Prediction of postoperative complications in patients undergoing head and neck surgery with free-flap reconstruction.

Head and neck surgery with free-flap reconstruction (FFR) is associated with postoperative complications (PCs), which may affect the short and long-term prognosis of patients. This study aimed to propose a new model to predict the risk of PCs after head and neck surgery with FFR. A retrospective analysis was performed on 528 patients undergoing head and neck surgery with FFR between March 2020 and June 2022 in a tertiary head and neck centre. All information assessed included patient characteristics, tumour, laboratory, and surgical and anaesthetic factors. The primary outcome was a composite of severe non-surgical PCs during hospitalisation. All patients were grouped into those with PCs and those without PCs. Based on univariate and multivariate logistic regression analysis, a model to predict the risk of PCs was constructed and validated with bootstrap resampling. The incidence of PCs was 23.30% (123/528). Five predictors of PCs were identified: age, flap size, blood loss, maximum lactic acid value, and type of airway. The model incorporating these variables had moderate predictive discrimination, with an area under the receiver-operating characteristic curve (AUC) of 0.74 (95% CI: 0.69 to 0.79) for PCs, and maintained satisfactory calibration performance (Brier score of 0.157). Meanwhile, the model demonstrated satisfactory clinical significance. We propose a new simple PCs prediction model that shows promise for the guiding of perioperative management after head and neck surgery with FFR.