Journal of Sports Sciences

ISSN: 0264-0414 (Print) 1466-447X (Online) Journal homepage: http://www.tandfonline.com/loi/rjsp20

Are gait characteristics and ground reaction forces

related to energy cost of running in elite Kenyan
runners?

J. Santos-Concejero, N. Tam, D. R. Coetzee, J. Oliván, T. D. Noakes & R. Tucker

To cite this article: J. Santos-Concejero, N. Tam, D. R. Coetzee, J. Oliván, T. D. Noakes

& R. Tucker (2016): Are gait characteristics and ground reaction forces related to
energy cost of running in elite Kenyan runners?, Journal of Sports Sciences, DOI:
10.1080/02640414.2016.1175655

To link to this article: http://dx.doi.org/10.1080/02640414.2016.1175655

Published online: 07 May 2016.

Submit your article to this journal

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at

http://www.tandfonline.com/action/journalInformation?journalCode=rjsp20

Download by: [Orta Dogu Teknik Universitesi] Date: 07 May 2016, At: 10:09
 JOURNAL OF SPORTS SCIENCES, 2016
http://dx.doi.org/10.1080/02640414.2016.1175655

Are gait characteristics and ground reaction forces related to energy cost of running
in elite Kenyan runners?
J. Santos-Concejeroa,b, N. Tamb, D. R. Coetzeeb, J. Olivánc, T. D. Noakesb and R. Tuckerb,d
a
Department of Physical Education and Sport, University of the Basque Country UPV/EHU, Vitoria-Gasteiz, Spain; bDivision for Exercise Science and
Sports Medicine, Department of Human Biology, University of Cape Town, Cape Town, South Africa; cDepartment of Physiology, European
University of Madrid, Madrid, Spain; dSchool of Medicine, University of the Free State, Bloemfontein, South Africa

ABSTRACT ARTICLE HISTORY

The aim of this study was to determine whether gait cycle characteristics are associated with running Accepted 1 April 2016
economy in elite Kenyan runners. Fifteen elite Kenyan male runners completed two constant-speed
KEYWORDS
running sets on a treadmill (12 km ·h−1 and 20 km ·h−1). VO2 and respiratory exchange ratio values were Energy cost of running;
measured to calculate steady-state oxygen and energy cost of running. Gait cycle characteristics and ground contact; stride
Downloaded by [Orta Dogu Teknik Universitesi] at 10:09 07 May 2016

ground contact forces were measured at each speed. Oxygen cost of running at different velocities was length; stride frequency;
192.2 ± 14.7 ml· kg−1· km−1 at 12 km· h−1 and 184.8 ± 9.9 ml· kg−1· km−1 at 20 km· h−1, which African runners
corresponded to a caloric cost of running of 0.94 ± 0.07 kcal ·kg−1·km−1 and 0.93 ± 0.07 kcal· kg−1·
km−1. We found no significant correlations between oxygen and energy cost of running and biome-
chanical variables and ground reaction forces at either 12 or 20 km· h−1. However, ground contact times
were ~10.0% shorter (very large effect) than in previously published literature in elite runners at similar
speeds, alongside an 8.9% lower oxygen cost (very large effect). These results provide evidence to
hypothesise that the short ground contact times may contribute to the exceptional running economy
of Kenyan runners.

Introduction attributed to gait cycle and ground contact characteristics

(Anderson, 1996; Kyröläinen et al., 2001).
Since the 1970s, running economy has been recognised as a
Kenyan runners are among the most economical runners in
crucial determinant of distance running performance, though it
the world (Santos-Concejero, Billaut, et al., 2015), and as a
has also been referred to as “relatively ignored in the scientific
result, numerous researchers have studied whether this factor
literature” (Foster & Lucia, 2007). This highlights how poorly the
may enable Kenyan runners’ outstanding performance (Larsen,
relationships between running economy, training and biomecha-
2003; Saltin et al., 1995; Tam et al., 2012). Although a dissocia-
nical and gait characteristics are understood (Santos-Concejero,
tion between running economy and performance in Kenyan
Oliván, et al., 2015).
runners could exist (Mooses, Mooses, et al., 2015), a deep
Running economy is defined as the steady-state oxygen uptake
analysis of the running kinematics and energy cost of running
(VO2) required at a given submaximal velocity (Nummela, Keränen,
in Kenyan runners may provide important insights into the
& Mikkelsson, 2007) or as the energy requirement per unit of
biomechanical determinants of their energy cost of running
distance run (Fletcher, Esau, & Macintosh, 2009). It is influenced
and thus, performance. One hypothesis is that the exceptional
by muscle fibre distribution (Bosco et al., 1987), age (Krahenbuhl &
running economy of Kenyan runners may be the result of
Pangrazi, 1983), sex (Bransford, & Howley, 1977) and anthropo-
relatively short ground contact times (Kong & de Heer, 2008).
metric factors (Lucia et al., 2006). Running economy is also influ-
This may be a consequence of a high efficacy in the use of the
enced by biomechanical variables, largely attributed to gait cycle
recoil of elastic energy from the tendinous structures accom-
and ground contact characteristics (Kyröläinen, Belli, & Komi, 2001;
panied by shorter stretching and higher stretching to preacti-
Santos-Concejero et al., 2014).
vation ratio than has been measured in lesser performers
Shorter ground contact times (Nummela et al., 2007; Santos-
(Sano et al., 2013). This hypothesis has yet to be tested as
Concejero et al., 2013, 2014), lower stride frequencies (Tartaruga
the exceptional running economy of Kenyan runners has not
et al., 2012), longer swing times, greater stride angles (Santos-
been explored as an outcome of specific gait cycle and ground
Concejero et al., 2014) and longer strides (Santos-Concejero et al.,
contact characteristics in previous studies analysing these
2013, 2014; Tartaruga et al., 2012) have all been related to
parameters (Kong & de Heer, 2008; Sano et al., 2015).
improved economy in runners ranging in ability from elite to
Thus, the present study aimed to determine whether spa-
recreational. Ground reaction forces (GRFs) have also been
tio-temporal parameters and biomechanical variables of the
shown to influence economy, with lower peak GRFs associated
gait cycle may be associated with running economy in a group
with improved running economy biomechanical variables, largely

