2018 Section 5 - Rhinology and Allergic Disorders

Ahmed et al

Eligibility Criteria Studies were included if they utilized LDs as an adjunctive intervention to primary endonasal endoscopic repair of CSF rhinorrhea. LDs must have been placed intraoperatively at the time of surgical repair and maintained into the early postoperative period. Only studies that contained both arms—control (without LD) and intervention (with LD)— with at least 1 patient in each were included. Studies must have focused on repair of CSF fistulae within the anterior cranial fossa. Studies or cases from studies were included in our analysis only if the repair was performed solely endos- copically and if success rates of repair could be clearly dichotomized by LD placement. Only original research articles written in English and published in peer-reviewed journals were included. The patient population of the study must have been adult patients ( 18 years of age). Any included study must have contained at least 10 cases. Studies or cases from studies were excluded if (1) patients underwent open, combined open 1 endoscopic, or microscopic approaches to repair; (2) focus was on skull base defects of the middle or posterior cranial fossae; (3) ventriculoperitoneal shunts were placed concomitantly at the time of repair; (4) nonrhinologic leaks were included; and (5) success rates of repair by LD placement could not be clearly determined. Unpublished studies (eg, conference abstracts) were excluded. To provide a contemporary per- spective in light of the substantial advances in techniques and materials used in endoscopic repair of CSF leaks over the past few decades, studies published before 1990 were excluded. Studies with . 25% of the study population undergoing secondary repair of CSF rhinorrhea were excluded. Data Extraction and Examination Studies were identified and reviewed independently by 2 authors (O.H.A. and S.M.). Using the Covidence systematic review software (Veritas Health Innovation, Melbourne, Australia; www.covidence.org), abstracts and titles of all articles retrieved from our search were screened. Thereafter, full-text articles were carefully reviewed and selected for inclusion if the aforementioned criteria were met. References of all included studies in our full-text review were also examined to identify additional studies that may have not been identified by our initial search. Any discre- pancies in decisions to include or exclude studies were resolved through discussion. The information collected from each study included the sample size, study design, surgical repair methods, use of LDs, etiology of skull base defects, success rates of primary repair, and LD-associated complications. Etiology of skull base defects were categorized as spontaneous, traumatic (accidental), iatrogenic secondary to endoscopic sinus sur- gery (includes septoplasty cases), and iatrogenic secondary to anterior skull base resections. In the event of reported postoperative CSF leaks that recurred in a site other than

following skull base repair. In fact, 67% of surveyed American rhinologists routinely use LDs in the management of CSF leaks. 6 Their indications, however, are not clearly defined, and their use is quite variable among surgeons and centers. Accordingly, the reported use of LDs in patients undergoing endoscopic repair of anterior skull base defects is highly variable in the literature. 7 Furthermore, given the effi- cacy of endoscopic repair and the risks associated with lumbar drainage, its use has become controversial. There are, however, certain contexts where CSF diversion is thought to have greater utility. For instance, CSF diversion is thought to be particularly useful when one encounters a high-flow CSF leak, defined by Patel et al 8 as occurring when there is viola- tion of a cistern or ventricle. However, recent data have chal- lenged the use of LDs even in this difficult setting. 9,10 The complication rates associated with LDs are not insig- nificant, estimated to be 3% for major and 5% for minor complications. 10 These complications include headache, cel- lulitis at the puncture site, meningitis, pneumocephalus, and in rare instances, uncal herniation. 7,11,12 CSF leaks are typically classified according to their etiol- ogy: traumatic (accidental), iatrogenic (secondary to skull base resection or endoscopic sinus surgery), or spontaneous. Traumatic leaks are the most common etiology, followed by iatrogenic causes. 13 There is increasing evidence and senti- ment that LDs are not required for routine CSF leak repairs regardless of etiology. 10 Furthermore, to date, there has been no dedicated analysis examining the efficacy of lumbar drainage by CSF leak etiology. LD placement in the setting of endoscopic CSF leak repair is a well-established practice; however, its indications and efficacy remain unknown and controversial. This meta- analysis examines the existing body of literature to determine if adjunct lumbar drainage impacts the rate of postoperative CSF leak recurrence following endoscopic repair of anterior skull base CSF leaks. A subanalysis on the impact of lumbar drainage by CSF leak etiology is also performed. Methods A meta-analysis examining the use of LDs as an adjunct therapy to endonasal endoscopic repair of CSF leaks was conducted according to PRISMA guidelines (Preferred Reporting Items for Systematic Reviews and Meta- analyses). 14 As our study involved examining only the exist- ing peer-reviewed literature, Institutional Review Board approval was deemed unnecessary. Search Strategy A database search of EMBASE (1974 to November 2015), the Cochrane Review, and PubMed (1990 to November 2015) was performed to identify articles published through December 1, 2015. Search terms included ‘‘cerebrospinal fluid leak’’ or ‘‘cerebrospinal fluid rhinorrhea’’ and ‘‘lumbar drain.’’ The Medical Subject Headings database— a controlled vocabulary system used for indexing articles for MEDLINE—was used to maximize our search yield.

265