Reasons for performing study: Although the main cause of injury to the superficial digital flexor tendon (SDFT) is assumed to be high intensity loading of the tendon, to date the forces exerted on the SDFT during cantering have never been measured.
Objective: To measure the force exerted on the SDFT at walk, trot and canter on a treadmill.
Methods: Arthroscopically implantable force probes (AIFP) were implanted in the SDFT of the left and right forelimbs of 8 Thoroughbred horses (480–565 kg). The output of the AIFP was calibrated using the SDFT force calculated by inverse dynamics and an in vitro model of the lower forelimb at trot. The AIFP output was recorded at 1000 Hz at the walk, trot and canter (9 m/s) on a flat treadmill.
Results: AIFP data were analysed successfully in 13 measurement sessions at the walk and trot, in the leading forelimb in 8 sessions at canter and in the trailing forelimb in 5 sessions at canter. The mean ± s.d. maximal force in the SDFT was 3110 ± 1787 N at the walk, 5652 ± 2472 N at the trot, 7030 ± 2948 N in the leading forelimb and 6453 ± 2940 N in the trailing forelimb at canter.
Conclusions: The force in the SDFT increases with running speed from the walk to the canter. The force in the SDFT could not be measured at the gallop. Further study is needed to determine the force in the SDFT at high speed, because it is important for preventing injuries to the SDFT to limit overloading of this tendon.
Rest and controlled exercise is the main treatment for SDFT injuries (Gillis 1996; Dowling et al. 2000). However, it takes about one year to complete rehabilitation and return to race status. Furthermore, there is a high risk of developing recurrence or new injury in the contralateral SDFT before and after return to racing (Gibson et al. 1997).
It has been reported that the risk factors for SDFT injuries are gender (male), ageing, heavy bodyweight, hurdle or steeplechase and long-distance racing (Williams et al. 2001; Kasashima et al. 2004; Takahashi et al. 2004; Perkins et al. 2005). However, those risk factors are difficult to control. It is important to prevent SDFT injuries because it takes a long time to treat them and there is no effective treatment.
It has been assumed that the pathogenesis of SDFT injuries is the degenerative change that usually occurs in the central core of the mid metacarpal region (Dowling et al. 2000), although the exact mechanism remains uncertain. Hyperthermia and repeated stretching during running, causing damage to collagen or other matrix components, seem to cause degenerative alteration in the SDFT (Wilson and Goodship 1994; Birch et al. 1997; Patterson-Kane et al. 1998). It is also assumed that the peak force in the SDFT during running, which may be as high as the physiological limit, is related to occurrence of injuries. Therefore, reducing the force exerted in the SDFT is key to prevent injuries of this tendon. It has been reported that the force exerted in the SDFT decreased at trot and canter on an 8% sloped treadmill compared to the force at 0% slope (Takahashi et al. 2006). However, the magnitude of the decrease in absolute force exerted in the SDFT was unknown, because only the relative value was measured in that report. Although the force or strain, which is proportional to force, in the SDFT at the walk and trot has been reported (Lochner et al. 1980; Riemersma et al. 1988, 1996a,b; Jansen et al. 1993; Platt et al. 1994), only a few observations of the force or strain in the SDFT at high speed running have been reported (Stephens et al. 1989; Butcher et al. 2007; Crevier-Denoix et al. 2009). In these reports, the strain or speed of ultrasound were measured to estimate the force in the SDFT. Because it is impossible at present to know the true value of the force in SDFT during high speed running, the forces in the SDFT measured by several different methods need to be compared in order to provide more accurate estimates.
The purpose of this study was to measure the absolute force exerted in the SDFT at the walk, trot and canter on a treadmill. An arthroscopically implantable force probe (AIFP), which could measure the relative force in the SDFT, was calibrated based on the force calculated at trot using an in vitro forelimb model; then the force in the SDFT was measured on a treadmill from the walk to canter.
Materials and methods
Eight Thoroughbred horses (4 males 2 females and 2 geldings, bodyweight 480–565 kg, age 3–7 years) were used. All horses were examined physically for soundness and were not shoed for the experiments. Their hooves were regularly trimmed by farriers. The horses were familiar with exercise on the treadmill. Each horse was used in 2 experiments, each conducted at different times. The left forelimb was used in the first experiment and, after the insertion site of the force probe healed, the right forelimb was used in the second experiment.
