2026 Poster Presentations
P252: IMAGING FINDINGS ASSOCIATED WITH PERSISTENT IIH AFTER CSF LEAK REPAIR
Manick Saran, DO; Corinna Levine, MD; University of Miami
Introduction: Spontaneous cerebrospinal fluid (CSF) leaks are increasingly recognized as a manifestation of idiopathic intracranial hypertension (IIH). These leaks may serve as a natural release valve for chronically elevated intracranial pressure (ICP), but require timely surgical repair to prevent complications such as meningitis.
While endoscopic repair is effective, postoperative outcomes vary: some patients recover fully, while others continue to experience IIH symptoms or develop new leaks. This variability underscores the need for prognostic markers to refine risk stratification and inform longitudinal management. Although imaging features of elevated ICP—such as posterior globe flattening, optic nerve tortuosity, meningocele, and empty sella—are well established, their predictive value of a recurrent leak or persistent ICP elevation after CSF leak repair remains unclear. This study compares preoperative imaging and clinical factors in spontaneous CSF leak patients with and without a second CSF leak>3 months postoperatively to assess if these factors can predict outcomes.
Methods: A retrospective cohort study of patients diagnosed with spontaneous CSF rhinorrhea was performed from 2017-2025. Cohorts defined as “no recurrence” compared to those with a “second CSF leak” that developed >3months after initial repair. The database was queried for demographics, BMI, comorbidity score, Diuretic use, surgical details, including defect diameter [CL1] in millimeters (mm). Radiologic studies and postoperative outcomes were reviewed. Statistical analysis used the Wilcoxon rank sum test, Fisher’s exact test, and Odds ratios were calculated.
Results: Overall, 37 patients had spontaneous CSF rhinorrhea, with the following demographic distribution: 70% Female, 66% white, 27% Black, 44% Hispanic, mean Age 58 years (SD=11). Average follow-up was 17 months. Average time to second CSF leak was 64 weeks (SD=51). Five subjects (13%) developed a second spontaneous CSF leak >3 months after initial repair.
Demographics and leak duration were similar between cohorts. [CL2] The “second CSF leak cohort was more frequently male (60% vs 16% “no recurrence”), and tended to have fewer comorbidities (0 [0–0] vs. 2.5 [1.0–5.0] (higher score= more comorbidities, p = 0.064), and lower mean BMI (28.8 [28.6–31.1], 32.9 [29.8–35.4] p = 0.077). All subjects in the second CSF leak cohort were on diuretics compared to 55% of those in the “no recurrence” cohort. Optic nerve sheath dilation and empty sella tended to be more frequent in the second CSF leak cohort. All subjects in the 2nd CSF leak cohort had meningoencephaloceles as compared to 31% in the “no recurrence” cohort (OR=23, p=0.04) Defect size was larger in second CSF leaks and reached statistical significance due to the large difference between cohorts (17.5mm [12.5–22.5] vs. 4.0 [4.0–8.0], [CL3] [MS4] p=0.014).[CL5] Among those with a second CSF leak, 60% of the leaks occurred in a new skull base site.
Conclusion: Following initial repair of spontaneous CSF leaks, patients who experienced recurrence were more likely to have lower BMI, optic sheath dilation, and empty sella. Recurrent cases also showed a higher incidence of meningoencephalocele and larger initial defects. These findings suggest that meningoencephalocele and defect size may be key predictors of recurrence, warranting prospective validation in larger cohorts.