© 2015 John Wiley & Sons A/S.

Scand J Med Sci Sports 2015: ••: ••–•• doi: 10.1111/sms.12455

Published by John Wiley & Sons Ltd

Achilles tendon loading patterns during barefoot walking and slow running on a treadmill: An ultrasonic propagation study M. Wulf1, S. C. Wearing2,3, S. L. Hooper4, J. E. Smeathers2, T. Horstmann1,5, T. Brauner1 Faculty of Sports and Health Sciences, Technische Universität München, Munich, Germany, 2Institute of Health and Biomedical Innovation, Queensland University of Technology, Brisbane, Australia, 3Centre of Excellence for Applied Sport Science Research, Queensland Academy of Sport, Brisbane, Australia, 4Office of Health & Medical Research, Queensland Health, Brisbane, Australia, 5 MEDICAL PARK Bad Wiessee St. Hubertus, Bad Wiessee, Australia Corresponding author: Scott C. Wearing, PhD, Institute of Health and Biomedical Innovation, Queensland University of Technology, 60 Musk Avenue, Kelvin Grove, Qld 4059, Australia. Tel: +61 7 3138 6444, Fax: +61 7 3138 60330, E-mail: [email protected]

1

Accepted for publication 22 February 2015

Measurement of tendon loading patterns during gait is important for understanding the pathogenesis of tendon “overuse” injury. Given that the speed of propagation of ultrasound in tendon is proportional to the applied load, this study used a noninvasive ultrasonic transmission technique to measure axial ultrasonic velocity in the right Achilles tendon of 27 healthy adults (11 females and 16 males; age, 26 ± 9 years; height, 1.73 ± 0.07 m; weight, 70.6 ± 21.2 kg), walking at self-selected speed (1.1 ± 0.1 m/s), and running at fixed slow speed (2 m/s) on a treadmill. Synchronous measures of ankle kinematics, spatiotemporal gait parameters, and vertical ground reaction forces were simultaneously measured. Slow

running was associated with significantly higher cadence, shorter step length, but greater range of ankle movement, higher magnitude and rate of vertical ground reaction force, and higher ultrasonic velocity in the tendon than walking (P < 0.05). Ultrasonic velocity in the Achilles tendon was highly reproducible during walking and slow running (mean within-subject coefficient of variation < 2%). Ultrasonic maxima (P1, P2) and minima (M1, M2) were significantly higher and occurred earlier in the gait cycle (P1, M1, and M2) during running than walking (P < 0.05). Slow running was associated with higher and earlier peaks in loading of the Achilles tendon than walking.

Tendon injury is a common and often debilitating condition, which typically requires lengthy and difficult clinical management plans (Maffulli et al., 2004). Although the etiology of Achilles tendinopathy is likely multifactorial in nature, tendon injury is thought to reflect a failure of the tendon to adapt to the prevailing loading conditions (Maffulli et al., 2004). Measurement of the loading patterns of tendons during dynamic activities, such as running and jumping, therefore, is key to understanding the pathogenesis of tendon injury (Pourcelot et al., 2005b). Direct measurement of tensile force in tendons during dynamic activities was first made possible with introduction of the buckle-type force transducer (Salmons, 1969). Although not without technical limitations (Bey & Derwin, 2012), various direct measurement methods have since been applied to the measurement of tendon loads in both animal and human experiments (Gregor et al., 1988; Komi, 1990; Finni et al., 1998). Important insights into the function and adaptation of the muscle– tendon unit to variable conditions (Biewener & Gillis, 1999), load sharing between muscles (Herzog et al., 1994), elastic energy storage (Gregor et al., 1988), and neural control of locomotion (Walmsley et al., 1978)

have been made by studies using direct recordings. Direct measurement of muscle–tendon loading, however, is highly invasive and its application is not possible in many human experiments or practical in studies involving large cohorts. Consequently, tendon forces have been more commonly estimated from indirect techniques, including inverse dynamics, and Electromyography (EMG)-to-force processing (Prilutsky & Zatsiorsky, 1994; Reinschmidt & Nigg, 1995). Inverse dynamic models, however, have been shown to overestimate tendon loads by as much as 50% compared with direct measures of tendon load (Gregor et al., 1991). Hence there is a need for a more direct but noninvasive technique for monitoring tendon loading. Although ultrasound imaging has been widely used for diagnosis of tendon disorders since it was first introduced in clinical practice in the 1970s, quantitative ultrasound techniques have only emerged, relatively recently, as a potentially useful noninvasive measurement tool for characterizing the mechanical properties of tendon in vivo. The most common technique involves manual tracking of the muscle–tendon junction using cine B-Mode images (Maganaris & Paul, 2002); however, several semi-automated methods for tracking motion of

