Acute effects of footwear on running impact loading in the preschool years

ABSTRACT

The aim of this study was to assess acute effects of footwear conditions (barefoot, minimalist and standard running shoes) on running impact loading in the preschool years. Fourty-eight habitually shod preschool children (26 males and 22 females) were divided into four age groups (3-, 4-, 5- and 6-year-old). Children performed a simple running game in three different conditions. Three-dimensional biomechanical analysis was carried out during overground running. Statistical parametric mapping was performed on the vertical ground reaction force profiles during the stance phase and one-way repeated measures ANOVA on the vertical instantaneous loading rate. Three-year-old children displayed significantly lower vertical ground reaction force values in the barefoot condition compared to minimalist (3-7% stance) and standard running shoes (7-11% stance). There were also differences in vertical instantaneous loading rate, where 3-year-olds had lower loading when barefoot than in minimalist (p = 0.010, d = 1.19) or running shoes (p = 0.045, d = 0.98), despite no differences in the footstrike pattern (mostly rearfoot-midfoot strike). No differences were found for the older children. Running in minimalist shoes did not imitate barefoot running in 3-year-old children. On the contrary, increased loading was observed in minimalist shoes in early running developmental stages.   

KEYWORDS:

• Barefoot
• children
• ground reaction force
• minimalist shoes
• standard running shoes

Introduction

It has been shown that running is the most common physical activity in preschool age children (Wood et al., 2020). Keeping children physically active should be of great interest to parents. The importance of physical activity has been shown for brain development (Meijer et al., 2020), academic success (Kall et al., 2015) and long-term health benefits (e.g., skeletal, mental, cardio-respiratory, etc.) with physical activity continuing from childhood to adulthood (Loprinzi et al., 2012). Although running routines lead to highly commendable health benefits, there also remains a high incidence of running-related injuries in youth and adult runners (Kluitenberg et al., 2015; Krabak et al., 2016). Therefore, some researchers suggested the importance for focusing research on children (Hollander et al., 2014; Pedro Á. Latorre-Román et al., 2018; Plesek et al., 2021), to potentially elucidate the way children interact with the environment during running from a mechanical loading perspective (Krabak et al., 2021, 2019; Plesek et al., 2021).

In the past two decades, many researchers have focused on investigating impact loading during running as represented by the magnitude and shape of the ground reaction force (Davis et al., 2016; Davis & Hollander, 2019; Hollander et al., 2019; Lieberman et al., 2010). In particular, high vertical loading rates have been associated with bony and soft tissue injuries in runners (Davis et al., 2016; Van der Worp et al., 2016; Zadpoor & Nikooyan, 2011). Previous studies on adults and adolescents showed that the type of footwear and footstrike pattern could affect impact loading of the lower limb during running (Davis & Hollander, 2019; Davis et al., 2021; Hannigan & Pollard, 2020; Hollander et al., 2019; Lieberman et al., 2010; Rice et al., 2016). In addition, running barefoot or running in minimalist shoes with a forefoot strike pattern may lead to lower impact loading and the visual absence of the impact peak on the vertical ground reaction force component (VGRF; Lieberman et al., 2010; Rice et al., 2016). Davis and Hollander (2019) confirmed that the interaction of footwear and footstrike patterns is an important issue associated with loading in specific musculoskeletal sites. They suggested that, if all children would start to use a minimalist shoe early in their lives, this would likely lead them to adopt forefoot strike running patterns and consequently reduce impact loading. Additionally, Krabak et al. (2019) also proposed that perhaps toddlers and young children could start to wear footwear that would promote a forefoot strike pattern during running. These researchers suggested that this would potentially lead to strengthening of the foot and ankle, along with the aforementioned lowering of impact forces with a future possible reduction of running-related injuries. On the other hand, running barefoot or in minimalist shoes does not necessarily mean that most people will run with forefoot strike patterns. Evidence of this has been shown in some habitually barefoot populations (Hatala et al., 2013; Hollander et al., 2018; Pontzer et al., 2014) and also in 3–4-year-old habitually shod children (Plesek et al., 2021).

