Key words
performance - footwear - running technique - cadence
Introduction
Over the last ten years, renewed interest in more natural running footwear has seen
the development of the minimalist footwear (MFW) sector. Minimalist footwear can be
described as “providing minimal interference with the natural movement of the foot
due to its high flexibility, low heel to toe drop, weight and stack height, and the
absence of motion control and stability devices” [14]. Self-report surveys have reported that 76% of runners in 2012 and 53% in 2014 have
been interested in this novel footwear condition at some point [21]
[33], due to suggestions of improved performance and reduced injury risk, despite the
fact that this has not yet been supported by consistent high-level evidence in the
literature.
Whilst there are numerous purported benefits of running in minimalist shoes, perhaps
the only consistent observation in the literature is an improvement in running economy
(RE) [3]
[13]
[24]
[31]
[32]
[36]
[42]
[43], which is supported by a recent systematic review [7]. RE can be described as the energy cost of running at a given submaximal, steady-state
velocity [35], and is often reported as the oxygen cost of running [34]. RE has been shown to be the most reliable indicator of endurance performance in
a similarly trained group of runners [22]
[23], explaining up to 65% of race performance over 10 km [10]. In addition, recent research has suggested RE above the lactate threshold may be
a reasonable predictor of 1500 m performance [39].
The improvements in RE reported in MFW have largely been associated with a reduced
shoe mass when compared to conventional running shoes (CRS) [7]
[13]
[14]. However, the magnitude of improvement in RE cannot be explained by shoe mass alone
[7] and may be attributed to a number of other variables. In this regard, it has been
suggested that running biomechanics, which are consistently observed to change when
switching to MFW [4]
[24]
[37], may also influence RE [11]
[16]
[31]
[44]. One biomechanical variable that has received attention in this regard is step frequency.
There is limited evidence that suggests increasing step (or stride) frequency can
improve RE [11]
[40]. However, although somewhat dated research, increasing step frequency can also negatively
influence RE [6]
[20] or have no effect [26]. An increased step frequency has been well documented when running in minimalist
footwear (MFW) when compared to CRS [24]
[41]
[44]
[43], and it is possible that this difference could influence RE. However, the natural
changes to step frequency in relation to footwear type are typically small (~2–4%)
[24]
[41]
[44]
[43]. In addition, previous studies that have deliberately manipulated step frequency
have done so at a greater magnitude (e. g. 8–10%) [17]
[20], which may therefore not be applicable to this small “footwear-related’ change in
step frequency reported above. The small spontaneous changes in step frequency associated
with different footwear remain to be examined with respect to its impact on RE.
Therefore, the aim of this study was to examine if small enforced changes in step
frequency, of a magnitude typically experienced by changing from CRS to MFW footwear
types, will influence RE in trained males. A secondary aim was to provide practical
advice for those using different footwear in an attempt to increase step frequency
in the belief that it might enhance running performance.
Methods
Participants
Twelve male club-level runners with eight weeks prior minimal footwear experience
(up to approx. 30 km/week) were recruited for the study (age, 41±9 years; stature,
177.2±10.4 cm; body mass, 72.6±10.2 kg; V˙O2max, 52.1 ±7.5 mL·min−1·kg−1). Participants typically ran 4–6 days per week with a mean weekly running distance
of 52 (±11) km at the time of the study. Participants were excluded if they had reported
any running-related injuries in the last three months or had previous barefoot or
minimalist running experience before the eight-week MFW transition. This transition
has been previously reported [42]. All participants had previous experience with treadmill running. The participants
gave informed consent at the beginning of testing. Ethical approval for this study
was granted by the Dublin City University Research Ethics Committee. The present study
meets the ethical standards required of this journal [18].
Experimental design
In a single testing session, subjects underwent two 6-min RE tests: one in MFW and
one in CRS (balanced randomisation), during which the subject’s naturally selected
step frequency was recorded. For testing, foot size was measured and participants
were provided with one pair of MFW (Vibram® Five Finger “KSO”; ~150 g), and a neutral
CRS (Asics® “GEL-Cumulus” 2012; ~400 g). The same RE tests were then repeated in both
types of footwear (again randomised) but with enforced changes in step frequency.
This was controlled by a metronome (“Mobile Metronome” Android software) set at the
corresponding tempo of the opposite condition being tested (when participants ran
in MFW, their recorded step frequency in CRS was enforced and vice versa), denoted
“revSF” [reversed step frequency].
