BIOMECHANICS OF SHOD AND BAREFOOT RUNNING: A LITERATURE REVIEW

This study aims to analyze and summarize the biomechanical (kinematics, kinetics and neuromuscular) differences between shod and barefoot running, through a literature review. Searches were conducted for complete articles published between 2013 and November 2018 in the Web of Science, PubMed, Scopus and SPORTdiscus databases. The search terms used were Biomechanics, Kinetics, Kinematics, Electromyography, “Surface Electromyography”; and Unshod, Barefoot, Barefeet and Running. The search resulted in 687 articles; after excluding duplicates and selecting by title, abstract and full text, 40 articles were included in the review. The results show that there are important differences in the biomechanics of running when shod or barefoot. In general, studies indicate that in barefoot running: a) individuals present forefoot or midfoot foot strike patterns, while in shod running the typical pattern is the rearfoot strike; (b) greater cadence and shorter stride length are observed; and (c) there is greater knee flexion, lower peak vertical ground reaction force and greater activation of the medial gastrocnemius. In addition, barefoot runners contact the ground with greater plantar flexion, possibly as a strategy to reduce impact when stepping without footwear. These differences, as well as runners’ individual characteristics, should be considered in the prescription of the barefoot running, in order to minimize injuries resulting from the practice . Level of Evidence II; Review. estudo Kinematics, Unshod,


INTRODUCTION
Health benefits, low cost and improving coaching professionalization are some reasons to support road running popularity.In 2017, there were almost 20 million participants enrolled in road races in Brazil and in the United States of America. 1,2Along with such popularity, a high injury incidence ranging from 19.4% to 79.3% was observed. 3Injury-related factors are improper running shoes, improper training and biomechanical factors. 4Therefore, to deal with running injuries, comfort and performance running shoes were developed, 5 while barefoot running became popular. 6,7hy did these two actions become popular in outdoor running?
Historically, neolithic man already wore rudimentary footwear and the athlete Abebe Bikila, running barefoot, won the marathon of the 1960 Olympic Games in Rome.Although wearing shoes is an established human habit, people still have nice performance running barefoot.Robbins and Gouw 8 proposed the hypothesis modern sport shoes would be unable to protect individuals from running injuries because they attenuate plantar sensory information, which is necessary to change how they run.Two decades later, Liebermann et al. 9 showed the transient ground reaction force (GRF) 10 was absent in runners used to barefoot running.In this context, the discussion about running with shoes or barefoot is controversial.The sport's shoes technological development attracts consumers and, conversely, studies suggesting the barefoot running can eliminate the transient GRF pattern, leading to lower heel impact during running. 9For different reasons, also guided by the need to reduce the risk of injury, people run barefoot or with some type of footwear.
Hall et al. 7 reviewed the biomechanical differences of running with and without shoes.How does footwear affect joint kinematics in running?Moderate evidence indicates running barefoot is associated with higher cadence and lower dorsiflexion at first foot contact with the ground.Limited evidence suggests the forefoot pattern is more common in barefoot running.Does GRF decrease in barefoot running?Hall et al. 7 indicated the decreasing in peak GRF during barefoot running has moderate evidence.Different foot strike patterns modify how the foot first touches the ground and receives the impact and changes how foot muscles will attenuate such mechanical impact. 11oes running barefoot alter the muscle activation patterns?Only two studies, with limited evidence, describe how muscles act in running with shoes and barefoot.
All these issues continue to diverge opinions.Are new studies pointing out to the same direction as Hall et al. 7 findings, concerning biomechanical comparisons between shod and barefoot running?Recent investigations indicate that are differences between these conditions in kinematics (spatiotemporal parameters or joint kinematics) [12][13][14] , kinetics (peak GRF and moments) 15,16 and neuromuscular aspects 11,17 .Identifying these differences is important to minimize the injury risk, 6 and they should be summarized to guide running practice.In this sense, the aim of the study was to analyze and summarize the biomechanical (kinematics, kinetics and neuromuscular) differences between barefoot and shod running, through a literature review.

