Introduction
In most situations, lameness in horses causes asymmetries that can be detected when assessing movement at a symmetric gait such as the trot.1 Objective gait analysis is needed to accurately detect and quantify lameness.2,3,4,5,6,7,8 Traditional approaches for objective gait analysis such as kinematic evaluation with multiple high-speed cameras and kinetic evaluation with 1 or more stationary force plates are expensive and labor-intensive (ie, not practical for clinical use).9,10 To overcome these limitations, PISBSs have been developed to objectively detect and quantify asymmetric lameness in horses.11,12
One PISBS for lameness detection and quantification in horses12,a is composed of 3 small wireless sensors. Two of these sensors consist of uniaxial accelerometers that are attached to the dorsal midline region over the poll and over the sacral tuberosities to measure vertical acceleration of the head and pelvis, respectively. The third is a uniplanar gyroscope that is attached to the dorsal aspect of the right forelimb pastern region to measure the angular velocity on the sagittal plane. The gyroscope detects extension and flexion of the instrumented digit. Extension of the digit yields a negative signal, and flexion of the digit yields a positive signal. Data transmitted in real time to a portable computer are used to calculate the vertical excursion of the head associated with the stance of each forelimb and the vertical excursion of the pelvis associated with the stance of each hind limb. Because, at the trot, the limbs move as diagonal pairs, the phase of stance of the right forelimb determined by data collected with the right forelimb gyroscope is used to infer the phase of stance of the other 3 limbs (for the left hind limb, the same phase of stance of the right forelimb; for the left forelimb and right hind limb, the opposite phase of stance relative to the phase of stride of the right forelimb).
Readings obtained with the PISBS will be inverted (ie, the features associated with one limb will be attributed to the contralateral limb) under 2 circumstances. The first circumstance is when the gyroscope is attached to the dorsal aspect of the right forelimb pastern region but flipped 180° on the frontal plane relative to the standard position. The second is when the gyroscope is attached on its usual orientation but on the dorsal aspect of the left forelimb pastern region. In summary, either of these 2 alternative approaches for gyroscope instrumentation will lead to inversion of the results of the evaluations with the PISBS between right and left limbs for both forelimbs and hind limbs.11,12 Conversely, the authors’ experience has been that if both of these circumstances simultaneously exist when attaching the gyroscope of the PISBS (ie, placing the gyroscope on the left forelimb pastern region, flipped 180° on the frontal plane relative to the standard position), the PISBS will correctly detect and quantify lameness.
A cause of concern for horse owners, trainers, and veterinarians is that the sensor attached to the right forelimb may produce gait changes that could be interpreted by the PISBS as lameness. The authors’ experiences using this tool in hundreds of horses suggest otherwise. However, in a study13 investigating the effect of a lightweight (55-g) tactile stimulation device loosely attached around the pastern region of each forelimb, exaggerated elevation of both forefeet was observed. In that same study,13 it was observed that habituation minimized the effect of tactile stimulation of the forelimb pastern region and that the exaggerated elevation of the forefeet could no longer be detected after 6 trials in which the horses were trotted in hand along a 30-m runway.
The objective of the study reported here was to investigate whether the wrap with the gyroscope attached to a forelimb pastern region can interfere with forelimb motion and lead to errors in lameness detection and quantification with the PISBS in horses evaluated at the trot. We hypothesized that alternately attaching the gyroscope to the right forelimb pastern region (the standard position as recommended by the manufacturer) or attaching the gyroscope upside down (ie, flipped 180° on the frontal plane relative to the standard position) to the dorsal aspect of the left forelimb pastern region would affect the results of PISBS evaluations. We hypothesized that limb instrumentation would change the values of the main outcome variables generated by the PISBS (MINDIFFhead, MAXDIFFhead, SignedVShead, MINDIFFpelvis and MAXDIFFpelvis) by their respective threshold values (6.0, 6.0, 8.5, 3.0, and 3.0 mm) or greater.
