Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2022 Sep 30.
Published in final edited form as: Eur J Sport Sci. 2018 Nov 3;19(6):774–783. doi: 10.1080/17461391.2018.1541256

Running speed-induced changes in foot contact pattern influence impact loading rate

Bastiaan Breine 1, Philippe Malcolm 2, Samuel Galle 1, Pieter Fiers 1, Edward C Frederick 3, Dirk De Clercq 1
PMCID: PMC9524621  NIHMSID: NIHMS1826025  PMID: 30394188

Abstract

Purpose.

We aimed to determine the effect of speed-induced changes in foot contact patterns on the vertical instantaneous loading rate (VILR). We hypothesized that transition runners, i.e. runners that shift towards a mid- (MF) or forefoot contact pattern (FF) when running speed increases, show smaller increases in VILR than non-transition runners. i.e. runners that remain with a rearfoot contact pattern (RF).

Methods.

Fifty-two male and female runners ran overground at 3.2, 4.1, 5.1 and 6.2 m·s−1. Ground reaction forces, lower limb sagittal plane knee and ankle kinematics and plantar pressures were recorded. Multi-level linear regression models were used to assess differences between transition and non-transition runners.

Results.

Non-transition runners experienced larger speed-induced increases in VILR (48.6 ± 2.6 BW·s−1 per m·s−1) than transition runners (−1.4 ± 7.6 BW·s−1 per m·s−1). Transition runners showed higher VILRs and a more flat foot touch down at the same pre-transition speed than non-transition runners.

Conclusions.

When running speed increases, some runners transition towards more anterior foot contact patterns. This reduces or even eliminates the speed-induced increase in VILR. This result is especially the case for those RF runners who already have relatively high VILRs and flat foot positioning at slower running speeds.

Keywords: Biomechanics, Injury & Prevention, Kinetics

INTRODUCTION

Running is a popular leisure time activity with well-established benefits for mental and physical health. However, there are high rates of running-related injuries ranging from 3 to 59 running related injuries per 1000 hours of running (van Gent et al., 2007). Despite the multitude of research, there is still no consensus on which biomechanical parameters are the most important risk factors. A high vertical instantaneous loading rate (VILR) of the ground reaction force (GRF) can be considered a risk factor for the development of tibial stress fractures (Hamill & Gruber, 2017; Milner, Ferber, Pollard, Hamill, & Davis, 2006; van der Worp, Vrielink, & Bredeweg, 2016). Stress fractures can account for 15–20% of all musculoskeletal related injuries, with the tibia being the most common site of injury (35–49%).(Barnes, Wheat, & Milner, 2008) Moreover, VILR has been shown a reliable parameter to assess impact intensity, especially when different foot strike patterns are studied (Ueda et al., 2016).

When running speed increases, VILR increases (Breine, Malcolm, Frederick, & De Clercq, 2014). In addition to running speed, foot contact patterns influence VILR (Boyer, Rooney, & Derrick, 2014; Breine et al., 2016; Knorz et al., 2017). Compared to mid- (MF) and forefoot contact patterns (FF), rearfoot contact patterns (RF) generally show higher VILRs (Boyer et al., 2014; Giandolini, Arnal, et al., 2013; D. S. Williams, McClay, & Manal, 2000). Foot contact patterns are classified by determining the strike index which expresses the location of the initial point of contact on the foot as a percentage of the total foot length (Cavanagh & Lafortune, 1980). In a previous study we refined this strike index method using high-frequency plantar pressure measurements (Breine et al., 2014). The plantar pressure measurements revealed an occurrence of atypical RF. These atypical RF are characterized by an initial fast anterior center of pressure movement along the lateral shoe margin, after which the center of pressure moves medially into the midfoot zone with an early first metatarsal contact (Figure 1). In typical RF, the initial center of pressure movement is generally slower, and moves towards the foot midline within the most proximal 1/3 of foot length. The atypical RF showed the highest VILR and the FF showed the lowest VILR (Breine et al., 2014). When running speed increases, some RF runners transition towards a more anterior foot contact pattern (MF or FF). However, about half of the runners remain with an RF in a wide range of tested velocities (Breine et al., 2014). We previously compared VILRs between foot contact patterns at different speeds from 3.2 to 6.2 m·s−1 using a between-subject approach. The question remains how within-subject speed-induced changes from an RF to an MF/FF affect VILR. When instructing habitual RF runners to run with an MF/FF at a chosen speed, some studies found a decrease in VILR (Cheung & Davis, 2011; Giandolini, Arnal, et al., 2013; Giandolini, Horvais, et al., 2013).

Figure 1:

Figure 1:

Open in a new tab

Representative center of pressure trajectories in different foot contact patterns when running at 3.2 m.s−1.

When running with an MF or FF, an initial ankle dorsiflexion strategy is used, i.e. after initial contact eccentric work of the plantar flexors attenuate part of the impact (Lieberman et al., 2010; D. S. B. Williams, Green, & Wurzinger, 2012). When RF runners make a speed-induced transition towards a more anterior foot contact pattern, they also transition towards a different kinematic impact-reducing strategy. It is expected that a more explicit transition from an RF to an FF, i.e. a larger change in foot-ankle geometry at initial contact towards a more plantar flexed foot and ankle will have a larger influence on VILR. However, about 50% of runners do not make such a transition (Breine et al., 2014). These runners maintain an initial RF and use an initial ankle plantar flexion strategy, i.e. after initial contact eccentric work of the dorsiflexor muscles and deformation of the heel fat pad and shoe attenuate part of the impact (De Clercq, Aerts, & Kunnen, 1994; Gerritsen, van den Bogert, & Nigg, 1995).