METHOD
The literature search was performed on the Web of Science, PubMed, Scopus, and SportDiscus databases.The descriptors, Boolean operators and the search strategy are presented in table 1, clustered into biomechanical analysis, barefoot condition and running.Inclusion criteria were: complete research articles with biomechanical variables (kinematics, kinetics and/or neuromuscular aspects) comparing running with and without shoes, with persons older than 18 years old, with at least one group of healthy individuals.Only articles published in English between 2013 and 2018 in peer-reviewed journals were selected.Review studies, studies without full text available, and conference proceeding papers were excluded.
The literature search occurred in November 2018.We have only included studies from 2013 because another review about the same subject has already been published in that year 7 .Thus, this is an updated review about this topic.
For the literature synthesis, data about a) participants (sex, age, sample size, and level of experience in running); b) kinematics; and c) kinetics and d) neuromuscular results were presented.Only variables described for at least two studies and with statistically significant differences between shod and barefoot conditions were depicted.

RESULTS
Six-hundred-eighty-seven studies were found, of these, 185 were duplicated, resulting in 502 studies to read the title.One-hundred-fifty--four papers were selected, and, after reading their abstract, 71 articles were chosen for full reading.Applying the eligibility criteria, 40 articles were included for the analysis, according to Figure 1.
Participants' characteristics are in Table 2.These participants were experienced in running, but rarely in barefoot running or wearing minimalist footwear.Twelve studies included rearfoot runners, while the others did not have foot strike pattern as inclusion criteria.Sample size ranged from six to 241 participants, from 18 to 55 years old.Regarding sex, studies have included only males, one study included only females, and 22 studies included both sexes.Most studies have included amateur runners with varying weekly race volumes.

Kinetics
Fourteen studies measured any kinetic variables, such as external moment, impact, and loading rate (Figure 3C).Considering the GRF, most studies showed the loading rate is higher in barefoot condition, 20,28,31,34,47,48 and the impact peak has opposite results. 22,24Regarding articular moments, plantar flexion moment was higher in the barefoot condition, 15,16 while knee extension moment was higher in shod condition. 15,33

DISCUSSION
This study aimed to analyze and summarize the biomechanical (kinematics, kinetics and neuromuscular) differences between barefoot and shod running, through a literature review.This discussion is divided in four question-guided sections.Running kinematics is discussed in the first two sections; while the running kinetics and its neuromuscular aspects are discussed in sections 3 and 4, respectively.
The number of studies that showed significant differences for kinematic, kinetic and neuromuscular variables between shod and barefoot running are depicted in Figure 3.These studies compared barefoot and shod running, and some of them did also compare different footwear types/models. 14,43,44One study 41 tested running on different slopes, other two 14,45 compared different running speeds and another 16 compared different stride lengths.
The three-dimension motion analysis was applied to every study about ankle and knee kinematics.Running barefoot 18,27,34,37,40,50 presented larger ankle range of motion (ROM) during stance phase.Regarding knee ROM, results are variable 18,34,39,46 and only one study 21 compared knee kinematics at different gait phases.

How different are the motion patterns in shod and barefoot running?
There are different ways to define the running pattern.For example, by the foot strike (forefoot, midfoot or rearfoot) pattern, or by the running strategy using stride length or cadence.In this review, we found stride length and cadence differs between shod and barefoot running.Stride length is shorter in barefoot running, and consequently, cadence is faster.Hall et al. 7 has described similar results.The foot strike differs between shod and barefoot running.Barefoot running is associated to mid and forefoot strike patterns, while shod running is associated with rearfoot strike.Stance time in mid/forefoot strike is shorter.Faster some people run, more typical such strike patterns are, 51,52 suggesting subgroups, that is, some runners change, while others do not, their foot strike pattern when they run faster.Besides, barefoot running might provide more sensory information during foot strike. 5Shoes might reduce the foot sensory information during the stance phase. 53In addition, the first attempts to run barefoot are usually associated to higher GRF, leading to change foot strike pattern.Stride length is shorter in barefoot running, and it depends on swing phase.A shorter swing phase can be associated with a lower GRF and a shock attenuation strategy. 54arefoot running has a faster cadence.This is a consequence of kinematic changes during the stance phase. 55In the barefoot condition, De Wit, Clercq and Aerts 55 suggest the leg stiffness is higher, inducing higher cadence.The leg stiffness is affected by how the GRF vector is oriented to ankle, knee and hip joints. 4Cadence is associated with swing leg properties as an inverted pendulum and stance time also influences it.Stance time shorts when foot strike is not with the rearfoot, and, for barefoot condition when foot strike is mid/forefoot strike.Differences on spatiotempral parameters, such as cadence, 21,22,24,25,28,45 are evident even when the speed in both conditions (shod and barefoot running) is similar.