Recording was conducted as previously reported with some modifications (Takahashi et al. 2002). In brief, an arthroscopically implantable force probe (AIFP)1 with customised lead wires was used to measure the force in the SDFT. The AIFP was implanted in the left or right forelimb, at a position 8 cm below the inferior extremity of the accessory carpal bone, while the horse was under sedation with medetomidine (Domitor)2 6–7 µg/kg bwt i.v., and received local infiltration anaesthesia together with a lateral and medial palmar nerve block, using 0.5% bupivacaine (Marcain injection)3. A strain gauge (N11-MA-10-1000-VSE5)4 was attached to the front wall of both forehoofs to determine the timings of ground contact or lift-off of the hoof. After attachment of the AIFP and the strain gauge, the horses were injected with a sedation reversal drug, atipamezole (Antisedan)2 30–60 µg/kg bwt i.v. Recordings were started as soon as the horses could walk soundly, usually within 1 h after reversal. The AIFP was connected to the signal conditioner (DEMOD-DVRT-TC, MB-STD-4)1 and a transmitter of a telemetry system (WEB 5000)5, and hoof strain gauges were also connected to the transmitter. Markers were attached to the hoof (lateral at widest hoof width) and the centres of rotation of each joint. Positions of the distal interphalangeal (coffin) joint, metacarpophalangeal (fetlock) joint and carpal joint relative to markers were confirmed by radiographs. Kinematics was measured by use of a 2D motion analysis system with one camera (HSV-500C3 and MOVIAS)6, and GRF measured by a force plate (DPM-612B and EFP-396ASA21)7. The skin movement artefact was not corrected because the skin displacement in distal parts of the limb is small. The horses trotted at a comfortable speed (about 3 m/s) during kinematics and GRF measurements. All signals and kinematics data were synchronised and recorded at 250 Hz (except the left limb of Horse 7; 125 Hz) using a data recording system (ADM-686z PCI and LaBDAQ PRO)8. After the data were obtained, the horses were moved to the treadmill (SÄTO I)9 and walked (1.7 m/s), trotted (3.5 m/s) and cantered (9.0 m/s), at 0% slope. Then, another data set with the contralateral lead canter was measured. The signals from the AIFP and strain gauges were recorded at 1000 Hz on a personal computer (DI-720 and WinDaq/Pro+)10.
After the recordings, the AIFP was removed from the SDFT, and the skin was sutured under sedation with medetomidine (Domitor)2; 5 µg/kg bwt i.v., and local anaesthesia 0.5% bupivacaine (Marcain injection)3. An antibiotic mixture (benzylpenicillin procaine; 8000 u/kg bwt/day and dihydrostreptomycin sulphate 10 mg/kg bwt/day (Mycillin Sol)11 i.m. for 2 days), and a nonsteroidal anti-inflammatory drug (diclofenac sodium, Blesin2512; 1 mg/kg bwt/day per os for 21 days) were administered to each horse. More than 3 weeks after the first experiment, the second experiment was carried out. All procedures were approved by Equine Research Institute's Animal Care and Use Committee.
Calculation of flexor tendons and suspensory ligament forces and calibration of AIFP
The superficial and deep digital flexor tendons and suspensory ligament forces were calculated using a method previously reported (Meershoek and Lanovaz 2001; Meershoek et al. 2001). However, the in vitro model of the forelimb was modified using Thoroughbred morphometrical data (Table 1). The moment arm of the deep digital flexor tendon (DDFT) in the coffin joint was 30 mm, and those of SDFT, DDFT and suspensory ligament (SL) in the fetlock joint were 48, 41 and 30 mm, respectively. The relationship between the force in the suspensory ligament and the angle of the fetlock joint was 311.5 N/degree of fetlock extension. The time when GRF reached 200 N was defined as ground contact, and the force in the SL was assumed to be zero at that time. To calibrate the AIFP, the relationship between AIFP output and SDFT force was calculated from the corresponding graphs (calculated tendon force vs. AIFP output). The slope of the calibration curve was calculated with the least-square method using the data corresponding to the SDFT loading phase during stance, from the first positive value to the peak SDFT force.
Table 1. Mass, centre of mass (CoM), and moment of inertia (Izz) for the segments of the Thoroughbred model
CoMx and CoMy are the x and y locations of the percentage of segment length of the centre of mass in a local reference frame. Izz is the segment moment of inertia about the z-axis at the segment's CoM.
The SDFT force on the treadmill was calculated by use of the following equation:Force in the SDFT = (VAIFP - Vswing) × slope of calibration curve, where VAIFP is AIFP output voltage and Vswing is the average of the AIFP voltage during the swing phase at each speed. The baseline voltage of the AIFP sometimes shifted during a measurement. To compensate for baseline shift, Vswing at each gait was treated as baseline.