Materials and Methods
Horses
This study protocol was approved by the Institutional Animal Care and Use Committee of the University of Missouri. Twenty adult horses (14 mares and 6 geldings) from the research and teaching herd without lameness noticeable at the walk (American Association of Equine Practitioners grades 0 to 3) were used in the study. Horses ranged in body weight from 410 to 650 kg (median, 503 kg) and in age from 6 to 25 years (median, 15 years). The group included 11 Quarter Horses, 2 Paints, 2 Hanoverians, 2 Thoroughbreds, 1 Saddlebred, 1 warmblood, and 1 unknown breed.
Study design and data collection
Each horse underwent 2 lameness trials. Horse instrumentation was performed according to the PISBS manual.12 The head sensor (uniaxial accelerometer) was attached to a neoprene cap secured to the halter to allow positioning of the sensor over the poll. The pelvic sensor (uniaxial accelerometer) was attached to a plastic plate that was positioned over the sacral tuberosities and secured to the hair with 2 metallic clips. The forelimb sensor (uniplanar gyroscope) was attached to a neoprene strip that was wrapped around the right forelimb pastern region in the standard position for one of the trials and wrapped upside down (to flip the forelimb sensor 180° on the frontal plane relative to the standard position) around the left forelimb pastern region for the other trial.
As recommended by the PISBS manufacturer,11,12 for each trial, data collection was performed while the horses were trotted in hand over a 30-m-long straight asphalt aisle, back and forth twice (total length of the trot, approx 120 m). Measures were taken to ensure consistency of data collection, including data collection with the same unit (portable computer and software version), trotting of horses on the same aisle, and trotting of horses by the same individual during both trials. The order of trials was assigned by use of random numbers generated with the aid of spreadsheet software,b ensuring that the order of trials (left forelimb followed by right forelimb or vice versa) was distributed evenly among the horses. After the first trial, the gyroscope was moved to the opposite forelimb and the second trial data were collected within 5 minutes after the first trial ended.
Data analysis
Immediately after data collection, data were analyzed by means of the PISBS with default settings (remove outliers function set to remove all strides where MINDIFFhead or MINDIFFpelvis was outside the 95% CI for the mean [ie, mean ± 1.96 SD]). As recommended by the PISBS manufacturer, horses were considered to have forelimb lameness if the SD of MINDIFFhead or MAXDIFFhead was smaller than the mean value of each of these variables and the absolute value of SignedVShead was above threshold (8.5 mm). The sign of MINDIFFhead indicated the side of lameness (negative value, left forelimb; positive value, right forelimb). Horses were considered to have impact hind limb lameness if the absolute value of MINDIFFpelvis was above threshold (3 mm) and the SD of MINDIFFpelvis was smaller than the mean. Horses were considered to have push-off hind limb lameness if the absolute value of MAXDIFFpelvis was above threshold (3 mm) and the SD of MAXDIFFpelvis was smaller than the mean. The sign of MINDIFFpelvis and MAXDIFFpelvis indicated the side with impact or push-off lameness (negative value, left hind limb; positive value, right hind limb).
Statistical analysis
After normal data distribution was confirmed with the Shapiro-Wilk test, the effects of forelimb instrumentation approach and the effects of the order of the 2 lameness trials on the main continuous variables of interest (MINDIFFhead, MAXDIFFhead, SignedVShead, MINDIFFpelvis, and MAXDIFFpelvis) were investigated by means of graphical representation, repeated-measures ANOVA, and Pearson product-moment correlation analysis. Furthermore, the presence of consistent bias attributable to forelimb instrumentation with the gyroscope or the order of lameness trials was investigated by means of Bland-Altman analysis.
