A Prediction model for moderate-to-severe postoperative pain in patients undergoing oral and maxillofacial ambulatory surgery

BACKGROUND: To enhance hospital efficiency, ambulatory surgery (AS) has emerged and flourished in China. Oral and maxillofacial surgeries, particularly procedures such as wisdom tooth extraction in the field of alveolar surgery, are well-suited for AS. However, in clinical practice, it has been observed that patients undergoing oral and maxillofacial AS have a relatively high incidence of acute postsurgical pain (APSP). Therefore, the primary objective of this study is to develop and validate a predictive model for moderate-to-severe APSP following oral and maxillofacial surgery. METHODS: Four hundred forty-five patients who received AS (e.g., extraction of impacted tooth, enucleation of jaw cyst, arthroscopic disc repositioning, excision of superficial oral lesion, etc.) at Nanjing Medical University's Affiliated Stomatological Hospital between June 2024 and January 2025 were retrospectively evaluated. The least absolute shrinkage and selection operator (LASSO) regression and univariate analysis were employed to identify predictors in the training set. A multivariate logistic regression analysis was utilized to construct the predictive model. Bootstrap resampling was applied for internal model validation, and the temporal validation cohort included 73 July 2025 AS patients at the same center. RESULTS: In this study, sex, age, smoking history, and intraoperative use of tropisetron were identified as independent risk variables. The area under the curve (AUC) for the prediction model was 0.7443, 0.7308, and 0.725 in training, internal validation and temporal validation sets. The AUC value ranged from 0.73 to 0.75, indicating that the model has moderate discriminative ability. CONCLUSION: Our findings indicate that age, gender, smoking history, and intraoperative tropisetron use were associated with moderate-to-severe APSP within 24 h after ambulatory oral and maxillofacial surgery. A well-performing prediction model was developed and validated to estimate the risk of APSP in these patients.