Materials and Methods

Animals—Five Greyhounds (3 males and 2 females; mean age, 6.5 years; mean ± SD body weight, 30.2 ± 6.1 kg) with no recent history of lameness or other gait defects were used in the study. None of the dogs had a history of lameness of the hind limbs, and their gait was judged to be sound and clinically normal by a veterinary surgeon (CJF) on the days they were tested. The methods involved no invasive measurements, and the study was approved by the local ethical review committee.

Gait measurements—The measurement protocol has been described in detail elsewhere.¹ Briefly, small reflective markers were attached to the skin over the centers of rotation of the hip, stifle, tarsal, and MTP joints and the tuber coxae and distal portion of the foot on the right and left hind limbs. Kinematic data were collected by use of a 4-camera system^a synchronized with data from a force platform.^b Force and kinematic data were collected at 200 Hz for 3 s/trial. The 2 systems were calibrated in the kinematic frame with the origin of the kinematic reference located at the center of the force platform. A calibrated volume of space measuring 4 m in length, 1 m in width, and 1 m in height was recorded to enable tracking of markers through the stride cycle of the hind limbs with a dog trotting in either direction. Cameras were arranged in a semicircle on 1 side of the runway at heights varying from 0.8 to 2 m, such that each marker was visible to at least 3 cameras for resolution of a 3-dimensional image of the hind limbs during the stance and swing phases. Accuracy of the marker locations in the sagittal plane was determined to be ± 3 mm.

A handler initially guided each dog on the runway at a preferred velocity selected by each dog (medium velocity). This velocity was established during 8 to 10 warm-up trials that habituated the dogs to the testing area and runway. When a dog was trotting alongside the handler with minimal guidance, data were collected from a minimum of 5 strides of the right and 5 strides of the left hind limbs. Accuracy of the calculation for the center of pressure under the foot is crucial to the inverse-dynamics method and is dependent on the force transducers having a reasonably even load; thus, only trials in which the foot of each hind limb landed near the middle of the force platform were recorded.

The handler then encouraged each dog to run (fast velocity) for an additional 5 trials for each hind limb. The handler then walked beside each dog as it trotted on the runway (slow velocity). Actual trotting velocity was calculated for each trial from the stride length and stride time of the hind limbs by use of the following equation: velocity = stride length/stride time. Stride length was determined from the horizontal displacement of the marker on the paw of a hind limb between subsequent ipsilateral ground contacts, and stride time was computed from the elapsed time between ground contacts.

Kinetic calculations—Breed-specific morphometric data¹ were used in a link-segment model consisting of pelvis (tuber coxae-hip), thigh (hip-stifle), crus (stifle-tarsus), metatarsus (tarsus-MTP), and foot (MTP-fourth digit). Force and kinematic data were imported into a custom-designed computer program^c and combined with the breed-specific morphometric data. An inverse-dynamics method^4,5 was used to calculate joint angular excursions, net joint moments, and net joint powers in the sagittal plane for the hip, stifle, tarsal, and MTP joints. Moments on the cranial or dorsal side of the joints were assigned positive values, and moments on the caudal or palmar side of the joints were assigned negative values. Output from right and left hind limbs were combined within each velocity group after stride durations were adjusted on the basis of mean stride time and forces were adjusted on the basis of body weight, which yielded mean values for each dog. The values for each dog were then used to calculate the mean values for all dogs; mean values for all dogs were plotted as mean curves for comparison of velocity effects.

Total moment was calculated as the sum of the joint moments for the hip, stifle, tarsal, and MTP joints, after reversing the stifle moment so that the extensor moment would have negative values, similar to the convention used for the other 3 joints. It was necessary to convert these values because of our decision to assign negative values to moments on the caudal or palmar side of the joints, irrespective of whether the associated muscle groups were anatomically flexors or extensors. As a result, the moments generated by muscle groups contributing to extension of the limb joints during the stance phase, and therefore support of the trunk against gravity, could be summed arithmetically. Total power was computed as the arithmetic sum of the various joint powers.

