Lameness is one of the most common medical issues of horses, affecting 8 to 14 horses/100 horses examined.1,2 In addition, mild and subclinical lameness can result in reduced and suboptimal performance by equine athletes.3,4 Because mild lameness may indicate the start of a serious injury or the presence of an existing injury, early identification is important to prevent exacerbation.
Subjective lameness examination is the most common diagnostic method used for detection and monitoring of lameness.5 Even though common scoring systems (eg, American Association of Equine Practitioners lameness scale) used for such evaluations have specific criteria, there is variability in characteristics within each grade, making longitudinal assessment of an animal challenging when only slight improvement in lameness is detected.6,7 In addition, observers may have bias7 during assessment of improvements in lameness following perineural anesthesia. Thus, more accurate, objective methods are needed to supplement subjective lameness examinations both for the detection and monitoring of mild lameness and for assessment of improvements following perineural anesthesia.
Objective methods have been developed for examination of gaits of horses; such methods can be used to supplement subjective lameness examinations. Stationary force platform and optical kinematic analyses are effective for detection of mild lameness.8–13 However, use of these methods is limited by availability, expense, and time for data collection and analysis. These methods are also typically used in a laboratory setting, which limits their usefulness in a clinical setting.
Because of the deficiencies of these kinetic and kinematic techniques, other motion analysis systems that are attached to horses have been designed to objectively characterize lameness in both research and clinical practice settings. Such systems use multiple microelectromechanical components (eg, accelerometers, gyroscopes, and GPS tracking devices) and have wireless and telemetric components for data transmission.5,14–24 Several of these horse-mounted systems are sensitive enough for detection of mild single forelimb or hind limb lameness at a trot5,15,22 and for objective evaluation of hind limb flexion tests.17,23 Recently, optical methods have been used to detect significant changes in sagittal plane kinematics of the hoof at a walk and trot following induction of lameness in horses.12,13 A hoof-mounted motion analysis system could also be used to detect lameness.
The hoof is a suitable location for mounting of a small sensor because sensors can be rigidly attached, which prevents motion artifacts. As smaller and lighter sensors become increasingly available, they can be placed on the distal aspects of limbs in horses without inducing large alterations in motion. These hoof-mounted sensors can collect data for multiple strides and can be used outside a laboratory environment, making them desirable for clinical use. Inertial measurement units have been used to determine linear and angular kinematics in all 3 planes of motion. Such IMUs have been used on the poll (ie, over the occipital protrusion at the caudal aspect of the skull), withers (ie, most dorsal aspect of the shoulders), pelvis, distal aspect of the metacarpal and metatarsal regions, and hoof to evaluate clinically normal and lame horses.15,16,19–24 Results of a study25 of humans indicate IMU data have good agreement with data obtained with an optical system for the evaluation of 3-D angular kinematics. Results of another study16 in which a hoof-mounted IMU was used indicate that although sagittal plane data have the highest correlations with data determined with an optical system, the IMU also provides swing phase data that are similar to data determined with that method for the frontal and transverse planes of motion.16 Results of another study24 in which an IMU was mounted on the metacarpal or metatarsal regions of horses indicate accurate and precise determination of cranial-caudal and vertical displacement at the walk; results also indicate the difficulty of monitoring rapid motion in the distal aspects of limbs of horses.
Although information has been published regarding sagittal motion of the hoof following induction of lameness in horses, the effect of lameness on 3-D kinematics of the hoof is not well understood. The distal aspects of limbs in horses have small frontal and transverse plane motions, compared with the amount of sagittal rotation at the walk and trot.26,27 Because a hoof-mounted IMU can be used to determine 3-D hoof kinematics in clinically normal horses16 and this method is not restrained to use in a laboratory environment, IMUs may be used for detection of changes in hoof orientation following induction of lameness.
The objective of the study reported here was to determine lameness-induced alterations in hoof orientation in horses as measured with an IMU. We hypothesized that, after induction of lameness, there would be significant intralimb differences in hoof orientation in all 3 planes of rotation for lame forelimbs at both a walk and trot. We also hypothesized that differences in hoof orientation could be detected for horses with the mildest grade of lameness at both a walk and trot. In addition, we hypothesized that, following perineural anesthesia of medial and lateral palmar nerves, hoof orientation values would not be significantly different from baseline values.
Materials and Methods
Horses—Six clinically normal Quarter Horses (age range, 2 to 9 years; mean ± SD body weight, 364 ± 19 kg; mean ± SD height at the withers, 1.46 ± 0.03 m) that had no perceptible lameness at a walk or trot (grade 0 of 5 [modified American Association of Equine Practitioners lameness scale6]) were used in this study. These same 6 horses had been used in other related studies.12,13 Data for the present study and those other 2 studies12,13 were collected simultaneously.
For all horses, feet were trimmed and balanced, a steel keg shoe (mean ± SD weight, 324.8 ± 23.5 g) was placed on the left forelimb hoof, and a similar shoe was placed on the right forelimb hoof with nuts welded to the inner web of the medial and lateral branches of the shoe between the third and fourth nail holes (mean ± SD weight, 333.7 ± 25.6 g). The nuts on the right hoof were welded to the shoe so that they were flush with the solar aspect of the shoe and did not contact the horse's sole during weight bearing, as described.12,13 The median weight that the 2 screws added to the right forelimb shoe (lame limb) was 7.8 g (range, 6.8 to 10.6 g [depending on screw length]). Prior to initiation of the study, all horses were acclimated to the Gait Analysis Laboratory where data were collected. All procedures were approved by the Institutional Animal Care and Use Committee of Colorado State University.
Induction of lameness—Lameness was induced in the right forelimb hoof of each horse by insertion of a 6-mm-diameter threaded screw (blunt or with a 2-mm-diameter tapered end) into the medial and lateral nuts of the shoe. The length of the screws ranged from 11 to 17 mm. Blunted screws were used to induce all lameness grades in the first 2 horses and to induce the mildest lameness (grade 1 of 5) in the other 4 horses. The tapered screws were used to induce more severe lameness (grades 2 and 3 of 5) in those latter 4 horses. The screws were fully inserted into the nuts, and the heads of the screws were in contact with the ground when the horse was bearing weight. If a screw did not cause the desired degree of lameness, it was exchanged for a longer or shorter screw, as needed. The screw length that induced each grade of lameness was recorded for each horse.