The reasons why some runners transition to a more anterior foot contact pattern when running faster and others do not are still unknown. If speed-induced transitions to an MF/FF result in smaller increases in VILR, avoiding high VILRs could be a trigger for such transitions.

The general aim of this study is to determine the interrelation between running speed, foot contact pattern and VILR. To describe this interrelation we have constructed three separate, but related hypotheses. A first hypothesis is that transition runners who transition towards an MF or FF show smaller increases in VILR with increasing running speed compared to non-transition runners who remain with an RF. A second hypothesis is that within the group of transition runners, there is a linear relationship between the change in VILR and the change in foot and ankle angle. That is, greater speed-induced changes in foot and ankle angle (towards MF/FF) relate to smaller increases in VILR. A third hypothesis is that transition runners already show a more flat foot touchdown and less dorsiflexed ankle contact angle, which is associated with higher VILRs, at the same pre-transition speed than non-transition RF runners.

As both spatiotemporal characteristics and foot strike can influence impact loading we will also assess if there is a difference in how transition and non-transition runners increase their speed through increasing step frequency. (Hobara et al., 2012; Yong, Silder, Montgomery, Fredericson, & Delp, 2018)

METHODS

Participants

Thirty-nine male and thirteen female runners were recruited for this study and gave their written informed consent. All participants were free from injury at least 3 months prior to the study and ran ≥15 km per week. For the male participants, mean ± SD age was 28.5 yrs. ± 8.2; body mass 72.1 kg ± 5.7; height 1.80 m ± 0.05, self-reported training speed 12.8 ± 1.2 km·h−1 and years of running was 8.6 ± 6.0. For the female participants, mean ± SD age was 27.6 yrs. ± 7.9; body mass 59.1 kg ± 4.8; height 1.67 m ± 0.05, self-reported training speed was 11.1 ± 1.1 km·h−1 and years of running was 9.3 ± 5.2. Leg length was measured as height of the greater trochanter in a neutral standing position. This study was approved by the Ghent University ethics committee.

Protocol and experimental setup

After a warm-up of ten minutes, participants performed running bouts over a 25 m runway at 4 speeds: 3.2, 4.1, 5.1 and 6.2 m∙s-1. These speeds were chosen to represent a range from training pace up to race pace of experienced endurance runners (https://medium.com/runkeeper-everyone-every-run/how-long-till-the-finish-line-494361cc901b). Three left foot measurements were selected per running speed. Running speed was practiced by following pacing lights attached to the side of the runway. After each trial, the speed across the measurement zone (at about 15m of the runway) was checked with 4 pairs of infrared timing gates (NIR-100-RX, EMX Industries Inc., Cleveland, OH, USA) placed 2 m apart alongside the runway. Only running trials within a ± 0.2 m∙s−1 margin of the selected speed were retained for analysis. All participants wore the same running shoes (Magne, ARHF041, Li Ning, Beijing, China) to counter a possible bias of running shoe type. These shoes were modified for optimizing plantar pressure measurements by substituting a flat outsole and filling the gap in the midfoot region of the midsole with a standard EVA foam. For more details on these adjustments we refer to a previous study (Breine et al., 2014). Photos from the original and adjusted shoe are presented in supplemental digital content (SDC 1: Photo: Original shoe; SDC 2: Adjusted shoe).

GRFs (1000 Hz, AMTI, Watertown, MA, USA) and plantar pressures (500 Hz, Footscan®, rs scan international nv, Olen, Belgium) were recorded by a 2-m force plate with a pressure plate on top. Retro-reflective markers, placed on anatomical landmarks of the thigh, shank and foot, were tracked with 14 infrared cameras (Qualisys AB, Gothenburg, Sweden) to calculate three-dimensional lower body kinematics (200 Hz). Holes in the shoes enabled placement of markers directly on the skin to track foot motion, rather than shoe motion. The three-segment (thigh, shank, foot) six degrees of freedom kinematic model (Visual 3D, C-motion, Germantown, MD, USA), marker set and segment coordinate system definitions are described in detail in supplemental digital content 3 (SDC 3: Kinematic model). Knee, ankle and foot segment angles were calculated using Cardan sequence, but only the first resolution (sagittal plane) was used. We know from an earlier study that the frontal plane ankle and foot inversion angles at initial contact and frontal plane ankle and foot range of motions did not relate to VILR (Breine et al., 2016). Joint and segment angles were normalized to the standing position, which was measured during a static trial before the running trials. The neutral stance joint and segment angles can be consulted in supplemental digital content 4 (SDC 4: Table: Joint and foot segment angles in neutral stance). The selected kinematic variables are sagittal plane foot and ankle angle at initial contact. Initial contact was defined as the first kinematic frame (sampled at 200 Hz) when GRF (sampled at 1000 Hz) rose above a 20N threshold.

Frequency analysis of the GRF data showed resonance frequencies higher than 80 Hz. Residual analysis of GRF signals and a qualitative assessment of the over/under filtering effect of different cut-off frequencies, ranging from 50 to 100 Hz, were performed to determine the optimal cut-off frequency. GRF and kinematic marker coordinate data were filtered with a Butterworth 2nd order low pass filter with a cut-off frequency of 80 and 20 Hz respectively. Knee and ankle eccentric work was calculated using a Newton-Euler inverse dynamics approach. VILR was calculated as the maximal value of the first derivative of the vertical GRF and normalized to bodyweight (BW).

