• 1. Bockstahler BA, Vobornik A, Müller M, et al. Compensatory load redistribution in naturally occurring osteoarthritis of the elbow joint and induced weight-bearing lameness of the forelimbs compared with clinically sound dogs. Vet J 2009;180:202212.

    • Search Google Scholar
    • Export Citation
  • 2. Brebner NS, Moens NMM, Runciman JR. Evaluation of a treadmill with integrated force plates for kinetic gait analysis of sound and lame dogs in a trot. Vet Comp Orthop Traumatol 2006;19:205212.

    • Search Google Scholar
    • Export Citation
  • 3. Fanchon L, Grandjean D. Accuracy of asymmetry indices of ground reaction forces for diagnosis of hindlimb lameness in dogs. Am J Vet Res 2007;68:10891094.

    • Search Google Scholar
    • Export Citation
  • 4. LeQuang T, Maitre P, Roger T, et al. Is a pressure walkway system able to highlight a lameness in dogs? J Anim Vet Adv 2009;8:19361944.

    • Search Google Scholar
    • Export Citation
  • 5. Gillette RL, Angle TC. Recent developments in canine locomotor analysis: a review. Vet J 2008;178:165176.

  • 6. Lee DV, Bertram JE, Todhunter RJ, et al. Force overlap in trotting dogs: a Fourier technique for reconstructing individual limb ground reaction force. Vet Comp Orthop Traumatol 2002;15:223227.

    • Search Google Scholar
    • Export Citation
  • 7. Quinn MM, Keuler NS, Lu Y, et al. Evaluation of agreement between numerical rating scales, visual analogue scoring scales, and force platform analysis in dogs. Vet Surg 2007;36:360367.

    • Search Google Scholar
    • Export Citation
  • 8. Jevens DJ, Hauptman JG, DeCamp CE, et al. Contributions to variance in force-plate analysis of gait in dogs. Am J Vet Res 1993;54:612615.

    • Search Google Scholar
    • Export Citation
  • 9. Nordquist B, Fischer J, Kim SI, et al. Effects of trial repetition, limb side, intraday and inter-week variation on vertical and craniocaudal ground reaction forces in clinically normal Labrador Retrievers. Vet Comp Orthop Traumatol 2011;24:435444.

    • Search Google Scholar
    • Export Citation
  • 10. Rumph PF, Lander JE, Kincaid SA, et al. Ground reaction force profiles from force platform gait analysis of clinically normal mesomorphic dogs at the trot. Am J Vet Res 1994;55:756761.

    • Search Google Scholar
    • Export Citation
  • 11. Bertram JE, Lee DV, Case HN, et al. Comparison of the trotting gaits of Labrador Retrievers and Greyhounds. Am J Vet Res 2000;61:832838.

    • Search Google Scholar
    • Export Citation
  • 12. Bockstahler BA, Skalicky M, Peham C, et al. Reliability of ground reaction forces measured on a treadmill system in healthy dogs. Vet J 2007;173:373378.

    • Search Google Scholar
    • Export Citation
  • 13. Clements DN, Owen MR, Carmichael S, et al. Kinematic analysis of the gait of 10 Labrador Retrievers during treadmill locomotion. Vet Rec 2005;156:478481.

    • Search Google Scholar
    • Export Citation
  • 14. Fanchon L, Valette JP, Sanaa M, et al. The measurement of ground reaction force in dogs trotting on a treadmill. Vet Comp Orthop Traumatol 2006;19:8186.

    • Search Google Scholar
    • Export Citation
  • 15. Fanchon L, Grandjean D. Habituation of healthy dogs to treadmill trotting: repeatability assessment of vertical ground reaction force. Res Vet Sci 2009;87:135139.

    • Search Google Scholar
    • Export Citation
  • 16. Kapatkin AS, Arbittier G, Kass PH, et al. Kinetic gait analysis of healthy dogs on two different surfaces. Vet Surg 2007;36:605608.

  • 17. Keebaugh AE, Redman-Bentley D, Griffon DJ. Influence of leash side and handlers on pressure mat analysis of gait characteristics in small-breed dogs. J Am Vet Med Assoc 2015;246:12151221.

    • Search Google Scholar
    • Export Citation
  • 18. LaFlamme D. Development and validation of a body condition score system for dogs. Canine Pract 1997;22:1015.

  • 19. Faul F, Erdfelder E, Lang AG, et al. G*Power 3: a flexible statistical power analysis program for the social, behavioral, and biomedical sciences. Behav Res Methods 2007;39:175191.

    • Search Google Scholar
    • Export Citation
  • 20. Allen K, DeCamp CE, Braden TD, et al. Kinematic gait analysis of the trot in healthy mixed breed dogs. Vet Comp Orthop Traumatol 1994;7:148153.

