In recent years, veterinary medicine has increasingly used gait analysis as an objective measure of limb function.1–6 Since their introduction in the late 1970s, gait analysis systems have become more user-friendly and have become the gold standard for objectively evaluating lameness.7,8 Ground reaction forces (GRF) such as peak vertical forces and vertical impulse (VI) are routinely analyzed to evaluate treatment results of orthopedic disease.8 Historically, force plate analysis systems have been used primarily in academic settings due to cost, space, and time requirements. Recently, gait analysis systems like pressure sensitive walkways (PSW) (Strideway; Tekscan and Gait4Dogs; CIR Systems) and inertial measurement unit (IMU) based gait analysis (GaitKeeper; Vetsens and Whistle; MARS) have been introduced. IMU-based gait analysis systems measure temporal gait characteristics such as stride length and stance time.9
PSWs allow for the collection of stride length, stance time, gait velocity, and measurement of pressure changes through sensors that are used to calculate ground reaction force equivalents. Additionally, PSW measures successive footfalls while walking and simultaneously evaluates weight distribution. This gives rise to additional data regarding symmetry and weight distribution, which in conjunction with its ease of use has made them popular for gait evaluation in orthopedic practice.10
Previous literature on ground reaction forces was focused mainly on large breed dogs with hip, stifle, or elbow diseases with fewer papers evaluating gait of normal small or medium size dogs.6,8,11–13 Fahie et al. evaluated 66 dogs including all sizes with various trotting and walking speeds using the GaitRite System. Fahie suggested that parameters generated by varies walkway systems are quite different and cannot be compared for clinical research. Kim et al. evaluated 6 small and 6 large breed dogs at a walk using the Tekscan system. Amimoto et al. evaluated the recovery of small breed dogs using GRF after unilateral tibial plateau osteotomy (TPLO) surgery and Ichinohe et al. normal Beagles at different velocities.12,14 In these last 2 papers a force plate system instead of a PSW was used.
Velocity and acceleration have been reported to affect GRF.6,8,11–13 With the PSW we can standardize stride velocity and accurately measure GRF in succession by use of calibration. With a uniform calibration technique within study groups, the PSW has the ability to consistently measure gait data in animals as small as a few hundred grams despite these animals’ short stride lengths.1–3,15,16 The purpose of this study was to establish normal GRF for small to medium size dogs (< 25 kg) using a PSW and set walking and trotting speeds. We hypothesize that, at a given speed, small dogs have shorter stance time than medium size dogs. We hypothesize that DH influences GRF.
Materials and Methods
Animals
The study parameters for this research were in line with the animal welfare act and the study was approved by the Animal Care and use committee. All owners signed a consent form before trials. In addition, each owner filled out a Canine Brief Pain Inventory (CBPI) form.17,18 Thirty healthy client-owned adult dogs were enrolled in this study. Dogs were divided into 2 groups: small, which included dogs that weighed less than 15 kg, and medium, with dogs that weighed between 15 and 25 kg. To be considered clinically normal, the following criteria had to be met: Dogs had to have no history of musculoskeletal or neurological abnormalities, to have undergone physical examination by either 1 observer (A.L. or M.W.B.) and deemed normal, to have a complete Canine Brief Pain Index survey with a score of 0, and to have a score of 0 on the Colorado Canine Acute Pain Scale as assessed by 1 of the 2 observers.19 In addition, limb girth and goniometry of appendicular joints were obtained to further qualify the dogs as normal. Limb girth for the pelvic limbs was analyzed via 2 methods. The first method was measuring the girth of the pelvic limb at 70% of the femur length (from the greater trochanter to lateral condyle) while the patient was laterally recumbent. The second method was measuring limb girth at the level of the greater trochanter in a standing position as described by Fox.20 The thoracic limb girth was measured at the level of the acromion.20 Limb girth was measured with a Gulick tape measure (Gulick Tape Measure; Baseline Evaluation Instruments). ROM (range of motion) was obtained via a goniometer for the shoulders, elbows, carpi, hips, stifles, and tarsi.21 Dog height (DH) was measured at the withers with a yardstick for each dog and recorded.
Gait analysis
Two types of gait evaluations were performed for each dog over the pressure sensitive walkway (PSW) with multiple trials of each. One evaluation was performed at a walk and the other at a trot. One of 2 evaluators (J.A. or J.B.) walked each dog on the PSW. These evaluators were blinded to the other findings to limit observer bias.
PSW gait analysis consisted of 4-Tile high-resolution system embedded in a 12.8 ft X 3ft walkway (Strideway, Tekscan). In that 12.8 X 3 ft the active sensors were 8.5ft X 2.1ft with 3.88 sensels/cm2. Two non-functional tiles, constructed of the same materials and dimensions as the one containing the active pressure sensors, were positioned directly in front and behind the active pressure sensors with no gap, such that dogs trotted or walked over the non-functional tiles before contacting the active pressure sensors and again immediately after. The sensors were covered with a floor grip, a slip-resistant vinyl-based film with a textured grip about 0.008″ thick (FloorGrip; Mactac). A camera (Lifecam; Microsoft) was arranged at the end of the walkway, directed at the patient to capture each of the trials. Foot placement was confirmed using this camera (Lifecam; Microsoft) aimed at the PSW, with each trial recorded for slow-motion playback. Data analysis capabilities include automated foot segmentation to calculate temporal parameters such as heel contact and propulsion time, automatic generation of strike boxes to identify areas of foot contact, force vs time curves, pressure vs time curves, peak pressure vs time curves, plantar pressure profiles, force-time integrals, automatic stance detection, peak stance pressures, the center of force, and center of force trajectory. Strike boxes were validated by J.A. or J.B. by looking at the footfalls in slow motion. Automatic generation of gait tables with the display of gait cycle, step-stride, differential, and symmetry parameters was obtained after every walking or trotting trial (Strideway; Tekscan). Data including velocity, foot placement, and vertical forces were collected and analyzed using proprietary software (Strideway Software Version 7.8).
