Abstract
This study examined pre-operative measures to predict post-operative biomechanical outcomes in total knee arthroplasty (TKA) patients.
Twenty-eight patients (female=12/male=16, age=63.6±6.9, BMI=29.9±7.4 kg/m2) with knee osteoarthritis scheduled to undergo TKA were included. All surgeries were performed by the same surgeon (GD) with a subvastus approach. Patients visited the gait lab within one-month prior to surgery and 12 months following surgery. At the gait lab, patients completed the knee injury and osteoarthritis outcome score (KOOS), a timed up and go (TUG), maximum knee flexion and extension strength evaluation, and a walking task. Variables of interest included the five KOOS sub-scores, TUG time, maximum knee flexion and extension strength normalized to body weight, walking speed, and peak knee biomechanics variables (flexion angle, abduction moment, power absorption). A Pearson's correlation was used to identify significantly correlated variables which were then inputted into a multiple regression.
No assumption violations for the multiple regression existed for any variables. Pre-operative knee flexion and extension strength, TUG time, and age were used in the multiple regression. The multiple regression model statistically significantly predicted peak knee abduction moment, post-operative walking speed, and post-operative knee flexion strength. All four variables added statistically significantly to the prediction p<.05.
Pre-operative KOOS values did not correlate with any biomechanical indicators of post-operative success. Age, pre-operative knee flexion and extension strength, and TUG times predicted peak knee abduction moment, which is associated with medial knee joint loading. These findings stress the importance of pre-surgery condition, as stronger individuals achieved better post-operative biomechanical outcomes. Additionally, younger patients had better outcomes, suggesting that surgeons should not delay surgery in younger patients. This delay in surgery may prevent patients from achieving optimal outcomes. Future studies should utilize a hierarchical multiple regression to identify which variables are most predictive.
