Materials and Methods

Sample population—Twelve government-owned German Shepherd Dogs housed at the MWD Center at Lackland Air Force Base and under the medical care of the DoD MWD Veterinary Service were used in the study. All dogs were males and ranged from 13 to 22 months of age (mean, 16 months). Body weight ranged from 25 to 35 kg (median, 32 kg). The dogs were assigned to the study by DoD MWD Veterinary Service staff on the basis of a randomized listing of all MWDs purchased during the preceding month.⁹ Dogs were eligible for inclusion in the study when they had no history of orthopedic problems and did not have evidence of orthopedic anomalies, including osteoarthritis, during orthopedic examination and evaluation of radiographs taken of their limbs.

Data collected from 16 purebred Labrador Retrievers and reported elsewhere² were used for comparison with results obtained for the German Shepherd Dogs.

Study design—Angles for flexion and extension of the carpal, cubital (elbow), shoulder, tarsal, stifle, and hip joints of 1 randomly selected^b forelimb and the ipsilateral hind limb were measured by use of an EG and a UG in each dog. The side chosen for measurement alternated for each subsequent dog. Data were collected during processing of MWDs, which involves procedures that include physical and dental examinations and tattooing of the left pinna. All dogs were evaluated within a 10-day period. The protocol was reviewed by the Clinical Review Committee, DoD MWD Veterinary Service and Chief, US Air Force Animal Research Programs, Headquarters US Air Force Surgeon General, Research and Compliance Office.

Radiography—Each dog was medicated with an IM injection (0.11 mL/kg) of a combination of butorphanol tartrate^c (2 mg/mL), acepromazine maleate^d (0.5 mg/mL), and glycopyyrolate (0.05 mg/mL).^e Ten minutes after injection, a catheter was aseptically inserted into a cephalic vein, and anesthesia was induced by IV administration of propofol^f (approx 4 mg/kg, administered as needed to achieve the desired effect). Anesthesia was maintained with a constant rate infusion of propofol (1 to 3 mg/kg/min) administered by use of a syringe pump. Dogs were monitored by direct observation; measurement of body temperature, heart rate, respiratory rate, and oxyhemoglobin saturation; and assessment of an ECG, rhythm and quality of an arterial pulse, capillary refill time, and mucous membrane color.

The forelimb and ipsilateral hind limb selected for goniometric assessment in each dog were radiographed^g with the limbs held in maximal flexion and extension. Flexion and extension of the carpal, elbow, shoulder, tarsal, stifle, and hip joints on radiographs were digitally measured^h by use of a method described elsewhere.²

Goniometry and electrogoniometry—Measurements were obtained in triplicate by use of a UG and an EG by 1 investigator (TMT). The initial measurement was obtained by use of a UG; the measurement was conducted in accordance with a method reported elsewhere.² Measurements were obtained by use of an EG.^i–k The EG measurements were obtained by placing 2 sensors along the long axes of the shafts of the metacarpal, ulnar, humeral, scapular, metatarsal, tibial, femoral, and pelvic bones and moving the limbs until the joint being measured was in maximal flexion or extension (Figure 1). Placement of the sensors for all EG measurements was performed by 1 investigator (TMT). Sensors were placed on the cranial aspect of the bones for the carpal, shoulder, and stifle joints and on the lateral aspects of the bones for the other joints. Sensors were removed and joints were flexed and extended between each subsequent EG measurement.

View Full Size

Photograph depicting use of an EG to measure the cubital (elbow) joint during flexion. The 2 sensors are manually held in position (1 along the long axis of the shaft of the ulna [bottom sensor] and the other along the long axis of the shaft of the humerus [top sensor]). The EG device displays the relative motion of these sensors in 2 perpendicular planes.

Citation: American Journal of Veterinary Research 67, 12; 10.2460/ajvr.67.12.1974

• Download Figure
• Download figure as PowerPoint slide

The amount of time required to obtain UG and EG measurements was recorded. In addition, intraobserver repeatability of EG measurements was evaluated in 1 randomly selected dog by repeating EG measurements 5 times with an interval of at least 15 minutes between subsequent measurements.

After measurements were obtained, anesthesia was discontinued. Dogs were visually observed in a recovery kennel every 5 minutes until they were able to stand and were ambulatory. Body temperature, heart rate, respiratory rate, mucous membrane color, capillary refill time, and rhythm and quality of an arterial pulse were assessed at least once every 15 minutes until the dogs were fully recovered from anesthesia and returned to their separate kennels.

Statistical analysis—The amount of time required to obtain EG and UG measurements was compared by use of a 2-tailed t test.^b Values for UG, EG, and radiographic measurements were compared by use of an ANOVA and least square means.^l Variance of UG and EG measurements were compared by use of F tests^b and 95% limits of agreement.^10,11,l Variances of UG, EG, and radiographic measurements were compared by use of F tests and 95% limits of agreement.^11,l Values for UG measurements and range of motion of all joints measured in the German Shepherd Dogs were compared with those same values measured in 16 Labrador Retrievers by use of unpaired 2-tailed t tests.^b Significance was set at P < 0.05.