Statistical analysis—The mean for each of the 3 trotting velocities (slow, medium, and fast) was calculated from the stride length and stride time values of each trial, and means of each calculated velocity were tested by use of a 1-way ANOVA to detect significant differences. Local peak amplitudes of angles, moments, and power for each joint were identified and tested by use of a 1-way ANOVA and Tukey post hoc test to detect significant differences among the 3 velocities. Results were considered significantly different at values of P < 0.05.

Results

Mean trotting velocity for each of the 3 speeds (slow, medium, and fast), along with mean values for stride length, stride time, stance time, and swing time, was calculated (Table 1). Values differed significantly (P = 0.01) among the 3 velocities. Stance time was shortened more than swing time with increases in stride length and velocity.

Mean joint angular excursion, mean net joint moment, and mean joint power for the hip, stifle, tarsal, and MTP joints for each of the 3 velocities were plotted (Figures 1–4). Data were plotted to the same time scale, which enabled us to detect that stride time shortened with increases in velocity. The stride began at contact of a hind limb on the force platform and ended with contact of the ipsilateral hind limb on the runway. For the moment graphs, negative values indicated a net palmar moment and positive values indicated a net dorsal moment. For the power graphs, positive values indicated that the muscles responsible for the measured net moment were performing concentric (shortening) contractions, whereas negative values indicated eccentric (lengthening) contractions. Local peak amplitudes of angles, moments, and powers were sequentially identified and labeled (1, 2, 3, etc) and statistically compared among velocities (Table 2).

View Full Size

Mean joint angular excursions (A), net joint moments (B), and net joint powers (C) of the hip joint for 5 Greyhounds trotting at slow (dotted line), medium (dashed line), and fast (solid line) velocities. Discrete local peak amplitudes are sequentially labeled 1, 2, 3, and 4.

Citation: American Journal of Veterinary Research 67, 8; 10.2460/ajvr.67.8.1293

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Mean joint angular excursions (A), net joint moments (B), and net joint powers (C) of the stifle joint for 5 Greyhounds trotting at slow (dotted line), medium (dashed line), and fast (solid line) velocities. See Figure 1 for remainder of key.

Citation: American Journal of Veterinary Research 67, 8; 10.2460/ajvr.67.8.1293

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Mean joint angular excursions (A), net joint moments (B), and net joint powers (C) of the tarsal joint for 5 Greyhounds trotting at slow (dotted line), medium (dashed line), and fast (solid line) velocities. See Figure 1 for remainder of key.

Citation: American Journal of Veterinary Research 67, 8; 10.2460/ajvr.67.8.1293

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Mean joint angular excursions (A), net joint moments (B), and net joint powers (C) of the MTP joint for 5 Greyhounds trotting at slow (dotted line), medium (dashed line), and fast (solid line) velocities. See Figure 1 for remainder of key.

Citation: American Journal of Veterinary Research 67, 8; 10.2460/ajvr.67.8.1293

• Download Figure
• Download figure as PowerPoint slide

The hip joint extended through the stance phase and then flexed into the swing phase before reverting to extension for contact with the ground at the end of the swing phase (Figure 1). An extensor moment dominated the stance phase, which was followed by a flex or moment at the end of the stance phase to restrain hip extension and then to initiate flexion of the hip for the swing phase. At the end of the swing phase, a net extensor moment reversed the protraction motion of the hip joint. Power of the hip joint was positive throughout most of the stance phase, which indicated that the hip extensors were the net active muscle group and contracted concentrically to generate an extensor moment. A small burst of negative (ie, eccentric) work was performed by the hip flexors as they restrained extension at the end of the stance phase, and positive work was then performed by the flexors to initiate flexion of the hip into the swing phase. At the end of the swing phase, the extensors did positive work to extend the hip for placement of the limb on the ground. Hip extensor power for the fast velocity was significantly greater than the corresponding power for the slow velocity during the early portion of the stance phase.

The stifle joint flexed after contact of the hind limb and then was extended through the last part of the stance phase before flexing into the swing phase (Figure 2). There was a net flexor moment through the first half of the stance phase followed by a net extensor moment through the second half of the stance phase, which persisted into the swing phase as the joint was passively flexing. At the end of the swing phase, a net flexor moment dominated to restrain extension. Power was positive through the stance phase, initially coinciding with the flexor moment generated by the stifle flexors as they shortened and then coinciding with the extensor moment generated by the stifle extensors in the second half of the stance phase. Energy was absorbed by the stifle extensors as they restrained the passive flexion of the stifle joint during the early portion of the swing phase and by the stifle flexors as they restrained overextension of the stifle joint at the end of the swing phase. There was no significant difference in the moment peaks among velocities, but stifle power during the early portion of the stance phase was significantly greater for the fast velocity than for the slow velocity.