1

Wulf et al. the midsubstance of soft tissues have more recently been developed (Gennisson et al., 2013; Ooi et al., 2014). Collectively termed elastography, these various approaches provide semi-quantitative or quantitative indices of human tendon stiffness and are all, with the exception of shear wave elastography, predicated on an assumed speed of sound in the target tissue (Gennisson et al., 2013). The speed of transmission of ultrasound waves (S) is known to be dependent on the bulk modulus (E) and density (ρ) of the material through which it propagates, and in tendon, is governed by the classic Newtonian–Laplace equation with some adjustment for Poisson’s effects (ν) in elastic media (Vergari et al., 2012a):

S=

E (1 − ν ) ρ (1 + ν ) (1 − 2ν )

Studies in animal tendon have confirmed that the transmission speed of axial ultrasonic waves is also proportional to the applied load in tendon and can be effectively modeled as an exponential function (Hoffmeister et al., 1994; Miles et al., 1996; Kuo et al., 2001; Pourcelot et al., 2005a; Crevier-Denoix et al., 2009; Vergari et al., 2012b). Thus, while the variation in the transmission speed of ultrasound with the application of mechanical load to tendon represents a limitation of most elastographic approaches (Ooi et al., 2014), ultrasound transmission techniques take advantage of this relationship, to afford a direct noninvasive method of quantifying the mechanical loading of human tendon. To date, the ultrasound transmission technique has also been applied to the noninvasive measurement of human Achilles tendon in vivo during locomotion (Pourcelot et al., 2005b) and has been shown to be sensitive in detecting changes in tendon loading with footwear (Wearing et al., 2014). Thus far, application of the technique, however, has been limited to measurement of tendon loads during walking at relatively slow gait speeds. While the technique has the potential to noninvasively evaluate loading of superficial human tendons under dynamic conditions in near-real time, application of ultrasound transmission techniques to noninvasively evaluate loading of the human Achilles tendon has not been investigated during running. It is particularly important to evaluate Achilles tendon loading during running, since Achilles tendon injury often occurs in running-related sports (Leppilahti et al., 1996). Therefore, the aim of this study was to determine the ultrasonic velocity in the Achilles tendon during barefoot walking and slow running on a treadmill using an ultrasound transmission technique. Since ultrasonic velocity is proportional to the tensile load in tendon (Crevier-Denoix et al., 2009), and running has been previously shown to increase ground reaction force, ankle

2

kinematics and muscle activity (Cappellini et al., 2006; Arnold et al., 2013), it was hypothesized that axial ultrasonic velocity in the Achilles tendon would be higher during running than walking. Methods Participants Twenty-seven healthy adults (11 females, 16 males) were recruited from university faculty over a 1-month period to participate in the study. The mean (± SD) age, height, and body mass of participants were 26 ± 9 years, 1.73 ± 0.07 m, and 70.6 ± 21.2 kg, respectively. No participant reported a medical history of diabetes, inflammatory arthritis, familial hypercholesterolemia or Achilles tendon injury or pathology. Participant numbers were based on previously published data for human Achilles tendon (Wearing et al., 2014) and were sufficient to detect a 3.5% difference in the peak velocity in the tendon (α = 0.05, β = 0.20). The study received approval from the Institutional Ethics Committee and all participants gave written informed consent.

Equipment A flexible twin-axis strain-gauge electrogoniometer (Penny and Giles, Biometrics, Gwent, UK) was used to measure sagittal plane ankle movement during walking and running. The end-blocks of the device were fixed to the skin overlying the medial calcaneus and the distal aspect of the medial tibia of the right ankle using double-sided adhesive tape and further secured with surgical tape. Once secured, the goniometer was then zeroed during erect bipedal stance. The electrogoniometer has a resolution of 1°, and has been shown to be accurate to within 1.5% over a measurement range of 100° (Rowe et al., 2001). Vertical ground reaction force and spatiotemporal gait parameters were determined during gait using an instrumented treadmill system (FDM-THM-S, Zebris Medical GmbH, Isny, Germany). The system is comprised of a capacitance-based pressure platform housed within a motor-driven treadmill. The pressure platform had a sensing area of 108.4 × 47.4 cm and incorporated 7168 sensors, each approximately 0.85 × 0.85 cm. The treadmill had a contact surface of 150 × 50 cm and its belt speed could be adjusted between 0.2 and 22 km/h, at intervals of 0.1 km/h. The coefficients of variation for spatiotemporal gait parameters measured during walking are typically < 10% for repeated measures (Reed et al., 2013). Axial ultrasonic velocity was measured in the right Achilles tendon using a custom-built ultrasonic device and a five-element ultrasound probe. The probe consisted of a 1 MHz broadband pulse emitter and four regularly spaced receivers (range, 33.0– 40.5 mm). The probe was positioned over the midline of the posterior aspect of the Achilles tendon, with the emitter located 1 cm above the calcaneus and maintained in close contact with the skin by means of coupling medium and elasticized straps (Fig. 1). Received ultrasonic signals were digitized at 20 MHz and the time of flight of the first arriving transient between receivers was estimated using the first zero crossing criterion (Bossy et al., 2002). Average ultrasonic velocity was subsequently calculated given the known distance between receivers and the measured time of flight. Measurement precision for ultrasonic velocity is better than 3 m/s and the error in predicting applied tensile force in tendon from direct measures of ultrasonic velocity is typically < 2% (Crevier-Denoix et al., 2009).