Testing procedure
Resting blood lactate (Lactate Plus, Nova Biomedical, Waltham, MA, USA) was sampled
from the earlobe prior to the testing sessions. Respiratory data were measured using
a Viasys Vmax Encore 299 online gas analysis system (Viasys Healthcare, Yorba Linda,
CA, USA). The system was calibrated according to the manufacturer guidelines, including
atmospheric pressure and temperature, before each new test. For this system, accuracy
has been reported at 0.02% for oxygen measures, following a 15-min warm-up period
and calibrated within 5% of absolute operating range. A treadmill (Cosmed T170, Sport
Med, Weil am Rhein, Germany) RE test was then conducted in the assigned footwear,
either MFW or CRS in random order. Treadmill incline was set at 1% to account for
air resistance [25]. Participants ran four trials lasting 6 min at 11 km·h−1, which has previously been considered an appropriate steady-state “endurance running”
velocity [19]. At the end of each 6-min stage, participants were asked to stand to the side of
the treadmill and a blood lactate sample was collected within 30 s. At minute 5 in
each stage, step frequency was collected by counting the left foot contact with the
treadmill belt for 60 s duration. This procedure was repeated by the same investigator
in each subject and also filmed for a second assessment and accuracy (R2=0.95; Sony HDR-CX210, 60FPS; Sony, San Diego, CA, USA). Rudimentary foot strike pattern
(FSP) analysis was undertaken using this low-cost video camera, in which participants
were filmed in the sagittal plane at foot level over a 15 s period during minute 4
of testing. The video footage was then used to assign 1, 2, or 3 (1=forefoot strike,
2=midfoot strike, 3=rearfoot strike) to the participants’ foot strike pattern by the
principal investigator using Dartfish video analysis software (Dartfish 5.5, Fribourg,
Switzerland). A midfoot strike was assigned when there was no clear initial forefoot
or heel contact. The validity of this method has been previously examined and highly
correlated to the strike index (R2=.85) [1]. The next test in the opposite footwear was started after 3 min of passive rest
to allow the shoe type to be swapped over. After each shoe test was completed once,
both tests were repeated after three minutes of passive recovery, but this time with
the step frequency dictated by a metronome as reported above.
A V˙O2max test was completed at the end of the day for subject characterisation. This involved
a ramped treadmill protocol at 12 km·h−1 for a 5-min warm-up before increasing to 14 km·h−1 at 1% incline. The incline was then increased every minute until volitional exhaustion
and correlated with participants achieving a respiratory quotient (RQ) of 1.1 or above.
Participants conducted this test in their own shoe choice. V˙O2max was recorded as the highest breath-by-breath value averaged over 60 s.
Data processing
The RE values were determined from the mean data over the last 2 min of each stage
when participants had reached a true steady-state V˙O2. This was verified by less than a 1 mmol increase in blood lactate (post-trial minus
resting lactate) because this is considered well below maximal lactate steady state
[38], and an RQ of less than 1.0 [5].
Data analysis
Direct comparisons between RE and RErevSF, were completed in the same footwear (MFW vs. MFWrevSF and CRS vs. CRSrevSF) using repeated measures ANCOVA following establishment of parametric assumptions.
Because it is possible that the foot strike pattern can influence V˙O2 values [16]
[31], the foot strike pattern was included as a covariate in the analysis. Difference
in step frequency between MFW and CRS was also examined with a paired t-test. Statistical
significance was accepted at α≤0.05 (Statistical Package for the Social Sciences data
analysis software V22.0, SPSS Inc., Chicago, IL, USA).
Effect sizes are reported as Cohen’s d [8]. The smallest standardised change that is considered meaningful was assumed to be
an effect size of 0.20 for Cohen’s d [8].
Results
No participants were excluded based on any visual slow component for submaximal V˙02 consumption, an increase in blood lactate of>1 mmol (mean change from resting=0.44 mmol),
or an RQ greater than 1.0. Because this was an acute study, no dropouts occurred due
to injury or other reasons.
There was no difference whatsoever in the foot strike classification between the normal
and reversed stride frequency for either shoe condition. For the MFW footwear, the
distribution of foot strike patterns was 4 rearfoot strikes, 3 mid-foot strikes, and
5 forefoot strikes. For the CRS footwear, the distribution of foot strikes was 6 rearfoot
strikes, 3 mid-foot strikes, and 3 forefoot strikes. The mean increase in step frequency
for minimal footwear vs. conventional running shoes was 7.3±2.3 steps per minute (3.9%
change; p≤0.001; 95% CI of difference [5.86 to 8.80]; MFW 184.2±10.6 vs. CRS 176.8±10.5
steps per minute).