Each foot contact was categorized as an FF, MF, typical RF or atypical RF based on a combination of strike index, time of first metatarsal contact and a qualitative assessment of the center of pressure trajectory as described in earlier research (Breine et al., 2014). This qualitative assessment of the center of pressure trajectory allows to discern between typical and atypical RF which is impossible with only strike index (Figure 1).

Statistics

All selected variables of the three recorded trials per running speed were averaged per participant. If all trials could not be assigned to the same foot contact pattern, the average value for analysis was calculated with the trials of the most frequent foot contact pattern. Intraclass correlation coefficients (two-way mixed model, average measures) across trials for strike index and VILR were above 0.8, indicating low variability across trials. For each increasing step in running speeds the changes in the selected variables (VILR, foot angle at initial contact, ankle angle, ankle eccentric work and knee eccentric work) were calculated and normalized to a change per m·s−1 because the selected speed steps were not equal. Subjects were categorized as transition (typical RF-MF, atypical RF-MF, RF-FF, atypical RF-FF, MF-FF) or non-transition runners per increasing step in running speed based on their change to another foot contact pattern or not. Two grouping variables were created. Grouping A compared (A)typical RF who transitioned to an MF or FF with non-transition runners who maintained a (A)typical RF. Grouping B compared MF who transitioned to an FF with non-transition runners who maintained an MF. Table I shows the number of transition and non-transition runners for the different steps in running speed.

Table I:

Number of subjects who perform each type of foot strike pattern transition per step of increase in running speed. In total 52 runners were tested. At 3.2 m·s−1 31 subjects had a typical rearfoot, 11 an atypical rearfoot, 10 a midfoot and no forefoot contact patterns.

3.2 to 4.1 m·s−1 4.1 to 5.1 m·s−1 5.1 to 6.2 m·s−1
remains typical rearfoot 28 23 17
typical rear- to midfoot 1 1 1
typical rear- to forefoot 0 2 1
typical rear- to atypical rearfoot 2 2 4
remains atypical rearfoot 10 7 9
atypical rear- to midfoot 1 4 1
atypical rear- to forefoot 0 1 0
remains midfoot 9 10 7
mid- to forefoot 1 0 8
mid- to atypical rearfoot 0 1 0
remains forefoot 0 1 4
Open in a new tab

Multi-level linear regression models (two levels: participant and speed step) were constructed in MLwiN 2.26 (University of Bristol, Bristol, UK) to determine the effect of transition group (transition vs. non-transition) on the selected variables. These models were separately constructed for each grouping variable. For all transitions to a more anterior strike pattern, Pearson correlation coefficients (r) between change in foot and ankle angle and change in VILR were calculated in SPSS 22 (SPPS Inc., Chicago, IL, USA). The significance level was set at p<0.05 for all analyses.

Although not a primary goal of this study, we constructed a linear regression model between foot angle at initial contact and VILR for each running speed. Similar analyses were done for the relationship between ankle angle at initial contact and VILR. These regression models are used in the discussion to interpret and explain the effect of a foot contact pattern transition on VILR. These analyses were done separately for two subgroups: the subgroup of MF and FF runners and the subgroup of atypical and typical RF runners. We chose these two subgroups as we know that an RF and an MF/FF use opposing ankle strategies to reduce impact, which would obstruct linear modelling (Breine et al., 2016). Additionally, the changes in ankle and knee eccentric work as parameters for the loading of muscles crossing the ankle and knee joint with increasing running speed are presented.

RESULTS

At the first speed step from 3.2 to 4.1 m·s1 about 90% of runners remained with the same foot contact pattern, while about 10% transitioned to a more anterior foot contact patterns. At the second speed step from 4.1 to 5.1 m·s1 this was about 76% non-transition and 24% transition runners. At the last speed step from 5.1 to 6.2 m·s1 this was about 63% non-transition and 37% transition runners. (Table I)

Effect of speed-induced foot contact pattern transitions on VILR

For each grouping of transition vs. non-transition runners, we found that runners who transition towards a more anterior foot strike pattern experience smaller increases in VILR (−1.4 ± 7.6 BW·s−1 per m·s−1) than non-transition runners (48.6 ± 2.6 BW·s−1 per m·s−1) (p<0.001) (Table II).

Table II:

Mean and standard error values, predicted by the regression models, for grouping A and grouping B of transition and non-transition runners. Grouping A: The transition runners shift from a typical or atypical rearfoot contact pattern (RF) towards a mid- or forefoot contact pattern (MF/FF). The non-transition runners remain with a typical of atypical rearfoot contact pattern. Grouping B: The transition runners shift from a midfoot contact pattern (MF) towards a forefoot contact pattern (FF). The non-transition runners remain with a midfoot contact pattern. χ2 and p-values are given for the main effect of transition group on the selected variables. The mean values pre-transition for the non-transition runners are based on the values at the same speed as the speed pre-transition for the transition runners. A positive foot and ankle angle represent a dorsiflexed position.