    • Search Google Scholar
    • Export Citation
  • 21. Hottinger HA, DeCamp CE, Olivier NB, et al. Noninvasive kinematic analysis of the walk in healthy large-breed dogs. Am J Vet Res 1996;57:381388.

    • Search Google Scholar
    • Export Citation
  • 22. Robinson RO, Herzog W, Nigg BM. Use of force platform variables to quantify the effects of chiropractic manipulation of gait symmetry. J Manipulative Physiol Ther 1987;10:172176.

    • Search Google Scholar
    • Export Citation
  • 23. Nakagawa S, Cuthill IC. Effect size, confidence interval and statistical significance: a practical guide for biologists. Biol Rev 2007;82:591605.

    • Search Google Scholar
    • Export Citation
  • 24. Hanley JA, Negassa A, Edwardes MD, et al. Statistical analysis of correlated data using generalized estimating equations: an orientation. Am J Epidemiol 2003;157:364375.

    • Search Google Scholar
    • Export Citation
  • 25. Bland M. An introduction to medical statistics. 3rd ed. Oxford, England: Oxford University Press, 2015;315317.

  • 26. Besancon MF, Conzemius MG, Derrick TR, et al. Comparison of vertical forces in normal greyhounds between force platform and pressure walkway measurement systems. Vet Comp Orthop Traumatol 2003;16:153157.

    • Search Google Scholar
    • Export Citation
  • 27. Voss K, Wiestner T, Galeandro L, et al. Effect of dog breed and body conformation on vertical ground reaction forces, impulses, and stance times. Vet Comp Orthop Traumatol 2011;24:106112.

    • Search Google Scholar
    • Export Citation
  • 28. Mölsä SH, Hielm-Björkman AK, Laitinen-Vapaavuori OM. Force platform analysis in clinically healthy Rottweilers: comparison with Labrador Retrievers. Vet Surg 2010;39:701707.

    • Search Google Scholar
    • Export Citation
  • 29. Voss K, Galeandro L, Wiestner T, et al. Relationships of body weight, body size, subject velocity, and vertical ground reaction forces in trotting dogs. Vet Surg 2010;39:863869.

    • Search Google Scholar
    • Export Citation
  • 30. Roy WE. Examination of the canine locomotion system. Vet Clin North Am 1971;1:5370.

  • 31. Drüen S, Böddeker J, Nolte I, et al. Bodenreaktionskräfte der caninen Hintergliedmaße: Gibt es Unterschiede beim Gang auf Laufband und Kraftmessplatte? Berl Munch Tierarztl Wochenschr 2010;123:339345.

    • Search Google Scholar
    • Export Citation
  • 32. Roush JK, McLaughlin RM Jr. Effects of subject stance times and velocity on ground reaction forces in clinically normal Greyhounds at the walk. Am J Vet Res 1994;55:16721676.

    • Search Google Scholar
    • Export Citation

Advertisement

Reliability of kinetic measurements of healthy dogs examined while walking on a treadmill

View More View Less
  • 1 1Clinic for Small Animal Surgery and Reproduction, Ludwig-Maximilians-Universität München, 80539 Munich, Germany.
  • | 2 2Department of Veterinary Sciences, Ludwig-Maximilians-Universität München, 80539 Munich, Germany.

Abstract

OBJECTIVE

To investigate whether an actual improvement in gait could be differentiated from physiologic differences or habituation effects during gait analysis of dogs.

ANIMALS

11 healthy dogs.

PROCEDURES

On 4 examination days, kinetic parameters were measured while dogs were walking on a treadmill. Differences in mean parameter values and habituation effects (ie, effect sizes) were quantified and compared among examination days. Coefficients of variation for repeated measurements were calculated to determine measurement reproducibility, and minimum differences were calculated to distinguish between physiologic fluctuation and an actual change in gait pattern.

RESULTS

Among the 4 examination days, mean absolute differences in peak vertical force and vertical impulse (VI) varied from 1.5% to 5.3% of body weight (BW) and 0.9% to 1.8% of BW·s, respectively. Mean absolute differences in the percentage of stance-phase duration (%SPD) and relative stride length (RSL) varied from 0.9% to 3.2% and 1.7% to 3.0%, respectively. Reproducibility of parameter measurements was good. Values for %SPD had the lowest amount of dispersion and largest effect size, suggesting a habituation effect for this parameter. Calculated minimum differences among the days for peak vertical force, VI, %SPD, and RSL did not exceed 9.9% of BW, 3.3% of BW·s, 5.8 percentage points, and 5.2 percentage points, respectively.

CONCLUSIONS AND CLINICAL RELEVANCE

The %SPD of healthy dogs walking on a treadmill was the most sensitive and diagnostically reliable of the measured kinetic parameters, in contrast to VI and RSL. Findings suggested that actual changes can be distinguished from random physiologic fluctuations during gait analysis of dogs.

Abstract

OBJECTIVE

To investigate whether an actual improvement in gait could be differentiated from physiologic differences or habituation effects during gait analysis of dogs.

