Background/Objectives: In our previous study, we reported that postoperative flexion ≥ 120° is associated with better knee function after TKA, whereas flexion gain is associated with higher patient satisfaction. However, preoperative determinants of achieving these clinically relevant targets remain unclear. This study investigated preoperative factors predicting (1) postoperative flexion ≥ 120° and (2) clinically meaningful flexion gain after TKA. Methods: We retrospectively reviewed prospectively collected data from 221 primary TKAs (171 patients) performed between 2014 and 2020. Passive knee range of motion (ROM) was measured preoperatively and at 1 year postoperatively. Preoperative variables included age, sex, body mass index (BMI), Pain Catastrophizing Scale, Knee Society Score, and Knee Injury and Osteoarthritis Outcome Score (KOOS). Patients were categorized as the good-flexion group (postoperative flexion ≥ 120°) or poor-flexion group (<120°), and as the improvement group (flexion gain ≥ 5°) or no-improvement group (<5°). Variables differing between groups were entered into multivariable logistic regression. Receiver operating characteristic (ROC) analysis (Youden index) identified optimal cutoffs. Results: Postoperative flexion was ≥120° in 63.3% of knees, and flexion gain ≥ 5° occurred in 48.4%. In multivariable models, preoperative flexion angle was the only independent preoperative predictor of achieving postoperative flexion ≥ 120° (p < 0.001) and flexion gain ≥ 5° (p < 0.001). ROC analysis showed that a preoperative flexion cutoff of 120° best discriminated both outcomes (AUC 0.78 for postoperative flexion ≥ 120°; AUC 0.80 for flexion gain ≥ 5°). Conclusions: A 120° preoperative knee flexion threshold provides a simple, clinically actionable marker for predicting postoperative flexion ≥ 120° and meaningful flexion gain after TKA. Incorporating preoperative flexion into shared decision-making may improve counseling by setting realistic expectations for postoperative knee function and satisfaction.
Kubo et al. (Thu,) studied this question.