CONTACT J. Santos-Concejero jordan.santos@ehu.eus Department of Physical Education and Sport, Faculty of Physical Activity and Sport Sciences,
University of the Basque Country UPV/EHU. Portal de Lasarte 71, 01007 Vitoria-Gasteiz, Spain
© 2016 Informa UK Limited, trading as Taylor & Francis Group
 2 J. SANTOS-CONCEJERO ET AL.
of elite Kenyan runners. We hypothesised that East African
runners’ exceptional running economy would be associated
with shorter ground contact times, lower stride frequencies,
longer swing times, longer strides and lower GRFs, in support
of previously published literature.
Methods
Fifteen elite Kenyan male runners from the Kalenjin tribe (age:
23.7 ± 4.2 years) with mean 10 and 21 km race times of
28.7 ± 0.4 min and 62.2 ± 1.0 min, respectively, were recruited
for the study. Of the 25 athletes initially contacted, 10 declined
the invitation to participate due to injuries or being focused
on main competitions in Europe and America. The Research
Ethics Committee of the University of Cape Town approved
this study (HREC ref 151/2013), which was conducted post hoc
in accordance with the principles of the Declaration of Helsinki
Downloaded by [Orta Dogu Teknik Universitesi] at 10:09 07 May 2016
(October 2013, Fortaleza). Informed consent was obtained
from all individual participants included in the study.
All participants were seasoned competitors and were
tested between November and May. During this period, they
were in peak condition for target competitions of the road
season (fall and spring). Forty-eight hours prior to the first test,
athletes abstained from hard training sessions and competi-
tion. They refrained from alcohol and caffeine ingestion for at
least 24 h before testing. All athletes completed a familiarisa-
tion session on the treadmill prior to the experimental
sessions.
Anthropometry
For descriptive purposes, height (cm) and body mass (kg) were
recorded using a high precision balance (Seca 899, Seca,
Germany) and a stadiometer (Charder HM200P, Charder
Electronic Co, Taiwan) and the body mass index (BMI) was
calculated. Eight skinfold sites (triceps, biceps, subscapular,
supra-iliac, supraspinale, abdominal, front thigh and median
calf) were measured in duplicate with skinfold calipers (Holtain
Tanner-Whitehouse, Crymych, UK) by the same researcher to
the nearest millimetre and body fat percentage was calculated
(Yuhasz, 1974). All measurements were taken following the
guidelines outlined by the International Society for the
Advancement of Kinanthropometry.
Running economy assessment
All athletes completed two constant-speed running sets of
6 min each, separated by 5 min recovery periods on a tread-
mill (H/P/Cosmos Saturn, Nussdorf-Traunstein, Germany) at a
gradient of 1%. A slow increase in VO2 during a constant-work-
rate exercise performed above the lactate threshold has been
described, also known as the slow component of the VO2.
Thus, to ensure steady-state measurements, the speeds
selected, 12 km·h−1, which was tested first (warm-up pace),
and 20 km·h−1 (~half-marathon pace), were slower than the
individual lactate threshold of each athlete [further confirmed
during the test by the respiratory exchange ratio (RER) being
below 1.0 during the whole running bout for all athletes]. RER
maximum, minimum and mean values were 0.81, 0.75 and
0.78 at 12 km·h−1 and 0.99, 0.96 and 0.98 at 20 km· h−1.
Gas exchange data were collected using an automated
breath-by-breath system (COSMED Quark CPET, Rome, Italy),
which was calibrated before each session according to the
instructions of the manufacturers. Volume calibration was per-
formed at different flow rates with a 3-L calibration syringe
(Cosmed, Rome, Italy) allowing an error <2%. Calibration of
the gas analyser was performed automatically by the system
using both ambient and precision reference gases (16% O2,
5% CO2). VO2 (ml· kg−1· min−1) and RER values collected during
the last 60 s of each stage were averaged and used to calcu-
late steady-state running economy. Running economy in each
runner was calculated as the energy cost per distance (kcal
·kg−1· km−1), using VO2 and known energy equivalents. An
energy equivalent of 4.8 kcal· l−1 was applied when RER was
0.82. A change of ±0.01 in RER-value caused a ±0.012 kcal
change in energy expenditure (Fletcher et al., 2009). Running
economy tests were performed under similar environmental
conditions (20–24°C, 45–55% relative humidity at 130 m of
altitude) for all athletes.
Biomechanical variables
Biomechanics assessment included measurements of gait
cycle characteristics and kinetic factors. Gait cycle character-
istics measured were ground contact time, stride length, stride
frequency and swing time. Kinetic forces included ground
contact force and loading rate. Biomechanics assessments
were conducted on a 40 m indoor synthetic running track.
Three-dimensional marker trajectories were captured using an
eight-camera VICON MX motion analysis system (Oxford
Metrics Ltd, Oxford, UK), sampling at 250 Hz. GRF data were
collected using a 900×600 mm force platform (AMTI,
Watertown, MA, USA), sampled at 2000 Hz. Sixteen reflective
markers were attached according to the standard PlugInGait
model marker set. The capture volume had a length of 9 m,
allowing adequate distance for the absence of acceleration
and deceleration.
All participants completed six clean overground running
trials at each speed (12 and 20 km· h−1). This was performed
in a randomised order. Trials were accepted if the velocity was
within ±2% of the target speed. During these runs, synchro-
nised collection of marker motion and force platform was
obtained with VICON Nexus (VICON, Oxford Metrics, Oxford,
UK). A successful trial was defined as one within the specified
velocity range, where all markers were in view of the cameras
and there was no visual evidence of force platform targeting.
Marker trajectory and force platform data were filtered
using a low-pass fourth-order Butterworth filter with a cut-off
frequency at 25 and 100 Hz, respectively. For each trial, one
complete gait cycle was analysed per trial. Subsequently, a
customised program calculated ground contact and swing
time, stride and frequency, active peak and braking forces
and vertical initial loading rate. Initial loading rate was quanti-
fied between 20% and 80% of the slope from initial ground
contact to the initial peak. If no initial peak was visible, the
same parameters were measured at the same percentage of