Through the type of footwear selected, parents buying footwear for their preschool children could indirectly affect their movement pattern (Pedro Á. Latorre-Román et al., 2018; Plesek et al., 2021). Moreover, underlying kinematic and kinetic differences have been observed in the development of running in comparing preschool and older children (P. Á. Latorre-Román et al., 2021; Fortney, 1983; Latorre Román et al., 2019; Plesek et al., 2021). A recent study reported that footstrike patterns change differently in age groups of preschool children (Plesek et al., 2021). This recent research showed that 3- and 4-year-old children did not change their footstrike pattern according to the footwear condition (maintained rearfoot-midfoot strike in all footwear conditions) as 5- and 6-year-old who displayed a similar acute response to adult runners (shifted to forefoot strike in barefoot and minimalist shoes). As we know, children are not little adults, and we should be very cautious with application of knowledge from research in adult populations and transfer it to children, particularly to the youngest when running skill is still developing.

There are very little data on the effect of footwear on running impact loading in children. Hollander et al. (2014) compared running mechanics in barefoot, minimalist and cushioned shoes in preadolescent children. They reported the highest impact forces in cushioned shoes compared to minimalist and barefoot conditions (with the lowest forces in barefoot). Barefoot running and minimalist shoes displayed the most similarities (in the shape of the VGRF curve). To our knowledge, no study has investigated the effect of footwear on impact loading during running in preschool age (crucial developmental period for running skill). Therefore, the aim of the current study was to compare impact loading in preschool years in different footwear conditions during running. We hypothesised that higher impact loading would occur in barefoot and minimalist shoes compared to standard running shoes in younger children (3- and 4-year-olds), but that impact loading would be lower in the older children (5- and 6-year-olds).

Methods

Study design

This study used a randomised crossover design with three randomly counter-balanced treatments (three footwear conditions—barefoot (B), minimalist (M) and standard running shoes (SRS)). This study was approved by the Ethical Board of the University of Ostrava, Faculty of Education (protocol ID: OU-54483/45-2019) and was conducted in accordance with the principles of the Declaration of Helsinki. Before participating in the study, the children’s parents or legal guardians granted informed written consent and the child’s assent was also required before the start of the measurement.

Participants

This is a further analysis from the data collection of children in the recent study (participants consisted of the same groups of children) where we reported changes in lower limb kinematics at footstrike in three different footwear conditions (Plesek et al., 2021). Forty-eight healthy, typically developing Caucasian preschool children (26 boys and 22 girls aged 3–6 years) participated in this study. They were habitually shod and came from the Moravian-Silesian region in Czech Republic. All children attended a pre-primary educational child care institution or the first grade of primary education. Exclusion criteria for the current study included developmental coordination disorders or any lower limb abnormalities diagnosed by podiatrist, orthopaedist or practitioner (e.g., flatfoot, leg length discrepancy, genu varum and genu valgum). In addition, all participants were tested using the Movement Assessment Battery for Children-Second Edition (MABC-2) to reveal possible developmental coordination disorders. As an inclusion criterion, the limit for this study was set for over the 15th percentile for a certain age according to Czech norms of the MABC-2 (15th percentile is considered as a boundary for typically developing children from the motor developmental perspective, for given age; Psotta, 2014). No stratification for sex was required because a recent study showed no differences in footstrike patterns between preschool boys and girls (Pedro Á. Latorre-Román et al., 2018). Nevertheless, both sexes in each age group were included in the current study (Table 1).

Table 1. Participants' Characteristics.

	Age group
	3 years	4 years	5 years	6 years	p-value
Number of participants	12	12	12	12
Sex male/female	5/7	7/5	5/7	9/3
Age (year)	3.6 (0.3)	4.6 (0.3)	5.6 (0.3)	6.5 (0.4)	<0.001
Antropometric data
Height (m)	1.00 (0.04)	1.07 (0.03)	1.15 (0.05)	1.22 (0.06)	<0.001
Leg length (m)	0.45 (0.02)	0.49 (0.03)	0.54 (0.03)	0.59 (0.03)	<0.001
Mass (kg)	15.5 (1.6)	17.3 (1.3)	21.1 (3.8)	24.1 (4.1)	<0.001
BMI (kg/m^2)	15.5 (1.4)	15.0 (0.7)	15.7 (1.6)	15.9 (1.6)	0.441
BMI (percentile)—CZ norms	48.0 (29.3)	38.8 (17.7)	54.7 (30.4)	54.9 (25.7)	0.404
BFQ
'Total score	13.4 (1.8)	14.3 (2.5)	14.7 (1.9)	14.6 (1.2)	0.329
Warm weather	6.4 (1.0)	6.9 (1.4)	7.0 (1.1)	7.1 (0.7)	0.336
Cold weather	7.0 (1.0)	7.4 (1.0)	7.7 (0.9)	7.5 (0.7)	0.451
MABC-2
Total score	84.8 (10.1)	83.1 (7.9)	85.4 (6.4)	82.6 (6.7)	0.791
Percentile	67.8 (28.4)	66.9 (26.2)	71.5 (19.6)	62.0 (22.1)	0.769