Calibration
Before testing, patients’ phantom models according to the manufacturer’s recommendation were created to calibrate for our small patient weights. Phantom weights were created to simulate different weights using 2.5 kg increments, from 0–25 kg. The phantom models were created with ¾ inch, 12 X 12 X 12-inch equilateral triangles made from red oak plywood (Home Depot). Each triangle had 3 sorbothane vibration bumpers (Sorbothane Vibration Mount, Isolate IT). The bumpers were ½ inch tall and ¾ inch in diameter. Similar sorbothane bumpers were used in a prior study mimicking paw pads to accurately calibrate the walkway to get accurate measurements of GFR. A step calibration technique was used for both study groups. For the step calibration, the investigator clicked on the calibration button and the computer directed the user to place a load on the sensor. In our case, an appropriate weight was placed on the phantom model and the phantom model with weight was placed on the PSW and held for 5–10 seconds until calibrated, then the phantom model was removed.11
Trial
For each testing day, each dog was weighed on the same electronic scale (Way Electronic Scale; VSSI) and allowed to become familiar with the PSW and testing area. Dogs were given 5 minutes to roam and acclimatize to the room. Dogs were walked across the PSW until they appeared to be comfortable and relaxed, generally 3 to 6 passes. All analyses were performed both at a walking gait and a trotting gait. Between 3 to 5 valid passes were collected for each patient (Supplementary Video S1). Data including velocity, foot placement, and vertical forces were collected and analyzed using proprietary software (Strideway Software Version 7.8). The camera captured each valid run. A valid run occurred when a dog moved in a straight line across the PSW without pulling to a side or turning its head. Walking trials were considered valid when the velocity was between 0.9 m/s and 1.3 m/s, and acceleration was not greater than 0.5 m/s2. Trotting trials were considered valid when the velocity was between 1.7 to 2.1 m/s, and acceleration was not greater than 0.5 m/s2 (Analytics Software Version 9.4; SAS). The PSW was calibrated before each trial using the previously described phantom models normalized to each patient’s body weight within 5 kg. Once the trials were completed, the recorded videos of the runs were individually analyzed to assure the correct footfalls were recorded and patients were not trotting during a walking trial or walking during a trotting trial. Proprietary software was used to generate gait analysis tables and graphs. These tables and graphs were subsequently recorded and analyzed. Gait Time – Front in seconds is the time it took for a study subject to walk across the entire length of the 4-Tile walkway. Longer walkways will take longer to walk across. Gait Distance is the entire study length that a subject traveled on the PSW. Gait Velocity is the speed with which a study subject traveled over the PSW. Gait Cycle Time is the time in seconds it takes for the paw of 1 gait cycle to touch the sensor and that same paw to touch the sensor again a second time for the next gait cycle. A gait cycle is when all 4 paws have touched the ground. Cycles per minute are the number of Gait Cycles per minute. The maximum force is peak pressure relative to % BW, or kg BW. Maximum Force (%BW): Maximum force is measured during the stance of the left front foot (LF), left hind foot (LH), right front foot (RF), and right hind foot (RH). Maximum force values are reported as a percentage of the patient’s body weight. If there are multiple stances, the maximum force values are averaged. The maximum force is also presented as the maximum force in kilograms relative to BW. Peak pressure is measured in kilopascal and is the maximum pressure for each foot during the stance time.
Statistical analysis
Data were analyzed using Analytics Software (SAS version 9.4). Outcomes included Stance Time, Swing Time, Stride Time, Stride Length, Stride Velocity, Stride Acceleration, Maximum Force in Percent Bodyweight, Maximum Force in kilograms, Force Time Integral in percent bodyweight per second, Force Time Integral in kilograms per second, and Maximum Peak Pressure. For each combination of dog, form of gait (trot or walk), and handler side (left or right) there were 1 to 4 trials (technical replicates) (most dogs had 2 or 3 trials). For each limb (LF, LH, RF, and RH) these within-subject outcome data were averaged and used for downstream analysis. Maximum force as a percent of body weight and stride acceleration were considered to be self-normalized. Four dynamically normalized variables were computed (using a modification of the method by Betram et al. as follows: Duty factor was the ratio stance time to stride time. Relative velocity was the ratio
Results
Thirty clinically normal dogs were included in the study. Sixteen dogs were in the small group (< 15 kg) and 14 in the medium group (15 to 25 kg). There was 8 female spayed, 2 male intact, and 6 male neutered dogs in the small group and 7 female spayed, 1 female intact, and 6 male neutered dogs in the medium group. Breeds in both groups included mixed (12), Australian Shepherd (2), Pitbull (2), Husky (2), Pomeranian (2), Norwich Terrier (2), Jack Russel Terrier (1), Dachshund (1), Italian Lagotto Romagnolo (1), Chihuahua (1), Pembroke Welsh Corgi (1), Beagle (1), American Staffordshire Terrier (1), and Basset Hound (1). The average age for dogs in the small group was 5.6 years old, while the average age for the medium group was 4.5 years old. There was no difference in age between groups. The average weight for the small group was 8.6 kg (SD ± 3.9), while the average weight for the medium group was 21.3 kg (SD ± 2.1). Medium dogs were heavier than small dogs (P < .001).
Using the proximal forelimb girth measurements of the thoracic limbs, the average limb girth for the small group in the thoracic limbs was 20.2 cm (20.2 cm Right Forelimb, 20.2 cm Left Forelimb) while in the medium group, the average was 29.2 cm (29.2 cm Right Forelimb, 29.2 cm Left Forelimb). Using the thigh circumference measurements of the pelvic limbs at the level of the greater trochanter, the average limb girth circumference for small dogs in the hindlimbs was 28.3 cm (28.3 cm Right Hindlimb, 28.3 cm Left Hindlimb) while in medium dogs the average was 41.2 cm (41.1 cm Left Hindlimb, 41.2 cm Right Hindlimb). Using the circumference at 70% of femur length, the mean left thigh circumference was 19.8 cm, and the mean right thigh circumference was 19.8 cm for small dogs, whereas it was 31 cm for the left and 31 cm for the right thigh circumference for medium dogs. There was no difference between the right and left sides. There were differences between the limb girth of small and medium dogs in either method used. Medium dogs had significantly more limb girth in comparison, to all measurement methods used (P < .001; Table 1). The mean joint angles in flexion and extension in all joints were examined and the varus and valgus of the hock and carpal joints were recorded. There were no differences in ROM data between small and medium dogs. Median DH was 24 cm (mean 24.81 cm) for small and 45 cm (mean 44.28 cm) for medium dogs. Small dogs were significantly shorter than medium dogs (P < .001).