stance ±1 standard deviation (SD) in trials with an initial
impact peak. These data from each trial were then averaged.
Statistical analysis
All values are expressed as mean ± SD and coefficient of
variation (CV). Statistical analyses of data were performed
using the Statistical Package for the Social Sciences 21.0 soft-
ware package (IBM Corp, Armonk, NY, USA). Data were
screened for normality of distribution and homogeneity of
variances using a Shapiro–Wilk normality test and a Levene’s
test, respectively. Pearson’s product-moment correlations
were used to assess the relationships between gait cycle,
ground contact characteristics and running economy across
each running speed (12 and 20 km· h−1). The magnitude of
differences or effect sizes (ES) were calculated according to
Cohen’s d (1988) and interpreted as small (>0.2 and <0.6),
Downloaded by [Orta Dogu Teknik Universitesi] at 10:09 07 May 2016
moderate (≥0.6 and <1.2) and large (≥1.2 and <2) according
to the scale proposed by Hopkins, Marshall, Batterham, and
Hanin (2009). Linear regressions were performed to analyse
the relationships between biomechanical variables and RE and
95% confidence intervals (95% CI) and the standards errors of
estimate (SEE) were calculated for each correlation.
Significance for all analyses was set at P < 0.05.
Results
Anthropometrical characteristics of Kenyan runners are listed
in Table 1. The homogeneity of the group is indicated by a CV
less than 2% for performance-related variables (10 km and
half-marathon best times).
Oxygen cost of running at different velocities was
192.2 ± 14.7 ml· kg−1· km−1 at 12 km·h−1 (CV = 7.6%) and
184.8 ± 9.9 ml ·kg−1· km−1 at 20 km· h−1 (CV = 5.4%), which
corresponded to a caloric cost of running of 0.94 ± 0.07 kcal·
kg−1· km−1 (CV = 7.6%) and 0.93 ± 0.07 kcal·kg−1·km−1
(CV = 8.0%), respectively.
Running economy (kcal·kg−1·km−1), gait cycle and ground
contact characteristics’ correlations, including the 95% CI, are
depicted in Figure 1 (12 km· h−1), Figure 2 (20 km· h−1) and
Table 2. We found no significant correlations between running
economy and ground contact time (Figures 1(a) and 2(a),
trivial effect), swing time (Figures 1(b) and 2(b), trivial effect),
stride length (Figures 1(c) and 2(c), trivial effect) or stride
frequency (Figures 1(d) and 2(d), trivial effect) at either 12 or
Table 1. Participant characteristics and maximal test results (n = 15).
Mean ± SD
10 km race time (min) 28.7 ± 0.4
Half-marathon time (min) 62.2 ± 1.0
VO2max (ml ·kg−1· min−1) 71.9 ± 5.1
Training (km· week−1) 128.9 ± 18.8
Height (cm) 170.5 ± 6.3
Mass (kg) 54.8 ± 6.3
BMI 18.8 ± 1.3
Body fat (%) 8.7 ± 0.5
∑ 8 skinfold (mm) 34.2 ± 6.7
Values are mean ± SD.
n, number of participants; CV, coefficient of variation; VO2max, maximum oxygen
uptake; BMI, body mass index, ∑ 8 skinfolds, biceps, triceps, subscapular,
supraspinale, abdominal, suprailiac, mid-thigh and medial calf.
JOURNAL OF SPORTS SCIENCES 3
20 km ·h−1. We found no relative interindividual changes
between 12 and 20 km ·h−1 at any of the analysed biomecha-
nical variables (ground contact time: CV = 7% vs. 7%; swing
time: CV = 9% vs. 9%; stride length: CV = 7% vs. 7%; and stride
frequency: CV = 6% vs. 6%).
Similarly, no correlations were found between RE and peak
vertical GRF (Figure 3(a) and Table 2, small effect), peak hor-
izontal braking force (Figure 3(b) and Table 2, trivial effect)
and vertical loading rate (Figure 3(c) and Table 2, small effect)
at a running speed of 20 or 12 km· h−1 (data not shown).
Discussion
This study documents elite athletes with an average oxygen
cost of running among the best reported in the literature
(184.4 ± 9.9 ml ·kg−1· km−1). We explored whether previously
documented factors accounting for efficient running would
also be related in this group of elite athletes. Contrary to our
hypotheses and previously reported findings (Nummela et al.,
2007; Santos-Concejero et al., 2014; Tartaruga et al., 2012), we
found no meaningful associations between gait cycle charac-
teristics or GRFs and running economy in Kenyan runners.
These results are, however, in agreement with previous
research on elite Eritrean runners, which also found no asso-
ciation between running economy and gait cycle characteris-
tics including stride length, stride frequency, swing time and
ground contact time (Santos-Concejero, Oliván, et al., 2015).
The absence of any significant correlation in both these
elite athlete cohorts may be the result of the homogeneity of
the groups. In the present study, oxygen cost of running
values ranged from 171.1 to 203.0 ml ·kg−1· km−1, with a CV
of 5.4%. Similarly, performance-related variables including
best 10 km performances (CV = 1.4%) and laboratory-mea-
sured variables such as oxygen cost of running (CV = 5.4%)
were highly homogenous in our cohort. Given that running
