J Neurol Surg B Skull Base 2016; 77 - P096
DOI: 10.1055/s-0036-1580042

Intraoperative and Postoperative Cerebrospinal Fluid Leak Rates during Extended Transsphenoidal Surgery for Sellar Lesions: A Meta-Analysis

Jazmin Cole 1, Daniel A. Donoho 2, Ryan J. Austerman 2, Gabriel Zada 2
  • 1Baylor College of Medicine, Houston, Texas, United States
  • 2University of Southern California, Los Angeles, United States

Background: CSF rhinorrhea is a common postoperative complication associated with extended transsphenoidal surgery (TSS). Many intraoperative strategies exist for managing CSF rhinorrhea and their relative efficacy is not known. Over the past decade, a vascularized pedicled nasoseptal flap (PNSF) repair technique has become increasingly popular due to its potentially greater efficacy at preventing postoperative CSF leak. We performed a meta-analysis comparing the postoperative leak rates between patients who received pedicled naso-septal flaps and patients who received other types of repairs.

Methods: A Pubmed search was performed for English language reports of patients undergoing ETSS. We included studies which described the rate of intraoperative and postoperative CSF leak as well as the repair type. 3350 results were found at initial keyword search; 2580 manuscripts met prespecified inclusion and exclusion criteria and underwent full review. 27 studies containing data on 1389 patients were included in data analysis. Chi squared testing was used to compare leakage rates.

Results: We found 17 studies containing 841 patients who underwent PNSF repair. The PNSF repair studies reported extremely high rates of intraoperative CSF leakage, ranging from 30.6–100%. In the PNSF group, 806 patients (95.8%) had intraoperative CSF leakage and 49 patients (5.8%) had postoperative CSF leakage, with a failure rate of 6.1%. We found 10 studies containing 338 patients who underwent non-PNSF repair using a variety of autograft, allograft and combined techniques. The non-PNSF repair studies also reported extremely high rates of intraoperative CSF leak, ranging from 44.9–100%. In the non-PNSF group, 297 patients (87.9%) had intraoperative CSF leakage and 35 patients (10.4%) had postoperative CSF leakage, with a failure rate of 11.8%. PNSF was associated with a lower rate of postoperative repair failure (6.1% versus 11.8%, p = 0.002).

Conclusions: We report the results of the first meta-analysis of the efficacy of intraoperative CSF leak repair in extended TSS. Intraoperative CSF leakage is an extremely common finding in extended TSS approaches. PNSF repairs are associated with a lower rate of postoperative CSF leak compared with other repair techniques. However, meta-analysis has significant limitations in this heterogenous patient population undergoing a variety of repair techniques. Further study comparing CSF leak repair techniques is warranted.