Protocol Participants reported to the gait laboratory wearing lightweight, comfortable clothing and having abstained from vigorous physical

Achilles tendon loading in walking and running

Fig. 1. Illustration of the placement of the ankle goniometer (G) and ultrasonic probe (P) over the midline of the right Achilles tendon. Note: elasticized strapping and surgical tape have been omitted for illustrative purposes. activity on the day of testing. Prior to the fixation of the ultrasound probe and the electrogoniometer to the right leg, the skin overlying the posterior Achilles tendon and medial aspect of the right shank was prepared and cleaned using standard alcohol abrading methods (Cram & Rommen, 1989). Participants were then afforded a treadmill acclimatization session that included a 6-min period of steady-state walking at a self-selected “comfortable” speed (Lavcanska et al., 2005). Each participant’s “comfortable” walking speed was determined at the outset of the acclimatization session using a previously outlined method in which the participant was blinded to the treadmill speed (Reed et al., 2013). In brief, “comfortable” walking speed was determined by gradually increasing the belt speed until the participant indicated that they had reached their natural pace. The treadmill speed was then increased by 1.0 km/h beyond their natural pace and, blinded to the participant, reduced in 0.1 km/h steps until they perceived their preferred “comfortable” walking pace had been reestablished (Reed et al., 2013). This speed was subsequently recorded as the participant’s preferred walking speed. Following acclimatization, barefoot walking was performed at each participant’s self-selected comfortable speed, while barefoot running was conducted at a standard speed of 2.0 m/s. Following 6 min of steady-state gait at each speed, ultrasonic velocity in the right Achilles tendon, sagittal plane ankle movement, vertical ground reaction force, and basic spatiotemporal gait data were synchronously sampled for 60 s at a rate of 120 Hz. Approximately 50 ± 1 and 80 ± 1 gait cycles were analyzed during each walking and running trial, respectively.

Data reduction and statistical analysis Force, time, and velocity data were exported from the treadmill system in ASCII format and custom computer code (Matlab R2012a, MathWorks, Natick, Massachusetts, USA) was subsequently used to determine spatiotemporal parameters, including cadence, stride duration, step length, stance, and swing phase durations. The magnitude and timing of peak vertical ground reaction forces during walking and running were also calculated (Reed et al., 2013; Wearing et al., 2013). It should be noted that walking was characterized by two ground reaction force peaks (F1, F2), whereas slow running showed one dominant peak (F2). Since the peak ground reaction force in slow running (F2) is more closely aligned to the midstance and push off phase of the gait cycle, it was compared with F2 in walking in subsequent statistical analyses. The peak loading rate (PLR), defined as the peak instantaneous

force differential during the stance phase of gait, was also calculated. Maximum dorsiflexion (DF1, DF2) and plantarflexion (PF1, PF2) of the ankle joint during each gait cycle were defined and identified according to the recommendation of the International Society of Biomechanics, in which dorsiflexion was represented by positive values (Wu et al., 2002). Similarly, local maxima (P1, P2) and minima (M1, M2) in ultrasonic velocity were analyzed over the gait cycle (Fig. 2). Values were derived for each gait cycle recorded within the 60-s data capture period and the mean values for all gait cycles were calculated and used in subsequent statistical analysis. The Statistical Package for the Social Sciences (SPSS Inc, Chicago, Illinois, USA) was used for all statistical procedures. Kolmogorov–Smirnov tests were used to evaluate data for underlying assumptions of normality. Outcome variables were determined to be normally distributed and therefore means and standard deviations have been used as summary statistics. Paired t-tests were used to evaluate potential differences in vertical ground reaction forces and ultrasonic, spatiotemporal, and ankle joint parameters. Effect sizes were estimated using Cohen’s d for repeated measures (Dz), in which the difference between walking and running means were standardized by the deviation of the difference (Cohen, 1988). As a general guideline, Dz in the range of 0.20–0.50 was considered to be a small effect, 0.50–0.80 a medium effect, and a value of more than 0.80 as a large effect. (Cohen, 1988) An alpha level of 0.05 was used for all tests of significance.

