Development and validation of risk factors prediction model for hypoxemia in the PACU patients after major abdominal surgery: a retrospective study
Objective Hypoxemia is a common and potentially harmful complication in patients undergoing major abdominal surgery (MAS). We aimed to develop and validate a nomogram prediction model for hypoxemia in the post-anesthesia care unit (PACU) after MAS. Patients and methods A total of 378 patients who underwent MAS and were admitted into the PACU were selected as the training group from June 1, 2024 to December 31, 2024. From April 1, 2025 to May 31, 2025, 189 patients who underwent MAS and were admitted to the PACU were enrolled as the validation group for temporal validation. Logistic regression analysis was employed to develop the nomogram model using age, body mass index (BMI), the ratio of arterial partial pressure of oxygen to fraction of inspired oxygen (PaO₂/FiO₂), and hemoglobin (HGB) as variables. The area under the receiver operating characteristic curve (AUC), calibration curves, and decision curve analysis (DCA) were used to evaluate the model accuracy. The primary outcome was defined as hypoxemia that occurred after tracheal extubation in the PACU. Results In the training group (n = 378), hypoxemia occurred in 112 patients (29.6%). The final nomogram included four variables: age >60.5 years, BMI ≥ 24.0 kg/m2, PaO₂/FiO₂ < 311.5 mmHg, and HGB < 106.5 g/L (all p < 0.05). The model showed good calibration (Hosmer-Lemeshow test, p = 0.244) and discrimination (AUC = 0.88, 95% CI: 0.83–0.92). At the optimal cutoff of 0.242 (Youden index), sensitivity was 0.76 and specificity 0.85. Decision curve analysis demonstrated a positive net benefit across threshold probabilities of 0.03–1.00, confirming clinical utility. In the validation group (n = 189), the predictive accuracy was 78.3%, indicating that the model maintained excellent performance in the validation group. Conclusion The nomogram model based on age, BMI, PaO₂/FiO₂, and HGB shows good predictive performance for hypoxemia. This model may help anesthesia staff identify patients at high risk of hypoxemia in the PACU.