Risk Factors for Postoperative Intracranial Infections during Endoscopic Endonasal Skull Base Surgery and the Role of Antibiotic Prophylaxis

Background: Endoscopic endonasal skull base surgery (EESBS) is a minimally invasive procedure that is performed through a non-sterile corridor, the nasal cavity and paranasal sinuses, likely posing an infection risk without appropriate antibiotic prophylaxis.

Hypothesis: We hypothesized that postoperative intracranial infections during EESBS are reduced by antibiotic prophylaxis and elevated by risk factors that increase exposure to bacterial flora.

Methods: We retrospectively reviewed 200 consecutive cases of patients who underwent EESBS at our institution. Patients treated at an affiliated pediatric hospital were excluded. Data extracted from patient charts were: demographics, history of EESBS, medical comorbidity, diagnosis and location of lesion, surgical approach, intraoperative cerebrospinal fluid (CSF) leak, duration of surgery, prophylactic antibiotics, nasal packing, CSF diversion, and development of postoperative intracranial infection. Univariate analysis (Fisher's exact test and t-test) and multivariable logistic regression were used to identify risk factors for postoperative infection. Current CDC criteria for diagnosis of meningitis were used. Current prophylactic antibiotic protocol in non-allergic patients is intravenous ceftriaxone for 1 to 2 days and oral cefuroxime until nasal packing is removed.

Results: Of 200 cases, 10 were excluded due to mentioned exclusion criteria. Average age was 53.9 (SD: 17.5), with 98 female patients (51.6%). Forty patients had prior EESBS (21.05%), among whom 16 (40%) had 32 surgeries. 14 (7.4%) had a documented history of sinus disease and 10 (5.3%) experienced a previous CSF leak. 162 (85.3%) surgeries were for neoplasm, of which 53 (32.7%) were malignant. The remainder included non-neoplastic conditions such as odontoid pannus and meningoencephalocele. 101 patients (53.6%) had an intraoperative CSF leak and 9 among all cases (4.7%) had one postoperatively. There were eight cases of postoperative intracranial infection (4.2%), all with an intraoperative CSF leak. The intracranial infections were meningitis (N = 6) and intracranial abscess (N = 2). Two of the six meningitis cases were classified as chemical meningitis. No pituitary adenomas (N = 67) developed intracranial infection. Sex, BMI > 40, previous EESBS, previous craniotomy, prior radiation therapy, diabetes mellitus, previous CSF leak, smoking, packing duration, Caldwell-Luc approach and duration of antibiotic prophylaxis were not associated with intracranial infection. Intraoperative CSF leak (p = 0.008), lumbar drain (p = 0.0006), postoperative CSF leak (p < 0.0001), Merocel packing (p = 0.002), vascular flap (p = 0.025), surgery longer than 4 hours (p = 0.022), and malignant pathology (p = 0.005) significantly increased risk of intracranial infection. While prophylactic treatment with aztreonam/vancomycin/doxycycline (due to penicillin allergy) was significant (p = 0.040) in a univariate analysis, it did not achieve significance in a multivariable regression model. In a multivariable logistic regression model including sex and age a priori, malignant lesions (odds ratio, 13.79, 95% CI, 1.61–117.7) and postoperative CSF leak (odds ratio, 74.2, 95% CI, 13.01–422.93) were significant for increased risk of postoperative intracranial infection.

Conclusion: With current antibiotic protocol, the intracranial infection rate is comparable to previous literature. While this study was limited by sample size and retrospective design, it shows that EESBS can be performed safely on patients with diverse medical histories, but underlines the importance of reducing CSF leaks. The antibiotic protocol in patients with penicillin allergy may bear further evaluation.

No conflict of interest has been declared by the author(s).

Publication History

Article published online:
12 February 2021