The study was conducted in accordance with the Declaration of Helsinki and approved by the King’s College London Research Ethics Committee (REC Protocol No: BDM/11/12–66). Subjects were recruited via email advertising and snowball sampling within the university population of King’s College London from July 2012 to September 2013. Subjects were provided with an information sheet. Thirty-seven healthy adults (19 female and 18 male) provided written consent to participate. Inclusion criteria were aged 18–50 years and with previous experience of running on a treadmill. Exclusion criteria were any medical conditions presenting a risk when running, lower extremity injuries requiring medical attention in the past 6 months, previous or current use of minimalist footwear and the use of medically prescribed orthotics. All subjects completed a questionnaire detailing running behaviours and preconceived opinions on minimalist footwear. They received a complimentary pair of Vibram FiveFingers (VFF) as compensation for their time, along with instructions for use in accordance with the manufacturer’s recommendations. 24

Subjects attended on two occasions at least 48 hours apart. The protocol for each session was the same apart from the footwear worn. At the first session, subjects were randomly assigned to wearing either VFF or SRS and the other type was worn at the second session. Subjects performed a 2-min warm up at 1.8 m/s.

After a 3-min rest period, subjects were given the standardised verbal instruction ‘we would now like you to run in a light, soft and quiet way’. This was in accordance with suggestions by the distributor and similar to those used in the study by Diebal et al 27 to encourage a FFS. Subjects ran for another 2 min attempting to follow the instructions, which were repeated after one and 2 min. Thereafter, data were recorded for 20 s.

Following marker placement, a 20 s static standing stance on the horizontal treadmill was recorded. Subjects then undertook a 2-min run at 3.58 m/s to allow for familiarisation. This speed ensured that subjects were running 25 but remained below the proposed threshold for sprinting. 26 If subjects verbally reported difficulty with maintaining that speed it was reduced to 3.13 m/s. Following this, subjects continued to run for 30 s while 20 s of kinematic data were collected.

Eleven infrared markers were applied to the right leg. Marker locations ( figure 2 ) were chosen in accordance with previous running kinematic studies for comparison. 9 17 Markers were attached to the skin after palpation using double-sided adhesive and overlay tape. Markers for the tuber calcaneum and the head of 55th metatarsal were attached following palpation over the footwear.

A specifically designed programme in Matlab (version 2006a, Mathworks, Natik, MA, USA) allowed for the demarcation of initial contact using the method described by Altman and Davis 13 and corroborated by identification of the point at which either the 5th metatarsal or calcaneus equalled the velocity of the treadmill. Toe-off was indicated by the point at which the 5th metatarsal started to decelerate. Symmetry between left and right lower extremities was assumed. 17 Kinematic profiles were derived after averaging a minimum of 10 consecutive strides. 28 Kinematic data were analysed at four stages of the gait cycle to allow for comparisons with other running trials; prior to initial contact, initial contact, mid-stance and toe-off. 9 11 17 Although mid-stance is not a discrete point in time, it was defined as occurring at 50% of stance. 14

Kinematic data was digitised and analysed using the motion segments shown in figure 2 . All subsequent angles were corrected against angles measured during the static standing stance. This method has been reported to have a measurement precision of ±0.26 ˚ . 9 Plantar foot angle was calculated as the difference in height between the 5th metatarsal and calcaneus marker on the z-axis. 13

Data were analysed using Statistical Package for the Social Sciences (SPSS) version 20 (IBM, USA). Data were tested for normality by observation of the Q-Q plots and by using Kolmogrov-Smirnov tests. A factorial two-by-two repeated measures analysis of variance was performed for parametric data. Friedman’s analysis of variance was performed to analyse ankle angle at mid-stance, as this variable was found not to be normally distributed. Effect size (r) was reported as partial eta square; small effect (r=0.10), medium effect (r=0.30) and large effect (r=0.50). 29

The characteristics of the subjects and data from the questionnaire regarding running behaviours and preconceived opinions on minimalist footwear are shown in table 1 . The majority were regular runners, performing a mean of 1.5 hours a week. The majority (n=29) had heard of minimalist footwear and 19 of these thought they would have a beneficial effect.

Thirty five subjects completed the study with no reported adverse events. Three requested a lower running speed of 3.12 m/s. One of the 37 subjects recruited withdrew from the study after the first session due to an injury sustained outside the study and one subject’s data were unusable due to a technical error.

Neither footwear nor instruction had a significant effect on the proportion of time spent in the stance phase during running. There was a main effect of footwear such that cadence was higher (p<0.001, r=0.43) and stride length lower (p<0.01, r=0.27) when wearing VFFs. There was no main effect of instruction on either cadence or stride length ( table 2 ).

Knee and ankle angles were essentially similar throughout the running cycle with some exceptions ( figure 3 ). In the last 20% of the cycle, the ankle was more dorsiflexed when there were no instructions in both types of footwear and there was a suggestion that the knee was more extended when wearing shoes only. When wearing VFFs there was a slight decrease in knee extension when there were instructions.

There was a main effect of footwear at initial contact minus 10% at both the ankle (p=0.03, r=0.14) and the knee (p=0.04, r=0.12), whereby in VFF the ankle was in greater plantarflexion and the knee was in less flexion than when wearing shoes. At initial contact minus 10%, there was a main effect of instruction at the ankle (p<0.001, r=0.37) where plantarflexion increased marginally (0.3°) with instructions ( table 4 ).

