Materials and Methods

Testing procedure

Two FPs (50.2 X 50.2 X 4.44 cm; Fz maximum load, 1,334 N; Fx/Fy maximum load, 450 N; AccuGait System; AMTI NetForce Advanced Mechanical Technology, Inc) placed next to each other were used for gait analysis. Two walking paths (90.9 X 159.0 X 4.44 cm; total length, 736.4 cm) were placed in front and behind the FPs. A frame with through-beam photoelectric sensors (5 pairs of photoelectric sensors; minimum detection diameter, 8 mm; distance between photoelectric sensors, 300 mm; E3Z-T81A; OMRON Corporation) was placed at both ends of the plate within 2.5 m of each other to measure walking time, velocity, and acceleration. The height of the transmissive photoelectric sensors from the ground was adjusted to the dorsal edge of the scapula when the dog was in a normal standing position.¹⁵ The sampling rate for the measurement using FPs was unified at 120 Hz per channel for 3 s. A video camera was fixed at a position where each dog’s limb on the 2 FPs could be observed, and the measurement process was recorded, identifying the forelimb that first contacted the plate and the hindlimb that last contacted the plate; forelimbs and hindlimbs that contacted the plate at the same time or partially contacted the plate were excluded from the data. The arrangement of the gait analysis system described above is shown in Figure 1. A handler walked the Beagles using a collar and lead on a stage. GRF data were collected and analyzed using motion analysis software (ToMoCo-FPm; Toso System, Inc).

View Full Size

The overview (A) and the enlarged view (B) around force plate (FP) of the arrangement of the gait analysis system. The enlarged view around FP showed the size of the animal relative to the plates and the plates relative positions. The areas bordered by the white lines were the FP. The dotted lines showed the beams of through-beam photoelectric sensors to measure walking time, velocity, and acceleration.

Citation: American Journal of Veterinary Research 83, 9; 10.2460/ajvr.22.03.0057

• Download Figure
• Download figure as PowerPoint slide

Results

Discussion

In the present study, the SI in the Fx direction was high, whereas the SIs in the Fy and Fz directions were low. As the gait velocity increased, the PVF increased and VI decreased in the forelimbs and hindlimbs. At all velocities, the PBF and BI were significantly higher than the PPF and PI in the forelimbs, whereas the opposite was observed in the hindlimbs. As the velocity increased, the PPF significantly increased in the forelimbs; this tendency was also observed in the hindlimbs. However, there was no difference among PIs at all velocities. As the velocity increased, the PBF significantly increased in the hindlimbs; this tendency was also observed in forelimbs. In contrast, there was no difference among the BIs at all velocities. There was no significant difference among all parameters in the Fx direction at all velocities. However, the PLF was significantly higher than the PMF in the forelimbs during trot 1 and trot 2.

Among different breeds, differences exist in body weight, limb length, height, physique, and gait.^2,19 Moreover, it is better to walk dogs with different physiques at similar relative velocities, rather than absolute velocities, when performing a comparative study of various breeds.^5,6,19 In previous studies,^2,5,6,15 the standardization of the velocity according to physique and other factors was considered. Although standardization of the velocity according to physique and other factors was not performed during the present study, the results may be useful when comparing healthy Beagles with Beagles with diseases when premorbid data are not available.

Since gait velocity affects GFR variability, the approximate gait velocity allows for more accurate GRF measurements.⁸ The number of walking trials decreases when 2 FPs, rather than 1, are used; therefore, it is possible to suppress the variation in GRF caused by trial repetition.⁴ Therefore, during this study, the fluctuations in GRF might have been small. Although a single FP is more widely used, the use of 2 FPs can reduce the number of trials and the measurement time required to collect consistent and reproducible data.²⁰ This would reduce the possibility of worsening of lameness or GRF variations caused by excessive repetition of trials.^4,8,20,21