Data are presented as a mean (SD). Abbreviations: BFQ–Barefoot Questionnaire, BMI–Body Mass Index. MABC-2–Movement Assessment Battery for Children-Second Edition. Bolded values represent significant differences between age groups (p < 0.050).

Instrumentation

Kinematic and kinetic data were collected by using a 10-camera 3D motion capture system (Oqus, Qualysis, Inc., Göteborg, Sweden) sampling at 240 Hz and recorded synchronously with ground reaction forces from three force platforms at 1200 Hz (one large—90 x 90 cm 9287CCAQ02 and two small—60 x 40 cm 9286AA, 9281CA, Kistler, Winterthur, Switzerland). Height and body mass of all children were measured using a stadiometer (In Body 370, Biospace, Seoul, South Korea) and a body composition analyser (Inbody 770, Biospace, Seoul, South Korea). Body mass index (BMI) was calculated by dividing the body mass (kilograms) by body height² (metres squared), and BMI percentile was evaluated for each child according to Czech norms (Vignerova et al., 2006).

Protocol and setting

The detailed protocol including a diagram of laboratory and game settings was described in the study by Plesek et al. (2021).

Briefly, prior to biomechanical measurements, parents returned a six-item Barefoot Questionnaire (BFQ) evaluating how frequently their children wore shoes in two different weather conditions (cold and warm) and three environments (educational institution, sports and in and around the house; Hollander et al., 2016). This BFQ was modified for preschool children and also described in the previous study (Plesek et al., 2021). All children accompanied by their parents or family members visited the Human Motion Diagnostic Centre Biomechanics Laboratory at the University of Ostrava. Parents chose the right shoe size for their child before the data collection began (sizes ranged from 24 to 35 EU, from 8 K to 3.5 US). The intervention shoes consisted of true minimalist shoes (Leguanito, Leguano, Buchholz, Germany; with minimalist index 96%; mass 49–90 g; stack height 6 mm; with 0 mm heel to toe drop, no stability or motion control technology and extreme longitudinal and torsional flexibility) and standard running shoes (FortaRun CF I and K, Adidas, Herzogenaurach, Germany; with minimalist index 30%; mass 115–177 g; stack height 12–21 mm; with 4–10 mm heel to toe drop, one device of stability and motion control technology = extended medial flare, moderate resistance to longitudinal bending and high resistance to torsion). Both types of shoes were commercially available in all children sizes. The minimalist index score ranges from 0% (least minimalist) to 100% (most minimalist; Esculier et al., 2015). The following step included marker placement on each child carried out by an experienced researcher in 3D movement analysis. Retroreflective markers (6.4-mm diameter) were placed on each child’s landmarks (18 right limb, 4 on pelvis, 1 on the left greater trochanter and 1 left heel) according to the recommendation of Visual 3D software (C-motion, Germantown, MD, USA; Hamill et al., 2013). Calibration markers on the right foot/shoe were positioned over the first and fifth metatarsal heads, and triads of tracking markers were attached in triangle on the heel over bare calcaneus or intact shoe (Figure 1).

Figure 1. Footwear and marker placement on the children's bodies: (a) barefoot, (b) minimalist shoe and (c) standard running shoe.