Discussion

Our hypothesis that both running in minimalist footwear and running in SRS plus instructions leads to changes in lower limb kinematics, is supported by the results of this study. Importantly, we note that providing simple instructions to run ‘lightly, softly and quietly’ when wearing conventional running shoes leads to similar lower limb kinematic changes to wearing minimalist footwear. This is the first study to demonstrate this finding. Both instruction and minimalist footwear can increase the extent of forefoot landing, while the combination of the two had the greatest effect. Wearing minimalist footwear without instruction induced a forefoot strike running pattern prior to landing (IC −10% ). This was not seen at initial contact (IC), which is in contrast to the findings of others9 11 12 17 presumably because the effect of footwear type was dwarfed by the much larger effect of instruction. The effect of footwear on the extent of plantarflexion prior to landing (IC -10% ) was small (r=0.15) when compared with the larger effect of instruction on this variable prior to landing (r=0.47) or at initial contact (r=0.35).

From the clinical perspective some authors have proposed that the more forefoot strike running pattern may be associated with reduced lower limb injury risk.1 9 30 Others have not found this to be the case7 16 and Tam et al 8 emphasise that the link between minimalist footwear and running injury or performance has yet to be established using long term prospective studies. Some authors have observed a high incidence of foot bone marrow oedema31 or metatarsal stress fracture in those transitioning rapidly (0–2 months) from conventional to minimalist footwear32 although it has been suggested that a gradual transition may reduce this risk.30 Others have reported that forefoot running is beneficial for those with chronic exertional compartment syndrome27 and may be associated with a lower incidence of overuse injuries.30 It seems that the clinical effect of footwear type and running style varies in different parts of the lower limb,33 which may increase the risk of certain injuries and reduce the risk of others.

We studied lower limb running kinematics during a 20 s period following 2 min of running, which was unlikely to produce fatigue. Longer bouts of running with a forefoot strike require strong contractions of the calf muscle complex, which could be associated with an increased incidence in injuries to the Achilles tendon.33 Furthermore, such fatigue could lead to heel strike running pattern in minimalist footwear leading to an increased propensity for calcaneal fractures.32

The greater foot plantarflexion during the landing phase in the VI+ condition (which equates to a mean change of 27 mm into the forefoot strike posture when compared with the standard running shoe and no instruction condition (effect size r=0.28) can be explained by altered kinematics at both the knee and the ankle. With minimalist footwear and instruction, there was a mean increase in ankle plantarflexion prior to landing of 8° when compared with ankle angle during running with SRS and without instruction. The effect size for this difference was small (r=0.14). Similarly, there was a small effect (r=0.12) of minimalist footwear, equating to a mean increase of 3° in knee extension prior to landing, when compared with running in SRS alone. As was found in this study, others have shown that knee flexion angle at initial contact does not differ depending on type of footwear.11 12 Although there was a medium effect size (r=0.37) for the effect of instruction on ankle plantarflexion angle prior to landing it must be noted that the mean difference in ankle angle as a result of instruction was 0.3°. The effect of such small kinematic differences at the knee and the ankle, in terms of performance, injury risk and clinical significance is unknown. The contribution of each joint to the plantarflexion angle varies between footwear and instruction; when in minimalist footwear ankle plantarflexion increases and the knee is more extended. When under instruction both ankle plantarflexion and knee flexion increase.

We monitored kinematics in the lower limb only although clearly the human body is a series of connected segments, so the effect of footwear and instruction on other segments is not known. This could be particularly important in clinical terms if the relative position of the trunk was influenced in a manner thought likely to increase the risk of back pain. Furthermore, a limitation of the work presented is that we did not investigate kinetics, such as ground reaction forces, during the running conditions studied.

Our hypothesis that running in minimalist footwear and running in SRS plus instructions, leads to changes in spatiotemporal parameters, was only true for the effect of minimalist footwear. The increased cadence (a mean increase of four steps per minute, with effect size r=0.43) and decreased stride length, which equates to a mean 5 cm reduction in stride length (effect size r=0.27) when in minimalist footwear corroborates the work of others.11 33 This is thought to occur to reduce vertical loading rate of the lower limb34 such as occurs when running barefoot or in less cushioned shoes.9 35 There have been reports of a reduced stance duration and stride length with increased cadence11 17 26 36 in studies using treadmills with larger belt dimensions than used here, which may explain the different findings.

Our finding that instruction had no main effect is in contrast to that of Diebal et al 27 who observed a significant reduction in stride length and stance time with an increase in cadence following a 6-week period of training combined with a package of instruction three times per week, for approximately 45 min, which included the verbal cue to ‘run quietly’, a digital metronome to stabilise step cadence at 180 steps per minute and visual instruction via a video camera recording to facilitate forefoot strike technique. The varying levels of instruction, and the attempt to control cadence in the study by Diebal et al 27 may account for the difference in findings.

Instruction to ‘run in a light, soft and quiet way’ had greater effects on kinematics than wearing minimalist footwear without instruction. The large interindividual variations indicate that a number of subjects responded differently to the same instruction. This may be explained by the fact that less experienced runners have been reported to adopt lower preferred running speeds13 and that treadmill speeds 10% greater than an individual’s preferred speed significantly increases ankle plantarflexion at IC.36 It is possible that our less experienced runners exceeded their preferred running speed, which is supported by three subjects requesting to run at a reduced speed. Trainers and clinicians should carefully evaluate individual responses to instruction to achieve the desired effect on running style.

We studied only one type of minimalist footwear although there are a variety of different makes and models. The findings of this study demonstrate that instruction alone can influence running style in a way that may reduce the risk of injury and aid injury management and that this can be enhanced by the type of footwear. These findings should be considered in the context of a subject cohort with variable amounts of running experience and in light of this data being unable to reveal whether the kinematic changes observed with instruction would be sustained with longer bouts of running.