Lameness trials—Following the collection of baseline data (no induced lameness), each horse underwent induction of 3 grades of lameness (grades 1 through 3/5 [modified American Association of Equine Practitioners lameness scale6]) starting with the mildest and proceeding to the most severe grade. Briefly, grade 1 was described as intermittently lame at a trot, grade 2 was mildly but consistently lame at a trot, and grade 3 was moderately and consistently lame at a trot. None of the lameness grades resulted in visibly detectable lameness at a walk. Walk data were collected prior to trot data for each lameness grade. Horses were allowed to rest for several minutes between each lameness grade to limit the effects of fatigue. After collection of data for grade 3 lameness trials, 3 mL of 2% mepivacaine was injected SC around the medial and lateral palmar nerves. After 10 minutes, if a horse did not have sufficient visually detectable improvement in lameness (defined as a return to baseline gait characteristics or very mild, inconsistent lameness [grade 1]) and skin desensitization, an additional 1.5 mL of 2% mepivacaine was injected SC around the medial and lateral palmar nerves. The horse was then visually reassessed after another 5 minutes.
Instrumentation of horses—The IMUa (5.1 × 3.8 × 1.6 cm; 58.6 g) was composed of a triaxial gyroscope (measurement range, −1,200 to 1,200°/s), triaxial accelerometer (−200 to 200 × g), triaxial magnetometer (−1.9 to 1.9 Gauss), and thermostat (0° to 70°C). Data were sampled at 800 Hz in real time and stored on a handheld computer mounted on the horse until the end of the data collection session. The IMU was attached to a marker triad on the right forelimb hoof, and a custom-machined piece of metal (3.6 × 3.1 × 1.2 cm; weight, 75.7 g) was attached to a marker triad on the left forelimb hoof (Figure 1). The application of marker triads to each hoof was performed as described previously.12 Total weight of the right forelimb marker triad, IMU, cable, and fixation materials was 113.8 g; weight of the left forelimb marker triad was 130.9 g. The cable from the IMU was loosely attached to the right forelimb of each horse with an elastic bandageb wrapped around the distal aspect of the metacarpus and the distal aspect of the antebrachium; the cable was attached to a laptop computer mounted on a surcingle placed around the horse at the level of the sixth or seventh rib. Strain gauges were glued on the hooves of both forelimbs; cables for these gauges were secured with the elastic bandages and terminated at a data collection unit mounted on the surcingle (total weight, 9.5 kg). Data collected with the strain gauges were not analyzed in the present study.
Photograph of marker triads attached to an IMU (A) and attached to a machined piece of metal with weight similar to that of the IMU (B) that were mounted on the right forelimb hoof of a horse during baseline conditions, during 3 grades of lameness (induced by application of pressure to the sole), and after perineural anesthesia. The marker triad with the attached IMU (combined weight, 113.8 g) was rigidly mounted on the right forelimb hoof and the marker triad with the machined piece of metal (combined weight, 130.9 g) was attached to the left forelimb hoof of horses for analysis of hoof orientation during various gait conditions. Marker triads were 3.6 × 3.1 × 1.2 cm. The marker triads included 3 reflective markers (diameter, 1 cm) placed 10 to 15 cm apart for simultaneous determination of kinematic data by use of an optical method described in other studies.12,13
Citation: American Journal of Veterinary Research 75, 9; 10.2460/ajvr.75.9.800
Lameness trials—Data were collected in the Gait Analysis Laboratory; all horses were walked and trotted over an asphalt runway (width, 1.2 m; length, 24.8 m) with a rubberized surface (thickness, 9.3 mm). A stationary force platform was located in the middle of the length of the runway, and the velocity of horses during each trial was measured by the use of 5 infrared timing gatesc located along the length of the force platform at intervals of 1.5 m. This area of the runway was referred to as the capture volume. In the capture volume, horses had a constant velocity. During the baseline trials, a mean velocity was calculated for each horse at each gait. Data for 5 to 9 acceptable trials were collected for each horse at a walk and trot for the right forelimb. An acceptable trial was defined as a trial in which a horse traveled straight and at a consistent velocity that was within 10% of its mean initial velocity for the gait.
Data processing—The IMU data were transferred from the handheld computer to another computer for processing and analysis. Orientation angles in all 3 planes were determined with scripts28 of a computer programd developed by the manufacturer,a which were modified for use with the accelerometers in the IMU unit. The data processing script calculated 3 Euler angle rotations through a series of time steps by use of rotation matrices calculated from accelerometer and magnetometer data obtained with the IMU. The order of rotation was around the vertical axis (x-axis), then around the cranial-caudal axis (y-axis), and then around the medial-lateral axis (z-axis). Presmoothing of sensor data was performed with a 5-point moving average within the script.
As mounted on the marker triad on the right forelimb hoof, positive was directed caudally in the y-axis, medially in the z-axis, and distally in the x-axis (Figure 2). Rotation in all axes was positive in a counterclockwise direction (ie, right-hand rule). Angles were determined in the following order: transverse plane rotation around the x-axis (internal and external [positive] rotation; Ψ), frontal plane rotation around the y-axis (abduction and adduction [positive]; Φ), and sagittal plane rotation around the z-axis (flexion and extension; Θ), with toe-down (ie, the toe of the hoof pointed at the ground) rotation defined as positive.
Photograph depicting the 3-D orientation of the IMU as mounted on the right forelimb hooves of horses. The cranial-caudal axis (y-axis) was positive in a caudal direction, the medial-lateral axis (z-axis) was positive in a medial direction, and the proximal-distal axis (x-axis) was positive in a distal direction (with the hoof placed flat on the ground). Rotation around the z-axis (sagittal plane; Θ), y-axis (frontal plane; Φ), and x-axis (transverse plane; Ψ) are indicated. Rotation in all axes was positive in a counterclockwise direction (ie, right-hand rule). The cable from the IMU was loosely attached to a horse's limb with an elastic bandageb placed around the distal aspect of the metacarpal region and distal aspect of the antebrachium; the cable was attached to a laptop computer mounted on a surcingle placed around the thorax of a horse at the level of the sixth or seventh rib.