Grouping A non-transition RF runners mean ± SE transition runners RF to MF/FF mean ± SE χ2 p
ΔVILR per m·s−1 (BW·s−1 per m·s−1) 48.6 ± 2.6 * −1.4 ± 7.6 37.518 <0.001
VILR at speed pre transition (BW·s−1) 141.1 ± 4.7 * 186.5 ± 12.6 11.485 <0.001
Foot angle at speed pre transition (°) 21.6 ± 0.8 * 14.6 ± 2.0 10.549 0.001
Foot angle post transition (°) n.a. 2.0 ± 2.1 n.a. n.a.
Ankle angle at speed pre transition (°) 6.2 ± 0.6 * 0.5 ± 1.6 11.181 <0.001
Ankle angle post transition (°) n.a. −11.1 ± 1.8 n.a. n.a.
Grouping B non-transition MF runners mean ± SE transition runners MF to FF mean ± SE χ2 p
ΔVILR per m·s−1 (BW·s−1 per m·s−1) * 32.3 ± 6.2 * 0.9 ± 11.0 6.233 0.013
VILR at speed pre transition (BW·s−1) 127.9 ± 10.6 ** 170.8 ± 19.2 3.812 0.051
Foot angle at speed pre transition (°) 2.8 ± 0.6 3.6 ± 1.1 0.456 0.500
Foot angle post transition (°) n.a. 1.8 ± 1.1 n.a. n.a.
Ankle angle at speed pre transition (°) −11.9 ± 1.1 −10.7 ± 2.0 0.254 0.614
Ankle angle post transition (°) n.a. −12.9 ± 1.9 n.a. n.a.
Open in a new tab
*

indicates significant difference between groups; p<0.05.

**

indicates significant difference between groups; p<0.1.

Relationship between change in foot and ankle contact angle and VILR in the transition runners

Within the entire group of transition runners with a transition towards a more anterior foot contact pattern (Typical or Atypical RF to MF or FF and MF to FF), we found a significant correlation between the change in foot angle at initial contact and the change in VILR per increase in speed (r=0.445, p=0.043). We also found a significant correlation between the change in ankle angle at initial contact and the change in VILR per increase in speed (r=0.457, p=0.037). These correlations indicate that for a more pronounced transition, i.e., more anterior inclined foot position and more plantar flexed ankle at initial contact, the increase in VILR is smaller.

Pre-transition differences in foot and ankle contact angle and VILR between transition and non-transition runners

We found that runners who transition from an RF (typical or atypical) to a more anterior foot contact pattern (MF or FF) started off from a more flat foot touchdown (14.6 ± 2.0° vs. 21.6 ± 0.8°) and less dorsiflexed ankle angles (0.5 ± 1.6° vs. 6.2 ± 0.6°) at initial contact than the runners who maintained an RF. (Table II)

We found that runners who transition from an RF (typical or atypical) to a more anterior foot contact pattern (MF or FF) already had higher VILRs at the speed before transition (186.5 ± 12.6 BW·s−1) than runners who maintained an RF (141.1 ± 4.7 BW·s−1). Additionally, a difference in VILR was found at the speed before transition between runners who maintained an MF and runners who made the transition from an MF to an FF contact pattern, although this result was not significant at the p<0.05 level (p=0.051). (Table II)

Changes in spatiotemporal characteristics

We found no difference in how step frequency, expressed as steps per minute (SPM), increased with increasing speed between transition runners (from 3.2 to 4.1 m/s: + 8 SPM; from 4.1 to 6.2 m/s: + 16 SPM; from 5.1 to 6.2 m/s: +20 SPM) and runners that remained with an RF (from 3.2 to 4.1 m/s: + 7 SPM; from 4.1 to 6.2 m/s: + 15 SPM; from 5.1 to 6.2 m/s: +22 SPM).

Changes in ankle and knee eccentric work with increasing running speed

A runner’s foot contact pattern also relates to the loading of the muscles crossing the ankle and knee joint, as approximated by the eccentric work performed at these joints ( Williams et al., 2012; Williams et al., 2000). At 3.2 m·s−1, RF showed smaller eccentric work at the ankle joint (24.6 ± 7.0 vs. 39.1 ± 6.7 J, p<0.001) and higher eccentric work at the knee joint (44.8 ± 9.8 vs. 30.8 ± 7.7 J, p<0.001) compared to the MF/FF. Transition runners showed greater speed-induced increases in ankle eccentric work than non-transition runners (14.1 ± 1.7 vs. 4.2 ± 0.5 J per m·s−1, p<0.001, χ2=29.693). However, transition runners showed a decrease in knee eccentric work as speed increases (−7.9 ± 2.2 J per m·s−1), and non-transition runners showed almost no change in knee eccentric work with an increase in speed (2.8 ± 0.8 J per m-·s−1, p<0.001, χ2=20.782).

DISCUSSION

A main novelty of this study is that we focused on spontaneous foot contact pattern transitions in runners, when increasing their running speed, in relation to impact intensity. Previous research has mainly investigated the differences in impact intensity between foot contact patterns at a certain running speed (Boyer et al., 2014; Kulmala, Avela, Pasanen, & Parkkari, 2013; Rice, Jamison, & Davis, 2016) or on the effect of running speed on impact intensity (Hobara et al., 2012; Nigg, Bahlsen, Luethi, & Stokes, 1987), but never on the interaction between both. We found that runners who transition towards a more anterior foot contact pattern experience smaller increases in VILR than non-transition runners. This can be explained by a change in initial ankle and foot kinematics.