ANIMALS

11 healthy dogs.

PROCEDURES

On 4 examination days, kinetic parameters were measured while dogs were walking on a treadmill. Differences in mean parameter values and habituation effects (ie, effect sizes) were quantified and compared among examination days. Coefficients of variation for repeated measurements were calculated to determine measurement reproducibility, and minimum differences were calculated to distinguish between physiologic fluctuation and an actual change in gait pattern.

RESULTS

Among the 4 examination days, mean absolute differences in peak vertical force and vertical impulse (VI) varied from 1.5% to 5.3% of body weight (BW) and 0.9% to 1.8% of BW·s, respectively. Mean absolute differences in the percentage of stance-phase duration (%SPD) and relative stride length (RSL) varied from 0.9% to 3.2% and 1.7% to 3.0%, respectively. Reproducibility of parameter measurements was good. Values for %SPD had the lowest amount of dispersion and largest effect size, suggesting a habituation effect for this parameter. Calculated minimum differences among the days for peak vertical force, VI, %SPD, and RSL did not exceed 9.9% of BW, 3.3% of BW·s, 5.8 percentage points, and 5.2 percentage points, respectively.

CONCLUSIONS AND CLINICAL RELEVANCE

The %SPD of healthy dogs walking on a treadmill was the most sensitive and diagnostically reliable of the measured kinetic parameters, in contrast to VI and RSL. Findings suggested that actual changes can be distinguished from random physiologic fluctuations during gait analysis of dogs.

Instrumented gait analysis allows objective lameness diagnosis in dogs,1–5 enabling precise recording of force development and movement sequences unappreciated by the naked eye.5–7 Repeated gait analysis on a treadmill can provide objective data before and after an intervention and for patient follow-up. An improvement in kinetic measurements may be attributed solely to the success of an intervention; however, the extent to which any observed differences may be attributable to intraindividual physiologic fluctuations or to systematic training or habituation effects is unknown.

Studies regarding the reproducibility of kinetic gait measurements of dogs have typically involved the use of stationary force plates8–11 and treadmills.3,12–15 In these investigations, PVFs were lower when a treadmill was used, whereas VIs were typically higher when stationary force plates were used.9,11,12 Variables such as soil conditions,16 the person leading the dog or changes in this person,8,17 and the side of leash guidance (left or right)17 have no significant effect on gait measurements. In addition, the influence of data collection several times within a brief period (eg, on consecutive days12 or on the morning and evening of the same day16) or within several days or weeks3,9,14,15 on results of gait analysis has been evaluated, and fluctuations in measurements were observed between daily and weekly intervals, some of which were of different magnitudes.3,14

Cutoff values remain to be determined that could distinguish actual changes in gait patterns from those due to habituation, individual experience, and physiologic fluctuations. In 2008, Fanchon and Grandjean3 determined a cutoff value for a combination of PVF and the stress curve of a limb, whereby reliable differentiation (ie, sensitivity of 85% and specificity of 95%) was possible between lameness and soundness at values > 19.5%. The purpose of the study reported here was to determine normal cutoff values for parameters measured by instrumented gait analysis after examination of these parameters for their reproducibility (ie, intraindividual variability) and for systematic habituation effects.

Materials and Methods

Animals

Eleven nonlame dogs of various breeds evaluated at the Clinic for Small Animal Surgery and Reproduction, Ludwig-Maximilians-Universität München were included in the study with owner consent. Dogs were required to have a body condition score of 4 or 5 on a 9-point scale.18 None of the dogs had prior experience walking on a treadmill, and each was required to voluntarily walk on the treadmill to be included.

Reasons for initial evaluation included routine physical examination or physical fitness evaluation (eg, for suitability for sport or agility training). Mean age was 4.5 years (range, 1 to 9 years), and mean BW was 23.1 kg (range, 11.9 to 31.0 kg).

The number of dogs included in the study had been predetermined on the basis of a sample size calculation19,a (minimum effect size [Cohen d] > 0.8; α = 0.05; β = 0.20; 1-sided test) and was within the range of numbers of dogs used in other gait analysis projects or studies.8,9,11–13,15–17,20,21 The study protocol did not require approval by an animal care and use committee according to the German Animal Welfare Act and Animal Experimentation Ordinance.

Study protocol

Each dog underwent 4 treadmill examinations, with the first 3 on consecutive days and the final examination 6 weeks later. Before the first examination, each dog received a general clinical examination and an orthopedic examination. Back length and height at the highest point of the shoulders (withers) were measured. Dogs were weighed before each treadmill exposure. After orthopedic examination, radiography of the of the hip, knee, tarsal, shoulder, elbow, and carpal (including the paw) joints and the vertebral column was performed to rule out the presence of orthopedic disease. At the final examination, general clinical and orthopedic examinations were repeated.