Results All participants adopted a rearfoot strike pattern during walking and slow running as determined from visual inspection of pressure–time data from the treadmill system. Compared with walking, running was associated with a significantly shorter step length (t26 = 30.8, P < 0.05), stance phase (t26 = 25.8, P < 0.05), and swing phase duration (F26 = 25.8, P < 0.05), but significantly higher peak ground reaction force (t26 = 17.9, P < 0.05) and PLR (t26 = 12.0, P < 0.05; Table 1). The range of ankle motion during stance increased by 6° during slow running (t26 = 9.1, P < 0.05) compared with walking (Table 2), with significant increases in both maximum plantarflexion (PF2; t26 = 9.4, P < 0.05) and maximum dorsiflexion (DF1, t26 = 10.3, P < 0.05; DF2, t26 = 9.8, P < 0.05). In both walking and slow running, the overall pattern of ultrasonic velocity in the Achilles tendon was highly reproducible from one gait cycle to the next. The mean within-subject coefficient of variation for ultrasonic maxima and minima (P1, P2, M1, M2) ranged between 0.2% and 1.7% during walking and between 0.5% and 1.6% during slow running. Ultrasonic velocity in the Achilles tendon was characterized by two maxima and minima during walking and slow running (Fig. 2). At both gait speeds, ultrasonic velocity decreased immediately after foot strike (M1) and then gradually increased during early stance phase to peak (P1) at ≈ 50% and ≈ 27% of the gait cycle in walking and running, respectively. Ultrasonic values then decreased to a second minimum (M2) during the early swing phase in walking and running (70% and 60% of the gait cycle,

3

Wulf et al. Table 1. Mean (SD) spatiotemporal and kinetic parameters during walking and running

n Velocity (m/s) Cadence (strides/min) Stride duration (s) Step length (m) Stance phase duration (%) Swing phase duration (%) F1 (N) Time of F1 (%) F2 (N) Time of F2 (%) PLR (N/s)

Walk

Run

Dz

27 1.06 (0.14) 54 (4) 1.11 (0.08) 0.74 (0.06) 64 (2) 36 (2) 770 (214) 17 (2) 774 (204) 47 (2) 170 (55)

27 1.99* (0.05) 81* (3) 0.75* (0.03) 0.66* (0.08) 44* (4) 56* (4) – – 1445* (350) 17* (2) 326* (89)

−6.04 −6.42 4.64 0.96 4.98 −4.98 – – −3.50 11.97 −2.35

*Indicates a statistically significant main effect for gait speed (P < 0.05). %, percentage of gait cycle; Dz, Cohen’s d for repeated measures; F1 and F2, ground reaction force peak at time points 1 and 2; PLR, peak loading rate; SD, standard deviation. Table 2. Mean (SD) ankle joint motion during walking and running

Walk n Stance Range of motion (°) PF1 (°) Time of PF1 (%) DF1 (°) Time of DF1 (%) Swing Range of motion (°) PF2 (°) Time of PF2 (%) DF2 (°) Time of DF2 (%)

Run

Dz

27

27

12 (4) −5 (2) 9 (3) 6 (2) 47 (4)

18* (4) −6 (4) 9 (7) 10* (3) 29* (9)

−1.76 0.19 −0.02 −1.98 1.88

8 (3) −7 (3) 67 (2) 1 (1) 89 (4)

16* (5) −13* (4) 57* (7) 4* (2) 92* (5)

−2.26 1.81 1.33 −1.88 −0.56

*Indicates a statistically significant main effect for gait speed (P < .05). %, percentage of gait cycle; DF, dorsiflexion; Dz, Cohen’s d for repeated measures; PF, plantarflexion; SD, standard deviation.

Fig. 2. Typical ensemble histories taken from one participant for vertical ground reaction force (a), sagittal plane ankle movement (b), and ultrasonic velocity in the right Achilles tendon (c) during barefoot walking (light gray line) and slow running (solid dark line). Ensembles represent the average of 50 ± 1 and 80 ± 1 gait cycles for walking and running, respectively. Note that vertical ground reaction force during walking was characterized by two maxima (F1, F2), whereas only one maxima occurred (F2) during slow running. Ultrasound velocity in the Achilles tendon, in contrast, was characterized by two maxima (P1, P2) and two minima (M1, M2) during each gait cycle of both walking and slow running and closely mirrored dorsiflexion (DF1, DF2) and plantarflexion (PF1, PF2) of the ankle joint. Dashed lines represent standard deviations.