The tarsal joint flexed during the early portion of the stance phase and then extended through the late portion of the stance phase and into the early portion of the swing phase before flexing for the remainder of the swing phase (Figure 3). The net tarsal joint moment was extensor throughout the stance phase. Power was negative during the first half of the stance phase as the tarsal extensors absorbed energy in an eccentric contraction. This was then partially returned in a burst of positive work at the end of the stance phase for an active push-off. The net tarsal joint moment was significantly (P = 0.01) greater for the fast velocity than for the slow or medium velocities. Net negative power during the early portion of the stance phase was significantly (P = 0.01) greater for the fast velocity than for the slow or medium velocities. In the second half of the stance phase, the positive power output from the tarsal extensors was significantly (P = 0.01) greater for the fast velocity than for the slow velocity.

The MTP joint dorsiflexed through most of the stance phase before plantarflexing rapidly at the end of the stance phase and then dorsiflexing through the swing phase. The net moment was flexor through the stance phase as the MTP flexors restrained dorsiflexion. Because of this, power had negative values. For the slow velocity, there was a small period of positive work as the MTP flexors pushed off against the ground at the end of the stance phase. Despite the incremental increases in moment and power curves with increases in velocity, there were no significant differences in either variable among velocities.

Total moment and power, which were calculated as the arithmetic sums of the joint moments and powers for each joint, were plotted (Figure 5). Extensors of the hip, stifle, and tarsal joints, along with the flexors of the MTP joint, contributed to the net negative moment supporting the trunk during the stance phase, and this increased incrementally with increases in velocity. The total moment became flexor during the early portion of the swing phase and reversed to extensor again at the end of the swing phase. Total moment of the stance phase for the fast velocity was significantly greater than for the slow velocity. Only extremely small differences in moment amplitudes of the swing phase were evident among the 3 velocities. With increases in velocity, the peak negative power during the first half of the stance phase increased, but this energy was absorbed over a shortened time course. Peak negative power for the fast velocity was significantly greater than for the slow velocity. Similarly, peak power generation during the second half of the stance phase increased but without much decrease in the time frame during which positive work was performed. Peak positive power was significantly (P = 0.01) greater for the fast velocity, compared with values for the slow or medium velocities.

View Full Size

Mean net total joint moments (A) and net total joint powers (B) of all joints of the hind limb for 5 Greyhounds trotting at slow (dotted line), medium (dashed line), and fast (solid line) velocities. See Figure 1 for remainder of key.

Citation: American Journal of Veterinary Research 67, 8; 10.2460/ajvr.67.8.1293

• Download Figure
• Download figure as PowerPoint slide

Discussion

In response to changes in velocity, the angular displacement, moment, and power curves largely had changes in timing and local amplitudes without changes in their overall patterns. Increased velocity was associated with longer stride length and shortened stride time, which made all curves more compressed in the time domain. Interestingly, kinematic patterns for the stance and swing phases were compressed in a fairly uniform manner, whereas the moment and power patterns were shifted to a greater degree during the swing phases. This was especially true for the moment and power curves of the hip and stifle joints. Stance phase peaks were not dramatically shifted in time but did have significant increases in burst amplitude. This indicated that with increases in velocity, the peak power amplitudes during the stance phase were relatively later in the stride at these joints.