Treadmill examination

All treadmill examinations were carried out in the clinic's gait analysis laboratory. The treadmillb consisted of 2 nonslip plastic sheets covering 4 modified force plates, which measured the vertical ground reaction forces of all 4 limbs synchronously at a sampling rate of 1,000 Hz. The treadmill was controlled by dedicated software,c and the speed could be adjusted in increments of 0.02 m/s.

The same person guided each dog by leash and collar to walk. The speed of the treadmill was adapted to the size and ability of each dog, and this speed as determined on the first day of examination was maintained throughout the study. Recordings were made of constant, regular steps, with a walking speed between 1.0 and 1.2 m/s (mean, 1.11 m/s).

During each examination, dogs were first allowed to acclimate to the treadmill for 2 minutes and then 2 trials of 3 uninterrupted minutes each were recorded, separated by a rest period of 2 minutes. As many steps as possible were evaluated from the first minute of each 3-minute trial. Every step was analyzed in graphic and numeric form. Only those steps indicating even and correct foot placement on the force plates were selected and statistically evaluated.

The PVF, VI, and %SPD were recorded for each limb. The RSL ([stride duration × treadmill belt speed]/withers height × 100) was measured for each dog. In addition, the SI was calculated as reported elsewhere22 for PVF and VI, and the weight distribution per limb (ie, the percentage of the total body load carried by each limb) was determined. The cutoff values used to indicate symmetric loading of the limbs were 9 for the SI of the PVF and 10 for the SI of the VI, as determined in other studies.10,12,14,15 If these values were exceeded, lameness was deemed to exist and the dog was to be excluded from the study.

Statistical analysis

Statistical softwared-f was used for data analysis. Mean (SD) values of each measured parameter were calculated for each examination, as were mean values of the differences between pairs of examination days. Normality of data distribution was evaluated with the Kolmogorov-Smirnov test. To determine any systematic training (ie, habituation) effect, values of Cohen d (effect size) were calculated from the mean and SD values of the individual examination days. An effect size < 0.2 was regarded as negligible, from 0.2 to 0.5 as slight, > 0.5 to 0.8 as moderate, and > 0.8 as a clear effect.23 Differences between the various examination days were identified with generalized estimation equations derived from a linear model.24 Values of P ≤ 0.05 were considered significant.

As a measure of interindividual variability in values for PVF, VI, %SPD, and RSL, the CV was determined for data pertaining to all 4 examinations as the SD divided by the mean value. In addition, as a measure of the reproducibility (ie, intraindividual variability) of parameter measurements, the CVRM was calculated per the method described by Bland,25 whereby a value < 3% was regarded as excellent and a value between 3% and 10% as good. The MD for which it could be assumed with 95% probability that an observed difference was due to an actual change in the parameter (and not only to a physiologic variance associated with the individual dog, a training effect, or a measurement inaccuracy) was calculated25 from the SD of the differences over all 4 examination days (ie, SD*) separately for the forelimbs and hind limbs, taking into account whether differences were positive or negative, as SD* × 2.77.

Results

Dogs

Characteristics and measurements of individual dogs were summarized (Supplementary Table S1 and Supplementary Figure S1, available at avmajournals.avma.org/doi/suppl/10.2460/ajvr.81.10.804). A mean of 72 steps/examination day/dog were evaluated. General and orthopedic examinations revealed no abnormalities in any dog both before and at the end of the evaluation period. Three dogs had slight radiographic orthopedic abnormalities, such as mildly pronounced spondylosis or insertion tendopathy of the fifth digit, but were included because they had no clinical signs. Values for SI of the PVF and VI consistently indicated symmetry of gait as defined (Supplementary Table S2, available at avmajournals.avma.org/doi/suppl/10.2460/ajvr.81.10.804). Mean ± SD distribution of BW was balanced between 20.5 ± 2.3% and 20.8 ± 2.4% for the hind limbs and between 29.2 ± 2.4°% and 29.6 ± 2.3°% for the forelimbs. Thus, all dogs were deemed orthopedically normal. Because of a lack of statistically and clinically relevant differences in measured values between the 2 hind limbs and between the 2 forelimbs, values for each limb pair were averaged in further analyses.

Kinetic parameters

Mean ± SD values of kinetic parameters were tabulated (Table 1). The CV for PVF (indicating interindividual variability) was 12.4% for the hind limbs and 10.2% for the forelimbs. Mean values of the absolute differences in PVF among the examination days were low, with a maximum mean value of 2.3% of BW for the hind limbs and 5.3% of BW for the forelimbs (Table 2). The calculated effect sizes (Cohen d; interpreted as habituation effects) were correspondingly low, ranging from 0.02 to 0.28. Measurements differed significantly (P = 0.04) only for the forelimbs on the first versus second examination day. The reproducibility (CVRM) of PVF measurements (indicating intraindividual variability) was also good. Accordingly, the MD in PVF measurements obtained through repeated gait analysis indicated a low amount of systematic change at 5.2% of BW for the hind limbs and 9.9% of BW for the forelimbs.

