2018 Section 5 - Rhinology and Allergic Disorders

Otolaryngology–Head and Neck Surgery 156(1)

surgery continues to change, so does the sentiment on the use of LDs in the setting of CSF leak repair. It is becoming increasingly understood, anecdotally and from the recent lit- erature, that routine use of LDs in the repair of CSF leaks is unnecessary. 10 Since the advent of the pedicled nasoseptal flap popularized by Hadad et al, 32 largely considered the workhorse for endoscopic skull base reconstruction, multi- ple studies have shown that LDs may not be needed even in the challenging settings of high-flow CSF fistulae 9,25 and large skull base defects. 5 Nonetheless, the precise indica- tions for perioperative CSF diversion remain unclear and controversial. To our knowledge, this is the first meta-analysis dedi- cated to the examination of perioperative lumbar drainage in the setting of endoscopic repair of CSF leaks. Hegazy et al 2 conducted the first meta-analysis on the topic of endo- scopic repair of CSF rhinorrhea and concluded that lumbar drainage did not appear to significantly affect repair rates, as included studies had good results despite not using LDs. Psaltis et al 4 conducted a systematic review on endoscopic repair of CSF leaks but were unable to closely examine the utility of LDs, as data at the time were limited. Oakley et al 33 published a systematic review on the management of CSF rhinorrhea in which they address perioperative lumbar drainage. From a review of 19 studies, they concluded that, despite the limitation in available evidence, LDs do not sig- nificantly contribute to successful repair. Most recently, Tien et al 34 examined 9 studies that utilized perioperative LDs following endoscopic anterior skull base tumor removal. Across all studies, the overall postoperative CSF leak rate in cases with LDs was 7.5%, compared with 3.4% in cases managed without LDs. Our analysis of the data from the included studies shows that perioperative lumbar drainage provides no significant benefit in terms of postoperative CSF leak recurrence following endoscopic CSF leak repair (OR: 0.89, 95% CI: 0.40-1.95). In subgroup analysis examining only those leaks associated with skull base resections, LDs were also found to have no significant benefit in reducing postoperative CSF leak recurrence (OR: 2.67, 95% CI: 0.64-11.10). Furthermore, through our literature review process, 12 stud- ies were identified that met all eligibility criteria except not having 2 arms in the study. Two of these studies contained only patients that had perioperative lumbar drainage, and their reported success rates of primary endoscopic CSF leak repair were 90.5% and 96.0%. 35,36 The remaining 10 studies contained only patients that did not undergo perioperative lumbar drainage, and their reported success rates of primary repair were 83.3% to 100%. 9,11,37-44 Although these reported rates were not incorporated into the meta-analysis, the success rates between the 2 groups in these 12 studies appear comparable, further corroborating our findings. Taken collectively, our data suggest that the current liter- ature does not provide sufficient evidence to demonstrate that LDs significantly influence postoperative CSF leak recurrence rates following endoscopic CSF leak repair. Thus, the use of LDs in patients undergoing endoscopic

Figure 2. Funnel plot of studies included in meta-analysis. The black dashed line denotes 95% confidence intervals.

had a mean follow-up period . 6 months; this information could not be obtained for 1 study. The number of cases across the studies ranged from 10 to 150, for a total of 508 cases. Across all studies, the success of CSF leak repair with adjunctive lumbar drainage ranged from 42.9% to 100% and without lumbar drainage, 75% to 100%. Meningitis and headache were the most consistently reported postoperative complications across all studies. From all included studies, the aggregate rate of meningitis in patients with LDs was 4.7%, compared with 3.9% in those without LDs. The aggregate rate of headache in patients with LDs was 21.0%, compared with 8.9% in those without LDs. Effect of Lumbar Drainage in Endoscopic CSF Leak Repair Overall, 4 of 12 studies found higher rates of CSF leak repair success with the use of adjunctive lumbar drainage, compared with repairs done without lumbar drainage. Study-specific effect estimates for lumbar drainage are given in the forest plot ( Figure 3 ). Overall, use of adjunc- tive lumbar drainage was associated with 0.89 times lesser odds of successful CSF leak repair (95% CI: 0.40-1.95); however, the confidence interval makes this association insignificant. The only etiology with sufficient data for subgroup anal- ysis was iatrogenic leaks secondary to anterior skull base resections: 153 cases had success rates based on LD place- ment and were thus suitable for subgroup analysis. There was no significant heterogeneity ( Q = 1.2, P . .10, I 2 = 0%) among the studies in this subgroup (n = 6). The use of adjunctive lumbar drainage in this subgroup was also not associated with significantly different rates of successful repair as compared with cases performed without lumbar drainage (OR: 2.67, 95% CI: 0.64-11.10; Figure 4 ). Discussion Techniques for reconstruction and materials available to the endoscopic skull base surgeon have substantially evolved over the past decade. 10,32 As the landscape of skull base

268