To investigate the agreement in lameness classification between the 2 lameness trials, for each trial, forelimb lameness, hind limb push-off lameness, and hind limb impact lameness were classified as absent (0), left limb lameness (–1), or right limb lameness (1). The results were represented graphically as previously described.14 The proportion of agreement between the 20 pairs of trials for forelimb lameness, hind limb impact lameness, and hind limb push-off lameness was evaluated with the general z test, whereby z = (p – pexp) / (pexp × [1 – pexp]/n)½ and the expected proportion (p) of agreement (pexp) by chance was 50%. Exact 95% Clopper-Pearson CIs for the observed proportions of agreement for the 3 types of lameness were also calculated.
Statistical analyses were performed with statistical software.b,c For all comparisons, α was set at 0.05. Because multiple comparisons were performed, α was adjusted with the Bonferroni correction method (adjusted α = 0.05/number of comparisons [0.05/13 = 0.004]).15 A post hoc power analysis was conducted to estimate the power of this study to detect a difference equal to the threshold value for each of the 5 main variables measured with the PISBS.
Results
Horses
Results of the 2 lameness trials with the PISBS indicated that all horses were lame. Specifically, 14 horses had both forelimb and hind limb lameness, 2 horses had only forelimb lameness, 4 horses had only hind limb lameness, 16 horses had push-off hind limb lameness, 11 horses had impact hind limb lameness, and 9 horses had both push-off and impact hind limb lameness.
Lameness variables
Graphical representation of the means of the main lameness variables of interest (MINDIFFhead, MAXDIFFhead, SignedVShead, MINDIFFpelvis, and MAXDIFFpelvis) revealed no consistent bias (ie, an effect of limb instrumentation or trial order) but did show a lack of perfect repeatability of consecutive lameness trials using the PISBS (Figure 1). No effect was detected of limb instrumentation approach (gyroscope attached to the right forelimb pastern region, in the standard upright position or to the left forelimb pastern region and flipped 180° on the frontal plane relative to the standard position; Table 1 and Figure 2; Supplementary Figure S1) or trial order (Table 2 and Figure 3; Supplementary Figure S2) for any of the 5 main variables of interest. Because values could be negative or positive, depending on the lame limb, the means were close to 0 and the SDs were relatively large. To confirm that these features did not compromise data analysis, repeated-measures ANOVA was also performed after the data were transformed by adding the absolute value of the smallest result to the results for each variable. As expected, data transformation produced marked increases in the means and made the SDs smaller than the means, but the P values were virtually the same.
Mean values for MAXDIFFhead, MINDIFFhead, SignedVShead, MAXDIFFpelvis, and MINDIFFpelvis generated by a PISBS for lameness detection and quantification in 20 horses evaluated with the gyroscope of the PISBS alternately attached to the right forelimb pastern region, in the standard upright position (black bars) or to the left forelimb pastern region and flipped 180° on the frontal plane relative to the standard position (white bars) in 2 consecutive lameness trials. The order of the trials was randomized.
Citation: Journal of the American Veterinary Medical Association 259, 8; 10.2460/javma.259.8.892
Mean values for MAXDIFFhead, MINDIFFhead, SignedVShead, MAXDIFFpelvis, and MINDIFFpelvis generated by a PISBS for lameness detection and quantification in 20 horses evaluated with the gyroscope of the PISBS alternately attached to the right forelimb pastern region, in the standard upright position (black bars) or to the left forelimb pastern region and flipped 180° on the frontal plane relative to the standard position (white bars) in 2 consecutive lameness trials. The order of the trials was randomized.
Citation: Journal of the American Veterinary Medical Association 259, 8; 10.2460/javma.259.8.892
Mean values for MAXDIFFhead, MINDIFFhead, SignedVShead, MAXDIFFpelvis, and MINDIFFpelvis generated by a PISBS for lameness detection and quantification in 20 horses evaluated with the gyroscope of the PISBS alternately attached to the right forelimb pastern region, in the standard upright position (black bars) or to the left forelimb pastern region and flipped 180° on the frontal plane relative to the standard position (white bars) in 2 consecutive lameness trials. The order of the trials was randomized.