Table 1—

Mean ± SD values of kinetic parameters and CV (interindividual variability) and CVRM (intraindividual variability) for 11 orthopedically normal dogs examined while walking on a treadmill on 3 consecutive days and again 6 weeks later (examination day 4).

  Examination day  
ParameterLocation12 34CV (%)CVRM (%) 
PVF (% of BW)Hind limbs44.1 ± 5.844.5 ± 5.444.6 ± 5.144.8 ± 5.912.44.2
 Forelimbs62.5 ± 6.464.3 ± 6.463.5 ± 6.263.2 ± 6.910.25.6
VI (% of BW·s)Hind limbs12.1 ± 2.911.6 ± 2.411.3 ± 2.511.8 ± 2.622.88.0
 Forelimbs19.6 ± 2.119.4 ± 1.919.4 ± 1.918.9 ± 2.010.26.0
%SPDHind limbs64.2 ± 4.162.6 ± 4.062.1 ± 4.063.1 ± 4.46.63.3
 Forelimbs68.2 ± 1.866.6 ± 1.966.4 ± 2.466.7 ± 2.03.02.4
RSL (%)All limbs52.6 ± 6.652.3 ± 6.952.6 ± 7.252.1 ± 7.013.63.5
Table 2—

Mean ± SD absolute differences between pairs of examination days and effect sizes (Cohen d; interpreted as habituation effects) for the kinetic parameters of Table 1.

  Day 1 vs 2Day 2 vs 3Day 3 vs 4Day I vs 3Day 2 vs 4Day I vs 4   
ParameterLocationMean ± SDCohen dMean ± SDCohen dMean ± SDCohen dMean ± SD Cohen dMean ± SDCohen dMean ± SDCohen d    
PVFHind limbs1.8 ± 1.20.071.5 ± 0.70.022.0 ± 1.70.041.9 ± 1.60.092.3 ± 1.80.052.3 ± 2.10.11   
(% of BW)Forelimbs3.3 ± 3.2*0.283.5 ± 2.50.134.0 ± 2.70.055.1 ± 2.60.165.3 ± 4.10.175.3 ± 2.90.10   
VIHind limbs0.9 ± 0.70.180.9 ± 0.80.121.4 ± 0.90.191.2 ± 0.90.291.3 ± 0.80.081.3 ± 1.10.11   
(% of BW·s)Forelimbs1.3 ± 1.40.091.2 ± 0.90.001.0 ± 0.70.251.1 ± 0.70.091.8 ± 1.60.251.3 ± 1.10.34   
%SPDHind limbs2.6 ± 1.7*0.390.9 ± 0.70.122.5 ± 2.3*0.232.8 ± 1.7*0.492.5 ± 2.40.113.2 ± 2.20.25   
 Forelimbs2.1 ± 1.3*0.861.0 ± 0.50.091.4 ± 1.30.142.1 ± 1.8*0.851.5 ± 1.10.052.3 ± 1.6*0.79   
RSL (%)All limbs1.7 ± 1.60.053.0 ± 3.40.051.8 ± 1.50.072.3 ± 1.70.102.0 ± 1.80.122.2 ± 1.60.07   

Indicated difference between days was significant (P ≤ 0.05).

A Cohen d value < 0.2 was regarded as a negligible effect, from 0.2 to 0.5 as a slight effect, > 0.5 to 0.8 as a moderate effect, and > 0.8 as a clear effect.

Values of VI for the hind limbs varied to a greater degree than did corresponding values for PVF, with the highest CV (22.8%) observed in the study. The CV for the forelimbs was considerably lower at 10.2%. Mean values of absolute differences in VI between pairs of examination days were low, similar to those for PVF, with a maximum mean value of 1.8% of BW·s (Table 2). Low mean differences in VI among examination days led to low effect sizes. This resulted in low effect sizes (maximum Cohen d, 0.34 in the forelimbs) for VI, with mean values statistically similar between pairs of examination days. Values for CVRM indicated good reproducibility of VI measurements (Table 1). The MDs in VI were relatively high at 2.6% of BW·s for the hind limbs and 3.3°% of BW·s for the forelimbs.

Maximum values of %SPD for the hind limbs and forelimbs were observed on the first examination day (Table 1). This parameter had the lowest CVs of the 4 evaluated kinetic parameters, and values of CVRM indicated good to excellent reproducibility. Mean values for %SPD differed significantly between several pairs of examination days (Table 2). The MDs were low at 5.8 percentage points for the hind limbs and 4.5 percentage points for the forelimbs. Compared with the other 3 parameters, values for %SPD generally had the lowest amount of dispersion and largest effect size for each limb pair, suggesting a habituation effect.

High interindividual variability was found in measurements of RSL (CV, 13.6; Table 1). The maximum difference in mean values between 2 examination days was 3.0% (Table 2). Reproducibility of measurements was also good (CVRM, 3.5%). The MD in measurements of RSL was 5.2 percentage points.

