AbstractA significant proportion of osteoarthritis (OA) patients continues to experience moderate to severe pain after total joint replacement (TJR). So far, preoperative factors related to pain persistence have been mainly studied using individual predictor variables and distinct pain outcomes, thus leading to lack of consensus in the field. In this prospective observational study, we evaluated knee and hip OA patients before, 3 and 6 months post-TJR searching for clinical predictors of pain persistence. We assessed multiple measures of quality, mood, affect, health and quality of life, together with radiographic evaluation and performance-based tasks, modeling four distinct pain outcomes. Multivariate regression models were built, and a network analysis was applied to pain related biopsychosocial measures and their change with surgery. A total of 106 patients completed the study. Pre-surgical pain levels were not related to post-surgical residual pain. Distinct pain scales were associated with different aspects of the pain experience. Multi-factorial models did not reliably predict post-surgical pain in knee OA across four distinct pain scales and did not generalize to hip OA; however, network analysis of pain related biopsychosocial measures showed significant changes post-surgery in both groups. Our results show that although tested clinical and biopsychosocial variables are reorganizing after TJR in OA, they do not present as a robust markers for post-surgery pain outcomes. A better understanding of mechanisms underlying pain persistence after TJR is necessary to derive clinical prognostic factors.