Abstract: Objective: To identify distinct latent classes and potential predictors of chronic pain trajectories following Total Knee Arthroplasty (TKA), to provide a basis for formulating individualized intervention strategies. Methods: Patients undergoing TKA were recruited via convenience sampling from a tertiary grade A hospital in Shanghai from May to December 2020. The knee pain was assessed using Knee injury and Osteoarthritis Outcome Score at admission, and at 3, 6, and 12 months postoperatively. Latent profile mixed models and multivariate logistic regression analyses were used to identify pain trajectories and their predictors. Results: A total of 336 patients who underwent TKA were included. Four distinct chronic pain trajectories were identified: moderate-to-severe pain with low improvement (5.4%); moderate-to-severe pain with high improvement (11.9%); moderate-to-severe pain with moderate improvement (80.7%); mild pain with no change (2.1%). Gender, preoperative Self-Administered Comorbidity Questionnaire score, postoperative widespread pain, postoperative acute pain, and preoperative pain catastrophizing score were predictors of trajectories membership (P<0.05). Conclusion: Chronic postsurgical pain trajectories following TKA exhibit significant heterogeneity. The identified predictors can assist healthcare providers in early identification of high-risk patients and enable personalized interventions.

Key words: Total Knee Arthroplasty; chronic postsurgical pain; trajectory; predictor; latent profile analysis