Citation: American Journal of Veterinary Research 75, 9; 10.2460/ajvr.75.9.800
Output of the magnetometer in the cranial-caudal direction (ie, along the y-axis) was also used to locate the force platform. As horses moved into the capture volume and over the metal force platform, magnetometer values increased in magnitude. Orientation and accelerometer data in all 3 axes for 3 strides in the vicinity of the force platform were extracted from the entire trial data set. These data were imported into a commercially available kinematics systeme for further processing.
Orientation angles and linear accelerations were low-pass filtered at 15 Hz with a recursive fourth-order Butterworth filter. The events of hoof contact, heel-off (ie, heel lifting off the ground), and toe-off (ie, toe lifting off the ground) were determined by evaluation of the vertical and cranial-caudal acceleration profiles of the stride, as described.12,13 These gait events were used to divide the stride into the following segments: stance (hoof contact to toe-off), breakover (heel-off to toe-off), total swing (toe-off to hoof contact), initial swing (toe-off to initial 25% of swing phase), and terminal swing (last 75% of swing phase to hoof contact). Because there was an offset in the data for each angular orientation during the stance phase resulting from the manner in which the IMU was mounted on the hoof, the orientation in each of the 3 planes during the middle half of the stance phase was subtracted from values of all variables within that particular stride. This ensured that all orientations were set to zero during the stance phase. Thus, during the remainder of the stride, orientations were expressed relative to a flat hoof during the stance phase.
Temporal variables and maximum, minimum, and mean values were determined for each variable during breakover, total swing phase, initial 25% of swing phase, and terminal 25% of swing phase. Instantaneous orientations in all 3 planes were determined at hoof contact, heel-off, and toe-off. Total ROM of the hoof during breakover, total swing phase, initial 25% of swing phase, and terminal 25% of swing phase was also determined. For each trial, a mean of each variable was determined for the 3 strides near the force platform.
Statistical analysis—A commercially available programf was used for statistical analysis. Data were assessed for normality by examining normality plots, and if results indicated that data were nonnormally distributed, they were logarithmically (base e) transformed. Data sets with negative values were rank ordered. A repeated-measures mixed-model ANOVA was performed with each variable of interest as the outcome variable. Intralimb comparisons were made with baseline walk or trot values as control values; these were compared with values for each treatment (lameness grades 1 through 3 and following perineural anesthesia). Lameness grade was a fixed effect with horse velocity included as a confounding variable, and horse was included as a random effect. Values of P < 0.05 were considered significant.
Because the SDs of the perineural anesthesia condition data seemed to be larger than those of the other conditions, homogeneity of variance was tested with a Levene test (for normally distributed data) or Brown-Forsythe test (for nonnormally distributed data). If the test of homogeneity had a significant (P < 0.05) result, individual comparisons were made between each condition (baseline or lameness) and the perineural anesthesia condition. Homogeneity of variance was tested for all 3 orientation planes.
Results
Animals—Lameness was successfully induced in all horses. Blunt screws were used to induce lameness in the first 2 horses. In one of these horses, the IMU stopped transmitting data to the data logger following the baseline trials, so no data for lameness or local anesthesia conditions were collected for that horse. Data were also not obtained for the last horse to undergo the study for the baseline condition or the most severe (grade 3) lameness condition at a trot. At a trot, baseline, grade 1 and 2 lameness, and local anesthesia conditions were obtained for 5 horses; data for the grade 3 lameness condition were collected from 4 horses at that gait. At a walk, baseline data were obtained for 6 horses; all lameness and local anesthesia condition data were collected for 5 horses at that gait. Within 24 hours after induction of lameness, none of the horses had perceptible lameness at a trot.
Trot—Significant intralimb changes were detected in various variables for all 3 angular orientations following induction of lameness during stance and swing phases of the stride and for individual hoof events at a trot (Table 1). Significant changes in angular orientation of the hoof were also detected for the mildest (grade 1) degree of lameness during hoof breakover and total swing and initial swing phases of the stride. Significant changes in frontal plane orientation were more commonly detected at the mildest degree of lameness than were sagittal and transverse plane orientation changes at that degree of lameness. Values of several variables for all 3 angular orientations were similar to baseline values following perineural anesthesia during breakover (transverse plane minimum angle; sagittal plane maximum angle, mean angle, and ROM), swing phase (frontal plane minimum angle), initial swing phase (frontal plane minimum angle), and toe-off (sagittal plane angle).
Hoof orientation angles (°) for various phases of the stride and hoof events during the stride in 6 clinically normal horses at a trot during baseline conditions, after experimental induction of various grades of lameness, and following perineural anesthesia.