4

respectively), prior to peaking for a second time (P2) late in swing (95% of the gait cycle in both walking and running). Topic maxima (P1, t26 = −2.9, P < 0.05 and P2, t26 = −3.0, P < 0.05) and minima (M1, t26 = −6.7, P < 0.05 and M2, t26 = −2.2, P < 0.05) in ultrasonic velocity were significantly higher in running (1–8%) than walking (P < 0.05). P1 (t26 = 14.6, P < 0.05), M1 (t26 = 7.7, P < 0.05) and M2 (t26 = 8.9, P < 0.05) also occurred significantly earlier in the gait cycle during running. The effect size for the analysis of the difference in stance phase minimum (M1, Dz = −1.28) and maximum (P1, Dz = −0.96) was found to exceed Cohen’s convention for a large effect (Dz = 0.80), while that for the swing phase minimum (M2, Dz = −0.41) and maximum (P2, Dz = −0.57) was consistent with a small to moderate effect, respectively (Table 3). Discussion This study employed a noninvasive ultrasonic transmission technique to investigate the dynamic pattern of

Achilles tendon loading in walking and running Table 3. Mean (SD) ultrasonic parameters for the Achilles tendon during walking and running

Walk n Stance Range (m/s) M1 (m/s) Time of M1 (%) P1 (m/s) Time of P1 (%) Swing Range (m/s) M2 (m/s) Time of M2 (%) P2 (m/s) Time of P2 (%)

27

Run

Dz

27

287 (163) 1998 (209) 9 (2) 2285 (113) 50 (5)

165* (96) 2146* (163) 5* (3) 2311* (123) 27* (8)

1.04 −1.28 1.49 −0.96 2.82

288 (220) 1870 (229) 70 (5) 2157 (125) 95 (4)

240 (160) 1950* (202) 60* (8) 2190* (120) 95 (5)

0.24 −0.41 1.70 −0.57 0.00

*Indicates a statistically significant main effect for gait speed (P < .05). %, percentage of gait cycle; Dz, Cohen’s d for repeated measures; M, minima; P, maxima; SD, standard deviation.

loading of the Achilles tendon during walking and slow running. Consistent with our hypothesis, peak ultrasonic velocities in the Achilles tendon were significantly higher during running than walking. Slow running was also associated with an advancing temporal shift in the ultrasonic profile compared with walking, reflecting an earlier pattern of loading in the tendon during running. The ultrasonic velocity profile in the current study was biphasic during treadmill walking and slow running, with a temporal pattern similar to the force profile previously reported with direct measures of Achilles tendon during overground walking and running (Komi, 1990; Finni et al., 1998). All participants in the current study adopted a rearfoot strike pattern during both walking and slow running, and ultrasonic velocity in the Achilles tendon initially decreased following foot strike, reaching its first minimum at around 5–10% of the gait cycle (Fig. 2). Ultrasonic velocity then increased in the Achilles tendon to peak late in the stance phase (50% and ≈ 27% of the walking and running gait cycle, respectively) before decreasing to its second minimum during the early swing phase of gait (70% and 60% of the walking and running gait cycle, respectively). While the present experimental setup did not allow for a mechanistic explanation for the increase in ultrasonic velocity of the Achilles tendon observed during slow running, Finni et al. (2001) proposed that internal loading of a muscle–tendon unit was primarily dependent on both the amplitude of movement of the joint and contraction intensity of the corresponding muscle. In the current study, ultrasonic velocity in the Achilles tendon was closely coupled with ankle joint movement and slow running was associated with almost double the range of ankle movement when compared with walking. Although the range of ankle motion in the current study was less than that reported by some studies (Arnold et al., 2013), the magnitude and overall pattern of ankle movement is comparable with that reported during bare-

foot walking and running at comparable speeds on a treadmill (Tulchin et al., 2010; Zhang et al., 2013). In walking, both ankle dorsiflexion (DF1) and ultrasonic velocity (P1) in the Achilles tendon peaked at about the same time as F2 (≈ 47% of the gait cycle), whereas in slow running, their peaks occurred ≈ 10% later than the corresponding F2 (17% of the gait cycle). These findings are generally consistent with those of Arnold et al. (2013), in which peak activation of the triceps surae was noted to coincide with peak ankle dorsiflexion in both walking and running. In the current study, ankle plantarflexion was coupled with a decrease in ultrasonic velocity suggesting ankle plantarflexion was associated with reduced loading in the Achilles tendon in both walking and running. While only ankle joint kinematics was measured in the current study, it is important to note that the gastrocnemius muscle is bi-articular, crossing both the ankle and knee joints. Orishimo et al. (2008) noted that force in the Achilles tendon was reduced by 45% when the knee was flexed to 50° and Arnold et al. (2013) showed that approximately 60° of knee flexion occurred during the terminal stance phase of walking (1.0 m/s) and 80° during running (2.0 m/s) at almost identical speeds to those of the current study. Previous research has also shown that coordinated flexion of the knee and ankle joints during walking is largely unaffected by changes in gait speed (Chiu & Chou, 2012). Thus reduced ultrasonic velocity in the Achilles tendon associated with ankle plantarflexion in the current study may reflect the coordinated flexion of the ankle and knee joints. While higher ultrasonic velocity in the Achilles tendon may reflect the greater range of ankle joint movement noted during slow running, the higher velocity would also be consistent with a greater contraction intensity of the triceps surae (Finni et al., 2001). Interestingly, Arnold et al. (2013), in evaluating the effect of gait speed on EMG activity of 11 lower limb muscles, noted that only the soleus and the gastrocnemius showed earlier activation in running compared with walking. A shift to an earlier pattern of muscle activation when moving from walking to running has also been reported previously by others (Cappellini et al., 2006). In the present study, there was a similar temporal shift in the pattern of ultrasonic velocity in the Achilles tendon during running, suggesting that peak loads in Achilles tendon occurred earlier in slow running than walking. The current study has a number of limitations. Firstly, variations in ultrasonic velocity during walking and running in the current study were assumed to be proportional to the elastic modulus of the tendon, which in turn is load-dependent (Miles et al., 1996; Pourcelot et al., 2005a). Propagation of ultrasound waves in tendon, however, is also influenced by other factors including the temperature, hydration, and density of the tissue (Miles et al., 1996). Given the duration of recording, tendon hydration and temperature may reasonably be assumed