economy is a complex, multifactorial characteristic determined
by numerous physiological and biomechanical parameters
(Foster & Lucia, 2007), it is not surprising that no correlations
were found given the small spread of running economy in our
sample of elite athletes. Indeed, previous research that has
found significant associations between running economy and
gait characteristics did so in athletes with an oxygen cost of
transport ranging from 171.8 to 221.0 ml· kg−1· km−1
(CV = 11%), variability that is significantly greater (double)
than in the present study (Santos-Concejero et al., 2014).
This study thus differs with respect to both the performance
calibre of athletes, the narrow spread of performance ability and
the similarly narrow range in physiological and biomechanical
CV (%) variables, including running economy. The combination of the
1.4 athletic standard and homogeneity of the group in the present
1.6
7.1 study does, however, enable us to compare our findings to those
14.6 of previous studies that have utilised similar methods to offer
3.7 insight into the lack of association found here, as well as con-
11.5
6.9 tributing to a broader understanding of the running economy
5.7 determinants that may explain the exceptional performance of
19.5 African runners.
Previous research on East African runners in particular
(Saltin et al., 1995; Wilber & Pitsiladis, 2012) has shown that
Kenyan runners are more metabolically economical than elite
 4 J. SANTOS-CONCEJERO ET AL.
Downloaded by [Orta Dogu Teknik Universitesi] at 10:09 07 May 2016
Figure 1. Relationships between energy cost of running and ground contact time (a), swing time (b), stride length (c) and stride frequency (d) in Kenyan runners
(n = 15) at 12 km· h−1. 95% confidence intervals are presented.
European runners, and consequently, running economy has
been proposed as one of the factors contributing to the East
African running phenomenon (Larsen & Sheel, 2015). However,
the factors underlying their exceptional running efficiency
remain unclear. Recent research suggests that lower limb
muscle activation patterns and muscle–tendon interactions
are optimised to favour efficient storage and recoil of elastic
energy (Sano et al., 2013). Sano et al. (2013) found that Kenyan
runners reached higher hopping performance despite an over-
all smaller range of fascicle and tendinous length changes
during the contact phase, suggesting a high efficacy in the
use of the recoil of elastic energy from the tendinous struc-
tures. In addition, these researchers found that improved hop-
ping performance was accompanied by shorter stretching,
higher stretching to preactivation electromyography ratio
and, more importantly, reduced total ground contact times.
We found ground contact times of 0.24 ± 0.02 s and
0.18 ± 0.01 s at warm up (12 km· h−1) and race pace (20 km
·h−1) running speeds, respectively. These values are similar in
magnitude to those measured in a previous study on elite
Kenyan runners (0.22 ± 0.02 s at 12.5 km· h−1 and 0.17 ± 0.02 s
at 19.5 km· h−1) (Kong & de Heer, 2008). In that study, running
economy was not measured, but the authors hypothesised
that these short contact times may contribute to this known
phenomenon.
Research in non-Kenyan runners, where running economy
has been measured at similar speeds to those in the present
sample, supports this possibility. Ground contact time in elite
North African runners has been found to be related to oxygen
cost of transport when running at 13.0 and 19.5 km· h−1, with
ground contact times of 0.27 ± 0.02 s and 0.20 ± 0.01 s,
respectively (Santos-Concejero et al., 2013). These ground con-
tact times were ~10.0% greater (ES ~ 2.0; very large effect)
than in the present study at similar speeds, alongside an 8.9%
greater oxygen cost of transport (ES ~ 2.0; very large effect).
Given the homogeneity of the respective populations, the
similarity in methods and this large ES between the groups,
it is intriguing to speculate that the theory for ground contact
time as a key determinant of running economy may indeed
have merit.
The same appears to be true when studying more heteroge-
neous group of runners. For instance, previous studies on recrea-
tional endurance runners (Paavolainen, Nummela, & Rusko, 1999;
Santos-Concejero et al., 2014) have found that shorter ground
contact times are strongly associated with lower oxygen cost of
running at speeds ranging from 12 to 23 km·h−1. In these studies,
both running economy and ground contact times were ~10.0%
greater than the values observed in the present group of elite
Kenyan runners. However, it has to be acknowledged that not all
previous studies’ results support this hypothesis. Nummela et al.
 JOURNAL OF SPORTS SCIENCES 5
Downloaded by [Orta Dogu Teknik Universitesi] at 10:09 07 May 2016