In each horse and for each speed, the mean values of 5 consecutive strides were selected as representative values.
Superficial digital flexor tendon forces obtained in the leading vs. trailing forelimbs at canter were compared with a t test using commercial software (JMP 5.01a)13. Values of P<0.05 were taken to be significant.
An example of the time-force relationships in the flexor tendons and suspensory ligament is shown in Figure 1. The relationship between the AIFP output and the force in the SDFT during the tendon loading phase was almost linear (Fig 2). The mean ± s.d. peak force in the SDFT calculated by the in vitro forelimb model was 4451 ± 1307 N. The slope of the AIFP outputs-SDFT force curve was 3019 ± 3471 N/V (Table 2).
Table 2. Characteristics of the 7 Thoroughbred horses included in the study results, data used to calibrate an arthroscopically implantable force probe (AIFP) and the maximum SDFT force calculated by inverse dynamics from an in vitro forelimb model at a trot of about 3 m/s
Calibration slope (N/V)
Maximum SDFT force (N)
Maximum vertical GRF (N)
Ratio of maximum SDFT force to maximum vertical GRF
Bodyweight in Horses 1, 3 and 7 and age in Horse 1 vary because data in the same horses were collected repeatedly at at least 3 week intervals.
Out of 16 trials in the 8 Thoroughbred horses, AIFP data were measured successfully in 10 trials of 7 horses from the walk to canter on the treadmill. Data after lead change could be measured in 3 out of 10 trials. Data could be measured at the walk and trot in 13 sessions, at the canter in the leading forelimb in 8 sessions, and in the trailing forelimb in 5 sessions. The typical time-force curves during the stance phase of 5 consecutive strides at each gait are shown in Figure 3. At each gait, the force in the SDFT returned to zero at about 80–90% of stance phase. The average maximal force in the SDFT was 3110 ± 1787 N at the walk, 5652 ± 2472 N at the trot, 7030 ± 2948 N in the leading forelimb and 6453 ± 2940 N in the trailing forelimb at canter (Table 3). There was no significant difference between the forces in leading and trailing forelimbs at the canter.
Table 3. Peak SDFT forces of the 7 Thoroughbred horses included in the study results, at walk, trot and canter
Walk (1.7 m/s)
Trot (3.5 m/s)
Canter (9.0 m/s, Lead)
Canter (9.0 m/s, Trail)
L, left front limb; R, right front limb.
The force in the SDFT at each gait measured on the treadmill showed a large variation, because AIFP output was calibrated by the force in the SDFT calculated by the in vitro forelimb model and noninvasive measurements (GRF measured by the force plate and 2D kinematic data). According to Meershoek and Lanovaz (2001), the largest inaccuracies in calculated tendon forces are introduced by measurement errors in the point of application of the GRF, position of the distal interphalangeal joint marker and, to a lesser extent, the relationship between the force in the suspensory ligament and the fetlock joint angle. In the present study, the error in the point of application of the GRF was confirmed to be <1 cm. The position of each marker at the centre of the corresponding joint was confirmed by radiography (under static conditions). Further, the relationship between the suspensory ligament force and fetlock joint angle was measured in the same breed (Thoroughbred). Despite these checks for GRF application point and marker positions and the breed-specific determination of suspensory ligament force and fetlock angle values of maximum SDFT force calculated by the in vitro forelimb model and noninvasive measurements still varied widely.
In addition, the AIFP output is sensitive to the position and direction of the device in the SDFT. It has been reported that the output of the AIFP varies by 1.5–202.8% when removed from the original site and then reimplanted in another location of the same tendon (Fleming et al. 2000). Properties (stiffness and elastic modulus) of the SDFT into which the AIFP is inserted also influence the output of the AIFP. These factors may explain the variation of calibration slopes of the AIFP. Furthermore, it has been reported that the output of the AIFP varies by 3.7–109.4% when removed and reimplanted in the same location (Fleming et al. 2000). Slight displacement of the AIFP within the SDFT during the measurement could therefore influence the output of the AIFP.
Several methods have been used to measure the force in SDFT during high speed running. Crevier-Denoix et al. (2009) used the speed of ultrasound. The advantage of this method is that it is noninvasive. However, it may be difficult to measure the force in the DDFT or SL, because those are deeper-lying structures than the SDFT. Butcher et al. (2007) used an invasive method and measured the strain in the SDFT. However, the zero length of the strain in the tendon and ligament is difficult to determine because the strain increases rapidly when the load in tendon or ligament is in the low ‘toe region’. The AIFP used in this study was inserted into the SDFT invasively. However, the output of the AIFP had a linear relationship even in a low range of load (Fleming et al. 2000; Takahashi et al. 2002), and this method may be applied to measure the force in DDFT or SL at high speed.