Participants performed a very simple running game (back and forth submaximal self-selected speed running between two posts). The game based on shuttle running locomotion was presented to the children in a random counterbalanced order due to three different footwear conditions. The distance between the two ends of the runway (hard, flat and non-slippery surface) increased progressively by 1 metre according to the age of participants (3-yr = 10 m, 4-yr = 11 m, 5-yr = 12 m and 6-yr-old = 13 m). One running game consisted of eight running trials and was accomplished two times in each footwear condition (with at least a 3-minute long break between games). There was also an approximately 10–15-minute break between footwear conditions, while markers were removed and replaced on the child’s foot or intact shoe, followed by a standing calibration. Separate calibrations were recorded before each running footwear condition. Before each game in the new footwear condition, each child started by four non-recorded running trials which served for familiarisation and habituation to the footwear. Sixteen recorded running trials/footwear condition/participant were collected for the 3D biomechanical analysis.

Data processing

We compared the main dependent variables in this study (vertical ground reaction force (VGRF) and maximal vertical instantaneous loading rate (VILR)) among three footwear conditions (B, M and SRS) in four age groups of preschool children (3-, 4-, 5- and 6-year-old). Frequency analysis of nominally classified footstrike patterns based on the strike index (rearfoot strike/midfoot strike/forefoot strike) was the secondary outcomes (Gruber, Boyer et al., 2013). To avoid the effects of running speed on the aforementioned dependent variables, we calculated a dimensionless speed as a control variable based on pelvic horizontal velocity ($v$) and lower extremity length ($l_o$; Hof, 1996),

(1) $\mathrm{d}\mathrm{i}\mathrm{m}\mathrm{e}\mathrm{n}\mathrm{s}\mathrm{i}\mathrm{o}\mathrm{n}\mathrm{l}\mathrm{e}\mathrm{s}\mathrm{s}\mathrm{s}\mathrm{p}\mathrm{e}\mathrm{e}\mathrm{d}=\mathrm{v}/(\sqrt{}{\mathrm{g}\mathrm{l}}_{\mathrm{o}}).$(1)

Biomechanical data were processed by using Qualisys Track Manager (QTM, Qualisys, Göteborg, Sweden) and Visual 3D software (C-motion, Germantown, MD, USA). A threshold of 15 Newtons in the resultant of ground reaction force was used to identify initial contact/footstrike and toe-off/take-off (stance phase). Six successful trials were selected and analysed based on pelvis velocity (the closest six trials to the median). A successful running trial was determined when the entire right foot of the child contacted the force plate. A low-pass Butterworth filter using a cut-off frequency of 10 Hz was applied for kinematics and 50 Hz for ground reaction force data.

The VGRF curve was normalised into 101 points (0–100% stance phase) for following statistical parametric mapping analysis (Pataky, 2010). The VILR was determined by calculating the first derivative of the corresponding VGRF with respect to time. Both VGRF and VILR were normalised to the body mass of each participant. The maximum (peak) of VILR value was obtained within the first 27 % of stance as a local maximum using the same approach as in Rice et al. (2016). In almost all analysed trials, VILR appeared in the first 50 ms after footstrike (except for 1 trial out of 864 trials). According to Ueda et al. (2016), VILR is preferable to the vertical average loading rate (VALR) for more consistent evaluation among methods. The strike index was analysed based on the location of the centre of pressure at the initial foot-ground contact. For this parameter, three zones were determined: 1) rearfoot strike (RFS) 0–33.3%; 2) midfoot strike (MFS) 33.4–66.6% and 3) forefoot strike (FFS) 66.7–100% of the shoe/foot length measured from the heel (Cavanagh & Lafortune, 1980; Gruber, Boyer et al., 2013).

Statistical methods and study size

A one-way ANOVA or Kruskal-Wallis test for non-normally distributed data was used to compare characteristics (age, height, leg length, weight, BMI, BFQ and MABC-2) of participants among the age groups.

The vertical ground reaction force component was analysed in Matlab (MATLAB R2017a, Mathworks, Inc., Natick, MA, USA), and an open-source statistical parametric mapping (SPM) software package was used to carry out the statistical analysis of the VGRF during the entire stance phase (Pataky, 2010). One-way repeated measures ANOVA was performed for an initial analysis. The critical threshold for statistical significance in the initial analysis was set at p = 0.05. When significant effects were found in a certain age group, then a post hoc analysis was carried out by two- tailed paired t-tests with an adjusted alpha level of significance (p = 0.017).