| Variable | Plane | Baseline | Grade 1 | Grade 2 | Grade 3 | Perineural anesthesia |
|---|---|---|---|---|---|---|
| Breakover | ||||||
| Mean | Φ | 10.1 | 13.0* | 12.4* | 14.9* | 12.1 |
| Θ | 17.7 ± 9.06 | 15.2 ± 8.16 | 8.7 ± 14.08 | 7.2 ± 12.57* | 11.5 ± 20.69 | |
| Minimum | Ψ | −2.2 ± 12.31 | 3.9 ± 6.84 | 5.2 ± 4.60* | 8.8 ± 5.25* | 3.3 ± 8.72 |
| Maximum | Θ | 62.1 ± 13.12 | 56.1 ± 10.45 | 49.4 ±14.87* | 44.0 ± 15.30* | 57.3 ± 21.07 |
| ROM | Ψ | 44.8 ± 10.37 | 44.0 ± 9.93 | 40.2 ± 8.64* | 34.9 ± 9.19* | 44.2 ± 12.79 |
| Θ | 70.5 ± 8.33 | 66.2 ± 7.76 | 63.9 ± 8.95* | 59.2 ± 11.53* | 70.1 ± 17.82 | |
| Swing | ||||||
| Minimum | Φ | −9.8 | −12.6* | −11.2 | −11.0* | −10.7 |
| Maximum | Ψ | 23.8 ± 17.32 | 30.8 ± 12.62 | 31.7 ± 12.18 | 35.9 ± 13.89* | 32.2 ± 11.15* |
| Θ | 102.8 ± 7.11 | 100.9 ± 9.12 | 109.5 ± 15.99* | 113.1 ± 14.96* | 121.5 ± 34.44* | |
| ROM | Θ‡ | 92.7 ± 1.13 | 93.5 ± 1.14 | 101.9 ± 1.21* | 106.1 ± 1.20* | 116.3 ± 1.49* |
| Initial swing | ||||||
| Minimum | Θ | 43.2 ± 18.98 | 36.9 ± 20.08 | 29.4 ± 22.84* | 25.2 ± 23.02* | 36.9 ± 21.27* |
| Minimum | Φ | −8.2 | −10.0* | −9.7* | −9.5* | −9.3 |
| ROM | Ψ† | 55.6 ± 1.67 | 61.0 ± 1.63 | 62.5 ± 1.63* | 75.7 ± 1.45* | 60.7 ± 1.59* |
| Φ | 41.1 ± 16.37 | 47.0 ±15.49* | 45.7 ± 15.08* | 51.6 ± 14.81* | 45.8 ± 14.12* | |
| Θ | 51.2 ± 25.14 | 58.7 ± 26.67* | 62.7 ± 21.30* | 69.7 ± 23.07* | 59.3 ± 21.98* | |
| Terminal swing | ||||||
| Minimum | Φ | 23.2 ± 11.21 | 17.7 ± 12.01 | 18.7 ± 10.38 | 18.8 ± 8.19* | 14.4 ± 9.72* |
| Maximum | Θ‡ | 44.9 ± 1.33 | 47.7 ± 1.38 | 60.2 ± 1.65* | 63.4 ± 1.55 | 70.2 ± 1.74* |
| ROM | Θ‡ | 32.8 ± 1.55 | 37.2 ± 1.55 | 47.3 ± 1.90* | 50.9 ± 1.70* | 61.1 ± 2.13* |
| Event | ||||||
| Hoof contact | Θ | 17.1 ± 7.95 | 17.2 ± 10.41 | 21.5 ± 9.43 | 24.2 ± 10.92 | 33.1 ± 34.00* |
| Toe-off | Ψ | 4.5 | 9.4* | 8.1* | 12.3* | 7.8* |
| Θ | 56.0 ± 22.72 | 46.6 ± 23.94 | 42.1 ± 21.70* | 33.4 ± 24.82* | 50.1 ± 17.08 | |
Normally distributed data are reported as mean ± SD. Nonnormally distributed data are reported as medians, and data were ranked.
Value is significantly (P < 0.05) different from value for baseline conditions.
Data were logarithmically (base e) transformed.
Angles were determined in the following order: transverse plane rotation around the x-axis (internal and external [positive] rotation; Ψ), frontal plane rotation around the y-axis (abduction and adduction [positive]; Φ), and sagittal plane rotation around the z-axis (flexion and extension; Θ), with toe-down (ie, the toe of the hoof pointed at the ground) rotation defined as positive.
Following perineural anesthesia, the SDs for sagittal plane orientation variables were significantly larger than the SDs for baseline and lameness conditions for that orientation at a trot (Table 2). This effect was detected during stance and swing phases and during individual hoof events. In total, 12 of 19 sagittal plane variables had heterogeneity of variance. For 8 of those 12 variables, the SD for the perineural anesthesia condition was significantly larger than the SDs for the baseline condition and all lameness conditions. In the frontal and transverse plane orientations, 6 of 38 variables had heterogeneity of variance, but there was a larger SD for the perineural anesthesia condition versus other conditions only in the transverse plane at hoof heel-off.
The SD of sagittal plane (Θ) hoof orientation data (°) for various phases of the stride in 6 clinically normal horses at a trot during baseline conditions, after experimental induction of various grades of lameness, and following perineural anesthesia.
| Variable | Baseline | Grade 1 | Grade 2 | Grade 3 | Perineural anesthesia | P value |
|---|---|---|---|---|---|---|
| Breakover | ||||||
| Mean | 9.06 | 8.16 | 14.08 | 12.57 | 20.69 | 0.058 |
| Minimum | 9.42 | 6.20 | 12.11 | 9.08 | 19.76 | 0.157 |
| Maximum | 13.12* | 10.45* | 14.87* | 15.30 | 21.07 | 0.005 |
| ROM | 8.33* | 7.76* | 8.95* | 11.53* | 17.82 | 0.003 |
| Swing | ||||||
| Mean | 7.52* | 8.72* | 9.30* | 8.33* | 14.75 | 0.036 |
| Minimum | 8.66* | 10.30* | 9.54* | 13.12* | 37.18 | < 0.001 |
| Maximum | 7.11* | 9.12* | 15.99* | 14.96* | 34.44 | < 0.001 |
| ROM | 11.01* | 12.16* | 20.97* | 20.43* | 68.74 | 0.003 |
| Initial swing | ||||||
| Mean | 7.55* | 7.84* | 17.28 | 14.30* | 21.50 | < 0.001 |
| Minimum | 18.98 | 20.08 | 22.84 | 23.02 | 21.27 | 0.419 |
| Maximum | 9.82* | 11.48* | 14.45* | 14.79* | 22.55 | 0.013 |
| ROM | 25.14 | 26.67 | 21.30 | 23.08 | 21.99 | 0.440 |
| Terminal swing | ||||||
| Mean | 8.71 | 10.83 | 23.51 | 18.90 | 19.74 | 0.027 |
| Minimum | 8.14* | 9.35* | 10.01* | 9.40* | 37.45 | 0.023 |
| Maximum | 13.33* | 15.84* | 38.68 | 34.67* | 51.26 | 0.002 |
| ROM | 14.81* | 16.55* | 36.99 | 32.86* | 79.81 | 0.001 |
| Hoof contact | 7.95* | 10.41* | 9.43* | 10.92* | 34.00 | < 0.001 |
| Heel-off | 8.62 | 5.50 | 10.71 | 7.90 | 19.21 | 0.129 |
| Toe-off | 22.72 | 23.94 | 21.70 | 24.82 | 17.08 | 0.456 |
Values of P were determined with the test for homogeneity of variance; values of P < 0.05 were considered significant.
