The aim of this study was to develop a model of the canine forelimb with special reference to the elbow. Retrospectively 12 dogs were evaluated, as well as a reference measurement. The retrospective measurement was conducted in walk, whereas the reference measurement was conducted in walk and trot. Based on this reference measurement, two different marker sets were compared. The first hypothesis was that the joint kinematics based on the simulations of the evaluated joints are similar to the in vivo measured values. The second hypothesis was that the Standard-Markerset provides as accurate results as the Reference-Markerset. Material and Methods A Morphometric model basis: entire amputated left forelimb of a private owned, non-neutered dog (26,5 kg ; 2,5 years) B Reference Measurement: clinically sound mixed breed female dog (21 kg; 10 years) C Retrospective Evaluation: 12 private owned dogs (24,1 ± 5,2 kg; 4,5 ± 1,8 years) The clinical and neurological examination and the orthopaedic examination of the dogs where within normal limits.
For excluding lameness of the measured dogs a treadmill with four integrated force plates was used for kinetic measurements. Kinematics were captured by recording either five (Standard-Markerset) or 16 (Reference-Markerset) reflecting markers in anatomical accurately defined places by ten cameras and the Eagle Digital Real Time System of the Motion Analysis Corp., Santa Rosa, CA, USA. For kinematic recordings CORTEX 126.96.36.1995 (Motion Analysis Corp., Santa Rosa, CA, USA) was used.
Results All data was normally distributed (Kolomogorov-Smirnov-Test). High correlation values were determined between the simulated movement patterns and the real measurements. Statistical analysis showed that there is a significant shift between the direct kinematics and the values acquired from simulations. In the simulations it was showed that the maximum extension of the Standard-Markerset significantly differed in all measurements from the reference measurements. While the Reference-Markerset only differed when evaluating the elbow and the carpus in walk. Regarding flexion simulation resulted in a different situation: the simulation showed comparable results regarding the Standard-Markerset in both gaits for the elbow, whereas for the Reference-Markerset the maximal flexion decreased significantly. When simulating the shoulder and carpal joint both markersets showed significant differences from the reference measurement. The temporal parameters showed a significant shift in respect to the extension when using either markerset: for the shoulder joint to the left (earlier onset in the movement cycle), for the other two joints to the right. The temporal occurence of the flexion was found to be slightly more stable, for the elbow were both markersets showed the same values as the reference measurement. The simulations which are based on the retrospective data analysis showed a similar statistical result as those which were collected during the reference measurement. A significant increase happened in the maximum flexion and extension, and the range of motion.
Discussion It can be assumed that both marker sets are giving a good representation of the basic movement pattern of the joints. Although significant differences were found, they should be interpreted regarding to their clinical relevance too. A majority of the described alterations differ only a few units from direct kinematics - these are significant, although the clinical significance is questionable. A possible explanation for the differences might be issues of scaling and skin movement. Therefore the first hypothesis must be rejected, whereas the second can be approved. As a conclusion the simulation, or modeling in general pose a great potential for gathering new information, but it should not be forgotten that the development and especially the testing of a model is difficult and can easily be biased.