All zero-dimensional scale data (VILR, dimensionless speed) were screened for normality by the Kolmogorov-Smirnov test. Non-normally distributed data were analysed by the non-parametric Friedman test which is recommended for testing-related samples with a Wilcoxon signed rank test post hoc tests where necessary. If a normal distribution of the data was assumed, then we used one-way repeated measures ANOVA with Bonferroni correction for all pairwise comparison in post hoc analysis. In addition, if sphericity was violated (p < 0.05) and ε ≥ 0.75, then a Huynh-Feldt correction was used. Practical significance (effect size) for the main effect in the one-way repeated measures ANOVA was assessed using η² (small 0.01–0.06), medium 0.07–0.14, and large >0.14) and Cohen´s d (small 0.2–0.5, medium 0.5–0.8, and large >0.8) in pairwise comparison of footwear conditions (Cohen, 1988). In addition, we also presented differences in means for comparisons (MD) and 95% confidence interval (CI) for difference as recommended by Harrison et al. (2020). An alpha level of significance was set at 0.05 for all tests. All analyses were performed in SPSS 24 (IBM, Armonk, NY, USA).

A chi-square goodness of fit test was used to assess frequencies of the categorised footstrike pattern (RFS/MFS/FFS), between three footwear condition in each age group separately (72 trials/footwear condition/age group). The alpha level was set at 0.05. Pairwise comparison between B-M, B-SRS and M-SRS was also performed with Bonferroni correction (p = 0.017).

An a priori sample size estimation was performed for running footstrike patterns and footwear in different age groups of preschool children analysis in the previous study by Plesek et al. (2021). In the aforementioned study, a statistical power analysis software programme (G*Power 3.1.9.7, Düsseldorf, Germany) suggested having 12 children per age group. A one-way within-subject repeated measures ANOVA with 12 participants used for the current study would be sensitive to effects of η² ≥ 0.33 with minimal statistical 80% power (alpha level = 0.05). However, to ensure correct data interpretation, we used practical significance (biological relevance) of the differences between mean values of footwear conditions in each age group (Cohen, 1988).

Results

There were significant differences in age, height, leg length and mass among age groups. No significant differences were found in BMI percentile (CZ norms), MABC-2 (CZ norms) and BFQ (Table 1).

In the SPM analysis of the VGRF, differences among footwear conditions in each year of preschool age were observed (Figure 2). Further post hoc analysis revealed significant differences in 3-year-old children between barefoot and both shod conditions during impact peak (B < M 3–7% stance, p = 0.012; B< SRS 7–11% stance, p = 0.009) and also in 5-year-old children between barefoot and standard running shoes (B< SRS 10–14%, p = 0.006). In addition, 3-year-olds were found to have higher values of VGRF in minimalist shoes compared to standard running shoes during 0–4% of the stance phase (p = 0.007). No differences were found between footwear conditions in the VGRF curve in 4- and 6-year-old children during 0–50% of stance phase.

Figure 2. Vertical ground reaction force (VGRF) during the stance phase: (A) 3-year-old, (B) 4-year-old, (C) 5-year-old and (D) 6-year-old children. Data are presented as mean and standard deviation (SD). Statistical parametric mapping (SPM)—vertical ground reaction force during first 50 % of the stance phase. Initial analysis: One-way repeated measures ANOVA (SPM (F)). Post hoc analysis: Paired, two-tailed t-test (SPM (t)). Abbreviations: B–barefoot, BW–body weight (%), M–minimalist shoes and SRS–standard running shoes.

Significant differences in VILR were observed among footwear conditions in 3-year-old children (p = 0.005, η² = 0.386). Further pairwise comparison indicated that 3-year-olds had a significantly lower VILR in the barefoot condition than in minimalist or standard running shoes with a large effect size (p = 0.010, d = 1.19, MD = −57.98, 95% CI for difference [−102.04; −13.92]; p = 0.045, d = 0.98, MD = −49.73, 95% CI for difference [−98.45; −1.02]), respectively (Figure 3). There were no footwear condition differences in VILR in the older age groups (4-year-old, p = 0.195, η² = 0.138; 5-year-old, p = 0.159, η² = 0.154; and 6-yr-old children, p = 0.159, η² = 0.154). Practically significant differences with a medium effect size were found among three footwear conditions in 4-year-old children, and pairwise comparison indicated a medium effect size between B < M (p = 0.290, d = 0.58, MD = −47.35, 95% CI for difference [−120.91; 26.21]), B < SRS (p = 0.378, d = 0.50, MD = −36.47, 95% CI for difference [−98.59; 25.64]) and a very small effect between M > SRS (p = 1.000, d = 0.12, MD = 10.87, 95% CI for difference [−75.03; 96.78]). In 5- and 6-year-old children, pairwise comparison showed practical differences between conditions either small or very small.

