2025 Poster Presentations
P254: POSTOPERATIVE CSF LEAKS IN 1,119 TRANSSPHENOIDAL SURGERY PATIENTS: A SINGLE INSTITUTIONAL EXPERIENCE OVER TIME
Kaasinath P Balagurunath, BA1; Christopher S Hong, MD1; Jakob V Gerstl, MBBS1; Ryan Chrenek, MD1; Sean Lyne, MD1; Noah L Nawabi, BS1; Rania A Mekary, PhD2; Timothy R Smith, MD, PhD1; 1Brigham and Women's Hospital; 2School of Pharmacy, Massachusetts College of Pharmacy and Health Sciences (MCPHS) University
Introduction: Postoperative CSF leaks remain a significant source of morbidity after transsphenoidal surgery (TSS). However, with improvements in Sellar reconstruction methods, and increased utilization of techniques such as fat grafting and vascularized nasoseptal flap construction to manage intraoperative leaks, the incidence of postoperative CSF leaks is expected to decrease.
Objectives: To characterize the clinical outcomes and factors predictive of postoperative CSF leaks in patients with any symptomatic Sellar pathologies, receiving TSS.
Methods: In this single-institution, retrospective study, a large cohort of 1,119 patients diagnosed with Sellar pathologies, who underwent TSS between 2008 and 2019 were examined. Lesion, surgical approach, clinical, and endocrinological characteristics at baseline and postoperatively were tracked. Hormone values were determined at baseline and up to 3 years postoperatively. Lesion size and location were determined using preoperative magnetic-resonance-imaging (MRI) and computerized-tomography-imaging (CT). T-tests and chi-square tests were not performed due to the high type-1 error rate.
Results: The overall postoperative CSF leak rate was 2.9% in this cohort of patients. Moreover, the CSF leak rate dropped from 7% in 2008 to below 4% by 2019. Pathologies with the highest postoperative CSF leak rates included craniopharyngiomas (22%), abscess/meningitis (50%), and meningiomas (14%). Nonfunctioning-pituitary-adenomas, despite having a high rate of intraoperative-CSF leaks (40%), experienced among the lowest rates of postoperative CSF leaks after an index surgery (2%).
Visual-loss, as defined by tumor induced acuity or visual-field deficits, were more prevalent in patients with postoperative CSF leaks (36.6% vs 63.6%). Patients with postoperative CSF leaks experienced lower rates of pituitary adenoma pathology (42.4% vs 71.5% respectively). Moreover, tumor-induced hypopituitarism rates were higher among the CSF leak group (63.6% vs 36.6%). Preoperative desmopressin prescription rates were higher in the group of patients experiencing postoperative leaks (12.1% vs 3.1%). The most common preoperative symptom was headache (50.1% vs 63.6% in patients without vs with postoperative CSF leaks).
Hyperprolactinemia and high-GH preoperatively were more commonly observed in the no CSF leak group (36.9% vs 12.5%, and 18.2% vs 0% respectively). However, preoperative hypocortisolism was more common in the postoperative CSF leak group (37.9% vs 19%). Multiple types of postoperative hormone dysfunction occurred in higher rates in the CSF leak group including low-GH, hypothyroidism, hyperthyroidism, hypoadrenalism, hypernatremia, hyponatremia, and hypocortisolism.
Patients with CSF leaks postoperatively had lesions with higher rates of suprasellar-extension and cystic changes (75.9% vs 57.9%, and 55.6% vs 27.5% respectively). Moreover, these patients experienced higher rates of microscopic approach utilization. Intraoperative leaks were more common in the group of patients with vs without postoperative CSF leaks (54.5% vs 34.9%), warranting increase utilization of closure methods such as fat grafting, fascia, and nasoseptal-flap construction. Postoperative rates of infection, reoperation, and visual deficits were unsurprisingly higher in the CSF leak group.
Multivariate-logistic-regression demonstrated that having a non-pituitary-adenoma (OR 0.13: 95% CI: 0.05-0.34), having received microscopic TSS (OR 3.89: 95% CI: 1.11-13.59), and having experienced intraoperative-CSF leaks (OR 2.99: 95% CI: 1.15-7.75) were predictive of postoperative CSF leaks.
Conclusion: Lesion type, surgical approach, and intraoperative-CSF leak incidence were predictive of postoperative CSF leak occurrence.