Value for the condition differs significantly (P < 0.05) from the value for perineural anesthesia.
Walk—Significant intralimb changes in all 3 angular orientations were detected following induction of lameness during stance and swing phases of the stride and for individual hoof events at a walk (Table 3). Significant changes in the frontal orientation of the hoof were also detected for the mildest (grade 1) degree of lameness during hoof breakover, hoof contact, and hoof heel-off. Only values for the sagittal orientation at hoof toe-off were similar to baseline values following perineural anesthesia; the other hoof orientation variables with significant changes following induction of lameness remained significantly different from baseline values at a walk after perineural anesthesia.
Hoof orientation angles (°) for various phases of the stride and hoof events during the stride in 6 clinically normal horses at a walk during baseline conditions, after experimental induction of various grades of lameness, and following perineural anesthesia.
| Variable | Plane | Baseline | Grade 1 | Grade 2 | Grade 3 | Perineural anesthesia |
|---|---|---|---|---|---|---|
| Breakover | ||||||
| Mean | Φ | 6.5 | 0.6* | −1.2* | −2.1* | 0.4* |
| Maximum‡ | Φ | 20.5 ± 1.83 | 15.9 ± 1.98* | 15.0 ± 2.05* | 14.2 ± 2.29* | 16.2 ± 1.96* |
| ROM | Ψ† | 36.2 ± 1.30 | 32.7 ± 1.27 | 30.7 ± 1.38* | 31.8 ± 1.45* | 30.2 ± 1.37* |
| Φ† | 30.1 ± 1.33 | 27.2 ± 1.34* | 25.9 ± 1.49* | 26.0 ± 1.50* | 25.7 ± 1.43* | |
| Θ | 54.2 ± 9.22 | 52.4 ± 9.52 | 45.5 ± 9.52* | 48.6 ± 10.21 | 41.5 ± 11.92* | |
| Swing | ||||||
| Mean | Θ | 42.9 | 42.5 | 46.4 | 43.9* | 53.4* |
| Maximum | Θ | 98.1 ± 7.71 | 98.4 ± 5.11 | 101.0 ± 6.92 | 104.4 ± 11.17* | 110.8 ± 16.30* |
| Initial swing | ||||||
| Mean | Φ† | 29.3 ± 1.67 | 35.3 ± 1.44 | 37.5 ± 1.38* | 36.6 ± 1.40 | 38.8 ± 1.30* |
| Minimum | Ψ | −60.2 ± 14.66 | −62.1 ± 14.46 | −64.3 ± 14.73* | −64.6 ± 15.24* | −67.5 ± 13.18* |
| Terminal swing | ||||||
| Mean | Θ† | 15.5 ± 1.49 | 16.6 ± 1.42 | 20.3 ± 1.65* | 23.5 ± 1.77* | 31.1 ± 2.07* |
| Maximum | Θ† | 37.9 ± 1.25 | 39.9 ± 1.28 | 40.1 ± 1.51 | 48.9 ± 1.49* | 63.2 ± 1.69* |
| ROM | Θ† | 39.2 ± 1.24 | 40.6 ± 1.34 | 40.4 ±1.57 | 46.7 ± 1.67 | 59.3 ± 1.60* |
| Event | ||||||
| Hoof contact | Φ | 22.1 | 38.3* | 46.0* | 42.2* | 49.5* |
| Θ | 5.2 | 6.7 | 8.6 | 8.4 | 8.7* | |
| Heel-off | Φ | 0.6 ± 6.78 | −1.6 ± 6.05* | −2.7 ± 5.99* | −2.3 ± 6.42* | −3.0 ± 4.19* |
| Toe-off | Ψ | 15.1 | 21.4 | 20.8 | 24.8* | 22.0* |
| Θ | 44.9 ± 25.11 | 41.9 ± 19.05 | 36.8 ± 16.74 | 29.8 ± 25.06* | 43.0 ± 19.74 | |
See Table 1 for key.
Similar to results for horses at a trot, the SDs of variables in the sagittal plane orientation following perineural anesthesia were larger than the SDs for such variables during baseline and lameness conditions at a walk (Table 4). This effect was detected during both stance and swing phases of the stride. Twelve of 19 sagittal plane variables had significantly larger SDs for the perineural anesthesia condition versus baseline and lameness conditions. In the frontal and transverse plane orientations, 5 of 38 variables had heterogeneity of variance; however, only the mean transverse orientation during breakover had a larger SD for the perineural anesthesia condition, compared with SDs for the baseline and lameness conditions.
The SD of sagittal plane (Θ) hoof orientation data (°) for various phases of the stride in 6 clinically normal horses at a walk during baseline conditions, after experimental induction of various grades of lameness, and following perineural anesthesia.