Groups of curves tended to have stepwise increases in local amplitudes with increases in velocity, but most of these were not statistically significant, probably in part because of the small number of subjects. There are relatively few opportunities to perform positive work during the gait cycle, and analysis of the results reported here indicates that most of the increases in power were during the early portion of the stance phase. Concentric work from the hip extensors and stifle flexors, which cross the palmar sides of both these joints, dominated the early portion of the stance phase. The tarsal extensors underwent an eccentric contraction during this period, and the positive flexor power measured at the stifle joint coincides with the negative power seen at the tarsus, which implies that the energy absorbed by the tarsal extensors through flexion of the tarsal joint contributed to the flexor moment at the stifle joint. In fact, the positive work subsequently performed by the tarsal extensors to extend the tarsal joint is less than the negative work preceding it, which indicates that some of the stored energy was released else-where (probably across the flexor aspect of the stifle joint). This power flow is a phenomenon attributed to 2-joint muscles^6,7 whereby energy is moved through segments to be conserved and to optimize the mechanical energy cost of gait. For example, the rectus femoris muscle may act in such a manner. As a flexor of the hip joint, it absorbs energy near the end of the stance phase while the femur is extending on the pelvis; this energy would then be transferred across the stifle joint as part of the concentric extensor moment of the quadriceps muscle. The large phase of negative work performed across the tarsus dominates the total power during the early portion of the stance phase in that the amount of total net work performed has a negative value (Figures 3 and 5). During the late portion of the stance phase, the positive amount of work performed for an active push-off is attributable to the energy returned from the concentrically contracting tarsal extensors and the quadriceps muscle across the tarsal and stifle joints, respectively.

A period of positive work is evident across the hip joint during the early portion of the swing phase to accelerate the hind limb into protraction. This burst of work is not directed against the ground, but the energy put into this motion is reclaimed during the late portion of the swing phase as it is transferred into the trunk when extensors of the hip joint contract eccentrically to reverse the limb's motion for placement. The slope of the kinematic pattern reveals that this period of hip flexion is more rapid with increases in velocity. The moment measured was similarly slightly increased, but the calculated power was not increased. The main increase in power across the stifle joint was during the early portion of the stance phase as the flexors of the stifle joint (which also cross the hip on its extensor side) contracted concentrically to generate a larger flexor moment. A second power burst for push-off during the late stance phase was detected earlier with increases in velocity, but the power amplitude was unchanged, which indicated that this push-off burst did not increase with increases in trotting velocity.

Despite variations in the amplitude and rate of angular motion during the swing phase, there were no incremental differences in amplitudes of kinetics during the swing phase. This was most likely attributable to the dependence of forces in the kinetic calculations. During the stance phase, when the reaction forces acting across joints are larger with increases in trotting velocity, the calculated moments and powers will similarly be larger. During the swing phase, the small increases in angular velocity of the limb segments concomitant with the shortened swing time and longer stride length were not sufficient to combine with inertial forces and cause larger joint moments or powers. In fact, in the most distal segments, the moments and powers were so small that they were inconsequential; thus, only the hip and stifle joints had measurable kinetics during the swing phase.

Total limb power was interesting in that the negative and positive power bursts during the stance phase varied with regard to time and amplitude characteristics (Figure 5). During the first half of the stance phase, amplitude increased with increases in velocity at the expense of the time during which negative work was performed, whereas during the second half of the stance phase, the amplitude increased with increases in velocity but the time during which the positive work was performed remained fairly constant. Because work performed across a joint is calculated as the area under each curve, this implies that the total amount of negative work performed during the first half of the stance phase remained relatively constant with increases in velocity. However, the amount of positive work performed through concentric contractions during the second half of the stance phase increased incrementally with increases in velocity. Energy generated by extensors of the hip joint and flexors of the stifle joint during the first half of the stance phase was coincidental with energy absorption by extensors of the tarsal joint and flexors of the MTP joint for the total net power. Energy stored in extensors of the tarsal joint is given back during the second half of the stance phase along with a small concentric burst from the extensors of the stifle joint, but these are the only positive contributions to push-off. Only the burst of power from the extensors of the tarsal joint varies with changes in velocity during this time period, indicating that they are the only muscle group that contributes to increases in velocity during the late portion of the stance phase.

An increase in trotting velocity was associated with increases in positive power from extensors of the hip joint and flexors of the stifle joint during the first half of the stance phase and with an increase in tarsal power during the second half of the stance phase. There were no incremental increases in power during the swing phase associated with increases in velocity. Qualitatively, the overall patterns of power at the various joints appeared to be conserved with increases in velocity because only local curve amplitudes increased as trotting speed increased.