Comparison of lower extremity kinematics and ground reaction forces between adolescents who run in motion control shoes with adolescents who run in neutral shoes

Abstract Differences in running biomechanics have been reported in adult long-distance runners when running in motion control or neutral shoes. The influence of growth and coordination refinement during puberty limits the generalisability of adult research to adolescents and it is unknown if biomechanics differ between adolescent who run in different types of shoes. The purpose of our study was to compare lower extremity kinematics and ground reaction forces between adolescents running in motion control or neutral shoes. We matched 18 adolescent long-distance runners who ran in motion control shoes with 18 adolescent long-distance runners who ran in neutral shoes according to running speed, sex, and physical maturation. We collected three-dimensional kinematic and ground reaction force data during overground running and performed statistical parametric mapping to compare joint angle and ground reaction force waveforms between the motion control and neutral shoe groups. We found no significant differences for hip, knee, or ankle joint angles (average differences <2°) or ground reaction forces (average differences <0.05 N/kg) between runners who ran in motion control or neutral shoes. Lower extremity kinematics and ground reaction forces were similar between adolescents who ran in motion control or neutral shoes.


Introduction
Shoe companies manufacture different models of running shoes, including motion control and neutral running shoes, to accommodate consumer requests for comfort and beliefs about the influence of shoes on running-related injuries (RRIs).Motion control shoes are designed to reduce the rate of shoe deformation and limit rearfoot motion as well as frontal and transverse plane movement at more proximal joints (Cheung & Ng, 2007).Conversely, neutral shoes are designed to allow for greater attenuation of ground reaction forces (GRFs) by aiming to not control foot motion (Rose et al., 2011).
Adult runners and healthcare providers believe running shoes are an important factor in reducing the risk of sustaining an RRI (Dhillon et al., 2020) and shoe properties (e.g., stability, cushioning, stack height/toe drop) have been observed to influence risk factors associated with RRIs (Dempster et al., 2021;Malisoux et al., 2022;Willems et al., 2021;Yu et al., 2021).Motion control shoes are designed to decrease peak rearfoot angles which theoretically reduce external moment arms and torques thereby reducing loads applied to musculoskeletal structures.However, prior research in adult runners questioned if there is a meaningful reduction in joint motion when running in motion control shoes.While studies reported that peak frontal and transverse plane rearfoot, ankle, and knee joint angles were smaller when running in motion control shoes compared to neutral shoes (Butler et al., 2007;Cheung & Ng, 2007;Hutchison et al., 2015;Langley et al., 2019;Lilley et al., 2013;Rose et al., 2011), most of the differences were below recently established minimal detectable change (MDC) values for running (Bramah et al., 2021).The lack of differences that exceed MDC values suggests that motion control shoes do not contribute to clinically meaningful reductions in peak joint angles.
Age may influence the interaction between running shoes and running biomechanics.Differences in running patterns were reported between mature (ages 40-60) and young females (ages 18-25) and mature females demonstrated a greater response to motion control shoes than the young adult group (Lilley et al., 2013).Like adults, adolescent long-distance runners may choose to run in motion control shoes with approximately 25% of adolescents running in motion control shoes (Enke et al., 2009).Most adolescent runners (>70%) reported arch type and shoe design were the most important factors for running shoe selection, but nearly half (43%) reported not knowing their arch type (Enke et al., 2009).While shoe type is important to adolescent long-distance runners, it is unknown what implications footwear has on their biomechanics, injury risk, or performance.
The interaction between footwear and running biomechanics has been investigated in adult runners, but the potential confounding effects of growth and maturation on musculoskeletal structure properties (e.g.increases in external moment arms [limb length]) and neuromuscular control (Lloyd et al., 2014) may influence movement patterns for younger, maturing runners (Garcia et al., 2022;Taylor-Haas et al., 2022).Since differences in running mechanics have been observed among adolescents of different stages of physical maturation, it may not be appropriate to generalise results from adult-based research to adolescent runners.Type of preferred footwear design varies among adolescent long-distance runners (Enke et al., 2009), yet no biomechanical footwear research has been conducted on adolescent runners.Investigations comparing footwear for adolescent runners are needed to identify if biomechanical factors associated with RRIs differ among adolescents who run in different types of footwear.Therefore, the purpose of our study was to compare running biomechanics between adolescent long-distance runners who ran in motion control shoes with those who ran in neutral shoes.We hypothesised adolescent long-distance runners who ran in motion control shoes would demonstrate similar joint angles and GRFs compared to adolescents who ran in neutral shoes.