Figure 2. Relationships between energy cost of running and ground contact time (a), swing time (b), stride length (c) and stride frequency (d) in Kenyan runners
(n = 15) at 20 km· h−1. 95% confidence intervals are presented.

Table 2. Gait characteristics, ground contact forces and running economy

correlations results. tested by Nummela et al. (2007) and the present cohort of
R P Kenyan runners. Similarly, Sano et al. (2015) found no ground
value value SEE
contact differences between Kenyan and Japanese elite runners
Gait characteristics at 12 km·h−1
Ground contact time (s) vs. RE (kcal ·kg−1· km−1) 0.437 0.118 0.088
at slow and medium speeds (9 and 13.9 km ·h−1). As for Nummela
Swing time (s) vs. RE (kcal· kg−1· km−1) 0.142 0.629 0.087 et al. (2007), Sano et al. (2015) did not measure running economy
Stride length (m) vs. RE (kcal· kg−1· km−1) 0.385 0.174 0.090 at either speed, and therefore, we cannot draw definitive con-
Stride frequency (Hz) vs. RE (kcal· kg−1· km−1) −0.975 0.856 0.009
clusions of what that absence of differences may mean for the
Gait characteristics at 20 km·h−1
Ground contact time (s) vs. RE (kcal ·kg−1· km−1) −0.131 0.656 0.080
ground contact and running economy interaction.
Swing time (s) vs. RE (kcal· kg−1· km−1) 0.247 0.394 0.078 Mechanistically, short ground contact times may reduce
Stride length (m) vs. RE (kcal· kg−1· km−1) 0.420 0.119 0.070 the duration of the braking phase during stance. This is usually
−1 −1
Stride frequency (Hz) vs. RE (kcal ·kg · km ) −0.186 0.508 0.076
associated with greater preactivation of the shank muscles,
Ground contact forces at 20 km·h−1
Vertical initial loading rate (BW/s) vs. RE (kcal·kg−1· 0.396 0.143 0.081
which increases the sensitivity of the muscle spindle poten-
km−1) tiating stretch reflexes to enhance musculotendon stiffness,
Peak braking force (N kg−1) vs. RE (kcal· kg−1· km−1) −0.131 0.641 0.084 thus improving running economy (Nummela et al., 2007;
Vertical ground reaction force (N kg−1) vs. RE (kcal· 0.209 0.455 0.079
kg−1· km−1)
Paavolainen et al., 1999; Santos-Concejero et al., 2014).
However, this may not be the case in Kenyan runners as
RE, running economy; SEE, standard error of estimate.
Sano et al. (2015) suggested that Kenyan runners with a longer
Achilles tendon and moment arm may produce maximal hop-
ping performance and submaximal running in a way that does
(2007) found similar ground contact times to the values reported not conform to commonly accepted theories about the
in the present study at 20 km· h−1 in well-trained young stretch-shortening cycle. In any case, given the established
European athletes. However, in that study, running economy at relationship between ground contact time and running econ-
20 km· h−1 was not measured, so we can only speculate about omy (Nummela et al., 2007; Paavolainen et al., 1999; Santos-
the possible running economy differences between the athletes Concejero et al., 2014), our running economy measurements,
 6 J. SANTOS-CONCEJERO ET AL.
Downloaded by [Orta Dogu Teknik Universitesi] at 10:09 07 May 2016
Figure 3. Relationships between energy cost of running and peak vertical ground reaction force (a), peak braking force (b) and vertical Initial loading rate (c) in
Kenyan runners (n = 15) at 20 km· h−1. 95% confidence intervals are presented.
combined with the short ground contact times in the present
study and that of Kong and de Heer (2008), support the notion
that ground contact time may contribute to the exceptional
running economy of Kenyan runners compared to runners
with inferior running ability.
Previously, Nummela et al. (2007) also suggested that
reduced GRFs may reduce the oxygen cost or running.
However, all GRF parameters investigated in this study were
not found to be associated with running economy at either 12
or 20 km ·km−1. Here too, the homogeneity of the group may
influence the ability to associate two variables where one is
multifactorial. No previous research on elite or subelite Kenyan
runners has investigated ground contact time and running
economy, making cross-study comparisons impossible. It may
also be that metabolic costs are largely affected by neuromus-
cular activation patterns that influence both segment kine-
matics and kinetics rather than a discrete outcome such as
ground contact time (Moore & Dixon, 2014).
Similarly, we found no correlation between swing time and
running economy (Figure 2(b)). Santos-Concejero, Oliván, et al.
(2015) reported a significant association between swing time
and the oxygen cost of running in elite Eritrean runners at race
pace (21 km· h−1), which was attributed to the fact that a
majority of muscles are active during the swing phase, most
particularly for the acceleration of the limbs and preparation
for ground contact (Chumanov, Heiderscheit, & Thelen, 2011).
Swing time has also been proposed as an effective discrimi-
nator of efficient gait patterns (Santos-Concejero et al., 2014).
However, the absence of significant associations between run-
ning economy and swing time in these Kenyan runners,
together with the similar swing times found in this study
and others (Santos-Concejero et al., 2013, 2014), suggests
that efficient energy cost of running in Kenyans may not be
linked to this parameter.
Previous researchers have also found significant associa-
tions between stride length and frequency with running econ-
omy (Kyröläinen et al., 2001; Santos-Concejero et al., 2014).
However, this relationship was not found in this homogenous
group of elite Kenyan runners (Figures 1(c,d) and 2(c,d)).
Traditionally, it has been reported that relatively efficient run-
ning patterns, such as longer stride lengths, may be a result of
either a biomechanical adaptation during training or a phy-
siological adaptation for energy optimisation (Cavanagh &
Williams, 1982; Nummela et al., 2007; Tartaruga et al., 2012).
These adaptations may result in a negative relationship
between stride frequency and running economy and a posi-
tive correlation between stride length and running economy
(Tartaruga et al., 2012).

