An Algorithm for the Use of the Free Tissue Graft as a Reconstructive Technique In The Endoscopic Endonasal Approach for Pituitary Tumors

 

Introduction: Though the pedicled nasoseptal flap has become the workhorse in skull base reconstruction, it is not without donor site morbidity. This study aims to identify factors that are associated with successful use of free tissue grafting for the reconstruction of defects resulting from the endoscopic endonasal approach (EEA) to resection of pituitary tumors.

Methods: A retrospective chart review of 223 patients who underwent EEA for pituitary tumors over 3.5 years (January 1, 2013 and June 30, 2016) in a tertiary academic medical center was conducted. Variables assessed included surgical exposure (sellar, parasellar or extending past parasellar), intraoperative cerebrospinal fluid (CSF) leak, volume of leak, body mass index (BMI), and whether the procedure was a revision. Reconstructive techniques were divided into those with no reconstruction, free tissue grafts, defined as free mucosal grafts, fat grafts or commercial collagen matrix, and vascularized flaps, defined as nasoseptal flaps, inferior turbinate flaps, or revision vascular flaps.

Results: Free grafts were used in 115/223 cases (51.6%), including 71/106 (67%) of cases with sellar exposure, 37/77 (48.1%) of cases with parasellar exposure and 7/40 (17.5%) of cases with extended exposure. In cases that had intraoperative CSF leaks, free grafts were used in 23/46 (50%) of sellar cases, 10/34 (29.4%) of parasellar cases, and 3/24 (12.5%) of extended cases. Minimal leaks were commonly reconstructed with free tissue grafts when resection was limited to the sella (19/29, 65.5%), however this decreased as the resection defect increased, 8/17 (47.1%) in parasellar and ⅜ (37.5%) in extended approaches. A case with a robust CSF leak was less likely to be reconstructed with free graft (1/10 (10%) sellar, ⅙ (17.7%) parasellar, 0/4 (0%) extended). There was no difference in the average BMI when comparing reconstruction with free graft to vascularized flap (31.7 versus 31.8). This held true when stratified by exposure and intraoperative CSF leak status with the exception of cases with an extended approach with intraoperative CSF leak (27.83 versus 31.64). When examined by pathology, patients with recurrent pituitary adenomas were less likely to undergo reconstruction with free grafts, 5/24 (20.8%), compared with vascularized flaps, 18/24 (75%).

Postoperative CSF leak occurred in 7 patients (7/223, 3.1%). Four of the 7 cases (57.4%) had parasellar exposure (4/77, 5.2%) and 3/7 (42.9%) had extended (3/40, 7.5%). Two of the 7 (28.6%) were repaired with a free graft (2/115, 1.7%) while 5/7 (71.4%) were repaired with a vascularized flap (5/95, 5.3%). The average BMI was 29.6, compared with 31.6 for those without a leak. Three of the 7 tumors (42.9%) were recurrent and one of the 7 (14.3%) included clipping of an aneurysm off of the ophthalmic artery.

Conclusions: Given these data, we propose an algorithm whereby low flow CSF leaks in sellar and parasellar resections can by reconstructed with a free graft. Vascularized flaps are to be reserved for high flow, extended approaches or recurrent disease. BMI does not appear to impact the choice of reconstructive technique, though further study is needed.