Foot Morphology and Running Gait Pattern between the Left and Right Limbs in Recreational Runners

Participants

This study was approved by the Nanyang Technological University Institutional Review Board (IRB Number: IRB-2021-124). Female and male recreational runners who were aged between 18 and 45 years old (Dingenen, Barton, et al., 2018) were recruited. The inclusion criteria were that participants 1) were recreational runner and not competing at a club or national level, 2) ran at least once a week for the past three months, 3) had experience in treadmill running and did not require to hold onto the handrails for support, and 4) were able to run continuously at a minimum speed of 8 km/h (2.2 m/s) for at least 10 min. Participants were excluded if they 1) were pregnant during the time of the study (for female participants only), 2) answered ‘Yes’ to any of the questions in the Physical Activity Readiness Questionnaire (PAR-Q+) (Warburton et al., 2019), indicating the presence of any serious health conditions which may affect running performance, 3) received surgeries on the legs within the past year, or 4) sustained serious injuries to lower limbs which required more than 7 days of rest in the past 6 months. Forty-four healthy participants [20 females, 24 males; group mean (standard deviation); age 25.1 (6.5) years old; height 167.2 (7.0) cm; body mass 62.8 (8.1) kg] were selected for analysis.

Procedures

The participants were required to wear their own running shoes. Tight-fitting, short attires were worn to reduce the marker movements during running. Cloth tape was fixed on 13 anatomical points on body in each limb (and hence 26 in total, Figure 1), including anterior superior iliac spine (ASIS), base of patella, talus, point between second and third metatarsals, posterior superior iliac spine (PSIS), mid-shank, Achilles tendon, calcaneus, extended line from greater trochanter towards the knee, lateral femoral epicondyle, lateral malleolus, and 2 markers on the shoe sole (fibular trochlea and fifth metatarsal head) (Dingenen, Barton, et al., 2018; Dingenen, Staes, et al., 2018).

Figure 1

Data acquisition and analysis

A standard Brannock device (The Brannock Device Co., Liverpool, NY, USA; Figure 2) was used to measure the foot lengths, arch lengths, and foot widths for both feet for all participants (Kong et al., 2015). Foot lengths and arch lengths were both expressed in the US sizes. Foot widths were measured as AAA, AA, A, B, C, D, E, EE, or EEE, and coded as 1, 2, 3, 4, 5, 6, 7, 8, or 9; the bigger the number, the wider the foot (Kong et al., 2015; Luximon & Goonetilleke, 2004). Navicular drop and hallux valgus (bunion) angle measures were performed by the same research team member for consistency. To measure the bunion angle (Figure 3a), a top-view, static photo covering both feet was taken for each participant (Fong et al., 2021). The bunion angles were quantified using the Kinovea software (Version 0.8.27, Kinovea, Bordeaux, France; available for download at http://www.kinovea.org). Navicular drop was measured for both feet following the methods provided in the previous studies (Barton et al., 2010; Menz, 1998). A vertical distance of the navicular tuberosity drop from the neutral position (Figure 3b) to the relaxing condition (Figure 3c) was measured.

Figure 2

Figure 3

Prior to the running experiment, each participant was given a 5-min walk-to-run warm-up and familiarization period on the treadmill (h/p/ cosmos saturn, h/p/cosmos® sports & medical gmbh, Nusseldorf-Traunstein, Germany). Immediately after, treadmill speeds were slowly increased until a point whereby the participants felt the speeds were right for them. While running at these self-selected speeds [mean speed: 9.37 (1.12) km/h], the researchers verbally confirmed with the participants again to ensure that they felt comfortable with the running speeds. The speeds were maintained till the end of the running session (around 10 min). The participants were instructed to run using their usual running techniques (i.e., forefoot strike, rearfoot strike) for the entire running session.

During the running session, a digital camera (120 Hz, model EX-100, Casio Computer CO., LTD, Tokyo, Japan) was used to record the participants’ running at five different positions (Figure 4). The five camera views were A) frontal view, B) full body back view, C) lower back view (focusing the ankles and feet), D) left sagittal view, and E) right sagittal view. For each camera position, the camera was positioned approximately 1.5 m away from the participants.

Figure 4

For each participant, the first 30 consecutive strides for each limb, which could clearly show the joints of interest, were analyzed using the free software Kinovea. Ten kinematic variables were measured, similar to the video-based analysis methods that have been reported in the previous studies (Dingenen, Barton, et al., 2018; Maykut et al., 2015; Pipkin et al., 2016). At midstance, peak hip adduction, peak knee abduction, and peak foot adduction were obtained from the frontal view. In the same gait event, peak contralateral hip drop and foot eversion angle were taken from the back view, while peak knee flexion and peak ankle dorsiflexion were taken from the sagittal views. At the moment of footstrike, foot inversion angle from the frontal view was measured. The time taken to achieve peak foot eversion at midstance was also calculated and expressed as a percentage of the total the time taken from the point of footstrike to the moment of toe-off for the same foot. Lastly, foot eversion excursion (difference between foot inversion angle at footstrike and foot eversion angle at midstance, from the frontal view) was computed. The average values of all variables across the 30 strides were used for analysis.

Statistical analysis

The data were imported into JASP (version 0.16.3; JASP Team, 2022) statistical software for analysis. According to the results of the Shapiro-Wilk test, the assumption of data normal distribution was violated for some variables investigated and hence, non-parametric statistical tests were applied. Wilcoxon signed-rank test was performed to compare the foot morphological characteristics and running kinematics between the left and right limbs. Effect size (r) was calculated from the Z values and interpreted as small (0.1 ≤ |r| < 0.3), medium (0.3 ≤ |r| < 0.5), or large (|r| ≥ 0.5). All statistical tests were set at the 0.05 level.