Discussion

The aim of the present study was to quantify physiologic variations in vertical ground reaction forces of healthy dogs while walking on a treadmill so that such variations could be distinguished from actual changes in the gait pattern in the future. The study data were sufficient for identifying significant habituation effects and parameter variability as well as for determining MDs to distinguish real changes from fluctuations.

For this study, dogs of various breeds were selected, but only dogs with an ideal body condition were included. Ground reaction forces were normalized to body weight and stride length was normalized to height at the withers, allowing comparison of our results with those of other studies9,12,26 as well as a uniformly distributed body center of mass during gait analysis.27 Previous studies11,28 of gait patterns in different breeds of dogs initially showed an influence of conformation and body weight on kinetic parameters, particularly on the weight distribution between forelimbs and hind limbs.28 However, after normalization of BW, height, and walking speed, only negligible differences remained in the stance phase.28 On the other hand, another study27 revealed clearer influences of dog conformation on kinetic parameters measured during gait analysis, and such differences should be taken into account, especially when comparing dogs or groups of dogs.29 Owing to differences in the position of the center of mass,27 dog groups with similar body conformations are recommended for gait-related research. Nevertheless, absolute standardization of the dog breed or weight class in a study can limit generalizability of the results to clinical settings.17

In the present study, gait analysis was performed on 3 consecutive days and again 6 weeks later to consider whether short-term and medium-term effects of treadmill walking might exist. Until now, only studies with brief intervals between gait analyses of dogs (eg, once or several times a day12–14) or longer intervals (eg, once a week3,9) have been reported.

The percentage weight distribution per limb in the present study was almost exactly the same as described by Roy in 1971.30 Furthermore, values for PVF in the present study corresponded to previously reported results for dogs (eg, mean PVF of 39.3% to 42.3% of BW for the hind limbs and 61.6% to 63.5% of BW for the forelimbs12,26). Absolute differences in PVF among examination days were in the range of other reported results.9 We believe that the parameter values that we obtained could be used as a reference for clinical application or for other studies because the included dogs were confirmed to be orthopedically healthy on the basis of orthopedic and radiographic findings as well as kinetic parameters (ie, weight distribution, SI of the PVF, and SI of the VI).

Values for VI in the study reported here were within the range of the values obtained in another study31 involving orthopedically healthy dogs, in which the mean VI for the hind limbs was 10.2% of BW·s. For various reasons, including experimental and analytic conditions, values for VI and differences in VI among examination days can differ among studies,3,9,12,14 making comparisons among studies difficult.

The CVRM for PVF and VI measurements in the present study was consistently < 10% and for %SPD (hind limbs) and RSL was only slightly above 3% (with values < 3% regarded as excellent25). Hence, reproducibility of the measured parameters was good and sufficient for diagnostic purposes. Contrarily, the relatively high CVs indicated higher variability among dogs. Potential causes of data dispersion include effects of the investigator or observer, the subject, and the examination method. In the present study, treadmill examinations were always conducted in the same room and at the same time of day. Each dog was walked at its preferred speed as identified initially, and the person leading the dog was the same throughout the study. Therefore, the observed data dispersion was likely attributable to the dogs themselves, which would agree with results of other studies8,12,14,15 indicating that the dog is the most influential factor (up to 88% of influence) on the reproducibility of kinetic gait measurements.

Significant habituation effects were observed in the present study for %SPD and PVF. These findings suggested the presence of a short-term habituation effect after the first examination day, which persisted over the second and third examination days. In another study9 of clinically normal dogs, clearer differences from the first examination day were also found.

A numeric decrease in %SPD occurred over the first 3 examination days. The longer %SPD on the first examination day could have indicated that the dog held its paws on the ground longer than on subsequent days, imperceptible to observers, owing to the unfamiliar situation and leading to a normal step rhythm on the second examination day. The high dispersion of RSL values and the absence of significant effect sizes between the first and second, and third and fourth, examination days indicated that this parameter was probably least influenced by habituation effects.

Differences in values of individual parameters were observed between the hind limbs and forelimbs of dogs in the present study. Such differences might be explained by the different function of the limb pairs in that the forelimbs have predominantly a support function and the hind limbs have predominantly a thrust function.32

The MDs for PVF, VI, and %SPD in the present study pertained to a single limb evaluated over several examinations. Previous reports3,22 of qualitative diagnosis of lameness in dogs provide findings of side (right and left limb) comparisons rather than forelimb and hind limb comparisons, and the MDs determined in our study could supplement those findings. The lowest MDs were obtained for parameters with low measurement variation, such as %SPD. In contrast, VI in particular had relatively large variance and high MDs. Achieving the determined MDs as reference points would likely be possible for dogs with moderate lameness because those dogs have a certain potential for improvement and deterioration. However, for dogs with mild or severe lameness, a change in parameters that had higher MDs in our study, such as PVF and VI, would need to be even higher than those MDs for actual change to be appreciated. The combination of biomechanically completely independent parameters, such as the inclusion of kinematic parameters in the evaluation criteria, could help to increase confidence when gauging whether a given dog has had an improvement or deterioration of limb function between 2 or more examinations. Because the present study involved healthy dogs, additional research is needed to further explore the clinical usefulness of the findings reported here. To identify changes in kinetic parameters as genuine changes in the gait pattern, such changes must exceed physiologic variability, and such variability should be taken into account in follow-up examinations of gait in dogs.