Figure 3. Maximal vertical instantaneous loading rate. One-way repeated measures ANOVA. Pairwise comparison of three footwear conditions in each age group. Data are presented as mean standard deviation (SD). Abbreviations: B–barefoot, BW–body weight, M–minimalist shoes and SRS–standard running shoes. Significant differences: *B-M, #B-SRS, with Bonferroni correction. *p < 0.050, ** p < 0.010, *** p < 0.001, # p < 0.05, ## p < 0.010 and ### p < 0.001.

The results obtained from the frequency analysis of three types of footstrike patterns (RFS, MFS and FFS) are shown in Table 2. We found significant differences in frequencies of RFS trials when comparing B with SRS conditions in 4-year-old (SRS>B, p = 0.005) and 5-year-old children (SRS>B, p = 0.008). Moreover, there were differences in the number of FFS when comparing B with SRS and M with SRS, in 4-year-olds (B> SRS, p = 0.001; M> SRS, p = 0.012), 5-year-olds (B> SRS, p = 0.001; M> SRS, p = 0.001) and 6-year-olds (B> SRS, p = 0.001; M> SRS, p = 0.001), respectively. No differences were found in frequencies of any footstrike pattern between footwear conditions in 3-year-old children.

Table 2. Secondary outcome (footstrike pattern—categorisation according to the strike index) and control variables (dimensionless speed).

	Age group
	3 years			4 years			5 years			6 years
	Frequencies (%)
Secondary outcome variable	B	M	SRS	B	M	SRS	B	M	SRS	B	M	SRS
Rearfoot Strike	22 (31%)	38 (53%)	37 (51%)	19 (26%)^#	27 (38%)	41 (57%)^#	21 (29%)^#	25 (35%)	42 (58%)^#	18 (25%)	24 (33%)	31 (43%)
Midfoot Strike	34 (47%)	24 (33%)	27 (38%)	28 (39%)	30 (41%)	27 (38%)	18 (25%)	18 (25%)	24 (33%)	14 (19%)	14 (19%)	29 (40%)
Forefoot Strike	16(22%)	10 (14%)	8 (11%)	25 (35%)^#	15 (21%)^§	4 (5%)^#,§	33 (46%)^#	29 (40%)^§	6 (9%)^#,§	40 (56%)^#	34 (47%)^§	12 (17%)^#,§
Control variable	Mean (SD)
Dimensionless speed	1.20 (0.94)	1.26 (0.14)	1.24 (0.15)	1.34 (0.17)	1.32 (0.21)	1.26 (0.18)	1.25 (0.09)	1.26 (0.16)	1.25 (0.14)	1.26 (0.13)	1.28 (0.14)	1.29 (0.15)

Data are presented as frequencies (%) and mean (SD). Pairwise comparison of three footwear conditions in each age group for rearfoot strike, midfoot strike and forefoot strike. Dimensionless speed = v/(√gl_o). Abbreviations: B–barefoot, M–minimalist shoes, SRS–standard running shoes, v-pelvis velocity, g–gravitational acceleration, l_o–-leg length. Bolded values represent significant differences between *BxM, ^#BxSRS and ^§MxSRS.

No differences were found among footwear conditions in dimensionless speed (p > 0.05), analysed by one-way repeated measures ANOVA in each age group separately (Table 2).

Discussion and implications

The aim of this study was to compare impact loading during running with regard to different footwear in preschool years. We hypothesised that running barefoot and in minimalistic shoes would increase impact loading compared to standard running shoes in 3- and 4-year-old children because in these footwear conditions, they displayed a rearfoot-midfoot footstrike with a significantly less plantar flexed ankle compared to 5- and 6-year-olds (Plesek et al., 2021). The results of this study showed contrary findings than we hypothesised. However, the minimalist shoes displayed the highest impact loading in 3-year-old children.