In fact, it has been shown that elite distance runners
possess an optimal balance between stride length and
frequency when compared to good distance runners,
which may produce a better running economy and run-
ning performance (Anderson, 1996). However, our finding
is that neither stride length and frequency are associated
to running economy in elite Kenyan runners, nor do these
stride parameters appear to be responsible for the excep-
tional running economy of Kenyan runners as the present
group ran with the very similar stride lengths and frequen-
cies as those documented in lesser performing runners
with inferior running economy values (Santos-Concejero
et al., 2013, 2014). Similarly, it appears that stride length
and frequency may not be sensitive enough to discrimi-
nate between athletes when studying homogenous
groups, as demonstrated in this and previous studies
(Santos-Concejero, Oliván et al., 2015).
Downloaded by [Orta Dogu Teknik Universitesi] at 10:09 07 May 2016
In any case, some limitations have to be acknowledged.
These include a relatively small (n = 15) and very specific
sample, which makes generalisation of the obtained results
to all runners difficult. In addition, the biomechanical vari-
ables were assessed during overground running, whereas
running economy was measured on a separate occasion on
a treadmill. It is known that the lack of air resistance results in
a lower energy cost compared with running outdoors at the
same speed and some authors argue that a 1% treadmill
grade may be too high to reproduce similar efforts when
running on an outdoor track (Mooses, Tippi, Mooses,
Durussel, & Mäestu, 2015). However, it has also been
reported that a 1% treadmill grade, which was used in this
study, accurately reflects the oxygen cost of running out-
doors (Jones & Doust, 1996). Finally, the 5% speed variance
range accepted between each of the runs during the biome-
chanics assessments may have affected the ground contact
and gait cycle measurements, which may influence the asso-
ciations between RE and stride characteristics calculated in
this study. Therefore, our results should be interpreted with
caution.
Future research analysing neuromuscular activation profiles
in conjunction with three-dimensional biomechanical analyses
during overground running and ground contact force mea-
surements in East African runners is thus warranted. This may
provide valuable information for coaches and applied scien-
tists for the identification and ongoing management of elite
running training programmes.
In summary, this study aimed to characterise the factors
associated with running economy in elite Kenyan runners. Our
results suggest that gait cycle characteristics such as ground
contact time, swing time, stride length, stride frequency and
GRFs are not significantly associated with economical running
patterns when studying homogenous groups of Kenyan run-
ners. However, our unique population provides evidence to
hypothesise that the short ground contact times may contribute
to the exceptional running economy of Kenyan runners since we
found ground contact times that are considerably different
(~10%) from those in the published literature for non-Kenyan
runners with increased oxygen cost of transport and inferior
performances. While we cannot make direct comparisons since
no performance-matched control group for elite Kenyan runners
JOURNAL OF SPORTS SCIENCES 7
exists, this study does suggest possible explanations for the
exceptional running economy of elite Kenyan runners.
Acknowledgements
This study was supported by a National Research Foundation grant (ref.
CPR20110704000020012) and by a Faculty of Health Sciences Postdoctoral
Research Fellowship (University of Cape Town) to JSC. The authors would