5

Wulf et al. to be stable during the barefoot gait conditions tested in the current study; however, there is evidence, albeit in animal models, that the composition of the Achilles tendon may vary along its length (Curwin et al., 1994). Likewise, although the ultrasound probe was firmly fixed to the skin via an adhesive coupling medium and elasticated straps, relative movement, with respect to the tendon, undoubtedly occurred with elongation of the tendon during loading. Similarly, it should be noted that the Achilles tendon is a conjoined tendon of the soleus and gastrocnemius muscles and the ultrasonic measurement technique used in this study cannot separate between these muscles. Thus, it is unknown whether tendon loading observed in the current study is applicable to the entire tendon structure. Further, this study evaluated the axial transmission of ultrasound waves in the Achilles tendon of healthy, young adults during barefoot locomotion on a treadmill, in which a rearfoot strike pattern was adopted at preferred walking and a standardized (∼ 2.0 m/s) slow running speed. The running speed employed in the current study was standardized across participants and was close to the typical walk–run transition speed reported for overground gait (McNeill Alexander, 2002). Both self-selected and walk–run transition speeds are known to be lower in treadmill gait compared with overground locomotion (Van Caekenberghe et al., 2010) and all participants in the current study demonstrated a discrete flight phase at the 2.0 m/s speed. As such, the barefoot gait conditions used in the current study were representative of preferred walking and slow running gaits. While temporal and spatial gait parameters in the current study are comparable with those reported for treadmill walking and running at almost identical speeds (Nilsson et al., 1985), further research is required to elucidate the relative role of knee movement and muscle activity on Achilles tendon loading profiles over a range of gait speeds and foot strike patterns. Moreover, while the treadmill system employed in the current study allowed collection of gait data over a large number of steps and at consistent, reproducible speeds, treadmills are known to place spatial and temporal constraints on gait and may introduce potential “learning” effects without the provision of a sufficient acclimatization period. Finally, while several studies have shown that treadmill walking has negligible effect on basic gait parameters and on the fascicular behaviour of the triceps surae compared with overground walking (Cronin & Finni, 2013), treadmill walking has also been shown to alter neuromuscular coordination and subsequent joint moments (Lee & Hidler, 2008). The findings of this study, therefore, may not necessarily be representative of unconstrained walking and running in “real-world” settings outside of the laboratory. In conclusion, this study was the first to use ultrasonic transmission techniques to compare and contrast Achilles tendon loading during barefoot walking and slow

6

running. Consistent with our hypothesis, peak ultrasonic velocities in the Achilles tendon were significantly higher during running. While ultrasonic velocity was closely coupled with ankle joint motion, slow running resulted in a temporal shift in the ultrasonic profile, reflecting earlier, more rapid and higher loading of the Achilles compared with walking. These findings provide new insights into loading patterns of the Achilles tendon during barefoot walking at self-selected speed and slow running, and are important given that Achilles tendon injury is typically associated with running gait.

Perspectives Measurement of tendon loading patterns during walking and running is important for understanding the pathogenesis of tendon “overuse” injury. This is the first study to use sonographic transmission techniques to compare direct measures of the loading of the human Achilles tendon in the same participants during both walking and slow running. Axial ultrasonic velocity in the Achilles tendon was closely coupled with ankle joint motion and in slow running was associated with higher and earlier peak ultrasonic velocity in the Achilles tendon than walking. These findings are important given that the loading conditions that cause “overuse” tendon injury are poorly defined and that around 80% of Achilles tendon injuries involve running rather than walking. Key words: Ultrasound, speed of sound, ultrasonic transmission, locomotion, soft tissue, biomechanics.

Acknowledgements This research received financial support from the Queensland Government.