In 3-year-old children, barefoot running reduced impact loading compared to both shod conditions

This study showed that VILR was significantly lower in the barefoot condition compared to minimalist or standard running shoes in 3-year-old children. This evidence is in the agreement with results of statistical parametric mapping of the VGRF, whereas impact peaks (3–11% stance phase) reach significantly higher values in the minimalist and standard running shoes than in the barefoot condition. Barefoot running may be a painful physical activity, and it has been shown that only 35% of adult runners maintain rearfoot strike on the hard surface compared to 80% on a soft surface (Gruber, Silvernail et al., 2013). We also observed this in our study in the 3-year-old preschoolers (31% RFS, 47% MFS, 22% FFS) even though a previous study suggested insufficiency to forefoot strike landing in such a young children based on less plantar flexed ankle at footstrike (Plesek et al., 2021).

A midfoot strike in adult runners exposes the body to lower mechanical loading (loading rates) compared to rearfoot strikers (Almeida et al., 2015). In the current study, three-year-old children showed mostly a midfoot strike pattern, particularly in barefoot condition. A midfoot strike and higher nociceptive stimulation of lower limb flexion reflex could consequently reduce an abrupt impact loading in the short period of time after initial contact (Cornelissen et al., 2013; Spaich et al., 2004). On the other hand, from the perspective of cutaneous nociceptive stimulation, it seems that 3-year-olds in minimalist shoes with a rearfoot-midfoot strike immediately after foot contact are not able to react due to lower sensitivity of the location of stimulus (heel has higher pain threshold than rest of the foot; Spaich et al., 2004). In addition, the VGRF curve in the current study increased immediately after foot contact (0–4% of stance phase) in minimalistic shoes and reached higher values than in SRS. An explanation could be that the peak of VILR usually occurred earlier in minimalist shoes and its magnitude is relatively very high in the youngest children. An absence of a cushioning sole probably could lead to a higher mechanical loading immediately after foot contact. In previous studies, several occult fractures of calcaneus in toddlers without any history of significant acute injury have been reported (Laliotis et al., 1993; Starshak et al., 1984). Some researchers suggested that it could indicate some possible overuse of the calcaneal bone from locomotion (Zeininger et al., 2018). From this perspective, one should be careful and reconsider using minimalist shoes and excessive exposure to running activities on hard surfaces in 3-yr-old children because this could possibly contribute to risk factors of calcaneal injuries in such young children.

In 4-6-year-old children, footwear did not affect impact loading as in 3-year-olds

Due to a lack of studies investigating impact loading in preschool children during running, it is difficult to compare ground reaction forces or loading rate values with previous research. However, a study published by Hollander et al. (2014) investigated the effect of footwear on running biomechanics in children (aged 6–9 years). This aforementioned study reported significantly lower magnitude of impact peaks in barefoot running and in minimalistic shoes compared to standard running shoes during treadmill running (fixed velocity at the level of 10 km/h which is close to running pace in the current study based on dimensionless speed). In the current study, we showed that from age 4 to 6, children demonstrated more forefoot strike patterns during barefoot running and also in minimalist shoes compared to standard running shoes. Surprisingly, a forefoot strike pattern only in 5-year-old children led to lower impact forces when comparing barefoot and standard running shoes (according SPM). It seems that, from an impact loading perspective, 4- to 6-year-old children are not as sensitive to changing footwear or could possibly adjust their movement patterns via a lower joint limb re-alignment (e.g., more plantar flexed ankle and more eccentric capacity) when they are not fatigued (Plesek et al., 2021; Rose et al., 2009). However, children in the present study (4 to 6-year-olds) in minimalist shoes had the highest loading rate despite a midfoot-forefoot strike. Previously, in adult runners, different results regarding loading rates with respect to footwear and footstrike have been observed. For instance, Rice et al. (2016) found significantly lower loading when runners used minimalist shoes and a forefoot pattern compared to rearfoot and forefoot strikers running in SRS. Moreover, Hollander et al. (2019) demonstrated that adult runners had significantly lower impact loading during an acute response to barefoot running compared to running in SRS. This was not seen in the 4-year-old and 6-year-old children in the current study. Nevertheless, the differences in the VGRF curve (in the time of impact peak) between barefoot and SRS would be confirmed in all preschool years (3-, 4-, 5- and 6-year-olds) if the least significance difference technique of the post hoc analysis was used as previously recommended by several researchers (Meier, 2006; Rothman, 2014).