Citation: Journal of the American Veterinary Medical Association 259, 8; 10.2460/javma.259.8.892
Results of repeated-measures ANOVA showing the effects of forelimb instrumentation approach (gyroscope of the PISBS attached to the right forelimb pastern region, in the standard upright position or to the left forelimb pastern region and flipped 180° on the frontal plane relative to the standard position) on 5 lameness variables generated by a PISBS in 20 horses in 2 consecutive trials.
| Variable | Instrumented forelimb | Mean ± SD | Mean ± SD difference between limbs | P value* |
|---|---|---|---|---|
| MAXDIFFhead | Left | –2.39 ± 15.76 | –0.11 ± 4.65 | 0.92 |
| Right | –2.28 ± 13.67 | — | — | |
| MINDIFFhead | Left | –1.99 ± 28.13 | 1.63 ± 6.86 | 0.31 |
| Right | –3.63 ± 26.89 | — | — | |
| SignedVShead | Left | 22.55 ± 22.67 | –0.88 ± 6.41 | 0.66 |
| Right | 23.21 ± 19.03 | — | — | |
| MAXDIFFpelvis | Left | 1.90 ± 8.42 | 0.41 ± 2.24 | 0.41 |
| Right | 1.49 ± 7.80 | — | — | |
| MINDIFFpelvis | Left | –0.25 ± 5.85 | 0.05 ± 1.77 | 0.89 |
| Right | –0.30 ± 6.09 | — | — |
The sign of each variable except MAXDIFFhead indicates the side of lameness (negative value, left limb; positive value, right limb).
Values of P > 0.05 do not support the alternative hypothesis that there was an effect of forelimb instrumentation approach.
— = Not applicable.
The order of trials was randomized so that the right forelimb was instrumented first in 10 horses.
Bland-Altman plots of agreement between values obtained with the gyroscope on the right versus left forelimb for MAXDIFFhead, MINDIFFhead, SignedVShead, MAXDIFFpelvis, and MINDIFFpelvis in the horses of Figure 1. The sign of variables other than MAXDIFFhead indicates the side of lameness (negative value, left limb; positive value, right limb). Within each plot, solid lines parallel to the x-axis represent the 95% limits of agreement of the difference between the 2 conditions, the thick dashed line represents the mean difference between the 2 conditions, the thin dashed lines represent the 95% CI of the mean difference between the 2 conditions, and the dotted lines represent the lameness thresholds according to the PISBS manual.
Citation: Journal of the American Veterinary Medical Association 259, 8; 10.2460/javma.259.8.892
Bland-Altman plots of agreement between values obtained with the gyroscope on the right versus left forelimb for MAXDIFFhead, MINDIFFhead, SignedVShead, MAXDIFFpelvis, and MINDIFFpelvis in the horses of Figure 1. The sign of variables other than MAXDIFFhead indicates the side of lameness (negative value, left limb; positive value, right limb). Within each plot, solid lines parallel to the x-axis represent the 95% limits of agreement of the difference between the 2 conditions, the thick dashed line represents the mean difference between the 2 conditions, the thin dashed lines represent the 95% CI of the mean difference between the 2 conditions, and the dotted lines represent the lameness thresholds according to the PISBS manual.
Citation: Journal of the American Veterinary Medical Association 259, 8; 10.2460/javma.259.8.892
Bland-Altman plots of agreement between values obtained with the gyroscope on the right versus left forelimb for MAXDIFFhead, MINDIFFhead, SignedVShead, MAXDIFFpelvis, and MINDIFFpelvis in the horses of Figure 1. The sign of variables other than MAXDIFFhead indicates the side of lameness (negative value, left limb; positive value, right limb). Within each plot, solid lines parallel to the x-axis represent the 95% limits of agreement of the difference between the 2 conditions, the thick dashed line represents the mean difference between the 2 conditions, the thin dashed lines represent the 95% CI of the mean difference between the 2 conditions, and the dotted lines represent the lameness thresholds according to the PISBS manual.