Foot morphology

To quantify foot morphology, this study measured participants’ foot dimensions (foot size, foot width, and arch length), bunion angle, and navicular drop. No significant differences were observed for any foot morphological variables (Table 1). The bunion angle was measured using a recent method proposed by Fong et al. (2021) based on top-view photos and 2D analysis. This new method exhibited excellent intra-rater and inter-rater reliability and good agreement with the widely used Manchester scale (Garrow et al., 2001). However, it is acknowledged that the measured bunion angles should not be directly compared with other radiographic or magnetic resonance image measures typically used in clinical settings (Heineman et al., 2018; Ortiz et al., 2016). Navicular drop can objectively measure the degree of foot pronation and has been used to classify the runners in previous studies (Eslami et al., 2014; Eslami & Ferber, 2013). Recent studies used the navicular drop measures as the criterion (e.g., the cue-off value of 10 mm) to classify flat-foot and normal-foot participants (Koh et al., 2020; Ng et al., 2021). In this present study, the navicular drop values were largely similar between two feet [group median (interquartile range, IQR); left foot 6.0 (4.0) mm, right foot 7.0 (4.5) mm]. The results of relatively small navicular drop values were not surprising as the current study only included healthy, asymptomatic participants. Collectively, foot dimensions, bunion angle, and navicular drop measures indicated similar foot morphology between the left and right feet among a group of healthy recreational runners with no symptoms of running-related injuries in this study.

Running kinematics

Asymmetrical gait patterns may induce running-related injuries (Zifchock et al., 2008), though gait asymmetry is common even for healthy runners with no symptoms (Hanley & Tucker, 2018; Radzak et al., 2017). In this study, several between-limb differences in running kinematics at midstance were identified for peak knee flexion, peak hip adduction, and peak foot eversion (Table 1, Figure 5). Although the magnitudes of the differences were small, these biomechanical variables were reported to be risk factors for running-related injuries (Ceyssens et al., 2019). For example, a small peak knee flexion may lead to insufficient load absorption during the stance phase and hence injuries to the Achilles tendon (Hein et al., 2014). A great peak hip adduction was found related to iliotibial band syndrome (Noehren et al., 2007), while conflicting findings were found for the relationship between peak foot eversion and injury risks (Ceyssens et al., 2019). This study identified between-limb differences in running kinematics but similar foot morphological characteristics between the left and right feet. Many factors can contribute to the gait asymmetry among healthy, asymptomatic runners, for example, muscle fatigue (Arampatzis et al., 1999) and limb dominance (Sadeghi et al., 1997). The unbalanced lower extremity kinematics may contribute to injuries in the future but may not be associated with the foot morphology according to the results of this study.

One foot or two feet

Among the studies on the relationship between foot morphology and gait biomechanics, some researchers determined participants’ foot types (e.g., flat foot) of both feet while did not clearly state the use of one limb or the average of both for gait analysis (Koh et al., 2020; Ng et al., 2021). While participants might have similar foot types bilaterally, this approach cannot capture the possible between-limb differences in running biomechanics. These earlier studies (Koh et al., 2020; Ng et al., 2021) could share similar observations as the present study where there were bilateral similarity in foot morphology and biomechanical differences in gait between limbs. As shown in Table 1 and Figure 5, a few kinematic variables were significantly differed between limbs. Hence, using one limb (or the average of both) to represent the running kinematics for both limbs may be inappropriate due to potential gait asymmetry. Since relatively short running time (less than 10 min) were examined in the studies, similar to this present study, the effects of fatigue (Arampatzis et al., 1999) may be minimal but other factors such as limb dominance (Sadeghi et al., 1997) could possibly lead to asymmetrical gait patterns. On the other hand, some studies either used one foot to classify participants into different groups with different foot types (Eslami et al., 2014; Eslami & Ferber, 2013; Zhang et al., 2017), or did not clearly indicate the use of one or the average of both feet (Kruger et al., 2019). From a practical standpoint, this practice may also be problematic as it is possible for one person to possess a flat (or over-pronated) foot on one side and a normal foot the other. If only one foot was arbitrarily chosen for foot type classification, the participant could be potentially allocated into either a normal group or flat (over-pronated) group. Such grouping methods can confound the research study design, making it difficult to draw meaningful conclusions from the data. While this present study did not find significant differences in any foot morphological characteristics investigated, there were subtle differences between the left and right feet (Table 1). Thus, researchers and practitioners should be aware of the potential between-limb asymmetry in both running kinematics and foot morphology, and exercise caution when examining the relationship between these characteristics.

Limitations

There were a few limitations to the present study. The foot morphology measurements were performed by an undergraduate sport science student without formal clinical training. The measured results may not be comparable to those measured by physicians, physiotherapists, or podiatrists who are clinically trained. In addition, this study adopted a 2D video analysis approach involving manual digitization of peak joint angles at the moments of footstrike and midstance. While this approach increases practicality in terms of clinical applications, it does not provide as much information as an analysis of the entire stance phase. Finally, this study also did not assess the intra-rater or inter-rater reliability for the 2D photo/video analysis including the bunion angle measurement and running kinematic analysis due to limited manpower. Future studies are recommended to assess the intra-rater and inter-rater reliability to supplement the measurement accuracy.

Conclusions

In conclusion, the present study comprehensively examined the foot morphology and running kinematics of both limbs. While no significant differences were identified between both feet for the foot morphological characteristics, differences in several kinematic variables were found between the left and right limbs during running. The results indicate that runners with similar morphological characteristics between the left and right feet can display between-limb gait asymmetry during running. Future studies should avoid arbitrarily analyzing one limb to represent a runner’s gait or foot morphology. If it is necessary to select one limb to represent both, between-limb symmetry should be first confirmed.