2025 Poster Presentations
P278: CONSIDERATION OF DIFFERENT SELLAR RECONSTRUCTION TECHNIQUES BASED ON PRESENCE OF INTRAOPERATIVE CSF LEAK DURING ENDONASAL TRANSSPHENOIDAL SURGERY
Jensen Ang, BMedSci, MBChB, FRCSEd1; Sophie Jia Qian Koh, MBBS1; Kaijun Tay, MBBS, MMed, FAMS2; Jia Xu Lim, MBBS, FRCSEd1; Ramez W Kirollos, MBChB, MD, FRCS, FEBNS1; Beng Ti Christopher Ang, MBBS, FRCS, FAMS1; Neville Wei Yang Teo, MBBS, MRCS, MMed2; 1National Neuroscience Institute; 2Singapore General Hospital
Background: Cerebrospinal fluid (CSF) leak is a known complication after endonasal transsphenoidal surgeries. Various techniques, including the nasoseptal flap (NSF) and free mucosal graft (FMG), have been described to reconstruct the sellar defect and prevent postoperative CSF leak. NSF could result in increased donor site morbidity, whereas FMG could be just as effective in reducing CSF leaks without additional postoperative sinonasal discomfort.
Methods: All endonasal transsphenoidal surgeries in a single tertiary academic center between 1 January 2019 and 31 December 2023 were retrospectively reviewed. Data was collected in terms of patient demographics, presence of intraoperative CSF leak (graded according to Esposito and Kelly’s description), reconstruction technique, and postoperative nasoendoscopic follow-up. The presence of postoperative CSF leak, crusting, adhesions, purulent sinusitis and smell disruption were recorded.
Results: A total of 282 endonasal transsphenoidal surgeries were performed, of which 270 were standard transsellar approaches, and 12 were extended approaches. 61 (21.6%) had intraoperative CSF leaks. In grade 0 cases, defect reconstruction was not done: the dural leaflets were flapped back and covered with oxidized cellulose. 4 grade 0 cases had NSF replaced because they were tumour recurrences and NSF was already used in prior surgery. For grade 1 leaks, abdominal fat and FMG harvested from the anterior nasal space (Figure 1) (17, 89.5%) or sphenoid sinus (2, 10.5%) were applied in 19 (46.3%), and NSF was used in 22 (53.7%). Both FMG and NSF required good circumferential bone contact (Figure 2), and edges were enforced with oxidized cellulose and sealant glue. For grade 2 (16 patients) and grade 3 (2 patients) leaks, reconstruction with abdominal fat graft and pedicled NSF technique was performed. Rates of postoperative CSF leaks were 1 in 217 (0.46%) cases with no reconstruction done, and 0 in the 65 cases with FMG or NSF reconstruction. Nasoendoscopic follow-up showed significant differences in average duration of crusting (2.26 versus 5.49 months, mean difference 3.22, 95% CI 1.55 – 4.90, p = 0.003), and smell disruption (1.31 versus 3.07 months, mean difference 1.75, 95% CI 0.73 – 2.78, p = 0.007) between FMG and NSF groups.
Conclusion: Intraoperatively observed absence of CSF leak was highly reliable, and no defect reconstruction was necessary. For grade 1 CSF leaks, reconstruction with FMG was sufficient in sealing the leak, and afforded better sinonasal outcomes than NSF. NSF might only be required for higher-grade leaks.
Figure 1. Harvesting of FMG from anterior nasal floor.
Figure 2. Placement of FMG over sellar defect.
