Abstract: BACKGROUND: Postoperative knee joint stiffness after distal femoral fracture is a core problem affecting functional recovery of patients. Its pathological mechanisms involve factors such as extensor mechanism adhesion, joint capsule contracture, and abnormal articular surface morphology. 
OBJECTIVE: To explore the efficacy of arthroscopy combined with small incision adhesion release in the treatment of extension stiff knee after distal femoral fracture, and to analyze the risk factors affecting the functional recovery of knee joint. 
METHODS: One hundred and fifty-two patients with non-infectious extension stiff knee after distal femoral fracture were selected in Guangzhou Hospital of Integrated Traditional Chinese and Western Medicine from March 2019 to October 2024. Patients were divided into good recovery group (≥70, n=122) and poor recovery group (< 70, n=30) according to Hospital for Special Surgery score 6 months after operation. The clinical data including patient age, reduction quality and continuous passive motion exercise were collected. The risk factors were analyzed by univariate and multivariate Logistic regression. A prediction model was established and the effectiveness of the model was evaluated by receiver operating characteristic curve, calibration curve, and decision curve. 
RESULTS AND CONCLUSION: (1) Univariate analysis showed that age ≥60 years old, non-anatomical reduction, no continuous passive motion exercise, postoperative complications, intraoperative flexion and extension range < 120°, increased preoperative changes of medial tibial joint line and decreased preoperative Hospital for Special Surgery score were significantly related to poor postoperative functional recovery (all P < 0.05). (2) Multivariate analysis confirmed that age ≥60 years, no continuous passive motion exercise, postoperative complications, intraoperative range of motion < 120°, low preoperative Hospital for Special Surgery score, and large change of medial tibial joint line were independent risk factors (all P < 0.05). (3) Threshold effect analysis revealed that there was a nonlinear relationship between intraoperative range of motion and recovery risk (P < 0.001), with 120° as the inflection point. When the activity was less than 120°, the recovery risk decreased by 3.3% for each increase (OR= 0.967). When ≥ 120°, the risk was no longer significantly reduced (P=0.073). (4) The discrimination (area under the curve = 0.821, 95% CI: 0.760–0.953), calibration (Brier value =0.097), and clinical net benefit (decision curve threshold 0.1–0.9) of the prediction model were all good. (5) These findings suggest that the functional recovery of stiff knee patients after distal femoral fracture is closely related to the threshold of intraoperative activity (120°), age, rehabilitation management, and preoperative joint state. The prediction model based on six independent risk factors has good efficacy and can provide basis for clinical intervention.

Key words: femoral fracture, knee joint stiffness, release surgery, knee joint function, prognosis, prediction model