Knee instability is a common complaint in osteoarthritis (OA), and a common reason for revision following total knee arthroplasty (TKA). Despite this, assessment of instability is hampered by the lack of a validated method of objective classification or quantification, with most research relying upon patient reports of frequency of symptoms. The aim of this thesis is to define a theoretical framework for instability in the knee, and to develop a protocol for the classification and quantification of instability in the native and prosthetic knee. Instability of the knee in this thesis is understood as the failure of the joint to return to a zero-state following perturbation using all the available active and passive mechanisms available to it, resulting in system collapse. Symptomatic instability is the awareness of reaching the boundary between the stable and unstable state. The prevalence of subjective instability in the end stage OA knee was measured from a publicly available database of pre-operative knee scores from TKA patients, while the prevalence of instability as a cause of revision was assessed from case note review of TKA revision patients from a tertiary referral orthopaedic unit. A single channel, tibia mounted accelerometer was selected for assessment of frontal plane knee movement during normal walking and a protocol developed its use. This was assessed for its repeatability and compared with standard gait analysis in healthy volunteers, and subjectively stable and unstable post-operative TKA patients. Found to be repeatable with differentiation of output between subjectively stable and unstable TKA, the protocol was adapted and used to compare subjectively stable and unstable OA knees prior to TKA. Using patient subjective assessment as classifier, wavelet transforms, Principal Component Analysis and linear regression was used to produce a classification model from the accelerometer data. The single accelerometer was found to produce classification with an accuracy of 84.6%, sensitivity of 93.3% and specificity of 72.7%, with area under the curve (AUC) of 0.797. This classification model for instability produces the basis from which the protocol can be adapted and developed to improve performance and ultimate quantify instability in the knee for use in clinical and research settings.
|Date of Award||14 Feb 2020|
- University Of Strathclyde
|Supervisor||Phil Riches (Supervisor) & Andy Kerr (Supervisor)|