but the same was not observed for the braking forces (+2.0 ± 6.7%; P = 0.172). Braking impulse (pooled values: −7.7 ± 2.5 vs. −8.8 ± 2.8 N.s; + 22.9 ± 23.8%; P = 0.018; η² = 0.21) increased and propulsive impulse (pooled values: 51.0 ± 18.0 vs. 48.2 ± 18.3 N.s; −5.4 ± 2.1%; P = 0.027; η² = 0.24) decreased from the first to the last sprint. The braking impulse was not different between conditions (pooled values: −8.4 ± 2.8 N.s; P = 0.377; η² = 0.06). The propulsive impulse was globally smaller for both EVA (49.0 ± 7.9 N.s; −3.0 ± 8.0%; P = 0.035) and TPU (48.8 ± 8.3 N.s; −3.5 ± 7.9%; P = 0.025) than CON (50.5 ± 8.0 N.s).

Irrespective of conditions, vertical (pooled values: −11.7 ± 9.8%) and leg stiffness (−9.5 ± 8.8%) decreased from sprint 1 to sprint 8 (both P < 0.001) (Fig. 3). Peak vertical force averaged values for the eight sprints were also significantly larger for EVA (+1.9 ± 2.3%; P = 0.008), but not TPU (+1.2 ± 2.4%; P = 0.109), compared to CON.

Our main finding was that CFOs improve performance during repeated treadmill sprints, yet with similar positive effects (e.g. ~+ 4% on averaged sprints 1–8 distance covered) for EVA and TPU materials. For the first time with RSE, we investigated the effects of CFOs on the biomechanical manifestation of fatigue. Importantly, no significant interaction was found between sprint number and footwear conditions for any stride mechanical parameter. Despite different resilience characteristics but similar configuration of tested CFOs, the overall results of the present study show that participants who wore CFOs, either made of EVA or TPU materials, produced essentially similar fatigue-induced adjustments in running mechanics during RSE.

Substantial increases in contact time occurred as fatigue developed, leading to a monotonic large decrease in step frequency, while there were non-significant increases in both flight time and step length. Such apparent decreased ability of the lower extremities to tolerate the ground impact observed across sprint repetitions is consistent with previous RSE observations (Girard et al. 2011b, c; Brocherie et al. 2016). Another unique observation was that, compared to CON, duration of propulsive phase was ~ 4–5% shorter for EVA and TPU trials, but was similar for braking phase in all conditions. While peak propulsive forces were similar, for the first time, we demonstrated that wearing insoles during RSE primarily influence the propulsive versus the braking component of the horizontal GRF by shortening the duration of propulsive force application.

There was no significant interaction between time and condition for any of the kinetic and kinematic variables studied. Consistent with previous RSE studies (Morin et al. 2011; Girard et al. 2015; Brocherie et al. 2016), our data showed that reduction in averaged horizontal force production was substantial, while no significant change occurred in the vertical direction. Remarkably, alterations in mean horizontal force production displayed a biphasic pattern with more marked decreases after the third sprint repetition. Following the first few repetitions of a sprint series, the time course of neuromuscular fatigue causing the reduction in performance is due mainly to peripheral fatigue, while central fatigue generally appears later (Collins et al. 2018). Altogether, this reinforces that producing large amounts of horizontal forces to the ground is paramount to better preserve sprint capacity as fatigue develops during RSE, yet wearing CFOs had no effect on this fatigue pattern.

Another novel aspect of this study was to specifically and continuously monitor braking versus propulsive peak forces and phase durations in the horizontal direction, as previously done for single running sprint (Morin et al. 2015; Nagahara et al. 2018). Globally, lengthening of braking versus propulsive phase duration was ~2.5 times larger, and thereby was mainly responsible for progressively longer contact times with sprints repetition. Additionally, peak propulsive but not braking forces decreased with sprint repetitions, with also higher braking and lower propulsive impulses, and followed a similar time course of changes than averaged horizontal force production. This implies that, with fatigue development, runners seem to slow down by pushing less forcefully forward as opposed to braking more in the early stance phase. Interestingly, lengthening of braking and propulsive phase durations (+11.4% and +4.2% from sprint 1 to 8, respectively) was of similar magnitude in elite female Rugby Sevens players undertaking the same RSE (Girard et al. 2020). However, this resulted in significantly lower values for propulsive (−11.9%), but not braking (+1.7%), impulses since alteration in peak braking and propulsive forces was of similar magnitude (~12–13%). This reinforces that biomechanical manifestation of fatigue during RSE are specific to the tested population.