Results

Results for UG, EG, and radiographic measurements were summarized (Table 1). In flexion, EG measurements were significantly smaller than UG measurements for the carpus and significantly larger for the elbow joint. Flexion measurements of the EG were larger, but not significantly different, than flexion measurements of the UG for the tarsus. Similarly, flexion measurements of the EG and UG did not differ significantly for the shoulder, stifle, and hip joints. In extension, EG measurements were significantly larger than UG measurements for the elbow joint; however, extension measurements of the EG and UG did not differ significantly for the other joints.

The amount of time required to obtain UG measurements (mean, 30 minutes; range, 21 to 35 minutes) was significantly (P < 0.001) longer than the amount of time required to obtain EG measurements (mean, 15 minutes; range, 8 to 22 minutes). Intraobserver repeatability of EG measurements was determined (Table 2).

The UG measurements for German Shepherd Dogs were lower than UG measurements for Labrador Retrieves during flexion and extension of the elbow, shoulder, tarsal, stifle, and hip joints (Table 3). The UG measurements of the carpus in flexion and extension did not differ between German Shepherd Dogs and Labrador Retrievers. Values for range of motion as determined by use of the UG were significantly lower in the tarsus of German Shepherd Dogs than in the tarsus of Labrador Retrievers; however, values for range of motion did not differ significantly between the breeds of dogs for the other joints.

Discussion

In the study reported here, we compared measurements obtained by use of an EG and a UG to each other and to measurements obtained from radiographs taken when dogs were anesthetized. We used the UG and radiographic methods described in another study² conducted to validate goniometry in Labrador Retrievers.

Analysis of results of the study reported here indicated that UG, EG, and radiographic measurements were not interchangeable. The EG measurements were more variable than were measurements collected by use of the UG in 8 of 12 joint positions evaluated. Variability of EG measurements was significantly higher than variability of radiographic measurements for 6 of 12 joint positions evaluated. Because we conducted EG measurements on live dogs, we cannot determine the fraction of EG variability resulting from the dogs, the device, and the methods used in this study. The fact that variability in EG measurements among dogs was not higher than intraobserver variability strongly suggests that variability attributable to the dogs had a low impact on variability of EG measurements. Comparing measurements made by use of an EG directly connected to a UG would have allowed us to assess variability of EG measurements resulting from the device.¹² Variability of another EG device was 2° to 4° (3% of measurements).¹²

Variability of EG measurements may have resulted from motion of the sensors in relation to the skin during repeated measurements. The decision was made to manually hold the sensors in place during EG measurements because we had the clinical impression that manually holding the sensors in place would minimize sensor motion. To our knowledge, optimal sensor stability in dogs has not been assessed. Sensor motion may have been decreased by obtaining EG measurements after clipping the haircoat in the region in which the sensors would be placed, by measuring values in dogs with short haircoats, by varying the sensor positions, and by strapping or gluing sensors to the skin. Analysis of results of a study¹³ in which investigators compared UG and EG measurements obtained for humans indicated that changes in clinical judgment based on angular changes of < 10° were invalid when rigid goniometric protocols were not used.

Measurements were collected more rapidly by use of the EG, compared with the amount of time required to obtain measurements by use of the UG, primarily as a result of the ease in reading the value on the EG once the sensors were properly positioned. That difference is unlikely to have a substantial impact on the practicality for use in clinical conditions.

The UG measurements in the study reported here were less variable than were the EG measurements. Nevertheless, the UG measurements differed significantly from radiographic measurements for 6 of 12 joint positions. The decreased variability of UG measurements, compared with the variability of EG measurements, may have resulted from the device, the methods used in the study, or differences in our experience with the UG and EG. Differences between UG and radiographic measurements were not identified in a study² in which investigators used similar methods for measurements of Labrador Retrievers. The UG measurements in the study reported here were lower than radiographic measurements for carpal, elbow, tarsal, and stifle joints during flexion and for stifle and shoulder joints during extension. This difference may have been influenced by the fact that the dogs in this study were anesthetized to enable radiographs to be taken, but the Labrador Retrievers in that other study² were only sedated. It could also have been influenced by the fact that variability of UG and radiographic measurements was lower in the study reported here than in the study² of Labrador Retrievers. This may have resulted from the fact that a single investigator performed UG measurements in this study (compared with measurements made by 3 observers in the Labrador Retriever study) and, possibly, because the German Shepherd Dogs in this study had less morphologic variability than the Labrador Retrievers enrolled in the other study. Finally, differences between UG and radiographic measurements may have been the result of the use of slightly differing anatomic landmarks for the 2 methods, particularly for flexion of the stifle joint in which there was a difference of 25° between UG and radiographic measurements.