Participants
As part of a larger study (Garcia et al., 2022), we recruited pre-adolescent and adolescent athletes between the ages of 8-19 years who participated in long-distance running activities (e.g., school or club cross-country, track distances !800 m, and/or road races ! 1 mile) for this secondary footwear analysis.We excluded participants if they selfreported a current RRI according to a consensus definition (Yamato et al., 2015).Participants provided signed informed consent (parental permission and assent if age < 18 years) following methods approved by the University's Institutional Review Board.Participants completed a single visit to our motion analysis lab where they underwent a three-dimensional running analysis.

Procedures
We instrumented participants with retroreflective markers and captured marker trajectory data at 120 Hz using a 12camera system (Raptor 4, Motion Analysis Corp.; Rohnert Park, CA).We used a modified Helen Hayes marker set and secured 28 markers on anatomical landmarks (cervical vertebra 7, acromion processes, sternal notch, olecranon processes, midpoint between radial and ulnar styloid processes, anterior superior iliac spines, posterior superior iliac spines, pelvic crests, greater trochanters, lateral and medial femoral epicondyles, lateral and medial malleoli, head of metatarsal 2, and calcaneal tuberosity).We placed the head of metatarsal 2 and calcaneal tuberosity markers on the surface of the participant's footwear.We also secured 21 markers on segments for tracking rigid-body motion (posterior trunk ).We captured an upright static calibration trial to establish joint centres, body segment coordinate systems, and local positions of tracking markers.We assigned an orthogonal axis system to each segment (three degrees of freedom) and defined the XYZ global coordinate system as X ¼ anterior/posterior (þanterior), Y ¼ axial (þup), and Z ¼ medial/lateral (þmedial).We defined hip, knee, and ankle joint angles as the distal segment relative to the proximal segment.
We measured joint kinematics and kinetics while participants ran over a 20 m runway with embedded force plates (Advanced Medical Technology, Inc.; Watertown, MA) and sampled force plate data synchronously at 1800 Hz.Participants wore their own shoes and completed a 5minute warmup period on a treadmill at a self-selected speed.Following the warmup, participants completed 5-10 familiarisation trials over the instrumented runway.We instructed participants to run at a comfortable speed that was neither too fast nor too slow and could be maintained for 20-30 minutes (Taylor-Haas et al., 2022).We monitored running speed with timing gates (TCi System, Brower Timing Systems; Draper, UT) placed 2.6 m apart and the average speed during the familiarisation trials was set as their self-selected speed.Participants ran across the runway until a minimum of 5 successful force plate strikes were recorded for the right limb.We only included trials within 5% of self-selected speed for final analysis.
We filtered marker trajectories and force plate data in Visual3D (C-Motion Inc.; Germantown, MD) using a 2 nd order low-pass filter with cut-off frequencies of 6 Hz and 15 Hz, respectively.We identified initial contact and toe-off events when the vertical force exceeded and dropped below 20 N, respectively.We calculated temporal-spatial parameters (running speed, cadence, step length, stance duration) according to initial contact and toe-off events and normalised GRFs to body mass (N/kg).