| Variable | Baseline | Grade 1 | Grade 2 | Grade 3 | Perineural anesthesia | P value |
|---|---|---|---|---|---|---|
| Breakover | ||||||
| Mean | 8.97* | 5.19* | 10.24* | 10.47* | 31.74 | < 0.001 |
| Minimum | 8.37* | 4.77* | 7.01* | 8.90* | 30.41 | < 0.001 |
| Maximum | 14.63* | 9.78* | 12.17* | 12.28* | 27.26 | < 0.001 |
| ROM | 9.22 | 9.52 | 9.52 | 10.21 | 11.92 | 0.757 |
| Swing | ||||||
| Mean | 5.46* | 7.73* | 11.70* | 14.98* | 23.03 | < 0.001 |
| Minimum | 14.39 | 6.85 | 8.54 | 11.51 | 19.69 | 0.083 |
| Maximum | 7.71* | 5.11* | 6.92* | 11.17* | 16.30 | < 0.001 |
| ROM | 18.16 | 8.37 | 12.98 | 19.85 | 16.55 | 0.170 |
| Initial swing | ||||||
| Mean | 6.71* | 5.97* | 6.21* | 6.72* | 12.08 | < 0.001 |
| Minimum | 24.56 | 17.22 | 14.97 | 24.43 | 21.06 | 0.198 |
| Maximum | 7.97* | 5.26* | 6.25* | 9.17* | 14.38 | < 0.001 |
| ROM | 28.11 | 19.10 | 16.16 | 27.93 | 20.64 | 0.113 |
| Terminal swing | ||||||
| Mean | 6.36* | 6.58* | 15.65* | 18.77* | 30.23 | < 0.001 |
| Minimum | 6.99* | 6.23* | 9.20* | 9.15* | 20.06 | < 0.001 |
| Maximum | 8.21* | 9.60* | 21.48* | 25.07* | 38.68 | < 0.001 |
| ROM | 8.07* | 10.66* | 27.92 | 32.58 | 30.59 | 0.002 |
| Hoof contact | 9.22* | 8.89* | 11.46* | 9.18* | 46.58 | < 0.001 |
| Heel-off | 8.77* | 5.16* | 6.66* | 7.33* | 31.27 | < 0.001 |
| Toe-off | 25.11 | 19.05 | 16.74 | 25.06 | 19.74 | 0.197 |
See Table 2 for key.
Discussion
By use of a hoof-mounted IMU, we detected significant changes in the 3-D orientation of hooves of horses after experimental induction of weight-bearing lameness at a walk and trot. Because the IMU was mounted on lame right forelimbs, only intralimb changes could be evaluated. Significant changes in the sagittal plane orientation (Θ) during both stance and swing phases of the stride were detected for hooves following induction of lameness. These changes in sagittal plane orientation of the hoof have been detected at a trot and walk by use of optical methods in other studies12,13; data for those studies were collected simultaneously from the same horses that were included in the present study. Sagittal plane (Θ) data obtained by use of optical methods in those other studies12,13 and with an IMU in the present study were similar; however, significant intralimb differences in values were detected after induction of lameness more commonly with the IMU method than they were with the optical method. Direct comparisons between the 2 methods were not performed because collection of data for those methods was not synchronized and IMU and optical data were not collected for the same strides within each trial. For the IMU data, strides on either side of the force platform were chosen for analysis. Results of another study29 indicate orientation data obtained with IMUs in the vicinity of a stationary force platform have high error until an IMU is at least 40 cm above the force platform. Because the magnetometer signal was used in the manufacturer processing script to determine hoof orientation and its magnitude was influenced by a horse moving over the force platform, data for strides directly over the force platform were not analyzed. For data determined with an optical system in those other studies,12,13 the stride of interest was always directly over the force platform.
As determined on the basis of optical methods, the sagittal orientation changes detected following lameness were more commonly interlimb, with most intralimb changes present in the nonlame limb.12,13 Although kinematic data were not collected with an IMU mounted on the nonlame forelimb hoof for horses in the present study, inter- and intralimb results similar to those determined with the optical method would have been expected. A larger number of strides were analyzed by use of the IMU (3 strides/trial) in the present study than were analyzed by use of the optical method in those other studies12,13; this may have improved the ability to detect significant intralimb differences for limbs after induction of lameness. Although multiple consecutive strides can be analyzed with an IMU, only 3 strides were analyzed for each condition in the present study because the runway was short in length. In each trial, horses had a few strides of acceleration prior to entering the capture volume and a few strides of deceleration after exiting the capture volume. Considering that it was a requirement for horses to move at a consistent velocity during data collection, we chose to analyze only a small number of strides for which horses were moving at a consistent velocity.
Although sagittal plane linear and angular kinematics of distal aspects of limbs were altered in horses with lameness in other studies,9–13 no studies have been conducted to evaluate changes in the orientation of the hoof in a lame limb. Results of another study27 indicate an increase in lateral roll of the hoof during breakover following contouring of the lateral branch of a shoe. Other studies30–32 have been conducted to determine 3-D orientation of distal aspects of limbs in clinically normal horses and in such horses with induced medial-lateral and cranial-caudal hoof imbalance. Results of the present study indicated significant changes in abduction-adduction (Φ) and internal-external rotation (Ψ) orientations of limbs at a walk and trot following induction of lameness. This included increased external rotation (Ψ) of the hoof during both breakover and toe-off. Additionally, during breakover, hooves were in a significantly more abducted position during the mildest lameness (grade 1) condition than during the baseline condition. During the initial 25% of swing phase, lameness resulted in an increased ROM in both transverse and frontal planes. These changes to the ROM may have been attributable to increased internal rotation (Ψ) and adduction (Φ) during the initial portion of the swing phase, which may have been compensatory for the external rotation and abduction of the hoof during breakover.
Following perineural anesthesia, values of several variables in all 3 planes of rotation were similar to baseline values at a trot. However, at a walk, only values for the sagittal orientation at hoof toe-off were similar to baseline values following perineural anesthesia. Because some horses had low-grade lameness after perineural anesthesia, it was not unexpected that all variables were not similar to baseline values. In addition, because lameness was only visually detectable at a trot in horses at that time and perineural anesthesia resulted in a substantial reduction of lameness, it would have been expected that a larger number of variables would return to baseline values at a trot, compared with results obtained at a walk. At a walk, lameness was not visually detectable in horses at that time. Therefore, it was expected that lameness would result in a smaller number of variables with significant orientation changes at a walk. In addition, because horses may have had mild lameness following perineural anesthesia, there would have been fewer variables with orientation values that returned to baseline values. Given that a larger number of variables returned to baseline values at a trot, compared with at a walk, these results supported the assessment of perineural anesthesia at a trot when examining changes to hoof orientation in a single forelimb lameness.