Citation: Journal of the American Veterinary Medical Association 259, 8; 10.2460/javma.259.8.892
Results of repeated-measures ANOVA showing the effects of trial order (first or second) on values for 5 lameness variables for the horses of Table 1.
| Variable | Trial order | Mean ± SD | Mean ± SD difference between trials | P value |
|---|---|---|---|---|
| MAXDIFFhead | 1 | –2.20 ± 13.37 | 0.27 ± 4.64 | 0.81 |
| 2 | –2.46 ± 16.02 | — | — | |
| MINDIFFhead | 1 | –2.72 ± 26.60 | 0.17 ± 7.06 | 0.91 |
| 2 | –2.90 ± 28.43 | — | — | |
| SignedVShead | 1 | 22.75 ± 18.83 | –0.27 ± 6.39 | 0.86 |
| 2 | 23.02 ± 22.85 | — | — | |
| MAXDIFFpelvis | 1 | 1.34 ± 7.67 | –0.70 ± 2.17 | 0.17 |
| 2 | 2.04 ± 8.53 | — | — | |
| MINDIFFpelvis | 1 | –0.62 ± 6.14 | –0.68 ± 1.63 | 0.85 |
| 2 | 0.06 ± 5.78 | — | — |
See Table 1 for key.
Bland-Altman plots of agreement between values measured during the first versus second trials for MAXDIFFhead, MINDIFFhead, SignedVShead, MAXDIFFpelvis, and MINDIFFpelvis in the horses of Figure 1. See Figure 2 for remainder of key.
Citation: Journal of the American Veterinary Medical Association 259, 8; 10.2460/javma.259.8.892
Bland-Altman plots of agreement between values measured during the first versus second trials for MAXDIFFhead, MINDIFFhead, SignedVShead, MAXDIFFpelvis, and MINDIFFpelvis in the horses of Figure 1. See Figure 2 for remainder of key.
Citation: Journal of the American Veterinary Medical Association 259, 8; 10.2460/javma.259.8.892
Bland-Altman plots of agreement between values measured during the first versus second trials for MAXDIFFhead, MINDIFFhead, SignedVShead, MAXDIFFpelvis, and MINDIFFpelvis in the horses of Figure 1. See Figure 2 for remainder of key.
Citation: Journal of the American Veterinary Medical Association 259, 8; 10.2460/javma.259.8.892
Results of correlation analysis of data from the 2 trials (grouped by instrumentation approach and trial order) suggested good agreement between the 2 instrumentation approaches and trial orders (Supplementary Figures S1 and S2). The slopes were close to 1 (range, 0.83 to 1.16), and the absolute value of the intercept was consistently < 30% of the threshold values recommended by the PISBS manufacturer for each of the evaluated variables. Furthermore, all coefficients of determination were > 0.91, indicating that > 91% of the variability of one trial could be explained by the variability of the other trial.
Bland-Altman plots revealed no evidence of consistent bias associated with forelimb instrumentation approach (Figure 2) or trial order (Figure 3) for any variable. For all variables, the mean difference between the 2 trials was within the normal range for the PISBS (ie, absolute value of the difference < threshold), and the 95% CIs of the mean difference between the trials included 0.
In 6 horses, mild hind limb lameness of 1 type (impact or push-off; MINDIFFpelvis or MAXDIFFpelvis slightly above threshold [ie, < 2 times threshold]) was found in one trial but not in the other. In 1 of those 6 horses, left forelimb lameness was found in one trial but not in the other (Figure 4). When the horses were classified as lame or sound, perfect agreement between the 2 trials was observed for 19 of 20 horses (95%; 95% CI, 75.1% to 99.9%; P < 0.001) regarding forelimb lameness, 17 of 20 horses (85%; 95% CI, 62.1% to 96.8%; P < 0.002) regarding hind limb impact lameness, and 17 of 20 horses (85%; 95% CI, 62.1% to 96.8%; P < 0.002) regarding hind limb push-off lameness. These proportions of agreement were all significantly (ie, P < 0.004 per adjustment for multiple comparisons) different from 50%. The same results were found when the effect of order of lameness evaluation was investigated (Figure 4). In all cases, disagreement between evaluations was the result of detection of lameness in only 1 trial, and in none of the horses did forelimb or hind limb lameness switch from one limb to the contralateral limb. The power to detect an effect of limb instrumentation approach or trial order on any variable of a magnitude equal to its threshold was at least 96% and 95%, respectively.