Statistical analysis
We retrieved shoe specifications (pronation control [neutral, stability, motion control], forefoot stack height [mm], and toe-drop [mm]) from an online running shoe website (www.runningwarehouse.com).Participants were excluded from the analysis if their shoe specifications were not provided on the website.We categorised running shoes as "motion control" or "neutral" according to the pronation control classification retrieved online and collapsed stability and motion control shoes into the same group.We matched participants who ran in motion control shoes with participants who ran in neutral shoes according to running speed, sex, and stage of physical maturation.We used the modified Pubertal Maturation Observation Scale (Davies & Rose, 1999) to classify participants as pre-, mid-, or postpubertal according to the number of secondary sex characteristics the participant presented with.We compared stance phase (initial contact to toe-off) ensemble waveforms from the right leg for hip, knee, and ankle kinematics, as well as anteroposterior, mediolateral, and vertical GRFs using one-dimensional statistical parametric mapping (SPM) independent t-tests.We conducted the SPM analyses using open-source MATLAB scripts (v.0.4.3, www.spm1d.org) (Pataky, 2010).We compared age, temporal-spatial parameters, and shoe specifications between groups using independent sample t-tests (RStudio v.1.2,RStudio, Inc., Boston, MA).We set statistical significance at p .05 for all analyses.

Discussion
To our knowledge, this is the first study to compare running kinematic and kinetic waveforms between adolescent long-distance runners who ran in neutral or motion control shoes.As expected, we observed similar joint angles and GRFs between adolescent long-distance runners who ran in neutral or motion control shoes.
Motion control shoes are manufactured to provide greater stability and reduce excessive rearfoot eversion (Cheung & Ng, 2007).We did not observe significant differences in shoe eversion between adolescents who ran in motion control or neutral shoes.However, our participants only ran in their preferred footwear.It is unknown if joint angles were similar between groups because footwear does not influence joint angles or because motion control shoes limited joint motion for participants who ran in motion control shoes and would have shown greater motion in neutral shoes (i.e., motion control shoes were recommended to them based on foot structure, running analysis, etc.) A crossover design is needed to investigate this further.Prior crossover studies in adult runners reported lower peak rearfoot eversion angles when running in motion control shoes compared to neutral shoes (Cheung & Ng, 2007;Langley et al., 2019;Lilley et al., 2013), but the clinical meaningfulness of these reported differences is questionable (Cheung & Ng, 2007;Langley et al., 2019;Lilley et al., 2013) as only one of these studies (Cheung & Ng, 2007) exceeded the MDC for peak rearfoot eversion (6.3 ) (Bramah et al., 2021).Our results are therefore similar to previous studies in adults and question the influence of shoe type on foot biomechanics.Most studies, including ours, measured foot motion by securing markers directly to the shoe (Cheung & Ng, 2007;Langley et al., 2019;Lilley et al., 2013), while other studies placed markers directly on the foot through windows cut into the shoe (Butler et al., 2007) or on bone pins (Stacoff et al., 2001).Peak rearfoot eversion angles measured by shoe markers were greater than rearfoot eversion angles measured by foot markers and bone pins (Alcantara et al., 2018;Gheluwe et al., 1995;Sinclair et al., 2013;Stacoff et al., 2001).It is possible we did not observe differences between shoe groups because we measured shoe motion and not foot motion.While it is assumed foot and shoe motion are similar, caution should be used when interpreting peak rearfoot eversion angles measured with shoe-based markers.
We did not observe significant kinematic differences at the knee or hip between adolescents who ran in motion control or neutral shoes.Reduced frontal plane motion at the foot has been proposed to reduce motion at more proximal joints (e.g., knee and hip) because internal rotation of the tibia may be coupled with rearfoot eversion (Dugan & Bhat, 2005).Running in motion control shoes reduced tibial rotation range of motion in adult runners but it was not reported if there was a difference in foot motion (Rose et al., 2011).However, the mean difference between motion control and neutral shoes for tibial rotation range of motion (1.4 ) fell below the MDC (5.2 ) so the differences are not clinically meaningful (Bramah et al., 2021).It is plausible that a reduction in frontal plane foot motion would influence more proximal joint motion.Since we did not observe significant differences in ankle/shoe kinematics between adolescents who ran in motion control or neutral shoes, it is not surprising that knee and hip motion were also not significantly different.Further research is needed to investigate if reducing frontal plane foot motion is coupled with a reduction in frontal plane proximal joint motion.
Adolescent runners in our study ran in a non-exerted state but the intended benefits of motion control shoes (i.e., reduced joint motion) may manifest during a prolonged run when a runner begins to tire.Peak rearfoot eversion and tibial internal rotation were shown to increase in adult runners when running in neutral shoes in an exerted state (Butler et al., 2007;Dierks et al., 2010).However, runners with low arches demonstrated significantly lower peak tibial internal rotation when running in motion control shoes than in neutral shoes at the end of a prolonged run (Butler et al., 2007).It is possible the desired effects of motion control shoes do not present until a runner tires and we may have observed different results had adolescent runners in  We observed no significant differences in GRFs between adolescents who ran in motion control or neutral shoes.Compared to motion control shoes, neutral shoes are often designed to allow for greater force absorption by allowing the foot to freely move (Rose et al., 2011).However, similar to our results, previous studies in adult runners reported no significant differences in vertical GRFs when running in motion control shoes (Nigg et al., 1988;Perry & Lafortune, 1995).As GRFs do not appear to differ among shoes offering different levels of motion control, shoe properties such as cushioning or stack height may have a greater influence on GRFs (Malisoux et al., 2022;Yu et al., 2021).It should be noted forefoot stack height and toe drop were similar between our motion control and neutral shoe groups, but further research is needed in adolescents to better understand how these younger runners respond to shoes with different properties.