Limb girth in front and hind leg measured at the level of the acromion for the front leg and measured at the level of the greater trochanter and at the level of 70% of the length of the femur for the hind leg.
| Baseline variable | Mean (±SD) (n = 16) small dogs | Mean (±SD) (n = 14) medium dogs | P-value |
|---|---|---|---|
| Limb Girth Left Forelimb at level of acromion | 20.2 (±5.0) | 29.2 (±2.3) | < .0001 |
| Limb Girth Right Forelimb at level of acromion | 20.2 (±5.0) | 29.2 (±2.3) | < .0001 |
| Limb Girth Left Hindlimb at level of greater trochanter | 28.3 (±7.1) | 41.1 (±2.7) | < .0001 |
| Limb Girth Right Hindlimb at level of greater trochanter | 28.3 (±7.0) | 41.2 (±2.7) | < .0001 |
| Limb Girth at level of 70% femur bone length Left Hindlimb | 19.8 (±4.9) | 31.0 (±3.9) | < .0001 |
| Limb Girth 70% femur bone length Right Hindlimb | 19.8 (±4.9) | 31.0 (±3.8) | < .0001 |
Stance Time, Swing Time, Stride Time, Stride Length, Maximum Force in kilograms, Force Time Integral in percent bodyweight per second, Force Time Integral in kilograms per second, and Maximum Peak Pressure all were higher in medium dogs compared with small dogs (P < .05). Inconclusive data was obtained for Stride Velocity and Stride Acceleration for both groups. Maximum Force normalized to Percent Body weight showed a difference at a trot in all legs and in the hindlimbs at a walk. No differences were noted at a walk in the forelimbs for maximum force in percent body weight. No differences were seen in the symmetry index between groups at either walk or trot (Tables 2 and 3). The duty factor was larger for the medium than for small dogs. Medium dogs traveled at a slower relative velocity than small dogs. When normalizing data for stride length using DH temporospatial parameters were larger in medium dogs at a walk and at a trot regardless of the handler side as seen in duty factor and stride length. When normalizing for relative stride time there was no longer a difference between the groups and when analyzing relative stride velocity smaller dogs walked significantly faster than medium dogs (Tables 4 and 5).
Ground reaction variables of the group small dogs compared with the group medium dogs at walk and trot.
| Small | Medium | |||
|---|---|---|---|---|
| Mean (SD) | Mean (SD) | |||
| Ground reaction variable | Limb | Walk/trot | Walk/trot | P-value |
| Stance Time in s | LF | 0.22 (0.09)/0.12 (0.04) | 0.37 (0.04)/0.19(0.02) | < .001 |
| LH | 0.20 (0.09)/0.11 (0.04) | 0.34 (0.04)/0.18(0.02) | < .001 | |
| RF | 0.22 (0.09)/0.12 (0.04) | 0.36 (0.04)/0.19(0.02) | < .001 | |
| RH | 0.20 (0.08)/0.11 (0.04) | 0.34 (0.04)/0.18(0.02) | < .001 | |
| Swing Time in s | LF | 0.22 (0.02)/0.23 (0.02) | 0.25 (0.02)/0.27 (0.03) | < .001 |
| LH | 0.25 (0.02)/0.24 (0.03) | 0.30 (0.05)/0.31 (0.04) | < .001 | |
| RF | 0.23 (0.02)/0.23 (0.02) | 0.25 (0.02)/0.27 (0.03) | < .001 | |
| RH | 0.26 (0.02)/0.24 (0.03) | 0.28 (0.03)/0.31 (0.04) | .013/< .001 | |
| Stride Time in s | LF | 0.45 (0.10)/0.34 (0.06) | 0.62 (0.05)/0.46 (0.04) | < .001 |
| LH | 0.45 (0.09)/0.35 (0.06) | 0.64 (0.08)/0.50 (0.05) | < .001 | |
| RF | 0.45 (0.10)/0.34 (0.06) | 0.62 (0.05)/0.47 (0.04) | < .001 | |
| RH | 0.46 (0.09)/0.35 (0.07) | 0.62 (0.06)/0.49 (0.05) | < .001 | |
| Stride Length in cm | LF | 48.11 (11.47)/64.92 (13.56) | 72.52(7.85)/94.90 (11.02) | < .001 |
| LH | 48.69 (11.47)/71.23 (16.39) | 78.66 (10.44)/104.04 (13.41) | < .001 | |
| RF | 48.26 (11.65)/64.93 (13.21) | 72.52 (7.97)/94.84 (10.93) | < .001 | |
| RH | 51.39 (11.35)/71.02 (17.25) | 75.82 (9.50)/99.35 (11.39) | < .001 | |
| Stride Velocity cm/s | LF | 108.14 (9.31)/189.53 (10.45) | 118.25 (10.21)/204.75 (14.30) | .008/.003 |
| LH | 109.49 (10.41)/201.06 (19.55) | 123.32 (10.92)/210.82 (16.47) | .001/.159 | |
| RF | 108.57 (11.68)/189.04 (10.34) | 117.84 (10.31)/203.80 (12.83) | .03/.002 | |
