gms | German Medical Science

Objective: Diagnosis of cerebrospinal fluid (CSF) infection in patients with external ventricular drainage is often done by clinical symptoms and/or laboratory values such as cell count or glucose values. But there seems to be a mismatch between the number of clinically diagnosed CSF infections and microbiological pathogen detection in the CSF of these patients. This study was conducted to analyze the ratio between clinically diagnosed and treated CSF infections and the confirmation of bacteria in the CSF by microbiological examination.

Methods: Retrospectively we analyzed the charts of all patients treated for EVD-associated CSF infections between 01/2011-06/2016. We checked for clinical signs of CSF infection like nuchal rigidity, elevated cell count in CSF and low glucose. In addition we examined all cases with regard to microbiological examination of the CSF and detection of bacteria. We also checked for antibiotic treatment of the CSF infection.

Results: A total of 214 patients were treated for CSF infection between 2011 and 2016. Most of these patients showed typical clinic symptoms of a CSF infection. All patients had an increased cell count often combined with low glucose in the laboratory CSF examination. Nearly all of them were treated with intravenous antibiotics. Interestingly, in only half of the cases CSF or the tip of the ventricle drain was sent to the microbiological laboratory. The majority of these samples was sent for microbiological evaluation in the years 2013-2016. In only one third of these samples pathogen detection occurred. In some of these cases there was only microscopic pathogen detection without microbial growth after incubation.

Conclusion: We saw a significant increase of microbiological pathogen examinations of CSF since 2013 before starting antibiotic treatment. This reflects the fact that we started to rethink the definition of CSF infection by using only clinical signs as well as the potential use of antibiotics. Interestingly in less than half of the examined samples a bacterial infection could be detected. That means in times with increasing resistance to antibiotics criteria for definition and treatment of CSF infections should be clearly formulated. That rises the question if clinical symptoms of a CSF infection are sufficient enough or if a microbiological pathogen detection is necessary before starting an antibiotic therapy particularly in patients with subarachnoid hemorrhage?