The relationships depicted in Figure 2 show that the foot/ankle contact angles, which are inherently connected to the subsequent initial foot unroll (i.e. the unrolling of the foot from initial contact to a foot flat position) are strongly related to the VILR at all selected running speeds. In RF runners, a greater foot contact angle (or more dorsiflexed ankle), which relates to the amount of initial ankle plantar flexion and foot unroll, relates to lower VILR values (adj. R2 30–51%). In MF and FF, a more plantar flexed ankle (or smaller foot contact angle), which relates to the amount of subsequent initial ankle dorsiflexion, relates to lower VILR values (adj. R2 41–52%). In terms of distal kinematics these strategies are opposed to each other (respectively a positive and negative slope of the linear fit with VILR), leaving the flat foot at ground contacts with the highest VILR. A runner could limit the increase in VILR with speed by avoiding a flat foot touchdown by either transitioning to a more anterior foot contact pattern (~transition runner) or by more explicitly using the current RF distal kinematic strategy (~RF non-transition runner). Moreover, in a gait retraining study aimed at reducing tibial shock (as an impact intensity measure) without changing running speed, Bowser et al. also found two effective strategies (Bowser, Fellin, & Davis, 2011): either to decrease dorsiflexion at initial contact towards an MF or FF, or to increase dorsiflexion at initial contact towards a more pronounced RF. These two opposite strategies confirm the opposing linear relationship between ankle/foot contact angle and VILR (Figure 2) (Garofolini, Taylor, Mclaughlin, Vaughan, & Wittich, 2017).

Figure 2:

Figure 2:

Open in a new tab

Relationship between foot and ankle angle at initial contact and VILR at four running speeds (3.2, 4.1, 5.1 and 6.2 m·s−1). On the left side: mid- and forefoot contact patterns and the relationship between ankle angle at initial contact and VILR. On the right side: rearfoot contact patterns and the relationship between foot angle at initial contact and VILR. Note that this is a separate x-axis for both foot contact pattern groups as for both foot contact pattern groups a different kinematic parameter shows the best correlation with VILR. For each running speed, and per subgroup (left: mid- and forefoot contact patterns, right: rearfoot contact patterns) the linear regression and associated adj. R2 between ankle (mid-forefoot) or foot angle (rearfoot) at initial contact and VILR is shown. The black circles indicate the data points associated with a typical non-transition runner who maintains a typical rearfoot contact pattern at all running speeds. The black squares indicate the data points of a transition runner who shifts towards a midfoot contact pattern at 5.1 m·s−1 and towards a forefoot contact pattern at 6.2 m·s-1.

The relationship between initial foot-ankle configuration and VILR shows that the most effective option to limit a speed induced increase of VILR is to transition to a more anterior foot contact pattern. Approximately 33% of the runners performed such a transition at one of the selected speed intervals. Runners who transitioned towards a more anterior foot contact pattern showed smaller (or even no) speed-induced increases in VILR compared to non-transition runners. Some runners who transitioned towards a more anterior foot contact pattern even managed to achieve a decrease in VILR with an increase in running speed. Whereas all non-transition runners showed an increase in VILR with every increased speed interval. If a transition was more pronounced (i.e., a greater change in ankle and foot angle at initial contact), the increase in VILR was smaller. Essentially, transitioning towards a more anterior foot contact pattern is effective in limiting the speed induced increase in VILR when a sufficient plantar flexed ankle is adopted at initial contact (~moving sufficiently to the left side in Figure 2). Such a transition is visualized in Figure 2 by a shift from the right side of the figure to the left side (black squares). The transition runner who shifts towards an MF, and eventually an FF contact pattern, only showed an increase in VILR of 54 BW·s−1 from 3.2 to 6.2 m·s-1.

Another option to limit a speed induced increase of VILR is to make more explicit use of the RF distal kinematic strategy when speed increases. This non-transition RF strategy was rarely used by any of the runners in this study. In Figure 2, the non-transition runner maintains an RF at all running speeds and shows a total increase in VILR of 150 BW·s-1. This runner increases his foot angle in the first speed step (note that this is also the step with the smallest VILR increase) but does not further increase it in the following speed intervals. The average foot angle at initial contact for the RF runners was 22° at both 3.2 and 6.2 m·s-1. In other words, the non-transition runners did not tend to more explicitly use the distal kinematic strategy of initial ankle plantar flexion and foot unroll with deformation of the heel fat pad. It could be that a change towards a more pronounced RF might hamper the required spatiotemporal changes for faster running. Step frequency and/or length must increase to run faster. As a more pronounced RF is often associated with longer contact times, a change towards a more pronounced RF might prevent the required increase in stride frequency as running speed increases (Breine et al., 2014; Kulmala et al., 2013). However, we found no difference in how transition and non-transition runners increased their running speed through an increase in stride frequency (and as such stride length). At the fastest running speed of 6.2 m·s−1 we found no difference in step frequency between transition and non-transition runners. At 6.2 m·s−1 runners that remained with an RF showed an average step frequency of 206 steps per minute and the transition runners showed an average step frequency of 204 steps per minute. Furthermore, this comparison was not confounded by leg length effects; an independent samples T-test showed that there was no significant difference (p=0.158; t=−1.439) in leg length between transition runners (89.3 ± 5.5 cm) and runners that remained with an RF (91.6 ± 4.5 cm). It is also possible that a more pronounced RF might put too much strain on the m. tibialis anterior, which needs to work eccentrically during the initial foot contact phase.