Graphical representation of lameness as detected by the PISBS in the horses of Figure 1. The results of each lameness evaluation are represented as a square. Top quadrants represent forelimbs, bottom quadrants represent hind limbs, left quadrants represent left limbs, and right quadrants represent right limbs. The 2 components of hind limb lameness are represented as halves of each bottom quadrant (bottom half, impact lameness; top half, push-off lameness). Shaded fields indicate lameness detected by the PISBS. Pairs of lameness trials separated by an arrow indicate that a significant (P < 0.004) difference was identified between the first and the second trials. The direction of the arrow indicates the order of the paired lameness trials.
Citation: Journal of the American Veterinary Medical Association 259, 8; 10.2460/javma.259.8.892
Graphical representation of lameness as detected by the PISBS in the horses of Figure 1. The results of each lameness evaluation are represented as a square. Top quadrants represent forelimbs, bottom quadrants represent hind limbs, left quadrants represent left limbs, and right quadrants represent right limbs. The 2 components of hind limb lameness are represented as halves of each bottom quadrant (bottom half, impact lameness; top half, push-off lameness). Shaded fields indicate lameness detected by the PISBS. Pairs of lameness trials separated by an arrow indicate that a significant (P < 0.004) difference was identified between the first and the second trials. The direction of the arrow indicates the order of the paired lameness trials.
Citation: Journal of the American Veterinary Medical Association 259, 8; 10.2460/javma.259.8.892
Graphical representation of lameness as detected by the PISBS in the horses of Figure 1. The results of each lameness evaluation are represented as a square. Top quadrants represent forelimbs, bottom quadrants represent hind limbs, left quadrants represent left limbs, and right quadrants represent right limbs. The 2 components of hind limb lameness are represented as halves of each bottom quadrant (bottom half, impact lameness; top half, push-off lameness). Shaded fields indicate lameness detected by the PISBS. Pairs of lameness trials separated by an arrow indicate that a significant (P < 0.004) difference was identified between the first and the second trials. The direction of the arrow indicates the order of the paired lameness trials.
Citation: Journal of the American Veterinary Medical Association 259, 8; 10.2460/javma.259.8.892
Results for horses with at least 1 form of lameness detected in 1 trial but not in the other trial were summarized (Supplementary Table S1).
Discussion
In the present study, conducted in average-size horses, no effect of limb instrumentation approach or trial order could be detected on any of the evaluated variables generated by the PISBS or on the classification of horses as being sound or exhibiting forelimb lameness, hind limb push-off lameness, or hind limb impact lameness affecting the left or the right limb or limbs. Neither a consistent pattern of shifting lameness from left to right or vice versa (eg, as would occur if the sensor caused lameness on the instrumented limb) nor consistent lameness exacerbation or attenuation associated with the instrumented forelimb was observed. The inconsistent within-subject disagreement between the results of repeated lameness evaluations (ie, the 2 trials) in horses with mild lameness was not unexpected. Such variability has been previously documented when 2 consecutive evaluations were compared with the PISBS with the gyroscope positioned as recommended by the manufacturer.16
In the present study, results of correlation analysis indicated good agreement between results for the 2 limb instrumentation approaches or trial orders. Observed coefficients of determination for MINDIFFhead, MAXDIFFhead, MINDIFFpelvis, and MAXDIFFpelvis (all > 0.91) were comparable to the coefficients of determination observed in a previous study16 (range, 0.78 to 0.91) involving paired evaluations with the PISBS in 236 horses. Unfortunately, the correlation coefficient or the coefficient of determination for SignedVShead was not reported in the previous study.16
When results were grouped by instrumentation approach or trial order in the present study, the Bland-Altman plots revealed no evidence of any consistent bias associated with these 2 sets of groups. For all 5 of the evaluated variables, the absolute value of the mean difference between the groups was always close to 0 and < 28% of the threshold values recommended by the PISBS manufacturer for the variables. When the horses were classified on the basis of whether they had or did not have forelimb lameness, hind limb push-off lameness, and hind limb impact lameness, the disagreement observed for only 6 of the 20 horses was considered mild. Actually, when the proportions of agreement for each lameness type were investigated, all proportions were larger than the expected proportion of agreement by chance (ie, 50%).