The peak of the SDFT force as calculated by the in vitro model usually appeared earlier than the peak of the AIFP (Fig 1). However, the time of the SDFT force peak in the in vitro model used in the present study was similar to that presented in the previous report using the same forelimb model (Meershoek and Lanovaz 2001). Because only the calculated SDFT force during tendon loading was used to calibrate AIFP output, this time difference only had a small effect on the slope of the calibration curve for the AIFP.
The peak force in the SDFT at the walk was 3.1 kN. In previous studies performed in ponies (Jansen et al. 1993; Platt et al. 1994; Riemersma et al. 1996a,b), the SDFT peak force was 1.2–2.7 kN, after conversion taking into account the average bodyweight of the horses used in the present study (517 kg). However, in other studies performed in Thoroughbreds and other breeds (horses of about 500 kg bwt) (Lochner et al. 1980; Butcher et al. 2007), the peak SDFT force was reported as 3.6–3.8 kN, which is close to the results obtained here.
The peak SDFT force at the trot was 5.7 kN. In a previous study using inverse dynamics and noninvasive measurements on Warmbloods (473–560 kg bwt), it was about 7.0 kN at 3.4 m/s (Meershoek and Lanovaz 2001), i.e. higher than that measured in this report. However, similar results (about 6.0 kN) were reported in another study using 3 Thoroughbreds measured on a treadmill at 4.1 m/s (Butcher et al. 2007). In addition, the SDFT force measured in 2 trotters (about 550 kg bwt) was reported as 6.5–7.0 kN at high-speed trot (10 m/s) on training tracks (Crevier-Denoix et al. 2009).
The peak SDFT force at canter was 7030 ± 2948 N in the leading forelimb and 6453 ± 2940 N in the trailing forelimb. The peak strain in the SDFT was 4.8% in Thoroughbreds at the canter (7.0 m/s; Butcher et al. 2007) and the corresponding force, estimated from the SDFT elastic modulus (E = 1.28 GPa from Riemersma and Schamhardt 1985) and tendon cross-sectional area (CSA), was 5.1 kN. In a few reports, the SDFT force has been measured at speeds faster than the canter. It was reported that the peak strain was 11.5–16.6% at gallop (Stephens et al. 1989). Using the same method as above and CSA reported in another study (Crevier et al. 1996), that strain is equivalent to 16.2–20.5 kN. Unfortunately, it is impossible to know the true value of the force in the SDFT at a gallop. However, the approximate value can be estimated by comparing the results measured with the different methods described, including the method in this study.
It has been reported that the SDFT force was related to the vertical ground reaction force (VGRF) in vitro, and that the ratio between the 2 forces was 0.76:1 (Jerbi et al. 2000). In this study, the ratio between the calculated SDFT force and the VGRF measured with the force plate was 0.82:1 (Table 2). The peak forces in the SDFT at the walk, trot and canter were 3.1, 5.7 and 6.5–7.0 kN, respectively, in this study, and the VGRF at each gait in previous reports was equivalent to 3.2–3.7, 5.4–6.0 and 6.7–8.8 kN, respectively (Ueda et al. 1981; Niki et al. 1984; Merkens et al. 1986, 1993a,b; Kai et al. 2000; Weishaupt et al. 2002), taking into account the horses’ average bodyweight in the present study. The SDFT force–VGRF relationship calculated from these data (0.74–1.06) was similar to the previous results (Jerbi et al. 2000).
There was no significant difference between the forces in leading and trailing forelimbs at the canter. On the other hand, it has been reported that maximum VGRF of the trailing forelimb was 25% higher than that of the leading forelimb (Merkens et al. 1993a). It is possible that this relationship also applies to force in the SDFT; however, the large variation in this result may hide the difference.
The output of the AIFP is calibrated using the force in the SDFT calculated at the trot by inverse dynamics and an in vitro model of the distal limb. The force in the SDFT measured with the calibrated AIFP increased with increasing running speed on the treadmill. The maximum force in the SDFT was approximately 3.1 kN at the walk, 5.7 kN at the trot and 6.5–7.0 kN at the canter. The standard deviation was relatively high. Although the high peak force in the SDFT during high speed running may be related to SDFT injury, it was difficult in this study to measure the force in the SDFT at the gallop. Further studies are needed to determine the force in the SDFT at high speed, because it is important for preventing injuries to the SDFT to limit overloading of this tendon.