The question remains why some runners perform a transition while others do not. Runners with a high VILR might be more likely to perform a transition as a way to limit a further speed-induced increase in VILR. We indeed found that transition runners showed greater VILRs and associated smaller foot contact angles pre-transition compared to non-transition runners at the same speed. Hence, impact intensity (e.g. VILR) which is partially regulated through foot contact angle might be a trigger to transition towards a more anterior foot strike pattern. The mean VILR for the non-transition RF runners at 3.2 m·s−1 was 94.4 BW·s−1, whereas the mean VILR for the transition runners was 136.0 BW·s-1. If these transition runners had not transitioned towards a more anterior foot strike pattern and showed the same speed-induced increase in VILR as the non-transition runners (that is, 48.6 BW·m−1 per m·s−1), they would end up with a high VILR of 281.8 BW·s−1 at 6.2 m·s-1. If we calculate the Z-score of the VILR of the transition runners at 3.2 m·s−1 compared to the entire group, they have a Z-score of 0.97. In previous studies that compared the VILR of injured with uninjured runners the VILR of the injured runners showed a Z-score of 0.68 to 0.91 (Milner et al., 2006). If transition runners did not perform a transition, they would remain with VILRs associated with an increased risk of injury, especially tibial stress fractures.

Based on the finding that runners with high VILRs and small foot angles at initial contact are more likely to perform a transition, one would expect that more atypical RF runners would perform a transition towards a more anterior foot contact pattern. Indeed, we found that over all speed intervals, a greater percentage of atypical RF (21%) made a transition towards an MF/FF than typical RF (8%). Despite the high VILRs in most of the runners with an atypical RF, the majority of them (about 80%) maintain this atypical RF as running speed increases. Other factors than impact intensity also determine why a runner chooses a certain foot contact pattern. Such factors might be load-bearing capacities, motor control skills, ankle plantar flexor muscle strength or running economy. As shifting towards an MF or FF changes the distribution of joint loads more towards the ankle and away from the knee joint this could influence the required muscle force and as such running economy. Some studies indicated that an MF is more economical (Santos-Concejero et al., 2014), while others found that an RF is more economical (Ogueta-Alday, Rodríguez-Marroyo, & García-López, 2014), But most studies, including recent reviews, state that there is no evidence to suggest that running with an MF or FF improves running economy compared to an RF (Anderson, Barton, & Bonanno, 2017; Hamill & Gruber, 2017; Roper, Doerfler, Kravitz, Dufek, & Mermier, 2017). As no previous studies have, future research should focus on the interaction effect of running speed, foot strike pattern and running economy also at faster speeds.

The observed changes in ankle and knee eccentric work with increasing running speed indicate that, performing a speed-induced transition towards a more anterior foot contact pattern might be a way to minimize the increase in eccentric load at the knee joint, but it requires stronger ankle plantar flexion muscles and tendons. As not all runners perform a transition, we hypothesize that a certain amount of plantar flexor muscle strength is needed to be able to perform such a transition and runners who perform a transition would have greater ankle plantar strength than those who do not (Liebl, Willwacher, Hamill, & Brüggemann, 2014). A shift towards an MF or FF might reduce the risk of impact related injuries, but as it increases the stress upon the Achilles tendon it might induce other types of injuries (Hamill & Gruber, 2017).

An explicit consequence of the design of this study is that we did not instruct runners to change their running style. This means that we were only able to compare between-subject differences in speed-induced foot contact pattern transition behavior. Nevertheless, we studied spontaneously occurring transitions and thus ecologically valid adaptations in running style. Future research could investigate whether intra-subject differences in speed-induced foot contact pattern transitions concur with the current findings on increase in VILR. However, such an approach would require instructing runners to perform certain speed-induced foot contact pattern transitions or not, which seems less ecological. In the last ten years, many studies have focused on the effect of shoe cushioning, heel offset, sole thickness etc. on running biomechanics, but this study and its results only apply for trained habitual shod distance runners.

Another important direction for future research is to further investigate the role of impact intensity and foot contact patterns as a risk factor for injuries, which is still debated, through prospective longitudinal intervention studies (Davis, Bowser, & Mullineaux, 2016; Hamill & Gruber, 2017). There is also a need for more musculoskeletal modelling studies that link external loading to the internal loading of the bones, tendons and muscles (Chen et al., 2016). These studies would allow us to better understand to what extent impact intensity relates to the development of impact-related running injuries.

Conclusions

When running speed increases, some runners transition towards more anterior foot contact patterns. This reduces or even eliminates the speed-induced increase in VILR with increasing running speed. This result is especially the case for those RF runners who already have relatively high VILRs and flat foot positioning at slower running speeds. Therefore elevated VILR might be a trigger for performing a transition.

Supplementary Material

sup fig 2

Supplemental digital content 1 (SDC 2): Photo : Adjusted shoe

sup mat 3

Supplemental digital content 1 (SDC 3): Description of the kinematic model.

sup mat 4

Supplemental digital content 2 (SDC 4): Table: Joint and foot segment angles in neutral stance.

sup fig 1

Supplemental digital content 1 (SDC 1): Photo : Original shoe

Funding

Financial and product support was provided from the Li Ning Company, Ltd. P.M. received partial support from the Center of Research in Human Movement Variability of the University of Nebraska Omaha and the NIH (P20GM109090).

Footnotes

Disclosure of statement

The authors have no conflicts of interest.