| RH | 112.70 (10.06)/199.73 (19.63) | 121.98 (12.39)/204.59 (16.74) | .031/.48 | |
| Stride Acceleration cm/s2 | LF | 5.00 (8.78)/15.42 (27.94) | 7.02 (11.19)/-1.09 (16.63) | .585/.066 |
| LH | 14.45 (19.28)/4.41 (30.25) | 1.11 (12.99)/47.40 (69.90) | .037/.039 | |
| RF | 3.55 (11.99)/15.39 (27.75) | 10.70 (15.56)/2.16 (15.56) | .167/.129 | |
| RH | 11.84 (33.74)16.00 (31.36) | 4.87 (12.39)/5.73 (36.33) | .472/.421 | |
| Maximum Force (PctBW) | LF | 75.57 (13.03)/98.98 (16.46) | 78.69 (11.44)/117.43 (21.90) | .494/.016 |
| LH | 48.91 (6.68)/57.20 (7.66) | 55.47 (8.86)/74.89 (10.56) | .029/< .001 | |
| RF | 75.08 (13.60)/97.53 (17.10) | 78.83 (13.15)/116.07 (18.47) | .450/< .001 | |
| RH | 48.60 (6.93)/57.28 (8.97) | 56.92 (9.80)/74.04 (11.03) | .011/< .001 | |
| Maximum Force kg | LF | 6.22 (2.53)/8.76 (3.19) | 16.65 (3.01)/24.87 (5.49) | < .001 |
| LH | 4.15 (1.87)/5.23 (2.24) | 11.69 (2.11)/15.84 (2.90) | < .001 | |
| RF | 6.16 (2.47)/8.62 (3.10) | 16.71 (3.49)/24.61 (5.08) | < .001 | |
| RH | 4.15 (1.87)/5.26 (2.34) | 11.97 (2.17)/15.60 (2.61) | < .001 | |
| FTI (Pct BW *s) | LF | 10.92 (2.96)/8.43 (1.72) | 19.93 (4.38)/14.80 (3.50) | < .001 |
| LH | 6.25 (2.17)/4.38 (1.53) | 11.93 (2.27)/8.58(1.52) | < .001 | |
| RF | 10.70 (2.85)/8.18 (1.76) | 20.06 (5.00)14.37 (3.18) | < .001 | |
| RH | 6.27 (2.13)/4.38 (1.60) | 12.26 (2.28)/8.31 (1.37) | < .001 | |
| FTI (kg* s) | LF | 1.03 (0.63)/0.81 (0.42) | 4.24 (1.11)/3.14 (0.86) | < .001 |
| LH | 0.61 (0.41)/0.44 (0.30) | 2.53 (0.60)/1.82 (0.41) | < .001 | |
| RF | 1.00 (0.61)/0.78 (0.41) | 4.27 (1.26)/3.05 (0.81) | < .001 | |
| RH | 0.61 (0.40)/0.44 (0.30) | 2.60 (0.59)/1.76 (0.38) | < .001 | |
| Maximum Peak Pressure (kPa) | LF | 173.08 (45.68)/195.51 (43.85) | 328.08 (35.54)/341.14 (40.43) | < .001 |
| LH | 143.13 (41.16)/162.47 (45.21) | 262.15 (41.02)/303.47 (32.67) | < .001 | |
| RF | 168.66 (43.67)/192.78 (43.72) | 320.05 (38.19)/339.05 (35.03) | < .001 | |
| RH | 145.33 (40.76)/161.97 (46.07) | 276.55 (36.01)/306.35 (31.26) | < .001 |
Symmetry data for dog groups small and medium at walking speed and trotting speed.
| Symmetry rubric | Walk/trot | Group small (range) | Group medium (range) | P-value |
|---|---|---|---|---|
| Stance Time Front/Hind | WALK | 1.0 (1.0–1.1) | 1.1 (1.0–1.2) | 0.244 |
| TROT | 1.1 (1.0–1.2) | 1.1 (1.0–1.2) | 0.032 | |
| Stride Time Front/Hind | WALK | 1.0 (1.0–1.0) | 1.0 (0.9–1.0) | 0.648 |
| TROT | 1.0 (0.9–1.0) | 1.0 (0.9–1.0) | 0.119 | |
| Stride Length Front/Hind | WALK | 1.0 (0.8–1.0) | 1.0 (0.8–1.0) | 0.219 |
| TROT | 0.9 (0.8–1.0) | 0.9 (0.9–1.0) | 0.354 | |
| Stride Velocity Front/Hind | WALK | 1.0 (0.8–1.0) | 1.0 (0.9–1.0) | 0.445 |
| TROT | 1.0 (0.9–1.0) | 1.0 (1.0–1.0) | 0.099 | |
| Max Force Front/Hind | WALK | 1.6 (1.2–2.1) | 1.5 (1.0–1.9) | 0.222 |
| TROT | 1.8 (1.1–2.2) | 1.7 (1.2–1.8) | 0.028 | |
| Stance Time Left/Right | WALK | 1.0 (1.0–1.0) | 1.0 (1.0–1.1) | 0.249 |
| TROT | 1.0 (1.0–1.1) | 1.0 (1.0–1.0) | 0.350 | |
| Stride Time Left/Right | WALK | 1.0 (1.0–1.0) | 1.0 (1.0–1.1) | 0.250 |
| TROT | 1.0 (0.9–1.0) | 1.0 (0.9–1.1) | 0.930 | |
| Stride Length Left/Right | WALK | 1.0 (0.8–1.0) | 1.0 (1.0–1.1) | 0.015 |
| TROT | 1.0 (0.8–1.2) | 1.0 (0.9–1.1) | 0.431 | |
| Stride Velocity Left/Right | WALK | 1.0 (0.9–1.0) | 1.0 (0.9–1.0) | 0.018 |
| TROT | 1.0 (0.9–1.1) | 1.0 (0.9–1.1) | 0.861 | |
| Max Force Left/Right | WALK | 1.0 (0.9–1.1) | 1.0 (0.9–1.1) | 0.345 |
| TROT | 1.0 (0.9–1.1) | 1.0 (0.9–1.0) | 0.880 | |
| Stance Time LF/RF | WALK | 1.0 (0.9–1.0) | 1.0 (1.0–1.1) | 0.224 |
| TROT | 1.0 (1.0–1.1) | 1.0 (1.0–1.1) | 0.310 | |
| Stride Time LF/RF | WALK | 1.0 (1.0–1.0) | 1.0 (1.0–1.0) | 0.074 |
| TROT | 1.0 (1.0–1.0) | 1.0 (1.0–1.0) | 0.981 | |
| Stride Length LF/RF | WALK | 1.0 (1.0–1.0) | 1.0 (1.0–1.0) | 0.364 |
| TROT | 1.0 (1.0–1.0) | 1.0 (1.0–1.0) | 0.189 | |
| Stride Velocity LF/RF | WALK | 1.0 (1.0–1.0) | 1.0 (1.0–1.0) | 0.057 |