like to thank Lara Grobler (University of Stellenbosch) for assisting with
running economy measurements.
Disclosure statement
No potential conflict of interest was reported by the authors.
Funding
This study was supported by a National Research Foundation: [grant (ref.
CPR20110704000020012)] and by a Faculty of Health Sciences
Postdoctoral Research Fellowship (University of Cape Town) to JSC.
References
Anderson, T. (1996). Biomechanics and running economy. Sports Medicine,
22, 76–89. doi:10.2165/00007256-199622020-00003
Bosco, C., Montanari, G., Ribacchi, R., Giovenali, P., Latteri, F., Iachelli, G., . . .
Saibene, F. (1987). Relationship between the efficiency of muscular
work during jumping and the energetics of running. European Journal
of Applied Physiology and Occupational Physiology, 56, 138–143.
doi:10.1007/BF00640636
Bransford, D. R., & Howley, E. T. (1977). Oxygen cost of running in trained
and untrained men and women. Medicine & Science in Sports, 9, 41–44.
Cavanagh, P. R., & Williams, K. R. (1982). The effect of stride length
variation on oxygen uptake during distance running. Medicine &
Science in Sports & Exercise, 14, 30–35. doi:10.1249/00005768-
198201000-00006
Chumanov, E. S., Heiderscheit, B. C., & Thelen, D. G. (2011). Hamstring
musculotendon dynamics during stance and swing phases of high-
speed running. Medicine & Science in Sports & Exercise, 43, 525–532.
doi:10.1249/MSS.0b013e3181f23fe8
Cohen, J. (1988). Statistical power analysis for the behavioural sciences.
Mahwah, NJ: Erlbaum.
Fletcher, J. R., Esau, S. P., & Macintosh, B. R. (2009). Economy of running:
Beyond the measurement of oxygen uptake. Journal of Applied
Physiology, 107, 1918–1922. doi:10.1152/japplphysiol.00307.2009
Foster, C., & Lucia, A. (2007). Running economy: The forgotten factor in
elite performance. Sports Medicine, 37, 316–319. doi:10.2165/00007256-
200737040-00011
Hopkins, W. G., Marshall, S. W., Batterham, A. M., & Hanin, J. (2009).
Progressive statistics for studies in sports medicine and exercise
science. Medicine & Science in Sports & Exercise, 41, 3–13. doi:10.1249/
MSS.0b013e31818cb278
Jones, A. M., & Doust, J. H. (1996). A 1% treadmill grade most accurately
reflects the energetic cost of outdoor running. Journal of Sports
Sciences, 14, 321–327. doi:10.1080/02640419608727717
Kong, P. W., & de Heer, H. (2008). Anthropometric, gait and strength
characteristics of Kenyan distance runners. Journal of Sports Science &
Medicine, 7, 499–504.
Krahenbuhl, G. S., & Pangrazi, R. P. (1983). Characteristics associated with
running performance in young boys. Medicine & Science in Sports &
Exercise, 15, 486–490. doi:10.1249/00005768-198315060-00008
Kyröläinen, H., Belli, A., & Komi, P. V. (2001). Biomechanical factors affect-
ing running economy. Medicine & Science in Sports & Exercise, 33, 1330–
1337. doi:10.1097/00005768-200108000-00014
Larsen, H. B. (2003). Kenyan dominance in distance running. Comparative
Biochemistry and Physiology Part A: Molecular & Integrative Physiology,
136, 161–170. doi:10.1016/S1095-6433(03)00227-7
 8 J. SANTOS-CONCEJERO ET AL.
Larsen, H. B., & Sheel, A. W. (2015). The Kenyan runners. Scandinavian Journal
of Medicine & Science in Sports, 25, 110–118. doi:10.1111/sms.12573
Lucia, A., Esteve-Lanao, J., Oliván, J., Gómez-Gallego, F., San Juan, A. F.,
Santiago, C., . . . Foster, C. (2006). Physiological characteristics of the
best Eritrean runners-exceptional running economy. Applied Physiology
Nutrition & Metabolism, 31, 530–540. doi:10.1139/h06-029
Moore, I. S., & Dixon, S. J. (2014). Changes in sagittal plane kinematics with
treadmill familiarization to barefoot running. Journal of Applied
Biomechanics, 30, 626–631. doi:10.1123/JAB.2013-0239
Mooses, M., Mooses, K., Haile, D. W., Durussel, J., Kaasik, P., & Pitsiladis, Y. P.