Perspective

Based on the evidence from studies published on adult runners, it has been suggested that running toddlers and preschool children use true minimalist shoes along with barefoot when they start running. These suggestions have been made with the purpose of reducing impact loading and consequently minimise the risk of musculoskeletal system overload (Davis & Hollander, 2019; Krabak et al., 2019). Recently, it has been claimed that gait in minimalist shoes seemed to replicate many aspects of barefoot walking and running mechanics, while children's soles would have been protected (Davis et al., 2021). However, the authors also admitted that there is still missing research evidence to support recommendations of using minimalist footwear for running in young children (Davis et al., 2021; Krabak et al., 2021). The current study is the first to investigate effects of footwear on running impact loading in preschool children.

Our data showed that it could be potentially risky to recommend running in true minimalist shoes in early running developmental stages, at least at the time of first experience with this type of shoe (acute response), especially on hard surfaces and without sufficient rest periods for bone adaptation, given that previous studies indicate possible overuse (overloading) of the calcaneal bone from locomotion in toddlers (Laliotis et al., 1993; Starshak et al., 1984; Zeininger et al., 2018). Yet, it is very difficult to draw any conclusions with recommendations for parents or teachers. However, the findings of this study could mark a starting line for further investigations. Future research should be addressed to investigate longitudinal changes of impact loading and morphological changes of the preschoolers and older children's bodies regarding exposure to the different footwear conditions.

Strengths and limitations of the study

VGRF could have a different course and shape in preschool children than in adults as seen in the previous studies that revealed a higher VGRF impact peak (1st peak) than the active peak (2nd peak) and could occur sooner than in adults (Fortney, 1983; Lieberman et al., 2010; Rice et al., 2016). Therefore, we felt that it was more appropriate to analyse the VGRF component as a continuous variable and combine it with discrete variables such as VILR. We suggest that this approach could allow a more comprehensive understanding of the footwear/footstrike effect on the impact loading phenomenon in young children.

In terms of limitations, even though all participants were considered as habitually shod, we did not collect information about their specific footwear type that they usually wore (e.g., experience with standard running shoes). Uniform footwear used during testing in this study could increase internal validity, but concurrently decrease external validity. In addition, all statistical tests used a conservative approach with alpha level correction for multiple testing even though some researchers suggested no corrections for biological data (Rothman, 2014). Furthermore, we analysed only an acute response to changing footwear conditions and we did not analyse age-related changes of impact loading or morphological changes. Finally, we used only a unilateral model for the biomechanical analysis because it would not be feasible to attach more markers on the bodies of such young children for all three different footwear conditions (due to the time-consuming procedure and concurrently attentional constraints of the children).

Conclusions

This study revealed that running in minimalist shoes did not imitate barefoot running and did not lower impact forces compared to standard running shoes in 3-year-old children. On the contrary, increased impact loading was observed in minimalist shoes compared to barefoot and even standard running shoes in early years of running skill development. However, no differences in footstrike patterns were found between footwear conditions in that age (mostly rearfoot-midfoot strike). Those who work with children should be careful and consider specificity of the minimalist shoes effect on loading when their children are exposed to run on hard surfaces.

Acknowledgments

We appreciate all the participants in this study. We would also like to thank Gillian Weir for her helpful advice in the statistical analysis (statistical parametric mapping analysis), Marketa Rygelova and Jiri Skypala with measurement coordination and Roman Farana for his advice in the writing of this manuscript.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

This article has been posted as a preprint in Research Square (DOI: 10.21203/rs.3.rs-1054040/v1). The data that support the findings of this study are available from the corresponding author [JP] upon reasonable request.

Additional information

Funding

This research was co-funded by the Czech Ministry of Education, Youth and Sport and European Union (European Regional Development Fund) by the grant Healthy Aging in Industrial Environment Program 4 HAIE [CZ.02.1.01/0.0/0.0/16_019/0000798].