Acknowledgments

This manuscript represents a portion of a dissertation submitted by Mr. Pietsch to the Clinic for Small Animal Surgery and Reproduction, Ludwig-Maximilians-Universität München as partial fulfillment of the requirements for a Doctor of Veterinary Medicine degree.

No third-party funding or support was received in connection with this study or the writing or publication of the manuscript. The authors declare that there were no conflicts of interest.

ABBREVIATIONS

%SPD

Percentage of stance-phase duration

BW

Body weight

CV

Coefficient of variation

CVRM

Coefficient of variation for repeated measurements

MD

Minimum difference

PVF

Peak vertical force

RSL

Relative stride length

SI

Symmetry index

VI

Vertical impulse

Footnotes

a.

G*Power, version 3.1.9.2, Heinrich Heine University Düsseldorf, Düsseldorf, Germany.

b.

Simi Reality Motion Systems GmbH, Unterschleissheim, Germany.

c.

QuadruPedLocomotion, Simi Reality Motion Systems GmbH, Unterschleissheim, Germany.

d.

Excel, version 12.0, Microsoft Corp, Redmond, Wash.

e.

SPSS Statistics, version 24.0.0.1, IBM Corp, Armonk, NY.

f.

MedCalc, version 17.9.7, MedCalc Software, Ostend, Belgium.

References

  • 1. Bockstahler BA, Vobornik A, Müller M, et al. Compensatory load redistribution in naturally occurring osteoarthritis of the elbow joint and induced weight-bearing lameness of the forelimbs compared with clinically sound dogs. Vet J 2009;180:202212.

    • Search Google Scholar
    • Export Citation
  • 2. Brebner NS, Moens NMM, Runciman JR. Evaluation of a treadmill with integrated force plates for kinetic gait analysis of sound and lame dogs in a trot. Vet Comp Orthop Traumatol 2006;19:205212.

    • Search Google Scholar
    • Export Citation
  • 3. Fanchon L, Grandjean D. Accuracy of asymmetry indices of ground reaction forces for diagnosis of hindlimb lameness in dogs. Am J Vet Res 2007;68:10891094.

    • Search Google Scholar
    • Export Citation
  • 4. LeQuang T, Maitre P, Roger T, et al. Is a pressure walkway system able to highlight a lameness in dogs? J Anim Vet Adv 2009;8:19361944.

    • Search Google Scholar
    • Export Citation
  • 5. Gillette RL, Angle TC. Recent developments in canine locomotor analysis: a review. Vet J 2008;178:165176.

  • 6. Lee DV, Bertram JE, Todhunter RJ, et al. Force overlap in trotting dogs: a Fourier technique for reconstructing individual limb ground reaction force. Vet Comp Orthop Traumatol 2002;15:223227.

    • Search Google Scholar
    • Export Citation
  • 7. Quinn MM, Keuler NS, Lu Y, et al. Evaluation of agreement between numerical rating scales, visual analogue scoring scales, and force platform analysis in dogs. Vet Surg 2007;36:360367.

    • Search Google Scholar
    • Export Citation
  • 8. Jevens DJ, Hauptman JG, DeCamp CE, et al. Contributions to variance in force-plate analysis of gait in dogs. Am J Vet Res 1993;54:612615.

    • Search Google Scholar
    • Export Citation
  • 9. Nordquist B, Fischer J, Kim SI, et al. Effects of trial repetition, limb side, intraday and inter-week variation on vertical and craniocaudal ground reaction forces in clinically normal Labrador Retrievers. Vet Comp Orthop Traumatol 2011;24:435444.

    • Search Google Scholar
    • Export Citation
  • 10. Rumph PF, Lander JE, Kincaid SA, et al. Ground reaction force profiles from force platform gait analysis of clinically normal mesomorphic dogs at the trot. Am J Vet Res 1994;55:756761.

    • Search Google Scholar
    • Export Citation
  • 11. Bertram JE, Lee DV, Case HN, et al. Comparison of the trotting gaits of Labrador Retrievers and Greyhounds. Am J Vet Res 2000;61:832838.

    • Search Google Scholar
    • Export Citation
  • 12. Bockstahler BA, Skalicky M, Peham C, et al. Reliability of ground reaction forces measured on a treadmill system in healthy dogs. Vet J 2007;173:373378.

