Ligament injuries have been implicated as an important cause of osteoarthritis. Joint instability following ligament injuries results in abnormal joint motion patterns, however a quantitative relationship with the development of arthritis has not been demonstrated. In an attempt to relate these changes to the progression of arthritis, an unconstrained knee loading system was designed to measure the envelope of joint motion in an animal model. The design goal of this system was such that future applications could incorporate sequential in-vivo application, removal and re-application to the knee at different time intervals throughout the injury and healing process. Part of this thesis involved validating the system performance with sequential application, documenting the effects on accuracy and reproducibility.

The testing system consisted of an unconstrained loading frame which followed the knee through flexion-extension, a load application system consisting of a deadweight and cable and pulley mechanism, a load cell which measured the moments applied to the tibia, and a magnetic tracking device which measured the joint kinematics. The system was integrated through custom written Lab VIEW software using an A/D board and the serial port on a personal computer. Knee flexability was evaluated under prescibed varus-valgus and iniernai-extemal loads followed by measurement of the resulting tibial motion patterns. After testing, the femoral head and condyles were digitized using a stylus attached to the tracking device receiver. A curve fitting routine determined the centre of the bony landmarks to transform the kinematic data into the femoral coordinate system.

The system was verified on seven right femurs taken from mature female New Zealand White rabbits. The varus-valgus and internal-external envelopes of motion were measured for multiple runs of the system, examining the inter and intra-specimen reliability. The effects of pre-conditioning on the visco-elastic behaviour was also examined.

Remounting of the knee in the system was assessed in two ways. Re-application of the components was assessed individually by remounting each component and measuring the variability in the orientation for each. An overall evaluation of re-application was also measured by completely removing and re-applying one knee to the system and comparing the motion pathways for the two applications.

Pre-conditioning showed motion pathway convergence after the first four cycles. Qualitatively the motion pathways showed good inter-animal agreement between the seven specimens. The magnitude of the valgus pathway was observed to be larger than the varus, and the internal pathway was observed to be larger than the external. In addition, significant differences were seen in the varus-valgus and internal-external laxity as a function of flexion angle.

Loading data as recorded by the load cell showed fairly constant loading for each specimen in both the varus-valgus and internal-external configuration. This consistency was also demonstrated in the reproducibility of the kinematic envelopes, since the amount of knee rotation was dictated by the loading level.

Assessment of serial application of the system showed maximum variations of less than 1° in the varus-valgus and less than 3° in the internal-external motion envelopes. The largest individual source of variability in the system (5.02°) was the digitization of the femoral landmarks. A numerical calculation of the effect of a 5° error in the determination of the femoral axis system demonstrated a minimal effect on the motion envelopes.

In conclusion, the motion pathways for the seven in-vitro knees tested were found to be reproducible both intra- and inter-animal. Moderate consistency and good reproduciblity was observed in the knee loading patterns. This system should provide an accurate and reproducible method of determining in-vivo changes in kinematics with the progression of joint disease.