Objective To investigate the risk factors of acute post-surgical pain (APSP) after total knee arthroplasty (TKA) in elderly patients with knee osteoarthritis (KOA), and to construct a prediction model.

Methods A total of 236 elderly KOA patients treated with TKA were retrospectively included. At 72 hours after operation, pain was assessed by numerical rating scales (NRS), and the patients were divided into APSP group (n=94, 39.83%) and non-APSP group (n=142, 60.17%). The clinical data of the two groups were collected, the risk factors of APSP were screened by multivariate Logistic regression analysis, and the prediction model was built on this basis; the performance of the prediction model was evaluated using receiver operating characteristic (ROC) curves, and the results were expressed by area under the curve (AUC), sensitivity and specificity.

Results Univariate analysis showed that age older than 75 years, diabetes, preoperative sleep disturbance, anxiety and postoperative complications, preoperative Visual Analogue Scale (VAS), preoperative Western Ontario and McMahon according to Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC), tourniquet usage time in the APSP group were significantly higher than those of the non-APSP group (P < 0.05). Multivariate Logistic regression analysis showed that age >75 years old (OR=1.318, 95%CI, 1.030 to 1.686), diabetes mellitus (OR=1.489, 95%CI, 1.134 to 1.954), preoperative VAS score >4 points (OR=1.551, 95%CI, 1.095 to 2.197), sleep disturbance (OR=1.398, 95%CI, 1.093 to 1.789), anxiety (OR=1.709, 95%CI, 1.247 to 2.343) were independent risk factors for APSP after TKA in KOA patients. ROC curve results showed that the AUC of APSP predicted by the model was 0.894(95%CI, 0.843 to 0.945), the sensitivity was 82.98, the specificity was 85.92%, and the accuracy was 84.75% during internal validation; in external validation, the AUC was 0.858(95%CI, 0.797 to 0.919), the sensitivity was 78.22%, the specificity was 81.36% and the accuracy was 80.22%.

Conclusion Advanced age, diabetes mellitus, preoperative pain, sleep disorders, and anxiety are independent risk factors for APSP in elderly KOA patients treated with TKA, and the model could predict the risk of APSP.