Because the SI is used clinically to evaluate changes before and after improvement in lameness, it is recommended that variance should be minimized.⁴ The SI can be used to detect lameness because limb symmetry is expected to change with the presence of lameness attributable to the force distribution pattern.²² Moreover, the SI is lower when 2 FPs, rather than 1, are used, and it is possible to reduce the variation in GRFs attributable to the repetition of trials.⁴ During this study, the PVF and VI were within the symmetrical range (< 15%) at all velocities. The SI was also calculated for parameters included in Fx and Fy directions. The PPF, PI, PBF, and BI were also within the symmetrical range at all gait velocities. The SIs of GRFs in the Fx direction were outside the symmetrical range. The SI has been assessed most accurately with PVF, and the VI has been considered valid; moreover, it has been considered inappropriate to assess the SI with other parameters.¹¹ However, the results of the present study showed that GRFs included in the Fy direction were also within the symmetrical range. This was because the symmetrical range was set to be relatively wide (0% to 15%).⁶ If the sample size is further increased, it is possible to investigate the range in which the SI is closer to 0%, such as 6%.⁷ Even when the symmetrical range was set to 0% to 6%, the PVF, VI, PPF, and PI values obtained at all velocities were within the symmetrical range. It has been suggested that the SI is more accurately evaluated with the PVF, VI, PPF, and PI than with the PBF, BI, and GRFs in the Fx direction. Even for a healthy dog, equal weight loading to the left and right sides during walking is not considered easy; furthermore, if even one GRF parameter is asymmetrical, it is not regarded as abnormal,^4,11,23,24 even if the SI is high. However, it was also suggested that forces would be not distributed evenly in this study. When evaluating treatment outcomes of dogs with mild hindlimb lameness and healthy dogs, an assessment of the SI during trot is more sensitive and accurate than that during walk.⁷ The SI values were closer to 0 and more symmetrical during trot than during walk in this study. This is because trot 1 is an efficient gait for healthy dogs because it has a certain speed, which lowers the energy loss to the left and right, and is suitable for prolonged movement.^25,26 However, walking involves a relatively slow gait; therefore, the weight load is not stable and tends to be asymmetrical, and multiple limbs are in contact with the ground at any time. Since the animal’s mass is moved by multiple simultaneous ground contacts, it is important to compare GRFs from the same gait cycle,²⁶ which was not assessed in this study. Moreover, there are several forms of lameness that are not subjectively detected at a trotting speed, while the stance phase is much longer, with greater range of motion of the joints, so the dog tends to show more SI differences when an area of cartilage defect is under strain only during walk.

Subject velocity has a significant effect on the FP values and must be limited to a narrow range when data are obtained.^10,13,14 It was reported that a change in velocity of 0.6 m/s or more affected PVFs of the hindlimbs during trot.^7,10,14 During the present study, changes in velocity did not have an effect on PVF because all velocity ranges were within 0.6 m/s. The range of 1.3 to 1.5 m/s was excluded when setting the gait velocity because it represents a range of velocity that could be either walk or trot, depending on the dog. It is also important to consider whether the animal is accelerating or decelerating,²⁷ which creates a turning moment around the center of mass that changes the relationship between the front and hind vertical forces.²⁶

As the gait velocity increased, the PVF increased and VI decreased in the forelimbs and hindlimbs. The results were in accordance with the prediction that at higher speeds contact time decreases, so that impulse must be provided with higher peak force; therefore, as a stepping force, the force in the Fz direction must increase with increasing velocity, since the animal must support its body weight against gravity: thus, vertical force must increase so that the average vertical force over the stride matches the weight of the animal. These results are similar to those of previous studies.^10,13,14 This may also be true because the stance time shortens as the velocity increases; therefore, VI, the area enclosed by time and force curves, decreases, and the instantaneous force, or PVF, increases.^13,18

Comparisons of the forelimbs and hindlimbs showed them to be loaded at approximately 60% and 40%, respectively, at all velocities, which was consistent with previous reports.^13–15 This may be because the forelimb is more likely to be loaded than the hindlimb because of the weight of the head and the amount of muscle mass in the forebody; in other words, the animal’s center of mass is closer to the forelimbs than to the hindlimbs. At all velocities, the PBF and BI were significantly higher than the PPF and PI in the forelimbs; however, the opposite was observed in the hindlimbs. This suggests that the hindlimbs are mainly responsible for propulsive force and the forelimbs are mainly responsible for suppressive force, consistent with a previous report.^18,26 This explains why the hip retractors are substantially more massive than the protractors, suggesting a greater capacity for acceleration from the hindlimb, although the shoulder has a more equal distribution of retractor and protractor muscle mass with no obvious predominance of protractors.²⁶