Author contribution M. Wulf was responsible for participant recruitment, data collection, and contributed to study design, analysis and interpretation of data, drafting, revision, and final approval of the manuscript. S. C. Wearing is responsible for the overall content of the manuscript as guarantor and made contributions to project conception and design, data and statistical analysis, and interpretation of data; drafting, revision, and final approval of the manuscript. J. E. Smeathers contributed to study conception and design, analysis and interpretation of data, revision and final approval of the manuscript. S. L. Hooper contributed to study design, analysis and interpretation of data, revision and final approval of the manuscript. T. Horstmann contributed to study conception and design, analysis and interpretation of data, revision, and final approval of the manuscript. T. Brauner contributed to

Achilles tendon loading in walking and running study conception and design, analysis and interpretation of data, revision, and final approval of the manuscript.

Participant consent Obtained.

Ethical approval The study was approved by the regional ethics committee.

Provenance and peer review Not commissioned; externally peer reviewed.

References Arnold EM, Hamner SR, Seth A, Millard M, Delp SL. How muscle fiber lengths and velocities affect muscle force generation as humans walk and run at different speeds. J Exp Biol 2013: 216: 2150–2160. Bey MJ, Derwin KA. Measurement of in vivo tendon function. J Shoulder Elbow Surg 2012: 21: 149–157. Biewener AA, Gillis GB. Dynamics of muscle function during locomotion: accommodating variable conditions. J Exp Biol 1999: 202: 3387–3396. Bossy E, Talmant M, Laugier P. Effect of bone cortical thickness on velocity measurements using ultrasonic axial transmission: a 2D simulation study. J Acoust Soc Am 2002: 112: 297–307. Cappellini G, Ivanenko YP, Poppele RE, Lacquaniti F. Motor patterns in human walking and running. J Neurophysiol 2006: 95: 3426–3437. Chiu S, Chou L. Effect of walking speed on inter-joint coordination differs between young and elderly adults. J Biomech 2012: 45: 275–280. Cohen J. Statistical power analysis for the behavioral sciences. 2nd edn. Hillsdale, NJ: Lawrence Erlbaum Associates, 1988. Cram JR, Rommen D. Effects of skin preparation on data collected using an EMG muscle-scanning procedure. Biofeedback Self Regul 1989: 14: 75–82. Crevier-Denoix N, Ravary-Plumioen B, Evrard D, Pourcelot P. Reproducibility of a non-invasive ultrasonic technique of tendon force measurement, determined in vitro in equine superficial digital flexor tendons. J Biomech 2009: 42: 2210–2213. Cronin NJ, Finni T. Treadmill versus overground and barefoot versus shod comparisons of triceps surae fascicle behaviour in human walking and running. Gait Posture 2013: 38: 528–533. Curwin SL, Roy RR, Vailas AC. Regional and age variations in growing tendon. J Morphol 1994: 221: 309–320. Finni T, Komi PV, Lepola V. In vivo muscle mechanics during locomotion depend on movement amplitude and

contraction intensity. Eur J Appl Physiol 2001: 85: 170–176. Finni T, Komi PV, Lukkariniemi J. Achilles tendon loading during walking: application of a novel optic fiber technique. Eur J Appl Physiol Occup Physiol 1998: 77: 289–291. Gennisson JL, Deffieux T, Fink M, Tanter M. Ultrasound elastography: principles and techniques. Diagn Interv Imaging 2013: 94: 487–495. Gregor RJ, Komi PV, Browning RC, Jarvinen M. A comparison of the triceps surae and residual muscle moments at the ankle during cycling. J Biomech 1991: 24: 287–297. Gregor RJ, Roy RR, Whiting WC, Lovely RG, Hodgson JA, Edgerton VR. Mechanical output of the cat soleus during treadmill locomotion: in vivo vs in situ characteristics. J Biomech 1988: 21: 721–732. Herzog W, Zatsiorsky V, Prilutsky BI, Leonard TR. Variations in force-time histories of cat gastrocnemius, soleus and plantaris muscles for consecutive walking steps. J Exp Biol 1994: 191: 19–36. Hoffmeister BK, Verdonk ED, Wickline SA, Miller JG. Effect of collagen on the anisotropy of quasi-longitudinal mode ultrasonic velocity in fibrous soft tissues: a comparison of fixed tendon and fixed myocardium. J Acoust Soc Am 1994: 96: 1957–1964. Komi PV. Relevance of in vivo force measurements to human biomechanics. J Biomech 1990: 23: 23–34. Kuo PL, Li PC, Li ML. Elastic properties of tendon measured by two different approaches. Ultrasound Med Biol 2001: 27: 1275–1284. Lavcanska V, Taylor NF, Schache AG. Familiarization to treadmill running in young unimpaired adults. Hum Mov Sci 2005: 24: 544–557. Lee SJ, Hidler J. Biomechanics of overground vs. treadmill walking in healthy individuals. J Appl Physiol 2008: 104: 747–755. Leppilahti J, Puranen J, Orava S. Incidence of Achilles tendon rupture. Acta Orthop Scand 1996: 67: 277–279.

