Background The rates of postoperative cerebrospinal fluid (CSF) leak following transsphenoidal
resection (TSR) of sellar lesions has been reported to range from 1.1 to 9.6% in the
literature (Gendeh et al 2006). Management of CSF leak typically includes conservative
measures to decrease intracranial pressure, lumbar drain placement to divert CSF flow,
or operative repair. We reviewed our institution's experience with TSR to determine
the rate of postoperative CSF leak requiring lumbar drainage. Secondarily, we investigated
whether the use of endoscopy or patient demographics were influencing factors.
Methods A retrospective review of patients at a single institution between 2012 and 2016
who underwent TSR via open microscopic or endoscopic approach was performed.
Results A total of 148 patients were identified who underwent 149 TSRs during the specified
period with an average follow-up of 14 months. Average age of all patients was 54
(range: 12–86) with 62 males and 87 females. One hundred eight (72.5%) patients underwent
an open microscopic approach compared with 41 (27.5%) via an endoscopic endonasal
approach. All patients received postoperative antibiotics for infection prophylaxis
as well as strict sinus precautions and an aggressive bowel regimen to decrease straining.
Twenty-nine patients (19%) had intraoperative CSF leak and 9 patients (6%) experienced
a CSF leak an average of 41 days postoperatively (range: 0–323 days). Of the patients
who experienced postoperative CSF leak, 7 (78%) had lumbar drain placed for an average
of 5.9 days (range: 0–15). Five (55.6%) patients with postoperative leak required
surgical repair. Rate of postoperative CSF leak was 5/108 (4.6%) in the open group
and 4/41 (9.8%) in the endoscopic group (RR: = 0.47, 95% CI: 0.13–1.68). Three (2.8%)
patients required a lumbar drain in the open group compared with 5 (12%) in the endoscopic
group (RR = 0.23, 95% CI: 0.05–0.91). For both techniques combined, average BMI was
35.4 for patients who required lumbar drain compared with 31.6 for patients who did
not require a lumbar drain, and 1 (10%) patient in the drain group had a history of
diabetes mellitus compared with 33 (23.7%) patients without a drain (OR: 0.36, 95%
CI: 0.04–2.92).
Conclusion The rate of CSF leak is relatively low for all patients undergoing TSR. There was
a significant difference between open microscopic and endoscopic approaches and the
relative risk of lumbar drain placement. The use of lumbar drains at our institution
is infrequent but can contribute to prolonged hospitalization as well as the possibility
of increased rates of meningitis. BMI and surgical technique are potential factors
influencing the likelihood of lumbar drain placement in patients undergoing transsphenoidal
surgery.