Analysis for horses following perineural anesthesia indicated a significant increase in SDs for sagittal plane data, compared with SDs for baseline and lameness conditions. This increase in SDs was not detected for the frontal and transverse planes. We postulate that this effect may have been attributable to changes in proprioception following perineural anesthesia. The effect of perineural anesthesia on proprioception most likely had a greater effect on flexion and extension than it did on motion in other planes, considering that ligaments provide more support to the distal aspect of a limb in the frontal and transverse planes. Authors of another study33 found that horses have greater variability in stride length following perineural anesthesia and concluded that lame horses have less variance in that variable than do sound horses. However, investigators in that study33 evaluated only horses with naturally occurring lameness before and after performance of regional anesthesia and did not include nonlame horses. Thus, the conclusions of those authors33 are most relevant for comparisons of lame horses before and after perineural anesthesia. For clinically normal horses in the present study, we found that sagittal plane values had less variability before and after induction of lameness, compared with values after perineural anesthesia. Thus, an increase in sagittal plane variability may have indicated successful perineural anesthesia.
The ROM of hooves of horses in the present study in frontal and transverse planes seemed to be larger than expected, on the basis of previously reported16,27 data. However, ROM in the sagittal plane was similar to previously reported data. The range of sagittal plane rotation of the hoof during breakover reported by authors of another study27 was similar to values determined for horses in the present study during baseline conditions and following perineural anesthesia. However, results of that other study27 indicate a smaller ROM in the frontal and transverse planes during hoof breakover, compared with values determined in the present study. This discrepancy in data for the frontal and transverse planes may have been attributable to larger frequencies in the IMU data resulting from differences in sampling rate and filtering, which may have affected maximum and minimum values. Horses may have detected a tactile sensation from the wires connecting the IMU to the data logging device, which may have affected the motion of the forelimbs and resulted in altered rotation. Results of another study34 indicate that tactile stimulation at the level of the proximal interphalangeal joint (pastern) has short-term effects on sagittal plane kinematics of forelimbs. However, horses wore the system for an extended period before and during data collection; therefore, tactile stimulation may have subsided during data collection. The authors are unaware of any studies in which the effects of tactile stimulation of distal aspects of limbs on extrasagittal plane movements at the level of the hoof were evaluated. Data were collected in close proximity to a large ferrous object (force platform) in the present study; it was likely that disturbances in function of the magnetometer resulted in alterations of data for those orientations. Because the magnetometer was used during data processing to determine orientations, the close proximity of the force platform was the most likely reason for these large ROMs in the frontal and transverse planes.
Although IMUs can be highly accurate in all 3 planes of rotation,25 processing methods can affect the accuracy of data obtained with such devices.35 With additional data processing, the 3-axis gyroscope in the IMU used in the present study may have had higher accuracy, compared with that for the orientation angles calculated with the proprietary data processing routine. Preliminary results determined with other processing methods (data not reported) indicated the potential for improved accuracy and precision of IMU orientation data. However, drift within the gyroscopic signal is common with this type of device and was exacerbated by the force platform, which was within the capture volume where data were collected in this study. Performance of calibrations in addition to those recommended by the manufacturer is often needed to improve the accuracy and precision of an IMU.36 The high ROMs detected in abduction-adduction (Φ) and internal-external rotation (Ψ) planes in the present study suggested that calibration additional to that recommended by the manufacturer is required.
Results of the present study indicated that the IMU was able to detect significant intralimb orientation changes following induction of lameness in all 3 planes of motion at both a walk and trot, with most significant changes detected for mild lameness conditions in the frontal plane. The sagittal plane kinematic changes of hooves detected with the IMU were similar to those detected by use of an optical system in other studies12,13; however, the IMU seemed to be slightly more sensitive for detection of intralimb changes, compared with the ability to detect changes with the optical system. Following perineural anesthesia, values similar to baseline values were detected with the IMU for various orientation variables, primarily for data obtained at a trot. In addition, significant increases in the SDs for sagittal orientation data were detected with the IMU, which may have been a useful indicator for assessment of regional anesthesia. Thus, the IMU (as mounted on hooves of horses in this study) may be useful for differentiation of nonlame, experimentally induced single forelimb lameness, and perineural anesthesia conditions for horses at a trot and a walk.
Further studies are warranted to improve data processing of the IMU, including integration of the gyroscopic signals with a Kalman filter. This would increase the accuracy and precision of the orientation data, especially in the frontal and transverse planes. Development of algorithms to improve manufacturer calibrations should also improve the accuracy of this system. Further improvements to data processing are needed before this IMU can be of clinical use for evaluation of lameness in horses. In addition, calculation of the sensitivity and specificity of hoof orientation data for determination of cutoff values for identification of lameness in horses would be important prior to use of this device.
ABBREVIATIONS
| IMU | Inertial measurement unit |
| ROM | Range of motion |
HP200-1200F0400R, H3-IMU, MemSense, Rapid City, SD.
Vetrap, 3M, Saint Paul, Minn.
MEK 92-PAD photoelectric control, Mekontrol Inc, Northboro, Mass.
MATLAB, MathWorks Inc, Natick, Mass.
Vicon-Motus 9.2, Vicon Motion Systems Inc, Centennial, Colo.
STATA, version 11, StataCorp LP, College Station, Tex.
References
1. USDA. National animal health monitoring system equine study. Part I: baseline reference of 1998 equine health and management. Fort Collins, Colo: USDA APHIS Veterinary Services, 1998.
2. Kaneene JB, Ross WA, Miller RA. The Michigan equine monitoring system. II. Frequencies and impact of selected health problems. Prev Vet Med 1997; 29: 277–292.
3. Gaughan EM. Skeletal origins of exercise intolerance in horses. Vet Clin North Am Equine Pract 1996; 12: 517–535.
4. Parente EJ, Russau AL, Birks EK. Effects of mild forelimb lameness on exercise performance. Equine Vet J Suppl 2002;(34):252–256.
5. McCracken MJ, Krammer J, Keegan KG, et al. Comparison of an inertial sensor system of lameness quantification with subjective lameness evaluation. Equine Vet J 2012; 44: 652–656.
6. American Association of Equine Practitioners. Guide for veterinary service and judging of equestrian events. 4th ed. Lexington, Ky: American Association of Equine Practitioners, 1991; 19.
7. Arkell M, Archer RM, Guitian FJ, et al. Evidence of bias affecting the interpretation of the results of local anesthetic nerve blocks when assessing lameness in horses. Vet Rec 2006; 159: 346–349.