The main reason for the lack of a perfect agreement between 2 consecutive lameness trials with the PISBS was likely that horses did not move exactly the same way during each trial. This lack of perfect repeatability between consecutive trials with the PISBS is well known16 and was clearly demonstrated (Figure 1). Factors such as voluntary changes in motion by the horse in response to environmental stimuli (eg, noise or visual contact with another horse), inconsistencies in how the handlers trotted the horses (eg, inconsistent speed or distance), and changes in lameness over time in response to exercise (either lameness improvement or worsening) may explain this disagreement. Standardization of the approach used to trot horses, including trotting all horses on a straight line on the same aisle with uniform asphalt floor and having the same person trot each horse during both lameness trials as performed in the present study, may have helped to improve agreement between the results of consecutive lameness trials with the PISBS.
Because no effect of limb instrumentation approach was observed, we concluded that, if there was any change in vertical excursion of head or pelvis caused by limb instrumentation, this effect was likely very transient or simply small enough not to interfere with the PISBS evaluation. For the PISBS, the gyroscope attached to the forelimb does not have any function other than to quantify stride frequency and time the stance phases of the instrumented limb. This information is then used by the PISBS to infer the stance phase of each diagonal limb pair when the horse is trotting.12 Therefore, even if the gyroscope attached to a forelimb increases foot elevation as produced by tactile stimulation of the pastern region with a device with a weight comparable to the weight of the gyroscope, this should not be expected to affect lameness detection and quantification with the PISBS. Although a change in foot flight might have affected the vertical motion of the head or pelvis, which is used by the PISBS for lameness detection and quantification, if this happened in the present study, its effect was not large or consistent enough to produce any consistent changes on the findings of the PISBS.
The findings of the present study suggested that the use of a lightweight sensor attached to a forelimb for lameness evaluation with the PISBS does not produce false lameness or attenuate lameness in average-size horses. The authors have used the PISBS to evaluate hundreds of horses and have never observed any evidence of false lameness attributable to instrumentation of a forelimb with the sensor in any horse. Even signs of mild intolerance to the wrap holding the gyroscope to the right forelimb pastern region are rarely seen. When this happened, the signs resolved after the horse had taken a few steps.
In the study reported here, instrumentation of 1 forelimb with a lightweight gyroscope of a PISBS had no effects on lameness detection or quantification with the PISBS. Findings suggested that the gyroscope can be attached either to the right forelimb, on the standard position as recommended by the manufacturer of the PISBS, or to the left forelimb and flipped 180° on the frontal plane relative to the standard position. The validity of the findings of this study for horses of different body sizes, especially for horses of very small breeds and for foals, deserves investigation.
Supplementary Materials
Supplementary materials are posted online at the journal website: avmajournals.avma.org.
Acknowledgments
Funded by the E. Paige Laurie Endowed Program in Equine Lameness, College of Veterinary Medicine, University of Missouri.
The E. Paige Laurie Endowed Program in Equine Lameness is led by Dr. Kevin Keegan, who is also the leading inventor of the PISBS used in this study and a founder and shareholder of Equinosis LLC, the manufacturer of the PISBS used in this study. Dr. Keegan played no role in the study design; collection, analysis, and interpretation of data; or decision to submit the manuscript for publication.