(2015). Dissociation between running economy and running perfor-
mance in elite Kenyan distance runners. Journal of Sports Sciences, 33,
136–144. doi:10.1080/02640414.2014.926384
Mooses, M., Tippi, B., Mooses, K., Durussel, J., & Mäestu, J. (2015). Better
economy in field running than on the treadmill: Evidence from high-
level distance runners. Biology of Sport, 32, 155–159. doi:10.5604/
20831862.1144418
Nummela, A., Keränen, T., & Mikkelsson, L. O. (2007). Factors related to top
running speed and economy. International Journal of Sports Medicine,
28, 655–661. doi:10.1055/s-2007-964896
Paavolainen, L. M., Nummela, A. T., & Rusko, H. K. (1999). Neuromuscular
Downloaded by [Orta Dogu Teknik Universitesi] at 10:09 07 May 2016
characteristics and muscle power as determinants of 5-km running
performance. Medicine & Science in Sports & Exercise, 31, 124–130.
doi:10.1097/00005768-199901000-00020
Saltin, B., Larsen, H., Terrados, N., Bangsbo, J., Bak, T., Kim, C. K., . . . Rolf, C.
J. (1995). Aerobic exercise capacity at sea level and at altitude in
Kenyan boys, junior and senior runners compared with Scandinavian
runners. Scandinavian Journal of Medicine & Science in Sports, 5, 209–
221. doi:10.1111/j.1600-0838.1995.tb00037.x
Sano, K., Ishikawa, M., Nobue, A., Danno, Y., Akiyama, M., Oda, T., . . . Komi,
P. V. (2013). Muscle-tendon interaction and EMG profiles of world class
endurance runners during hopping. European Journal of Applied
Physiology, 113, 1395–1403. doi:10.1007/s00421-012-2559-6
Sano, K., Nicol, C., Akiyama, M., Kunimasa, Y., Oda, T., Ito, A., . . . Ishikawa, M.
(2015). Can measures of muscle-tendon interaction improve our
understanding of the superiority of Kenyan endurance runners?
European Journal of Applied Physiology, 115, 849–859. doi:10.1007/
s00421-014-3067-7
Santos-Concejero, J., Billaut, F., Grobler, L., Oliván, J., Noakes, T. D., &
Tucker, R. (2015). Maintained cerebral oxygenation during maximal
self-paced exercise in elite Kenyan runners. Journal of Applied
Physiology, 118, 156–162. doi:10.1152/japplphysiol.00909.2014
Santos-Concejero, J., Granados, C., Irazusta, J., Bidaurrazaga-Letona, I.,
Zabala-Lili, J., & Gil, S. M. (2014). Stride angle as a novel indicator of
running economy in well-trained runners. Journal of Strength &
Conditioning Research, 28, 1889–1895. doi:10.1519/
JSC.0000000000000325
Santos-Concejero, J., Granados, C., Irazusta, J., Bidaurrazaga-Letona, I.,
Zabala-Lili, J., Tam, N., & Gil, S. M. (2013). Differences in ground contact
time explain the less efficient running economy in North African run-
ners. Biology of Sport, 30, 181–187. doi:10.5604/20831862.1059170
Santos-Concejero, J., Oliván, J., Maté-Muñoz, J. L., Muniesa, C., Montil, M.,
Tucker, R., & Lucia, A. (2015). Gait-cycle characteristics and running
economy in elite Eritrean and European runners. International Journal
of Sports Physiology and Performance, 10, 381–387. doi:10.1123/
ijspp.2014-0179
Tam, E., Rossi, H., Moia, C., Berardelli, C., Rosa, G., Capelli, C., & Ferretti, G.
(2012). Energetics of running in top-level marathon runners from
Kenya. European Journal of Applied Physiology, 112, 3797–3806.
doi:10.1007/s00421-012-2357-1
Tartaruga, M. P., Brisswalter, J., Peyré-Tartaruga, L. A., Avila, A. O., Alberton,
C. L., Coertjens, M., . . . Kruel, L. F. (2012). The relationship between
running economy and biomechanical variables in distance runners.
Research Quarterly for Exercise and Sport, 83, 367–375. doi:10.1080/
02701367.2012.10599870
Wilber, R. L., & Pitsiladis, Y. P. (2012). Kenyan and Ethiopian distance
runners: What makes them so good? International Journal of Sports
Physiology & Performance, 7, 92–102.
Yuhasz, M. S. (1974). Physical fitness manual. London, Ontario: University of
Western Ontario.