REFERENCES

  1. Anderson L, Barton C, & Bonanno D. (2017). The effect of foot strike pattern during running on biomechanics, injury and performance: A systematic review and meta-analysis. Journal of Science and Medicine in Sport, 20(2017), e54. 10.1016/j.jsams.2017.01.145 [DOI] [Google Scholar]
  2. Barnes A, Wheat J, & Milner C. (2008). Association between foot type and tibial stress injuries: A systematic review. British Journal of Sports Medicine, 42(2), 93–98. 10.1136/bjsm.2007.036533 [DOI] [PubMed] [Google Scholar]
  3. Bowser BJ, Fellin R, & Davis IS (2011). Kinematic strategies used by runners to reduce tibial shock following gait retraining. Medicine & Science in Sports & Exercise, 43(5), 581. 10.1249/01.MSS.0000402713.82719.8c [DOI] [Google Scholar]
  4. Boyer ER, Rooney BD, & Derrick TR (2014). Rearfoot and midfoot or forefoot impacts in habitually shod runners. Medicine and Science in Sports and Exercise, 46(7), 1384–91. 10.1249/MSS.0000000000000234 [DOI] [PubMed] [Google Scholar]
  5. Breine B, Malcolm P, Frederick EC, & De Clercq D. (2014). Relationship between running speed and initial foot contact patterns. Medicine and Science in Sports and Exercise, 46(8), 1595–603. 10.1249/MSS.0000000000000267 [DOI] [PubMed] [Google Scholar]
  6. Breine B, Malcolm P, Van Caekenberghe I, Fiers P, Frederick EC, & De Clercq D. (2016). Initial foot contact and related kinematics affect impact loading rate in running. Journal of Sports Sciences, 35(15), 1–9. 10.1080/02640414.2016.1225970 [DOI] [PubMed] [Google Scholar]
  7. Cavanagh P, & Lafortune M. (1980). Ground reaction forces in distance running. Journal of Biomechanics, 13, 397–406. [DOI] [PubMed] [Google Scholar]
  8. Chen TL, An WW, Chan ZYS, Au IPH, Zhang ZH, & Cheung RTH (2016). Immediate effects of modified landing pattern on a probabilistic tibial stress fracture model in runners. Clinical Biomechanics, 33, 49–54. 10.1016/j.clinbiomech.2016.02.013 [DOI] [PubMed] [Google Scholar]
  9. Cheung RTH, & Davis IS (2011). Landing pattern modification to improve patellofemoral pain in runners: a case series. The Journal of Orthopaedic and Sports Physical Therapy, 41(12), 914–9. 10.2519/jospt.2011.3771 [DOI] [PubMed] [Google Scholar]
  10. Davis IS, Bowser BJ, & Mullineaux DR (2016). Greater vertical impact loading in female runners with medically diagnosed injuries: a prospective investigation. British Journal of Sports Medicine, 50(14), 887–892. 10.1136/bjsports-2015-094579 [DOI] [PubMed] [Google Scholar]
  11. De Clercq D, Aerts P, & Kunnen M. (1994). The mechanical characteristics of the human heel pad during foot strike in running: An in vivo cineradiographic study. Journal of Biomechanics, 27(10), 1213–1222. 10.1016/0021-9290(94)90275-5 [DOI] [PubMed] [Google Scholar]
  12. Garofolini A, Taylor S, Mclaughlin P, Vaughan B, & Wittich E. (2017). Foot strike classification: a comparison of methodologies. Footwear Science, 9, S129–S130. 10.1080/19424280.2017.1314377 [DOI] [Google Scholar]
  13. Gerritsen KG, van den Bogert AJ, & Nigg BM (1995). Direct dynamics simulation of the impact phase in heel-toe running. Journal of Biomechanics, 28(6), 661–668. 10.1016/0021-9290(94)00127-P [DOI] [PubMed] [Google Scholar]
  14. Giandolini M, Arnal PJ, Millet GY, Peyrot N, Samozino P, Dubois B, & Morin J-B (2013). Impact reduction during running: efficiency of simple acute interventions in recreational runners. European Journal of Applied Physiology, 113, 599–609. 10.1007/s00421-012-2465-y [DOI] [PubMed] [Google Scholar]
  15. Giandolini M, Horvais N, Farges Y, Samozino P, Morin J-B, Nicolas G, … Morin PSJ (2013). Impact reduction through long-term intervention in recreational runners : midfoot strike pattern versus low-drop / low-heel height footwear. European Journal of Applied Physiology, 2077–2090. 10.1007/s00421-013-2634-7 [DOI] [PubMed] [Google Scholar]
  16. Hamill J, & Gruber AH (2017). Is changing footstrike pattern beneficial to runners? Journal of Sport and Health Science, 6(2), 146–153. 10.1016/j.jshs.2017.02.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Hobara H, Sato T, Sakaguchi M, Sato T, Nakazawa K, & Functions M. (2012). Step frequency and lower extremity loading during running. International Journal of Sports Medicine, 33, 310–313. 10.1055/s-0031-1291232 [DOI] [PubMed] [Google Scholar]