As the velocity increased, the PPF significantly increased in the forelimbs; this tendency was also observed in the hindlimbs. This instantaneous increase in the PPF was considered attributable to the shortening of the stance time and PVF.¹³ This increase in the PPF is predictable because at higher speeds contact time decreases, so that impulse must be provided with higher peak force. The force in the craniocaudal direction should be constant under constant velocity, as in the present study. The increase in PPF with increasing velocity is thought to be due to the fact that the increase in PPF of one limb is balanced by the force in the craniocaudal direction in the whole animal. However, there was no difference among PIs at all velocities. In previous studies^10,13,14 of Greyhounds, the PI tended to decrease with increasing velocity; however, this was not consistent with the results of the present study. The PI is calculated by multiplying the propulsive directional force and stance time. It has been suggested that the stance time decreases when the gait velocity increases. This could be because, while the standing time is reduced, the propulsive directional force increased more for Beagles, with shorter limbs compared with Greyhounds, because the dogs walked at the same velocities, which required faster limb movement in the smaller breed.¹³ It was suggested that this tendency would be a feature of small-sized dogs.

As the velocity increased, the PBF significantly increased in the forelimbs and hindlimbs, due to the fact that limb contact is made at higher velocity and larger limb angle, leading to a larger contact collision at higher velocity. These results are similar to those of previous studies.^10,13,14 However, in the present study, no significant difference in the forelimbs during trot 1 and trot 2 was observed. The aforementioned previous studies evaluated Greyhounds. It has been suggested that Beagles, which have a smaller physique and need to move their limbs faster to attain the same velocity,^4,5 would not attain the velocity of trot 2 when the forelimbs are suppressed. In contrast, there was no difference among BIs at all velocities. Previous studies^13,14 of Greyhounds reported that the BI tended to decrease with increasing velocity; however, this was not consistent with the present results. It was suspected that the reason for the discrepancy in results of the BI was the same as that for the discrepancy in results of the PI.

There was no significant difference among all parameters in the Fx direction at all velocities, which was consistent with previous reports.¹⁰ The PLF was significantly higher than PMF in the forelimbs during trot 1 and trot 2; however, there was no significant difference between the PLF and PMF in the hindlimbs at all velocities. In other words, weight bearing at the grounding site of the forelimbs was significantly more lateral than medial. This is because the forelimb is mainly responsible for the PBF and BI, which are suppressive forces, and the force is applied in the lateral direction, which is the stable direction. As previous studies did not show a definite trend in GRFs in the Fx direction, this trend could be a new finding.

A limitation of this study was that we considered absolute velocity rather than relative velocity, which accounts for the physique. Bertram et al¹⁹ compared gaits of Labrador Retrievers and Greyhounds using functional limb length, body weight, and relative velocity to normalize the data.¹⁹ Differences between breeds were mainly attributed to differences in body size, and the dogs moved in a dynamically similar manner. Moreover, some studies⁹ that included GRF data of dogs have provided descriptions of the habituation of dogs to the gait analysis laboratory or data-gathering routine. It is possible that the habituation time of the Beagles included in this study was insufficient; therefore, excessive repetition of walking trials would have led to variations. Eighteen Beagles were included in this study, but that number was insufficient; however, the number of animals was comparable to that of previous studies. As we did not have control over stride length, we tried to keep the velocity of linear motion as constant as possible by strictly controlling the acceleration during GRF measurements. Any acceleration causes a turning moment around the center of mass that redistributes the vertical force between fore and hindlimb,²⁶ which was not considered in this study, in which the GRF occurring in each limb at the same time was not measured.

In summary, the results of the GRF analysis used in the present study may be useful for comparing healthy Beagles with those with diseases when premorbid data are not available. Because the forelimbs are mainly responsible for the suppressive force, it is suggested that weight bearing is more stable in the forelimbs than in the hindlimbs, which are mainly responsible for propulsive force, and that a greater force is generated laterally than medially.

Supplementary Materials

Supplementary materials are posted online at the journal website: avmajournals.avma.org