Risk factors and predictive model for postoperative cerebrospinal fluid leakage following endoscopic endonasal pituitary adenoma surgery: a retrospective study focusing on pneumocephalus and sellar floor bony window.

Objective: To identify risk factors for cerebrospinal fluid (CSF) leakage after endoscopic endonasal pituitary adenoma surgery and to develop a clinical prediction model aimed at facilitating early detection and reducing the risk of related complications. Methods: Clinical data were retrospectively collected from patients who underwent endoscopic endonasal pituitary adenoma resection between January 2021 and September 2024. Postoperative CSF leakage was diagnosed based on clinical manifestations, glucose testing, and expert consensus. Univariable and multivariable logistic regression analyses were performed to determine independent risk factors, which were subsequently incorporated into a nomogram. The predictive performance of the model was evaluated using receiver operating characteristic (ROC) curves, calibration curves, and decision curve analysis (DCA). Results: Among 254 patients, 19 (7.48%) developed postoperative CSF leakage. Compared with those without leakage, affected patients exhibited larger tumor diameters, wider sellar floor bony windows, higher intraoperative CSF leak grades, greater suprasellar extension, and higher postoperative pneumocephalus grades. Multivariate logistic regression identified postoperative pneumocephalus grade (OR = 12.90, 95% CI: 3.24-59.80, p < 0.001), sellar floor bony window size (OR = 1.20, 95% CI: 1.07-1.37, p = 0.004), suprasellar extension grade (OR = 5.87, 95% CI: 1.15-46.10, p = 0.049), and intraoperative CSF leak grade (OR = 4.71, 95% CI: 1.14-20.30, p = 0.032) as independent predictors of postoperative CSF leakage. The nomogram incorporating these factors achieved excellent predictive accuracy, with an AUC of 0.948 (95% CI: 0.906-0.991). Conclusion: This study confirms that a higher grade of pneumocephalus and a larger sellar floor bony window are significant risk factors for postoperative CSF leak following endoscopic endonasal transsphenoidal surgery for pituitary tumors. Building on these findings, we developed a clinical prediction model for postoperative CSF leak by integrating relevant preoperative, intraoperative, and postoperative factors. This model facilitates the early prevention, identification, and management of CSF leaks, which is crucial for reducing the risk of associated complications and improving patient outcomes.