Prediction of 2-Year Incident Clinical Knee Osteoarthritis From Extensor Mechanism Mechanical Properties in Asymptomatic Community-Dwelling Older Adults: Prospective Cohort Study

Abstract Background Knee osteoarthritis (KOA) lacks disease-modifying treatment, making early identification of modifiable risk factors in asymptomatic older adults a priority. Patellar tendon stiffness (PTS) and knee extensor strength (KESt) may independently influence KOA development, yet their prognostic value in individuals free of knee symptoms is unestablished. Objective This study aimed to determine whether baseline extensor mechanism mechanical properties predict 2-year incident clinical KOA in community-dwelling older adults who were asymptomatic at enrollment, and to quantify the relative contribution of each parameter. Methods This 2-year prospective cohort study enrolled 116 community-dwelling older adults (220 knees) free of knee pain at baseline. Incident clinical KOA was diagnosed according to the American College of Rheumatology criteria. Prognostic performance was evaluated across 7 survival algorithms using 5-fold nested cross-validation with the Harrell C-index. Predictor importance was quantified using Shapley Additive Explanations (SHAP) values from the best-performing model, and absolute KOA risk was estimated across tertiles of the linear predictor score. Results Over 2 years, 55 (25%) knees developed incident clinical KOA. PTS and KESt were consistently selected across all cross-validation folds, whereas quadriceps muscle passive stiffness was not selected in any fold. The gradient boosted survival model achieved the highest discrimination (C-index=0.720, 95% CI 0.686-0.754), significantly exceeding chance ( P <.05). SHAP analysis identified KESt as the dominant predictor (mean |SHAP| 0.314, 95% CI 0.251-0.388), with PTS as the second most important predictor (0.166, 95% CI 0.128-0.200). Knees in the highest-risk tertile showed a 24-month KOA incidence of 38.4% (28/73), which was more than twice that in the lowest-risk tertile (13/74, 17.6%). Conclusions PTS and KESt independently and consistently predicted 2-year incident clinical KOA in asymptomatic community-dwelling older adults. Extensor mechanism mechanical profiling may identify individuals at elevated KOA risk prior to symptom onset, although these exploratory findings require confirmation in larger cohorts before KESt and PTS can be considered for screening or preventive intervention.