| TROT | 1.0 (1.0–1.0) | 1.0 (1.0–1.0) | 0.722 | |
| Max Force LF/RF | WALK | 1.0 (0.9–1.1) | 1.0 (0.9–1.1) | 0.471 |
| TROT | 1.0 (0.9–1.1) | 1.0 (0.9–1.1) | 0.451 | |
| Stance Time LH/RH | WALK | 1.0 (0.9–1.1) | 1.0 (0.9–1.1) | 0.740 |
| TROT | 1.0 (0.9–1.1) | 1.0 (0.9–1.1) | 0.248 | |
| Stride Time LH/RH | WALK | 1.0 (0.9–1.0) | 1.0 (0.9–1.2) | 0.476 |
| TROT | 1.0 (0.9–1.1) | 1.0 (0.8–1.2) | 0.661 | |
| Stride Length LH/RH | WALK | 1.0 (0.7–1.0) | 1.0 (1.0–1.3) | 0.045 |
| TROT | 1.0 (0.7–1.3) | 1.0 (0.9–1.2) | 0.422 | |
| Stride Velocity LH/RH | WALK | 1.0 (0.8–1.0) | 1.0 (0.9–1.3) | 0.080 |
| TROT | 1.0 (0.8–1.3) | 1.0 (0.9–1.1) | 0.965 | |
| Max Force LH/RH | WALK | 1.0 (0.9–1.2) | 1.0 (0.8–1.1) | 0.296 |
| TROT | 1.0 (0.9–1.2) | 1.0 (0.9–1.1) | 0.604 |
See Table 2 for key.
Ground reaction variables at a walk.
| Variable | Limb | Group small mean (SD), L/R | Group medium mean (SD), L/R | Pr > |t|, L/R |
|---|---|---|---|---|
| Duty Factor | LF | 0.48 (0.11)/0.48 (0.11) | 0.60 (0.02)/0.58 (0.04) | < 0.001/< 0.001 |
| LH | 0.43 (0.11)/0.43 (0.11) | 0.54 (0.04)/0.52 (0.04) | < 0.001/0.004 | |
| RF | 0.47 (0.12)/0.48 (0.11) | 0.59 (0.03)/0.58 (0.03) | < 0.001/< 0.001 | |
| RH | 0.43 (0.10)/0.42 (0.11) | 0.55 (0.03)/0.54 (0.04) | < 0.001/< 0.001 | |
| Relative Velocity | LF | 0.71 (0.12)/0.71 (0.12) | 0.56 (0.04)/0.59 (0.03) | < 0.001/0.005 |
| LH | 0.71 (0.12)/0.73 (0.12) | 0.59 (0.05)/0.60 (0.03) | 0.005/0.004 | |
| RF | 0.71 (0.12)/0.72 (0.12) | 0.56 (0.04)/0.59 (0.03) | < 0.001/0.002 | |
| RH | 0.75 (0.13)/0.74 (0.11) | 0.59 (0.06)/0.59 (0.03) | < 0.001/< 0.001 | |
| Relative Stride Length | LF | 2.00 (0.32)/1.99 (0.24) | 1.65 (0.21)/1.73 (0.24) | 0.004/0.018 |
| LH | 2.00 (0.31)/2.05 (0.30) | 1.81 (0.26)/1.79 (0.27) | 0.115/0.017 | |
| RF | 1.99 (0.28)/2.01 (0.26) | 1.64 (0.22)/1.73 (0.23) | 0.005/0.012 | |
| RH | 2.17 (0.37)/2.12 (0.37) | 1.74 (0.29)/1.72 (0.22) | < 0.001/< 0.001 | |
| Relative Stride Time | LF | 2.84 (0.28)/2.82 (0.32) | 2.93 (0.36)/2.96 (0.41) | 0.431/0.421 |
| LH | 2.83 (0.25)/2.83 (0.28) | 3.08 (0.48)/2.99 (0.49) | 0.022/0.298 | |
| RF | 2.84 (0.29)/2.82 (0.30) | 2.94 (0.37)/2.97 (0.39) | 0.342/0.34 | |
| RH | 2.91 (0.24)/2.88 (0.29) | 2.99 (0.44)/2.90 (0.42) | 0.458/0.809 |
Ground reaction variables at a trot.
| Variable | Limb | Group small mean (SD), L/R | Group medium mean (SD), L/R | Pr > |t|, L/R |
|---|---|---|---|---|
| Duty Factor | LF | 0.34 (0.08)/ 0.33 (0.06) | 0.42 (0.03)/0.41 (0.02) | < 0.001/< 0.001 |
| LH | 0.30 (0.07)/ 0.30 (0.06) | 0.38 (0.03)/0.36 (0.04) | 0.004/0.013 | |
| RF | 0.33 (0.08)/ 0.33 (0.07) | 0.42 (0.02)/0.41 (0.03) | < 0.001/< 0.001 | |
| RH | 0.31 (0.06)/ 0.30 (0.06) | 0.37 (0.03)/0.37 (0.03) | 0.009/0.007 | |
| Relative Velocity | LF | 1.22 (0.14)/1.24 (0.14) | 0.96 (0.06)/1.01 (0.06) | < 0.001/< 0.001 |
| LH | 1.32 (0.18)/1.30 (0.21) | 0.98 (0.09)/1.07 (0.12) | < 0.001/< 0.001 | |
| RF | 1.22 (0.15)/1.23 (0.14) | 0.96 (0.07)/1.00 (0.06) | < 0.001/< 0.001 | |
| RH | 1.26 (0.19)/1.32 (0.20) | 0.96 (0.08)/1.01 (0.05) | < 0.001/< 0.001 | |
| Relative Stride Length | LF | 2.62 (0.33)/2.64 (0.35) | 2.14 (0.29)/2.22 (0.26) | < 0.001/< 0.001 |
| LH | 2.92 (0.48)/2.85 (0.62) | 2.27 (0.35)/2.49 (0.25) | < 0.001/< 0.001 | |
| RF | 2.62 (0.34)/2.64 (0.35) | 2.13 (0.29)/2.22 (0.27) | < 0.001/< 0.001 | |
| RH | 2.78 (0.44)/2.92 (0.47) | 2.26 (0.30)/2.30 (0.29) | < 0.001/< 0.001 | |
| Relative Stride Time | LF | 2.14 (0.18)/2.13 (0.17) | 2.23 (0.23)/2.19 (0.23) | 0.332/0.533 |
| LH | 2.21 (0.18)/2.19 (0.21) | 2.32 (0.25)/2.35 (0.25) | 0.211/0.142 | |
| RF | 2.14 (0.18)/2.15 (0.15) | 2.22 (0.23)/2.21 (0.23) | 0.348/0.548 | |
| RH | 2.22 (0.22)/2.21 (0.17) | 2.37 (0.31)/2.29 (0.30) | 0.114/0.431 |
Ground reaction variables normalized to withers height of the small dog group compared with the medium dog group at walk and trot. Mean values and standard deviation included for both handler sides.
