Association between IMU-derived Gait Parameters and 12-month Falls after Total Knee Arthroplasty: A Preliminary Study
Abstract
Introduction
Falls after total knee arthroplasty (TKA) remain a clinically important issue. Previously, we reported that inertial measurement units (IMUs) can quantify early postoperative alterations in gait characteristics, including variability, regularity, and smoothness, but their association with subsequent falls is unclear. This preliminary study examined whether IMU-derived gait parameters measured 3 weeks after TKA for knee osteoarthritis (OA) were associated with falls during the first 12 months after surgery.
Method
Forty patients aged ≥60 years undergoing primary TKA for knee OA were recruited. At postoperative week 3, participants walked twice at a comfortable speed over a 15-m walkway while wearing IMUs on the lower trunk (L3) and bilaterally on the tibial tuberosities. Gait parameters were gait speed, coefficient of variation of stride time, and trunk-acceleration harmonic ratio and autocorrelation coefficient, each calculated in the vertical, mediolateral, and anteroposterior directions. Preoperative fall history (past 12 months) and fear of falling at week 3 were also collected. Falls (≥1) were assessed by postal questionnaires at 6 and 12 months postoperatively. Fallers and non-fallers were compared using Student’s t-test or chi-square test, as appropriate.
Results
Thirty-seven participants (92.5%) returned at least one follow-up questionnaire (6 and/or 12 months), and three were lost to follow-up. Nine (24.3%) reported at least one fall during the 12-month follow-up. No between-group differences were observed in any IMU-derived gait parameters or fear of falling. Participants with a preoperative fall history had a higher fall rate than those without (7/13 [53.9%] vs 2/24 [8.3%]; χ²=9.3, p=0.004).
Conclusion(s)
In this preliminary cohort, IMU-derived gait parameters at 3 weeks after TKA were not associated with falls during follow-up. Fallers more often had a preoperative fall history. Larger studies with prospective fall monitoring and multivariable modelling are warranted.