In the present study, similar fatigue-induced changes on spring mass model characteristics occurred while sprinting repeatedly in our three conditions. In line with previous studies where participants did not wear CFOs (Girard et al. 2011b, c; Brocherie et al. 2016), reductions in vertical stiffness and leg stiffness to a lower extent resulted from greater vertical displacement of the centre of mass and leg compression, respectively, since peak vertical forces did not change. Even confronted with larger applied peak vertical forces (albeit only significant for EVA), neither vertical displacement of the centre of mass nor leg compression magnitudes changed when participants wore CFOs, so that vertical and leg stiffness values remained unchanged. Previously, an increase in energy return and/or longitudinal bending stiffness shoe features induced only subtle differences in stride mechanics during overground running at an intensity lower than the ventilatory anaerobic threshold (Flores et al. 2019). Contrastingly, we demonstrated for the first time that EVA, and TPU to a lower extent (non-significant trend), can shorten the contact time and increase peak vertical forces, thereby producing more forceful ground contacts during RSE.

In line with previous research on the effect of orthoses on aerobic cost during a 1-h run (Kelly et al. 2011), we found lower heart rates, but not oxygen uptake values, while wearing CFOs. Interestingly, RPE profile did not differ between trials, probably due the combined effects of improved performance with lower heart rates in EVA and TPU trials. While participants were blinded to shoe wear conditions, a “belief” or expectation effect cannot be overlooked to partly explain these results. This may have in turn, consciously or unconsciously, contributed to maintain motivation to perform maximally during repetition of “all out” efforts. In the absence of measures of task effort awareness (i.e. sense of effort; Christian et al. 2014) and shoe comfort, it is, however, difficult to accept or reject this hypothesis.

Improved performance while wearing CFOs occurred despite the increased footwear weight in EVA and TPU trials (~+ 8 and + 14% compared to CON, respectively). While heavier footwear can be detrimental to mechanical and/or energy cost of distance running (Frederick et al. 1986; Franz et al. 2012), at least under the present circumstances, this confounding factor did not negatively influence repeated sprint ability. Although the sprint-instrumented treadmill is a useful tool for assessment of running biomechanics, caution is needed since measurement of stride mechanical parameters during overground running may differ slightly (Riley et al. 2008). An unexplained finding of our study was that propulsive impulses were ~ 3% lower (braking impulses were comparable), despite better performance, when wearing orthotics. This implies that the present results would need to be confirmed under “real world settings”. Longer sprints, shorter rest periods or a combination of both as well as with environmental perturbations, all exacerbating performance decrements during RSE (Girard et al. 2011a; Brocherie et al. 2016), may derive a greater ergogenic benefit from the addition of CFOs. While no assessment was proposed, it cannot be ruled out that EVA and TPU trials may have been perceived as more comfortable footwear conditions, which would explain better performance maintenance across “all out” efforts. Finally, it remains to be verified if specific changes in GRFs patterns would occur, due to differences in running velocity, during early and late acceleration phase of repeated treadmill sprints (Girard et al. 2015).

Compared to shoe only, wearing CFOs improve repeated sprint ability, yet with similar positive effects for EVA and TPU materials. There was no noticeable difference in the time course and magnitude of mechanical adjustments for the kinematics, kinetics and spring-mass model parameters during repeated treadmill sprints. While lower extremities behave in a similar manner for the great majority of stride variables, EVA, and TPU to a lower extent, can globally shorten the contact time and increase peak vertical forces compared to CON.