See Table 4 for key.
The ratio of forelimb to hind limb girth ratio was 0.7 (0.6–0.9) for small, 0.7 (0.7–0.9) for medium dogs and did not differ between groups (P < .966). The forelimb-to-hind limb girth ratio was not significantly correlated with the front-to-hind limb GRF ratio at a walk and at a trot for the small group (r = 0.114, P = .671; r = 0.199, P = .477). The forelimb to hind limb girth ratio was not significantly correlated with front to hind limb GRF ratio at a walk or at a trot for the medium group (r = −0.001, P = .998; r = 0.351, P = .219).
Dogs were led across the PSW with similar speeds for both dog groups at walk and at trot (Table 6). At a walk, the small dogs traveled on average 107.3 cm/s and the medium dogs 110.1 cm/s. At a trot, the subjects traveled on average 188.94 cm/s for the small group and 203.67 cm/s for the medium group.
Gait analysis data from group small and medium when walking across a 4-Tile high resolution system embedded in a 12.8 ft X 3ft walkway.
| Average walk | Small | Medium |
| Number of Stances | 40.6 | 27.04 |
| Gait Time - Front (s) | 4.01 | 3.6 |
| Gait Distance - Front (cm) | 425.01 | 417.88 |
| Gait Velocity - Front (cm/s) | 107.3 | 110.1 |
| Gait Cycle Time (s) | 0.46 | 0.62 |
| Cycles/Minute | 141.06 | 98.04 |
| Average trot | Small | Medium |
| Number of Stances | 30.85 | 21.06 |
| Gait Time - Front (s) | 2.25 | 1.97 |
| Gait Distance - Front (cm) | 423.97 | 399 |
| Gait Velocity - Front (cm/s) | 188.94 | 203.67 |
| Gait Cycle Time (s) | 0.34 | 0.46 |
| Cycles/Minute | 180.47 | 130.14 |
Gait Time – Front is the time it took for the study subject to walk across the entire length of our walkway. Longer walkways will take longer to walk across. Gait Distance is entire study length that a subject traveled on the PSW. Gait Velocity is the speed with which the study subject traveled over the PSW. Gait Cycle Time is the time in seconds it takes for the paw of one gait cycle to touch the sensor and that same paw to touch the sensor again a second time for the next gait cycle. A gait cycle is when all four paws have touched the ground. Cycles/Minute is the number of Gait Cycles per minute.
The handler side did not affect the data. The smaller dog group had 50.15% more stances during the walk speed on the PSW compared with the medium dogs and 46.49% more stances during the trot. Small dogs had a higher number of stances per trial on the PSW compared with medium dogs (Table 6). The gait cycle was shorter for small-breed dogs than for medium size dogs. Small dogs had 43.88% more gait cycles per minute at a walk and 38.67% more gait cycles at a trot compared with medium dogs. The gait cycle time changed in both groups from walk to trot within the small and medium groups. As dogs traversed the PSW at a trot their gait cycle got shorter compared with a walk. BW in kilograms and stance time were significantly correlated at a walk and trot (r = 0.887, P = .001; r = 0.888, P = .001). DH and stance time were significantly correlated at a walk and trot (r = 0.854, P = .001; r = 0.876, P = .001). As BW or DH increased stance time increased as relative velocity decreased. When separating small and medium dog groups only small dogs had a correlation between weight in kilograms to stance time at a walk (r = 0.832, P < .001) and at a trot (r = 0.834, P < .001). Small dogs also had a correlation between stance time and DH at a walk (0.923, P < .001) and a trot (0.905, P < .001).
Discussion
The PSW provided repeatable stride information for clinically normal small and medium size dogs in our study and was easy to use, easy to assemble, and accessible. Small and medium sized dogs can be walked and trotted over the PSW and reference measurements could be obtained to be a guide for future studies.
The data obtained in this study suggests that there are differences in gait between small and medium size dogs. As expected, small dogs have lower GRF in comparison with medium size dogs. Small breed dogs weigh less than their larger counterparts; therefore, it is expected when applying the basic principle of force equal to mass times acceleration that smaller dogs generate lower GRF. When we normalized data according to body weight (Table 2) GRF was proportional. We chose DH at the withers as another morphologic factor to relativize data. When we normalized temporospatial data such as duty factor, velocity, stride length, and stride time according to DH we expected the difference to be similar. Relative stride time was similar, the duty factor was larger for medium and relative velocity was faster for small, and the relative stride length was longer for small dogs at a walk and at a trot. These differences may have remained or in part reversed due to additional factors associated with body type rather than just DH. DH is just 1 variable of volume (height, width, and length) that in turn is reflected in body weight (volume X density). Additional data may be needed to account for the remaining differences in temporospatial data. A variety of breeds were included in the study and many differences in leg lengths within the small and medium size dog groups were present. Even though we chose these 2 groups with different weights we did not separate body type, body condition score, or breed. A short-legged Dachshund and a more long-legged, lean, and less than 15 kg mix breed dog were in the same small dog group jet have different stride lengths and probably a different ideal velocity at walk and trot. It would be expected that a lean, long-legged mix breed dog has a longer stride length and longer stride time at the same absolute walking or trotting speed. A more uniform group of small and medium size dogs may have achieved clearer results in these aspects. Alternatively, separating the small and medium dogs and comparing them to large breed dogs (> 25 kg) may have accentuated the differences between sizes better. We chose small and medium dogs as these data points are still missing in the veterinary literature.