8. Ishihara A, Bertone AL, Rajala-Schultz PJ. Association between subjective lameness grade and kinetic gait parameters in horses with experimentally induced forelimb lameness. Am J Vet Res 2005; 66: 1805–1815.
9. Buchner HHF, Savelberg HHCM, Schamhardt HC, et al. Limb movement adaptations in horses with experimentally induced fore- or hindlimb lameness. Equine Vet J 1996; 28: 63–70.
10. Galisteo AM, Cano MR, Morales JL, et al. Kinematics in horses at the trot before and after an induced forelimb supporting lameness. Equine Vet J Suppl 1997;(29):97–101.
11. Keegan KG, Wilson DA, Smith BK, et al. Changes in kinematic variables observed during pressure-induced forelimb lameness in adult horses trotting on a treadmill. Am J Vet Res 2000; 61: 612–619.
12. Moorman VJ, Reiser RF, Peterson ML, et al. Effect of equine forelimb lameness on hoof kinematics at the trot. Am J Vet Res 2013; 74: 1183–1191.
13. Moorman VJ, Reiser RF, Peterson ML, et al. Effect of equine forelimb lameness on hoof kinematics at the walk. Am J Vet Res 2013; 74: 1192–1197.
14. Keegan KG, Yonezawa Y, Pai PF, et al. Evaluation of a sensor-based system of motion analysis for detection and quantification of forelimb and hind limb lameness in horses. Am J Vet Res 2004; 65: 665–670.
15. Pfau T, Robilliard JJ, Weller R, et al. Assessment of mild hindlimb lameness during over ground locomotion using linear discriminant analysis of inertial sensor data. Equine Vet J 2007; 39: 407–413.
16. Moorman VJ, Reiser RF, McIlwraith CW, et al. Validation of an equine inertial measurement unit system in clinically normal horses during walking and trotting. Am J Vet Res 2012; 73: 1160–1170.
17. Marshall JF, Lund DG, Voute LC. Use of a wireless, inertial sensor-based system to objectively evaluate flexion tests in the horse. Equine Vet J Suppl 2012;(44):8–11.
18. Keegan KG, MacAllister CG, Wilson DA, et al. Comparison of an inertial sensor system with a stationary force plate for evaluation of horses with bilateral forelimb lameness. Am J Vet Res 2012; 73: 368–374.
19. Church EE, Walker AM, Wilson AM, et al. Evaluation of discriminant analysis on dorsoventral symmetry indices to quantify hindlimb lameness during over ground locomotion in the horse. Equine Vet J 2009; 41: 304–308.
20. Starke SD, Witte TH, Mat SA, et al. Accuracy and precision of hind limb foot contact timings of horses determined using a pelvic-mounted inertial measurement unit. J Biomech 2012; 45;1522–1528.
21. Olsen E, Anderson PH, Pfau T. Accuracy and precision of equine gait event detection during walking with limb and trunk mounted inertial sensors. Sensors (Basel) 2012; 12: 8145–8146.
22. Starke SD, Raistrick KJ, May SA, et al. The effect of trotting speed on the evaluation of subtle lameness in horses. Vet J 2013; 197: 245–252.
23. Starke SD, Willems E, Head M, et al. Proximal hindlimb flexion in the horse: effect on movement symmetry and implications for defining soundness. Equine Vet J 2012; 44: 657–663.
24. Olsen E, Pfau T, Ritz C. Functional limits of agreement applied as a novel method comparison tool for accuracy and precision of inertial measurement unit derived displacement of the distal limb in horses. J Biomech 2013; 46: 2320–2325.
25. Rouhani H, Favre J, Crevoisier X, et al. Measurement of multisegment foot joint angles during gait using a wearable system. J Biomech Eng 2012; 134: 061006.
26. Clayton HM, Sha DH, Stick JA, et al. 3D kinematics of the interphalangeal joints in the forelimb of walking and trotting horses. Vet Comp Orthop Traumatol 2007; 20: 1–7.
27. Keegan KG, Satterley JM, Skubic M, et al. Use of gyroscopic sensors for objective evaluation of trimming and shoeing to alter time between heel and toe lift-off at end of the stance phase in horses walking and trotting on a treadmill. Am J Vet Res 2005; 66: 2046–2054.
28. Konvalin C. Technical document: calculating heading, elevation and bank angle. Rapid City, SD: MemSense, 2008. Available at: memsense.com/docs/MTD-0801_1_0_Calculating_Heading_Elevation_Bank_Angle.pdf. Accessed Nov 8, 2013.
29. de Vries WHK, Veeger HEJ, Baten CTM, et al. Magnetic distortion in motion labs, implications for validating inertial magnetic sensors. Gait Posture 2009; 29: 535–541.
30. Chateau H, Degueurce C, Jerbi H, et al. Normal three-dimensional behavior of the metacarpophalangeal joint and the effect of uneven foot bearing. Equine Vet J Suppl 2001;(33):84–88.
31. Chateau H, Degueurce C, Denoix J-M. Evaluation of three-dimensional kinematics of the distal portion of the forelimb in horses walking in a straight line. Am J Vet Res 2004; 65: 447–455.
32. Degueurce C, Chateau H, Jerbi H, et al. Three-dimensional kinematics of the proximal interphalangeal joint: effects of raising the heels or the toe. Equine Vet J Suppl 2001;(33):79–83.
33. Peham C, Licka T, Girtler D, et al. The influence of lameness on equine stride length consistency. Vet J 2001; 162: 153–157.
34. Clayton HM, White AD, Kaiser LJ, et al. Short-term habituation of equine limb kinematics to tactile stimulation of the coronet. Vet Comp Orthop Traumatol 2008; 21: 211–214.
35. Mazza C, Donati M, McCamley J, et al. An optimized Kalman filter for the estimate of trunk orientation from inertial sensors data during treadmill walking. Gait Posture 2012; 35: 138–142.
36. Brodie MA, Walmsley A, Page W. The static accuracy and calibration of inertial measurement units for 3D orientation. Comput Methods Biomech Biomed Eng 2008; 11: 641–648.
