Serum and Cerebrospinal Fluid Vancomycin and Gentamicin Concentrations during Ventriculoperitoneal Shunt Surgery: An Observational Study

Abstract

Background: Ventriculoperitoneal shunts become infected predominantly during surgery. Optimal antibiotic prophylaxis during cerebrospinal fluid (CSF) shunt surgery has not been fully determined. Objective: To quantitate the concentrations of vancomycin and gentamicin in the serum and CSF during 100 ventriculoperitoneal shunt surgeries at the time of shunt manipulation. Design: Descriptive, prospective, observational study with a 27- to 76-month postoperative follow-up period range. Setting: University hospital. Patients: Eighty-three hydrocephalic patients ranging from premature to adult ages. Interventions: Intraoperatively, children received intravenous vancomycin 15 mg/kg/dose and gentamicin 2.5 mg/kg/dose; adults received vancomycin 1 g/dose and gentamicin 1.5 mg/kg/dose. Antibiotics were continued for 24 hours postoperatively, with dosage and schedule adjusted for age. Antibiotic serum and CSF concentrations were sampled one hour after infusion, and measured via fluorescence polarization immunoassay. Meticulous surgical infection control measures were followed. Main Outcome Measures: Correlation of intraoperative serum and CSF antibiotic concentrations with postoperative infection rate. Results: Intraoperative vancomycin serum concentrations ranged from 10.8 to 53.8 μg/mL (28.0 ± 9.5 μg/mL, mean ± SD). Intraoperative gentamicin serum concentrations ranged from 3.9 to 9.4 μg/mL (4.8 ± 2.0 μg/mL). In all surgeries, intraoperative CSF vancomycin and gentamicin concentrations were low (vancomycin <5 μg/mL, gentamicin <2 μg/mL). The infection rate among surgeries was 2.0%, and the infection rate among patients was 2.4%. Two shunt infections occurred, one with Candida albicans and one with Citrobacter freundii. Conclusions: Serum concentrations of vancomycin and gentamicin were above the minimum inhibitory concentrations for likely pathogens in 84% and 74% of surgeries, respectively. Intraoperative bacteriostatic and bactericidal CSF antibiotic concentrations may not be necessary to achieve a low postoperative shunt infection rate. Larger confirmatory studies are needed.