We also accepted our hypothesis that smaller dogs had a shorter stance time at a given speed at a walk and trot than medium dogs. Smaller dogs had a higher cadence in our study as they were walked across the PSW at the same absolute speed as the medium dogs. Dogs with higher cadence have reduced stance time and reduced stride time. Kim at all evaluated 6 small dogs at a walk and found similarly that stance time was reduced compared with 6 large breed dogs > 25 kg BW).5 In this study, gait velocity was not the same between dogs and large breed dogs walked faster (velocity range, 0.98 to 1.07 m/s) than small dogs (velocity range, 0.68 to 0.88 m/s). When normalizing for DH relative stride time was no longer different between the groups. This confirms that DH is an important factor in normalizing data for small and medium dogs. Other body shape differences also play a role such as chest conformation, body condition score, and other factors.22–24
Overall, the medium dog group appeared more homogenous in our study since the weight in kilograms and DH did not correlate to stance time in medium dogs but did so in the small breed dogs. When calculating the correlation between BW and DH initially all dogs were grouped together, and the results were not found to be significant. But when separating small and medium size dog groups the small dog group only was significantly correlated for BW and DH.
Data for stride velocity and acceleration were inconclusive between groups. This was expected as within both groups dogs accelerated within a narrow window of the study’s walk and trot speed. The overall difference between small dog speed (average 107.3 cm/sec) and medium dog (average 110.1 cm/s) speed at walk was very low resulting in similar acceleration and velocity. The same was true for trot speed with small dog trot average of 188.94 cm/s and medium dog trot speed of 203.67 (Table 6). When analyzing relative stride velocity accounting for DH smaller dogs walked significantly faster than medium dogs at the absolute speeds chosen by us. We chose these common speeds, but when considering that the small dog group had 43% and 38% more gait cycles per minute at a walk and trot, respectively, one can consider reducing the walking and trotting speed by 40% to achieve a comfortable overall speed for smaller dogs. Additionally, there was a correlation between DH and stance time and also between BW and stance time within the small dog group. It may be necessary to further divide the small size dogs into subgroups to obtain more homogenous subgroups with DH and BW as variables. Small dogs were significantly shorter than medium dogs (median DH 24 vs 44 cm). Dividing dogs weighing less than 15 kg into those < 10 cm DH, 10–20 cm DH, and > 20 cm DH may give additional reference points.
To obtain stride information each patient was allowed to acclimate to the room with the PSW. This was followed by 3–6 test runs and 3–5 recorded passes were made. Because there are only 4 tiles in the system capturing the data, large long-legged dogs will require more passes to gain sufficient data for analysis. Our small breed dogs consistently got over 30 stances per cycle at walk as well as trot speed. This was a lot more than our medium dogs with on average 20 footfalls or stances per cycle at a walk and 27 at a trot (Table 6). It is our observation that small breed dogs may only need 1–2 valid trials to gain enough data for analysis. Ideally, a set of data with normal values for each breed could be created that takes breed differences, body type, and leg length into account. This was beyond the scope of this study and a much larger sample size would be needed as a follow-up study to break down the stride data in each breed.
Matching the video with the footfalls allowed for important accuracy of the data. Every footfall was verified by the observer later by matching videos with the computer data. This is time-consuming; however, once a routine is established, it is a feasible process in an orthopedic clinic setting.
The handler side has previously influenced gait data.4 In our study we did not see significance with the handler side. Care was taken to make sure patients walked looking straight ahead and not being influenced by the dog handler. Limb girth was larger in the medium group and symmetric. This was expected as these were normal dogs and with an increase in BW muscle mass is expected to increase.
ROM data was not different between the groups which was expected as normal dogs would achieve normal ROM regardless of the size of the pets. On the other hand, a larger dog would be expected to have a larger amount of muscle mass than a small dog, which was reflected in increased girth for the front and hind legs for the medium dogs compared with small dogs.
Limitations of this study include the data gathered being specific to the proprietary software. It also is important to maintain an even gait throughout the length of the sensors. If an even gait is not maintained the data gathered can be skewed. Influences on the dog can be reduced by placing the PSW in an area with few or no distractions and having an experienced handler walk the patient. In addition to an even gait, the direction, side of the handler, and different handlers themselves can introduce inconsistent data. Dogs may be trained to walk on one or the other side of their handler making it difficult to record even gait when walking on their unfamiliar side. Dog morphology was considered in part by measuring DH and by measuring BW. Additional features such as chest and abdominal circumference or separation by breed may have created a more uniform data set.
In summary, the PSW is a reliable device for small and medium-sized dogs to gather GRF. We accept our hypothesis that small in comparison to medium-sized dogs have shorter stance time at a given speed and that these results are influenced by DH. They have decreased limb girth and similar joint ROM. A data set for small and medium-sized dogs were created. There is a difference in GRF between normal small and medium-sized dogs suggesting that normal reference data for PSW need to take dog size into account.
Further studies to identify breed differences may be warranted.
Supplementary Materials
Supplementary materials are posted online at the journal website: avmajournals.avma.org.
Acknowledgments
We want to acknowledge Michael Biderman and Stephen Werre for their statistical analysis and contributions to this paper.
References
- 1.↑
Boyd BS, Puttlitz C, Noble-Haeusslein LJ, John CM, Trivedi A, Topp KS. Deviations in gait pattern in experimental models of hindlimb paresis shown by a novel pressure mapping system. J Neurosci Res. 2007;85(10):2272–2283. doi:10.1002/jnr.21366
- 2.
Budsberg SC, Chambers JN, Lue SL, Foutz TL, Reece L. Prospective evaluation of ground reaction forces in dogs undergoing unilateral total hip replacement. Am J Vet Res. 1996;57:1781–1785.