Preliminary results presented at the 38th Annual Phi Zeta Research Day, College of Veterinary Medicine, University of Missouri, May 2015.
Footnotes
Lameness Locator, Equinosis LLC, St Louis, Mo.
Excel, Office 365, Microsoft Corp, Redmond, Wash.
SAS university edition, version 2018, SAS Institute Inc, Cary, NC.
Abbreviations
| MAXDIFFhead | Mean of the differences between maximal head heights after stance of left forelimb and stance of right forelimb |
| MAXDIFFpelvis | Mean of the differences between maximal pelvic heights after stance of left hind limb and stance of right hind limb |
| MINDIFFhead | Mean of the differences between minimal head heights during stance of right forelimb and stance of left forelimb |
| MINDIFFpelvis | Mean of the differences between minimal pelvic heights during stance of right hind limb and stance of left hind limb |
| PISBS | Portable inertial sensor–based system |
| SignedVShead | Vector sum of MINDIFFhead and MAXDIFFhead with the sign of MINDIFFhead (ie, positive, right forelimb lameness; negative, left forelimb lameness) |
References
- 1. ↑
Caron JP. Objective and subjective gait analysis techniques. In: White NA, Moore JN, eds. Current techniques in equine surgery and lameness. 2nd ed. Philadelphia: WB Saunders Co, 1998;501–504.
- 2. ↑
Fuller CJ, Bladon BM, Driver AJ, et al. The intra- and inter-assessor reliability of measurement of functional outcome by lameness scoring in horses. Vet J 2006;171:281–286.
- 3. ↑
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.
- 4. ↑
Hewetson M, Christley RM, Hunt ID, et al. Investigations of the reliability of observational gait analysis for the assessment of lameness in horses. Vet Rec 2006;158:852–857.
- 5. ↑
Keegan KG, Wilson DA, Wilson DJ, et al. Evaluation of mild lameness in horses trotting on a treadmill by clinicians and interns or residents and correlation of their assessments with kinematic gait analysis. Am J Vet Res 1998;59:1370–1377.
- 6. ↑
Keegan KG, Dent EV, Wilson DA, et al. Repeatability of subjective evaluation of lameness in horses. Equine Vet J 2010;42:92–97.
- 7. ↑
McCracken MJ, Kramer 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.
- 8. ↑
Arkell M, Archer RM, Guitian FJ, et al. Evidence of bias affecting the interpretation of the results of local anaesthetic nerve blocks when assessing lameness in horses. Vet Rec 2006;159:346–349.
- 9. ↑
Riber C, Cuesta I, Munoz A, et al. Equine locomotor analysis on vet-gates in endurance events. Equine Vet J Suppl 2006;36:55–59.
- 10. ↑
Keegan KG. Evidence-based lameness detection and quantification. Vet Clin North Am Equine Pract 2007;23:403–423.
- 11. ↑
Kramer J, Keegan KG. Kinematics of lameness. In: Hinchcliff KW, Kaneps AJ, Geor RJ, eds. Equine sports medicine and surgery. 2nd ed: Saunders Elsevier, 2013;223–238.
- 13. ↑
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.
- 14. ↑
Lopes MAF, Eleuterio A, Mira MC. Objective detection and quantification of irregular gait with a portable inertial sensor-based system in horses during an endurance race—a preliminary assessment. J Equine Vet Sci 2018;70:123–129.
- 15. ↑
Curran-Everett D. Multiple comparisons: philosophies and illustrations. Am J Physiol Regul Integr Comp Physiol 2000;279:R1–R8.
- 16. ↑
Keegan KG, Kramer J, Yonezawa Y, et al. Assessment of repeatability of a wireless, inertial sensor–based lameness evaluation system for horses. Am J Vet Res 2011;72:1156–1163.
