CORRELATION OF BILATERAL FRONTAL PLANE LOWER LIMB ALIGNMENT DURING GAIT ACROSS SEVERITY OF KNEE OSTEOARTHRITIS AND POST- TOTAL KNEE ARTHROPLASTY

Markerless motion capture offers the potential for clinical assessments of orthopaedic patients that incorporate kinematic data. While much gait analysis has been performed unilaterally historically, bilateral data from markerless motion capture provides the opportunity to consider the status of both limbs in biomechanical analyses. The purpose of our analysis was to compare bilateral gait waveform data in participants with varying combinations of knee osteoarthritis (OA), post-total knee arthroplasty (TKA), and asymptomatic joints in their left and right knees using principal component analysis (PCA) for the knee joint angle waveforms in the frontal and sagittal planes. Orthopaedic patients were recruited from clinic visits. Two advanced practice physiotherapists provided a clinical status assessment (considering x-ray findings, function, and pain) for both knees (classified as mild, moderate, or severe OA, or asymptomatic, or previous TKA). Additional asymptotic participants were recruited from the community. Gait was assessed during over-ground walking at a self-selected speed in the clothes and shoes worn that day. Markerless motion capture was recorded using 8 commercially available synchronized video cameras (Sony RX0-II) recorded at 60 Hz with 1/125 shutter speed and processed using Theia3D (Theia Markerless Inc., version v2023.1.0.3161p9). PCA was applied to the gait waveforms and principal component (PC) scores, capturing the maximum variability in the waveform data, were calculated for knee joint angles in the frontal plane (knee varus/valgus, stance only) and sagittal plane (knee flexion/extension, whole stride). Simple linear regression was applied to determine correlations of the PC scores for the two sides for all combinations of knee joint statuses. Mild and moderate knee OA were collapsed into a single category (“MildMod”). 614 knees in 307 individuals were included. In individuals with a TKA on one side, the frontal plane first principal component (PC1, magnitude of knee varus/valgus alignment during stance) was not correlated with the other side, but for all other combinations of clinical status, there were statistically significant correlations, with stronger correlations for the same clinical status on both sides (Fig. 1). In the sagittal plane, PC2 (associated with stiff knee gait) was significantly correlated across all groups. Despite different degrees of knee OA severity, frontal plane gait waveform PC1 scores, capturing varus/valgus alignment, were significantly correlated for all combinations of unreplaced joints, including asymptomatic and severe knee OA. This indicates a level of patient-specific frontal plane alignment in both limbs that is retained across severities of knee OA. Frontal plane alignment, however, was not correlated between replaced and non-replaced joints, likely indicative of TKA placement in a standard neutral alignment. The correlations in alignment between sides have implications for TKA placement in patient-specific frontal plane alignments, as these findings suggest that both pre-operative and contralateral alignment measures could be considered in defining patient-specific TKA alignment targets. For any figures or tables, please contact authors directly.