What is the evidence for postoperative lumbar drains in endoscopic repair of <scp>CSF</scp> leaks?

The overall success rate for endonasal endoscopic repair of cerebrospinal spinal fluid (CSF) leaks is high, ranging from 90% for primary repairs to 97% for secondary.1 With the advancement of vascularized flaps, the rates of postoperative reoccurrence of CSF leaks have decreased.2 The complication rate is reportedly less than 0.03%1; however, when they occur, they can include CSF leak reoccurrence and bacterial meningitis. The role of postoperative lumbar drains (LD) in the endoscopic management of CSF rhinorrhea is currently controversial.1-4 Cerebrospinal spinal fluid diversion with a lumbar drain may prevent postoperative intracranial pressure (ICP) elevations that would disrupt the graft closure. However, LD utilization has been associated with increased complication rates of up to 12.3%.2 These complications include pneumocephalus, persistent headaches, meningitis, uncal herniation, and lumbar radiculopathy. Albu et al. reported a randomized controlled trial (RCT) on the endoscopic repair of 75 patients with a LD compared to 75 patients without a LD.3 High-flow CSF leaks subsequent to violating cisterns or ventricles during large tumor removals were excluded. Lumbar drains were set to drain at 5 to 10 mL/hour until postoperative day 3. Postoperative diuretics and ventriculoperitoneal (VP) shunts were not used in any cases. The success rate of the entire cohort was 93% (140/150), and the success rate of using a LD was not significantly different than without LD usage (95% vs. 92%, P = 0.2). Albu et al. noted that elevated ICP levels (77% elevated ICP vs. 97% traumatic and 96% iatrogenic leaks) correlated with higher recurrence rates, and the use of a LD made no difference. Two systematic reviews investigated the impact of LD usage. Psaltis et al. systematically reviewed two prospective and 53 retrospective trials involving a total of 1,778 CSF fistula repairs, with a mean follow-up of 35.9 months.1 The authors excluded leaks caused during initial tumor resections. Significant heterogeneity was found among the included studies. Forty-seven studies provided data on LDs, with the majority using the drains from 2 to 5 days. Not enough data was available to assess the impact of LD usage. On the basis of a prior meta-analysis of 14 studies comprising 289 CSF fistulae repairs, Hegazy et al. reported that lumbar drains did not significantly influence success rates. Stokken et al. reviewed two recent retrospective case series that investigated postoperative CSF leak rates after the repair of brisk to high-flow leaks.2 Garcia-Navarro et al. investigated gasket seal closure and DuraSeal (Integra, Plansboro, NJ) of 46 large diaphragmatic or dural defects. Twenty-one cases added a nasoseptal flap. Thirty-one patients had a LD placed intraoperatively, which was left in position for 1 to 2 days and drained at 5 mL/hour. Mean follow-up was 28.5 months, and two postoperative CSF leaks occurred. A significant relationship between LD usage and postoperative CSF leak rates was not observed. In the second case series, Eloy et al. investigated the use of Gelfoam (Pfizer, New York, New York) or fat with a pedicled nasoseptal flap in 59 cases of brisk-flow CSF leaks. Brisk-flow leaks involved the visualization of CSF egress without a Valsalva maneuver. Eloy et al. did not use a LD in any patient; no postoperative CSF leaks were observed in a mean follow-up of 14.6 months. Stokken et al. stated that the most common indication for LD use was in high-risk (e.g., intracranial hypertension, previous radiation, revision cases) or high-flow (defined as when a cistern or ventricle was violated) leak repairs.2 Caballero et al. retrospectively investigated 105 cases with CSF leak repairs, of which 68 cases had a LD.4 Five patients had a VP shunt placed, and one patient had an lumboperitoneal shunt placed. Diuretics were used in 34 patients, of which 24 patients received acetazolamide. With a mean follow-up of 13 months, 15 patients with a LD had a recurrent leak (22%), and five patients without LD had a recurrent leak (14%). The recurrence rate was not significantly different (P = 0.15). Furthermore, recurrence was not significantly different in the subgroups of spontaneous, traumatic, and iatrogenic leaks. The aforementioned retrospective studies are subjected to selection bias because LDs were likely placed in more complicated cases. Therefore, the evidence for high-risk cases is unclear. There is a high level of evidence (Table 1) demonstrating that lumbar drains do not reduce postoperative CSF leaks. Future multicenter RCTs designed to investigate the use of lumbar drains in higher risk repairs are needed. Lumbar Drain: The highest level of evidence was level 1b, a RCT.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.