    • Search Google Scholar
    • Export Citation
  • 13. Clements DN, Owen MR, Carmichael S, et al. Kinematic analysis of the gait of 10 Labrador Retrievers during treadmill locomotion. Vet Rec 2005;156:478481.

    • Search Google Scholar
    • Export Citation
  • 14. Fanchon L, Valette JP, Sanaa M, et al. The measurement of ground reaction force in dogs trotting on a treadmill. Vet Comp Orthop Traumatol 2006;19:8186.

    • Search Google Scholar
    • Export Citation
  • 15. Fanchon L, Grandjean D. Habituation of healthy dogs to treadmill trotting: repeatability assessment of vertical ground reaction force. Res Vet Sci 2009;87:135139.

    • Search Google Scholar
    • Export Citation
  • 16. Kapatkin AS, Arbittier G, Kass PH, et al. Kinetic gait analysis of healthy dogs on two different surfaces. Vet Surg 2007;36:605608.

  • 17. Keebaugh AE, Redman-Bentley D, Griffon DJ. Influence of leash side and handlers on pressure mat analysis of gait characteristics in small-breed dogs. J Am Vet Med Assoc 2015;246:12151221.

    • Search Google Scholar
    • Export Citation
  • 18. LaFlamme D. Development and validation of a body condition score system for dogs. Canine Pract 1997;22:1015.

  • 19. Faul F, Erdfelder E, Lang AG, et al. G*Power 3: a flexible statistical power analysis program for the social, behavioral, and biomedical sciences. Behav Res Methods 2007;39:175191.

    • Search Google Scholar
    • Export Citation
  • 20. Allen K, DeCamp CE, Braden TD, et al. Kinematic gait analysis of the trot in healthy mixed breed dogs. Vet Comp Orthop Traumatol 1994;7:148153.

    • Search Google Scholar
    • Export Citation
  • 21. Hottinger HA, DeCamp CE, Olivier NB, et al. Noninvasive kinematic analysis of the walk in healthy large-breed dogs. Am J Vet Res 1996;57:381388.

    • Search Google Scholar
    • Export Citation
  • 22. Robinson RO, Herzog W, Nigg BM. Use of force platform variables to quantify the effects of chiropractic manipulation of gait symmetry. J Manipulative Physiol Ther 1987;10:172176.

    • Search Google Scholar
    • Export Citation
  • 23. Nakagawa S, Cuthill IC. Effect size, confidence interval and statistical significance: a practical guide for biologists. Biol Rev 2007;82:591605.

    • Search Google Scholar
    • Export Citation
  • 24. Hanley JA, Negassa A, Edwardes MD, et al. Statistical analysis of correlated data using generalized estimating equations: an orientation. Am J Epidemiol 2003;157:364375.

    • Search Google Scholar
    • Export Citation
  • 25. Bland M. An introduction to medical statistics. 3rd ed. Oxford, England: Oxford University Press, 2015;315317.

  • 26. Besancon MF, Conzemius MG, Derrick TR, et al. Comparison of vertical forces in normal greyhounds between force platform and pressure walkway measurement systems. Vet Comp Orthop Traumatol 2003;16:153157.

    • Search Google Scholar
    • Export Citation
  • 27. Voss K, Wiestner T, Galeandro L, et al. Effect of dog breed and body conformation on vertical ground reaction forces, impulses, and stance times. Vet Comp Orthop Traumatol 2011;24:106112.

    • Search Google Scholar
    • Export Citation
  • 28. Mölsä SH, Hielm-Björkman AK, Laitinen-Vapaavuori OM. Force platform analysis in clinically healthy Rottweilers: comparison with Labrador Retrievers. Vet Surg 2010;39:701707.

    • Search Google Scholar
    • Export Citation
  • 29. Voss K, Galeandro L, Wiestner T, et al. Relationships of body weight, body size, subject velocity, and vertical ground reaction forces in trotting dogs. Vet Surg 2010;39:863869.

    • Search Google Scholar
    • Export Citation
  • 30. Roy WE. Examination of the canine locomotion system. Vet Clin North Am 1971;1:5370.

  • 31. Drüen S, Böddeker J, Nolte I, et al. Bodenreaktionskräfte der caninen Hintergliedmaße: Gibt es Unterschiede beim Gang auf Laufband und Kraftmessplatte? Berl Munch Tierarztl Wochenschr 2010;123:339345.

    • Search Google Scholar
    • Export Citation
  • 32. Roush JK, McLaughlin RM Jr. Effects of subject stance times and velocity on ground reaction forces in clinically normal Greyhounds at the walk. Am J Vet Res 1994;55:16721676.

    • Search Google Scholar
    • Export Citation

Contributor Notes

Address correspondence to Dr. Meyer-Lindenberg (andrea.meyer-lindenberg@chir.vetmed.uni-muenchen.de).