The variability in measurements in the study reported here did not appear to be influenced by the joint measured, with the exception of the hip joint. Variability was increased for this joint. Lack of an ability to detect specific differences in variability between joints was likely attributable to the relatively small sample size. In humans, joint mechanics influence the variability of goniometric measurements. For example, measurements of the elbow joint are more repeatable than are measurements of the carpus because the elbow joint is a single-axis hinged joint, whereas the carpus has motion in multiple directions.¹ The increased variability of EG measurements for the hip joint may have resulted from challenges in identifying the long axis of the pelvis or the increased cross-talk of the EG sensors resulting from specific placement of EG sensors during measurement of the hip joint.¹⁴ The increase in EG cross-talk is a phenomenon whereby motion of a joint in 1 plane (eg, abduction-adduction) influences the perceived motion of that joint in another plane (eg, flexion-extension).¹⁵ The phenomenon of EG cross-talk is particularly relevant in the carpal and hip joints of humans.¹⁴ The high EG variability for hip joints is a particular concern because motion of the hip joint is particularly important when screening German Shepherd Dogs for hip dysplasia.

Measurements collected for German Shepherd Dogs in the study were lower than measurements obtained for Labrador Retrievers for all joints but the carpal joints in another study² conducted by our laboratory group. This finding is in agreement with our clinical impression and data collected from clinically affected patients during the past 15 years. Differences between German Shepherd Dogs and Labrador Retrievers may have resulted from differences between evaluators, patterns of samples tested, and methods. Although male and female Labrador Retrievers were assessed, all dogs in the study reported here were males. The influence of sex on range of motion in dogs is unknown. In humans, the heavier muscle mass in males typically limits the motion of specific joints, particularly joint motion that has soft tissue approximation at the end of motion (ie, flexion of the stifle joint).¹⁶ Joint motion appears to be influenced by the increased weight for men, compared with that for women, rather than as a direct influence of sex.¹⁷ Mean weights were identical (32 kg) in the German Shepherd Dogs and Labrador Retrievers tested. If muscle mass was an important factor that influenced joint motion in the male German Shepherd Dogs, compared to the male and female Labrador Retrievers, it would have been expected that the German Shepherd Dogs would have less flexion for their stifle and shoulder joints. This is contrary to the findings of the study.

Differences between German Shepherd Dogs and Labrador Retrievers may have resulted from the slightly lower age of the German Shepherd Dogs (mean, 16 months), compared with that for the Labrador Retrievers (mean, 36 months). Although the effect of aging on joint motion in dogs is not known, joint motion decreases with aging in humans.¹⁸

German Shepherd Dogs included in the study reported here and the Labrador Retrievers in the other study² were screened for osteoarthritis; however, joint motion would possibly decrease with age, even in joints without clinical and radiographic signs of osteoarthritis. If age was a factor in the difference in range of motion between German Shepherd Dogs and Labrador Retrievers, it would be expected that older dogs would have less joint motion (ie, higher scores for flexion and lower scores for extension). Again, this is contrary to the findings of the study.

Differences between German Shepherd Dogs and Labrador Retrievers may have resulted from the fact that the German Shepherd Dogs were anesthetized, whereas the Labrador Retrievers were only sedated. The influence of anesthesia on joint motion has not been assessed in dogs. In humans, anesthesia influences specific joint motions, including the cranial-caudal motion of normal knee joints.¹⁹ Also, passive measurements are more variable than active measurements in humans.¹

On the basis that we considered the aforementioned factors to have little effect, the differences between German Shepherd Dogs and Labrador Retrievers most likely resulted from differences in joint shape, muscle mass, and biomechanics. Even though German Shepherd Dogs had lower values for angles during flexion and extension, the range of motion of their joints was similar to that of Labrador Retrievers for all joints, except the tarsus, where the range of motion was smaller in German Shepherd Dogs than in Labrador Retrievers. The smaller range of motion for the tarsal joint resulted from the significantly lower extension in the tarsal joint of German Shepherd Dogs. The combination of lower values for flexion and extension with a similar range of motion suggests that differences exist in bone shape but not in joint shape. Although the stance of German Shepherd Dogs and Labrador Retrievers differs significantly,^20,21 little is known about the difference in bone and joint shapes for various breeds of dogs. Differences in tibial plateau slopes were identified between Greyhounds and Labrador Retrievers in 1 study.²² Such differences may result in important differences in joint mechanics. A specific assessment of differences in bone and joint shape between German Shepherd Dogs and Labrador Retrievers was beyond the scope of the study; however, such an assessment in these and other breeds would be clinically beneficial because it could enhance our ability to manage animals affected by several pathologic conditions of bones and joints.

The study reported here provides reference UG measurements for German Shepherd Dogs. These measurements may be of value because joint disease, particularly osteoarthritis, is a major problem affecting MWDs and is responsible for a reduction in effective working life of these highly trained animals. In 1 report,^a 43 of 245 (18%) MWDs were discharged because of osteoarthritis. Early detection by use of serial objective joint measurements may enable early intervention to improve performance and increase time of service.

We conclude from the study reported here that EG measurements are more variable than UG measurements, particularly in hip joints. Therefore, we do not recommend the use of an EG to assess range of motion with the methods described in this study. German Shepherd Dogs have ranges of motion similar to those in Labrador Retrievers but have lower scores for angles during flexion and extension, most likely as a result of differences in bone shape.