No differences were identified between RE and RErevSF for minimal footwear when foot strike pattern was taken into account as a covariate
(p=0.55; 95% CI of difference [–1.71 to 0.97]; Cohen’s d=0.09; RE 40.72±4.08 vs. RErevSF 41.09±4.19 mL·min−1·kg−1), or conventional running shoes (p=0.55; 95% CI of difference [–0.78 to 1.37]; Cohen’s
d=0.06; RE 42.04±4.68 vs. RErevSF 41.74±5.09 mL·min−1·kg−1). Differences are displayed in [Fig. 1]. Removing the foot strike pattern as a covariate factor did not influence the outcome
in any way.
Fig. 1 Differences in running economy (RE) when comparing the minimalist footwear (MFW)
self-selected step frequency to reversed step frequency (SF), and the same comparison
in conventional running shoes (CRS). No significant differences were observed. Error
bars represent SD.
Discussion
The main finding of the present study is that changes in step frequency as a result
of footwear condition (~4%) are not large enough to have any significant impact on
RE, even when the foot striking pattern is controlled for in the analysis. Therefore,
we reject the alternate hypothesis for this study. If RE changes are indeed associated
with switching to MFW, this is most likely not due to the changes in step frequency
that are typical of changing into this footwear.
The results of this study support previous work suggesting that step frequency is
not an influencing factor for RE [2]
[9]
[22]
[44]. Naturally selected step frequency has been found to be close to that which optimises
running economy, with small deviations resulting in little or no change [6]
[30]. To support this observation, more recent studies have deliberately reduced stride
length by 3% (which will increase stride frequency) and have shown no change in RE
[12]
[28]. A possible reason for this absence of any difference in RE is most likely due to
the magnitude of the changes to step frequency observed, which were found to be well
below the changes imposed in experimentally imposed studies [17]
[20]. In one study, Franz, Wierzbinski and Kram [15] estimate that the ~3% greater step length observed during traditionally shod running
when compared to barefoot would account for less than a 0.4% metabolic saving. As
changes to step frequency in MFW were also in the region of ~3% reported by Franz,
Wierzbinski and Kram [15] in the barefoot condition, it could be suggested that the same conclusion applies.
In the study by Hamill, Derrick and Holt [17], the authors noted that the preferred step frequency was the optimal for oxygen
consumption. The authors also noted significantly greater oxygen consumption at –10%,
–20%, and +20% step frequency, but there was no significant difference at +10%. Therefore,
the mean +3.9% increase in step frequency in MFW in the present study is very small,
and as such unlikely to cause any changes in RE, as observed here.
In contrast to these findings, a significant negative correlation between RE and stride
frequency (r=–0.61) has been previously observed in 16 male long-distance runners
[40]. In addition, Connick and Li (2014) have suggested that RE was optimised at a stride
which was 2.9% (±2.4%) shorter than preferred [11]. In an earlier training study, 9 runners who presented with uneconomical freely
chosen step lengths underwent a 3-week biofeedback programme to reduce step length
by 10%; a marked reduction in freely chosen step length as well as an improvement
in RE was observed [29]. It is therefore possible that benefits to RE with changes in step frequency may
only be apparent for those runners with low step frequency and/or uneconomical step
length/frequency, and this should be examined further. Regardless, there appears to
be contradictory research in the association between step mechanics and RE, with the
current study confirming that no changes in RE are observed with small changes to
step frequency.
One limitation of the current study is that the changes in step frequency were experimentally
imposed in an acute study, and it is possible that acute, forced, unnatural changes
to running mechanics may limit any potential benefits to RE [27]
[42]. However, this theory needs to be investigated further, as well as long-term training
studies on self-optimisation of RE with changes in running mechanics. Finally, it
would have been beneficial to work with a larger sample size, so that groups could
be divided according to foot strike patterns in order to establish if there is any
interaction between foot strike and step frequency changes in RE. This question could
be examined in future research with a large cohort of runners with varied running
mechanics.
Conclusion
Changes in step frequency as a result of footwear condition (~4%) do not have any
significant effect on RE. Therefore, changes in RE associated with MFW are most likely
due to other factors not examined in this study.