- 3.↑
Fahie MA, Cortez JC, Ledesma M, Su Y. Pressure mat analysis of walk and trot gait characteristics in 66 normal small, medium, large, and giant breed dogs. Front Vet Sci. 2018;5. doi:10.3389/fvets.2018.00256
- 4.↑
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(11):1215–1221. doi:10.2460/javma.246.11.1215
- 5.↑
Kim J, Kazmierczak KA, Breur GJ. Comparison of temporospatial and kinetic variables of walking in small and large dogs on a pressure-sensing walkway. Am J Vet Res. 2011;72(9):1171–1177. doi:10.2460/ajvr.72.9.1171
- 6.↑
Light VA, Steiss JE, Montgomery RD, Rumph PF, Wright JC. Temporal-spatial gait analysis by use of a portable walkway system in healthy Labrador Retrievers at a walk. Am J Vet Res. 2010;71(9):997–1002. doi:10.2460/ajvr.71.9.997
- 7.↑
Kieves NR, Hart JL, Evans RB, Duerr FM. Comparison of three walkway cover types for use during objective canine gait analysis with a pressure-sensitive walkway. Am J Vet Res. 2019;80(3):265–269. doi:10.2460/ajvr.80.3.265
- 8.↑
Rytz U, Johnston SA, Budsberg SC. Effects of acceleration on ground reaction forces collected in healthy dogs at a trot. Vet Comp Orthop and Traumatol. 1998;11(01):15–19. doi:10.1055/s-0038-1632610
- 9.↑
Ladha C, O’Sullivan J, Belshaw Z, Asher L. GaitKeeper: a system for measuring canine gait. Sensors (Basel, Switzerland). 2017;17(2):309. doi:10.3390/s17020309
- 10.↑
Besancon MF, Derrick TR, Ritter MJ, Conzemius MG. Comparison of vertical forces in normal greyhounds between force platform and pressure walkway measurement systems. Vet Comp Orthop and Traumatol. 2003;16(03):153–157. doi:10.1055/s-0038-1632766
- 11.↑
Agostinho FS, Geraldo B, Justolin PL, et al. Influence of calibration protocols for a pressure-sensing walkway on kinetic and temporospatial parameters. Vet Comp Orthop and Traumatol. 2015;28(01):25–29. doi:10.3415/VCOT-14-05-0081
- 12.↑
Amimoto H, Koreeda T, Ochi Y, et al. Force plate gait analysis and clinical results after tibial plateau levelling osteotomy for cranial cruciate ligament rupture in small breed dogs. Vet Comp Orthop and Traumatol. 2020;33(03):183–188. doi:10.1055/s-0039-1700990
- 13.↑
Nelson SA, Krotscheck U, Rawlinson J, Todhunter RJ, Zhang Z, Mohammed H. Long-term functional outcome of tibial plateau leveling osteotomy versus extracapsular repair in a heterogeneous population of dogs. Vet Surg. 2012;42(1):38–50. doi:10.1111/j.1532-950X.2012.01052.x
- 14.↑
Ichinohe T, Takahashi H, Fujita Y. Force plate analysis of ground reaction forces in relation to gait velocity of healthy beagles. Am J Vet Res. 2022;83:9. doi:10.2460/ajvr.22.03.0057
- 15.↑
Lascelles BD, Roe SC, Smith E, et al. Evaluation of a pressure walkway system for measurement of vertical limb forces in clinically normal dogs. Am J Vet Res. 2006;67(2):277–282. doi:10.2460/ajvr.67.2.277
- 16.↑
Wustefeld-Janssens BG, Pettitt RA, Cowderoy EC, et al. Peak vertical force and vertical impulse in dogs with cranial cruciate ligament rupture and meniscal injury. Vet Surg. 2015;45(1):60–65. doi:10.1111/vsu.12419
- 17.↑
Brown D, Boston R, Coyne J, Farrar JT. A novel approach to the use of animals in studies of pain: validation of the canine brief pain inventory in canine bone cancer. Pain Med. 2009;10:133–142. doi:10.1111/j.1526-4637.2008.00513.x
- 18.↑
Brown DC, Boston RC, Farrar JT. Comparison of force plate gait analysis and owner assessment of pain using the canine brief pain inventory in dogs with osteoarthritis. J Vet Intern Med 2013;1:22–30. doi:10.1111/jvim.12004
- 19.↑
Hellyer, PW, Uhrig, SR, Robinson, NG. Canine Acute Pain Scale and Feline Acute Pain Scale. Colorado State University Veterinary Medical Center; 2006. www.cvmbs.colostate.edu/ivapm/professionals/members/drug_protocols/painscalecaninenobandagesPAH.pdf
- 20.↑
Fox E, Brunke MW. The ratio of pelvic limb muscle circumference to thoracic limb muscle circumference in 115 sound dogs. Scientific presentation abstracts: 2020 ACVS virtual sessions. Vet Surg. 2020;77.
- 21.↑
Jaegger G, Marcellin-Little DJ, Levine D. Reliability of goniometry in Labrador retrievers. Am J Vet Res. 2002;63(7):979–986. doi:10.2460/ajvr.2002.63.979
- 22.↑
Bertram JE, Lee DV, Case BS, Todhunter RJ. Comparison of the trotting gaits of Labrador Retrievers and Greyhounds. Am J Vet Res. 2000;61:832. doi:10.2460/ajvr.2000.61.832
- 23.
Hans EC, Zwarthoed B, Seliski J, Nemke B, Muir P. Variance associated with subject velocity and trial repetition during force platform gait analysis in a heterogeneous population of clinically normal dogs. The Vet J. 2014;202(3):498–502. doi:10.1016/j.tvjl.2014.09.022
- 24.↑
Mall SU, Steigmeier-Raith S, Reese S. Growing Beagles and Foxhound-Boxer-Ingelheim Labrador Retriever mixed breed show a forelimb dominated gait and a cranial shift in weight support over time during a kinetic gait analysis. Am J Vet Res. 2022;83(7):1–10.