Limitations
Our study should be reported within the context of its limitations.First, participants only ran in their own preferred shoe and we are unable to report if running biomechanics in adolescent long-distance runners change when switching from a neutral to a motion control shoe or vice versa.It is possible runners who ran in motion control shoes had similar joint angles to runners who ran in neutral shoes because the shoe reduced joint motion.Still, our results are similar to previous studies (Cheung & Ng, 2007;Langley et al., 2019;Lilley et al., 2013) and demonstrate running biomechanics are similar between motion control and neutral shoes.Second, most shoes worn by runners in the motion control group were deemed to moderately control foot pronation.It may be possible the shoes did not provide adequate stability to influence running biomechanics.Third, arch and foot characteristics were not measured for participants.There may be an interaction between arch structure and response to shoe type (Butler et al., 2007) that may need to be controlled for in future studies.Fourth, the mechanical properties of the shoes are unknown.Adult runners were found to modify running patterns to maintain constant external loads as the running shoe cushioning capacity decreases with more miles run in that pair of shoes (Kong et al., 2009).We did not test the mechanical properties of the participants' shoes nor collect the mileage of the running shoes and differences in shoe properties may need to be considered.However, running shoes have been observed to maintain functional stability for over 1000 km of use (Hennig, 2011).While we acknowledge our limitations, the results of our study begin to fill a knowledge gap in an understudied population.Further investigations are necessary to better understand the interaction between shoe type and running biomechanics in adolescent long-distance runners.
Lower extremity joint angles and GRFs were similar between adolescent long-distance runners who ran in motion control or neutral shoes.Our findings are consistent with previous research in adult runners and further support that footwear does not have a clinically meaningful influence on running biomechanics.
FOOTWEAR SCIENCEour study completed an exhaustive running session.Further research should investigate the role of shoes in adolescent running biomechanics during a prolonged run.

Figure 1 .
Figure 1.Mean hip, knee, and ankle/shoe kinematic waveforms (thick line), 95% CI (shaded area), and SPM output between adolescent long-distance runners who ran in neutral and motion control shoes.Solid red: motion control shoes group; Dashed blue: neutral shoes group; CI: confidence interval; SPM: statistical parametric mapping.

Figure 2 .
Figure 2. Mean ground reaction force waveforms (thick line), 95% CI (shaded area), and SPM output between adolescent long-distance runners who ran in neutral and motion control shoes.Solid red ¼ motion control shoes group; Dashed blue: neutral shoes group; CI: confidence interval; SPM: statistical parametric mapping; GRF: ground reaction force.

Table 1 .
Age and temporal-spatial parameters comparison between adolescent long-distance runners who ran in motion control or neutral shoes.