Biomechanical function in knee osteoarthritis and post-total knee replacement: comparing subjective and objective outcomes and predicting gait function post-total knee arthroplasty

De Vecchis, Marina 2023. Biomechanical function in knee osteoarthritis and post-total knee replacement: comparing subjective and objective outcomes and predicting gait function post-total knee arthroplasty. PhD Thesis, Cardiff University. Item availability restricted.

Preview	PDF - Accepted Post-Print Version Download (6MB) | Preview
	PDF (Cardiff University Electronic Theses and Dissertations Form - Memorandum of Understanding) - Supplemental Material Restricted to Repository staff only Download (287kB)

Abstract

Patient-reported outcomes (PROMs) have been widely utilised to evaluate the TKR outcome and to predict it. Although PROMs inform on the patient’s perception of function, they are mostly influenced by pain levels, and relate poorly to what patients can achieve objectively. For this reason, both subjective and objective function should be measured to comprehensively quantify the impairments pre-TKR, the improvement post-TKR and to aid in TKR outcome prediction. However, there is no gold standard to measure function objectively. This research aimed to advance the application of the Cardiff classifier, a measure of gait biomechanics, to subjects with severe knee OA and post-TKR, to compare the Classifier to similar measures of gait biomechanics, identify factors predictive of the gait biomechanics post-TKR, and investigate the relationship between biomechanics, patient-reported outcomes and physical performance pre- and post-TKR. 3D gait analysis was performed in two cohorts of non-pathological subjects (NPs) and patients pre and one-year post-TKR (Cardiff and Karolinska cohorts). The Cardiff classifier’s Belief of OA (BOA), Gait Deviation Index (GDI) and the GDI-kinetics were utilised to evaluate patients’ objective gait function at each time point in both cohorts. The BOA had a large responsiveness to change, which was greater than the GDI and GDI-kinetic in 39 patients from the UK and 29 from Sweden. While the correlation between BOA-GDI and BOA-GDI-kinetic was moderate pre-TKR in both cohorts, the two gait indexes and their change pre to post-TKR showed poor or mixed agreement with the BOA post-TKR or its change score. By comparing the outputs of the classifiers developed from each cohort, it was found that about 55% of the highest-ranking gait features discriminating patients pre-TKR to their references were the same or similar between Cardiff and Karolinska patients. Gait biomechanics improved in both patients’ groups but mostly did not return to normal one-year post-TKR. In the Cardiff cohort mentioned above, it was demonstrated that when comparing the patients gait function to NPs of similar age (NP50 classifier), the BOA was significantly lower (=better gait) pre- and one-year post-TKR versus comparing patients to a younger group of NPs (mixed-age classifier), but the change in gait function was comparable between the NP50 and mixed-age classifiers one-year post-surgery. Pre-surgical and surgical factors did not correlate to the change in BOA one-year post-TKR (NP50 classifier). A regression model revealed that the objective gait function pre-TKR, sex and BMI explained 56% of the variance of the gait function one-year post-TKR; there was a significant association between a worse gait function pre-surgery and a worse gait biomechanics one-year post-TKR, irrespective of sex and BMI. A patient sub-group analysis also showed that a greater knee ROM pre-TKR was associated with a better gait function post-TKR. 3D gait analysis data, performance-based tests (PBTs) (timed-up and go, 40m fast-paced walk test, stair climb test and 30s chair test), Oxford Knee Score and Knee Injury and Osteoarthritis Outcome Score were collected from patients pre, three and six months post-TKR. It was found that trunk kinematics in the frontal, sagittal and transverse planes were not relevant in aiding in the discrimination of gait biomechanics between 9 NPs (n = 18 knees) and 18 subjects with late-stage OA (n = 20 knees) within the Cardiff classifier. Results showed a correlation, or trends of association, between gait biomechanics and the core PBTs suggested by OARSI (40m fast-paced walk test, stair climb test and 30s chair test), pre, three and six months post-TKR. However, no correlation, nor a trend of association, could be found between PROMs and gait biomechanics or between PROMs and PBTs pre- or three, six and approximately twelve months post-TKR. Employing the Cardiff classifier to assess in vivo knee kinematics during a step-up motion showed an 83.8% accuracy in discriminating between severe knee OA (n = 18) and NP knee function (n = 19). The novel application of the Cardiff classifier to knee kinematics data collected via single-plane fluoroscopy showed that two years post-TKR, the knee function improved but was not comparable to NPs. The level of classification uncertainty was higher than previous studies employing the classifier, suggesting the need to include additional knee arthrokinematics features. Additionally, the results showed that knee kinematics was not associated with the OKS pre- or post-TKR or satisfaction, nor the patient’s perception of their knee “feeling like a normal knee” two years post-TKR. This work aided in expanding the application of the Cardiff classifier beyond the assessment of gait biomechanics, supported the use of PBTs for the assessment of function, reinforced the evidence that objective measures of function and PROMs measure different constructs and should be utilised together in evaluating OA or TKR outcomes.

Item Type:	Thesis (PhD)
Date Type:	Completion
Status:	Unpublished
Schools:	Engineering
Uncontrolled Keywords:	Cardiff Classifier Performance-based test Fluoroscopy Step-Up Trunk kinematics Gait Deviation Index
Date of First Compliant Deposit:	2 May 2024
Last Modified:	02 May 2024 13:38
URI:	https://orca.cardiff.ac.uk/id/eprint/168659

Actions (repository staff only)

Edit Item