How does footwear affect joint kinematics in running?
Comparing running with and without shoes, the lower limb joint kinematics is different.We highlight differences in the ankle, knee and hip.Hall et al. 7 showed limited evidence for the ankle and moderate evidence for the knee.In addition, we have found results that indicate changes in hip movements in running.
During barefoot running, ankle dorsiflexion decreases, plantar flexion increases, and the foot strikes the ground in ankle plantar flexion.When running with a rearfoot pattern, dorsiflexion occurs at the foot strike.Hall et al. 7 showed limited evidence suggesting greater eversion in barefoot running.Ankle ROM is larger during stance phase in barefoot running; however, for the initial contact and impulsion, such ROM is larger in shod running. 40Strauts et al. 40 found the runners used rearfoot strike in both shod and barefoot running.This result may be related to a lower ankle ROM in the barefoot condition compared with studies in which foot strike patterns varied among forefoot, midfoot and rearfoot patterns.
During stance phase, knee flexion is higher in barefoot running.Such knee flexion can reduce the vertical GRF and decrease injury risk. 56For barefoot runners with forefoot strike pattern, the knee ROM is shorter.Thus, barefoot running increases the maximum knee flexion, but decreases its ROM.
Barefoot running changes hip joint kinematics, and hip flexion ROM is lower during stance phase. 40Strauts et al. 40 suggested changes in hip and pelvis angles are coordinated with ankle joint movements.Muscle strength, muscle activation patterns, anthropometry, joint mobility and muscle length also affect hip, pelvis and trunk kinematics during running. 33

Does GRF decrease in barefoot running?
Kinetic gait analysis is mostly GRF-based and well-established in gait analysis.The GRF decreases with shorter stride length. 22The shorter the stride length, closer is the initial foot contact to the vertical projection of the center of mass, whereas the knee and hip flexion peaks decrease. 57he first peak of the vertical GRF is lower in barefoot condition and associated with shorter stride length. 34,47Such peak decrease may be related to GRF spectral change; 58 while forefoot strike pattern has more power spectrum up to 10 Hz, rearfoot strike pattern has more power spectrum above 10 Hz, including higher GRF frequency content like GRF transient.This implies the smallest first GRF peak is not due to barefoot running, but due to the forefoot strike pattern.
Runners used to run barefoot and with the forefoot strike pattern have only propulsion GRF peak, while runners used to run with shoes showed the impact and propulsion GRF peaks. 22Running with and without shoes produce similar GRF. 40The impact peak is three times lower in runners used to run barefoot with forefoot strike compared with runners used to run with rearfoot strike. 9But, the GRF peak was higher in barefoot condition with rearfoot strike compared to shod condition. 40Lieberman et al. 9 showed the GRF transient occurs during rearfoot strike pattern and it increases in barefoot condition compared with the shod condition.
The loading rate describes how fast GRF was applied during foot strike.In Hall et al., 7 the loading rate changes according to the foot strike pattern in barefoot running.In barefoot running, the forefoot strike reduces the loading rate and the rearfoot strike increases it. 28,34,41,42or the rearfoot strike pattern, the loading rate is higher without shoes than running with shoes; 28,31,34,47,48 but one study 48 found no difference between these two conditions.The loading rate is a high frequency GRF component, 59 and the high frequency GRF content is associated with the foot strike pattern, 58 and not necessarily due to the barefoot running.
There is little information about the moment patterns in running with and without shoes.Two studies 30,33 have shown the knee extension moment was higher in shod running than in barefoot running, and a study 25 showed no difference in ankle, knee and hip moments between running conditions.Hall et al. 7 did not evaluate the joint moments.Regardless of the shoe condition, the GRF peak and joint moments are not different for a similar speed condition. 20For shod and barefoot running, joint moment increases are associated with larger stride length.For the sagittal plane, increasing the stride length 10% longer augmented the knee and ankle peak moment, but no differences were found between running with and without shoes.