  18. Knorz S, Kluge F, Gelse K, Schulz-Drost S, Hotfiel T, Lochmann M, … Krinner S. (2017). Three-dimensional biomechanical analysis of rearfoot and forefoot running. Orthopaedic Journal of Sports Medicine, 5(7), 1–10. 10.1177/2325967117719065 [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Kulmala J-P, Avela J, Pasanen K, & Parkkari J. (2013). Forefoot strikers exhibit lower running-induced knee loading than rearfoot strikers. Medicine and Science in Sports and Exercise, 45(12), 2306–13. 10.1249/MSS.0b013e31829efcf7 [DOI] [PubMed] [Google Scholar]
  20. Lieberman DE, Venkadesan M, Werbel WA, Daoud AI, D’Andrea S, Davis IS, … Pitsiladi. (2010). Foot strike patterns and collision forces in habitually barefoot versus shod runners. Nature, 463, 531–536. [DOI] [PubMed] [Google Scholar]
  21. Liebl D, Willwacher S, Hamill J, & Brüggemann G-P (2014). Ankle plantarflexion strength in rearfoot and forefoot runners: a novel clusteranalytic approach. Human Movement Science, 35, 104–20. 10.1016/j.humov.2014.03.008 [DOI] [PubMed] [Google Scholar]
  22. Milner CE, Ferber R, Pollard CD, Hamill J, & Davis IS (2006). Biomechanical factors associated with tibial stress fracture in female runners. Medicine and Science in Sports and Exercise, 38(2), 323–328. 10.1249/01.mss.0000183477.75808.92 [DOI] [PubMed] [Google Scholar]
  23. Nigg BM, Bahlsen H.a, Luethi SM, & Stokes S. (1987). The influence of running velocity and midsole hardness on external impact forces in heel-toe running. Journal of Biomechanics, 20(10), 951–959. 10.1016/0021-9290(89)90081-X [DOI] [PubMed] [Google Scholar]
  24. Ogueta-Alday A, Rodríguez-Marroyo JA, & García-López J. (2014). Rearfoot striking runners are more economical than midfoot strikers. Medicine and Science in Sports and Exercise, 46(3), 580–5. 10.1249/MSS.0000000000000139 [DOI] [PubMed] [Google Scholar]
  25. Rice HM, Jamison ST, & Davis IS (2016). Footwear Matters: Influence of Footwear and Foot Strike on Load Rates during Running, (2), 2462–2468. 10.1249/MSS.0000000000001030 [DOI] [PubMed] [Google Scholar]
  26. Roper JL, Doerfler D, Kravitz L, Dufek JS, & Mermier C. (2017). Gait Retraining from Rearfoot Strike to Forefoot Strike does not change Running Economy. International Journal of Sports Medicine, 38(14), 1076–1082. 10.1055/s-0043-110225 [DOI] [PubMed] [Google Scholar]
  27. Santos-Concejero J, Tam N, Granados C, Irazusta J, Bidaurrazaga-Letona I, Zabala-Lili J, & Gil SM (2014). Interaction Effects of Stride Angle and Strike Pattern on Running Economy. International Journal of Sports Medicine, 1118–1123. 10.1055/s-0034-1372640 [DOI] [PubMed] [Google Scholar]
  28. Ueda T, Hobara H, Kobayashi Y, Heldoorn T, Mochimaru M, & Mizoguchi H. (2016). Comparison of 3 Methods for Computing Loading Rate during Running. International Journal of Sports Medicine, 37(13), 1087–1090. 10.1055/s-0042-107248 [DOI] [PubMed] [Google Scholar]
  29. van der Worp H, Vrielink JW, & Bredeweg SW (2016). Do runners who suffer injuries have higher vertical ground reaction forces than those who remain injury-free? A systematic review and meta-analysis. British Journal of Sports Medicine, 50(8), 450–7. 10.1136/bjsports-2015-094924 [DOI] [PubMed] [Google Scholar]
  30. van Gent RN, Siem D, van Middelkoop M, van Os a G., Bierma-Zeinstra S. M. a, & Koes BW (2007). Incidence and determinants of lower extremity running injuries in long distance runners: a systematic review. British Journal of Sports Medicine, 41(8), 469–80. 10.1136/bjsm.2006.033548 [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Williams DSB, Green DH, & Wurzinger B. (2012). Changes in lower extremity movement and power absorption during forefoot striking and barefoot running. International Journal of Sports Physical Therapy, 7(5), 525–32. [PMC free article] [PubMed] [Google Scholar]
  32. Williams DS, McClay IS, & Manal KT (2000). Lower extremity mechanics in runners with a converted forefoot strike pattern. Journal of Applied Biomechanics, 16, 210–218. [Google Scholar]
  33. Yong JR, Silder A, Montgomery KL, Fredericson M, & Delp SL (2018). Acute changes in foot strike pattern and cadence affect running parameters associated with tibial stress fractures. Journal of Biomechanics, 76, 1–7. 10.1016/j.jbiomech.2018.05.017 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

sup fig 2

Supplemental digital content 1 (SDC 2): Photo : Adjusted shoe

NIHMS1826025-supplement-sup_fig_2.jpg (5.8MB, jpg)
sup mat 3

Supplemental digital content 1 (SDC 3): Description of the kinematic model.

NIHMS1826025-supplement-sup_mat_3.docx (13.1KB, docx)
sup mat 4

Supplemental digital content 2 (SDC 4): Table: Joint and foot segment angles in neutral stance.

NIHMS1826025-supplement-sup_mat_4.docx (12.8KB, docx)
sup fig 1

Supplemental digital content 1 (SDC 1): Photo : Original shoe

NIHMS1826025-supplement-sup_fig_1.jpg (995.7KB, jpg)

RESOURCES