Maffulli N, Sharma P, Luscombe KL. Achilles tendinopathy: aetiology and management. J R Soc Med 2004: 97: 472–476. Maganaris CN, Paul JP. Tensile properties of the in vivo human gastrocnemius tendon. J Biomech 2002: 35: 1639–1646. McNeill Alexander R. Energetics and optimization of human walking and running: the 2000 Raymond Pearl memorial lecture. Am J Hum Biol 2002: 14: 641–648. Miles CA, Fursey GA, Birch HL, Young RD. Factors affecting the ultrasonic properties of equine digital flexor tendons. Ultrasound Med Biol 1996: 22: 907–915. Nilsson J, Thorstensson A, Halbertsma J. Changes in leg movements and muscle activity with speed of locomotion and mode of progression in humans. Acta Physiol Scand 1985: 123: 457–475. Ooi CC, Malliaras P, Schneider ME, Connell DA. Soft, hard, or just right?” Applications and limitations of axial-strain sonoelastography and shear-wave elastography in the assessment of tendon injuries. Skeletal Radiol 2014: 43: 1–12. Orishimo KF, Burstein G, Mullaney MJ, Kremenic IJ, Nesse M, McHugh MP, Lee SJ. Effect of knee flexion angle on Achilles tendon force and ankle joint plantarflexion moment during passive dorsiflexion. J Foot Ankle Surg 2008: 47: 34–39. Pourcelot P, Defontaine M, Ravary B, Lemâtre M, Crevier-Denoix N. A non-invasive method of tendon force measurement. J Biomech 2005a: 38: 2124–2129. Pourcelot P, van den Bogert AJ, Huang X, Crevier-Denoix N. Achilles tendon loads at walk measured using a novel ultrasonic technique. Comput Methods Biomech Biomed Engin 2005b: 8: 221–222. Prilutsky BI, Zatsiorsky VM. Tendon action of two-joint muscles: transfer of mechanical energy between joints during jumping, landing, and running. J Biomech 1994: 27: 25–34.

7

Wulf et al. Reed LF, Urry SR, Wearing SC. Reliability of spatiotemporal and kinetic gait parameters determined by a new instrumented treadmill system. BMC Musculoskelet Disord 2013: 14: 249. Reinschmidt C, Nigg BM. Influence of heel height on ankle joint moments in running. Med Sci Sports Exerc 1995: 27: 410–416. Rowe PJ, Myles CM, Hillmann SJ, Hazlewood ME. Validation of flexible electrogoniometry as a measure of joint kinematics. Physiotherapy 2001: 87: 479–488. Salmons S. The 8th International Conference on Medical and Biological Engineering – meeting report. Biomed Eng 1969: 4: 467–474. Tulchin K, Orendurff M, Karol L. A comparison of multi-segment foot kinematics during level overground and treadmill walking. Gait Posture 2010: 31: 104–108.

8

Van Caekenberghe I, De Smet K, Segers V, De Clercq D. Overground vs. treadmill walk-to-run transition. Gait Posture 2010: 31: 420–428. Vergari C, Pradon D, Ravary-Plumiöen B, Pourcelot P, Crevier-Denoix N. Achilles tendon force and axial speed of sound: a calibration method under clinical conditions. Comput Methods Biomech Biomed Engin 2012a: 15: 355–356. Vergari C, Ravary-Plumioën B, Evrard D, Laugier P, Mitton D, Pourcelot P, Crevier-Denoix N. Axial speed of sound is related to tendon’s nonlinear elasticity. J Biomech 2012b: 45: 263–268. Walmsley B, Hodgson JA, Burke RE. Forces produced by medial gastrocnemius and soleus muscles during locomotion in freely moving cats. J Neurophysiol 1978: 41: 1203–1216. Wearing SC, Reed L, Hooper SL, Bartold S, Smeathers JE, Brauner T. Running

shoes increase Achilles tendon load in walking: an acoustic propagation study. Med Sci Sports Exerc 2014: 46: 1604–1609. Wearing SC, Reed LF, Urry SR. Agreement between temporal and spatial gait parameters from an instrumented walkway and treadmill system at matched walking speed. Gait Posture 2013: 38: 380–384. Wu G, Siegler S, Allard P, Kirtley C, Leardini A, Rosenbaum D, Whittle M, D’Lima D, Cristofolini L, Witte H, Schmid O, Stokes I. ISB recommendation on definitions of joint coordinate system of various joints for the reporting of human joint motion – part I: ankle, hip, and spine. J Biomech 2002: 35: 543–548. Zhang X, Paquette MR, Zhang S. A comparison of gait biomechanics of flip-flops, sandals, barefoot and shoes. J Foot Ankle Res 2013: 6: 45–54.