Does running barefoot alter the muscle activation patterns?
The lower limb muscle activity, joint kinematics and kinetics are associated.We are presenting a larger muscle set and their behavior during running with and without shoes, compared with Hall et al. 7 The interaction between foot strike patterns and running with or without shoes can change the leg muscles activation patterns.The triceps surae muscle activation is associated to the foot strike pattern.The gastrocnemius lateralis and medialis muscles showed more activation in the barefoot condition with forefoot strike than with rearfoot strike in any shoes condition. 11,17,34,35,37For the rearfoot strike, 47 the soleus and gastrocnemius muscles showed a peak activity during the stance phase and greater activity at the end of the swing phase.Bonacci et al. 15 showed that for the barefoot running, these muscles were more active during ankle plantar flexion.The increased ankle joint moment during forefoot strike is associated with greater soleus, gastrocnemius lateralis and gastrocnemius medialis muscle activation.Soleus, gastrocnemius lateralis and tibialis anterior muscles were less active during barefoot running.
The tibialis anterior muscle activation occurs during the foot strike.This muscle was more active during running with shoes and rearfoot strike. 17,34This muscle is more active immediately before foot strike.Such pre-activation maintains the plantar flexion at the end of the swing phase and prepares the leg for impact.During impact, the tibialis anterior muscle should be relaxed to release the forefoot. 60he knee extensor/flexor muscle activations depend on the interaction between barefoot running and running phases.The activation levels of vastus medialis and biceps femoral muscles have decreased in barefoot running compared with shod running during foot strike and propulsion phases. 40In the stance phase, the biceps femoris and rectus femoris muscles showed no difference between footwear and barefoot conditions. 34Probably, such similarity was induced by not controlling the foot strike pattern previously, 40 and biceps femoral activation might respond more to foot strike pattern than shoe condition.
In general, this literature review and Hall et al. 7 have similar results.Hall et al. 7 is a meta-analysis with 18 studies published from 1979 to 2013.Our narrative review combines the results of 40 studies from 2013 to 2018.To include various biomechanical studies about running with and without shoes, we have decided not to do a meta-analysis.We believe the results we have summarized will support the current and growing discussions about running with or without shoes.Despite these different studies being described, many biomechanical aspects of running with and without shoes were similar between Hall et al. 7 and our review.Our review is an update for this subject.

CONCLUSION
This literature review showed important differences in the biomechanics of running with and without shoes.Such differences should be understood and considered in the decision-making process between running with shoes or barefoot.Running barefoot might present lower impact forces, but still has a high loading rate and high triceps surae muscle activation.More studies about variables few explored, e.g.joint kinetics, should be carried out to enhance the understanding of running with and without shoes.

Figure 1 .
Figure 1.Flowchart of the studies' selection process.

Figure 2 .
Figure 2. Number of studies according with the instruments and running speeds used in the analysis protocols.

Figure 3 .
Figure 3. Number of studies that showed significant differences in the comparison of the spatiotemporal parameters (A), joint kinematics (B), kinetic variables (C) and neuromuscular aspects (D) between shod and barefoot running.

Table 1 .
Descriptors used in searches.

Table 2 .
Number of participants, age group and level of running practice in the studies included in the review (continues).