Cerebrospinal Fluid Shunt Infections: A Multicenter Pediatric Study

Background: Infections complicate 5%-10% of cerebrospinal fluid (CSF) shunts. We aimed to describe the characteristics and contemporary pathogens of shunt infections in children in Canada and the United States. Methods: Descriptive case series at tertiary care hospitals in Canada (N = 8) and the United States (N = 3) of children up to 18 years of age with CSF shunt infections from July 1, 2013, through June 30, 2019. Results: There were 154 children (43% female, median age 2.7 years, 50% premature) with >= 1 CSF shunt infections. Median time between shunt placement and infection was 54 days (interquartile range, 24 days-2.3 years). Common pathogens were coagulase-negative staphylococci (N = 42; 28%), methicillin-susceptible Staphylococcus aureus (N = 24; 16%), methicillin-resistant S. aureus (N = 9; 5.9%), Pseudomonas aeruginosa (N = 9; 5.9%) and other Gram-negative bacilli (N = 14; 9.0%). Significant differences between pathogens were observed, including timing of infection (P = 0.023) and CSF leukocyte count (P = 0.0019); however, differences were not sufficient to reliably predict the causative organism based on the timing of infection or discriminate P. aeruginosa from other pathogens based on clinical features. Empiric antibiotic regimens, which included vancomycin (71%), cefotaxime or ceftriaxone (29%) and antipseudomonal beta-lactams (33%), were discordant with the pathogen isolated in five cases. There was variability between sites in the distribution of pathogens and choice of empiric antibiotics. Nine children died; 4 (44%) deaths were attributed to shunt infection. Conclusions: Staphylococci remain the most common cause of CSF shunt infections, although antibiotic resistant Gram-negative bacilli occur and cannot be reliably predicted based on clinical characteristics.

Automated summary

This study described the characteristics and pathogens of CSF shunt infections in children in Canada and the United States. Staphylococci remain the most common pathogen, but antibiotic resistant Gram-negative bacilli also occur. Empiric antibiotic regimens were inconsistent, highlighting the need for more accurate prediction of pathogens.