Cerebrospinal Fluid Shunt Infections: Established and Potential Risk Factors with a Focus on Tracheostomy and Percutaneous Endoscopic Gastrostomy

Background: Hydrocephalus is a condition in which fluid accumulates in the brain, ultimately causing an increased intracranial pressure. Patients requiring ongoing fluid drainage undergo permanent shunt insertion. However, 8-12% of all shunt patients subsequently develop a shunt infection. Patients requiring a tracheostomy and/or a feeding tube (PEG) are hypothesized to have an increased risk of infection due to the close anatomical proximity to the distal shunt catheter and the bacterial translocation between the surgically inserted devices associated with that.

Aim: The study identified the occurrence of shunt infection in this study cohort. Furthermore, the role of known risk factors (external ventricular drainage (EVD), etiology of hydrocephalus etc) and potential risk factors (tracheostomy, PEG, DM, Ward of hospitalisation) and their subsequent risk of shunt infection was investigated.

Method: This observational single center study retrospectively analyzed all adult patients undergoing primary shunt surgery at Klinikum Kassel between January 1st, 2012 and December 31st, 2021 with a follow-up of at least 3 months. The 278 patients were assessed according to such parameters as the incidence and timing of shunt infection, the relationship between known risk factors (EVD, etiology etc) and potential risk factors (tracheostomy, PEG, Ward of hospitalization etc) and their association with shunt infections. Statistical analysis included Chi-Square and Fisher-Freeman Halton test for categorical variables, logistic regression for multivariate risk factor analysis, and Kaplan-Meier survival analysis for timing of infection.

Results: The 278 patients presented with a mean age of 62 at shunt insertion of which 23 (8.2%) subsequently developed a shunt infection. In our study cohort no statistically significant association could be observed between PEG insertion and subsequent shunt infection (p = 1.00). However, a weak association between tracheostomy and shunt infection could be seen (p = 0.048, OR=2.45, 95% CI:1.02–5.86). Furthermore, short-term, external fluid drainage insertion (EVD) was a significant risk factor for shunt infection in this study cohort (p = 0.036, OR=2.49, CI:1.04–5.98). The etiology of hydrocephalus was a significant predictor of shunt infection, with post-hemorrhagic patients (PHH) being at higher risk for shunt infection (p = 0.035).

Conclusion: While univariate analysis suggested significant associations between some risk factors (EVD, tracheostomy and etiology) and subsequent shunt infection, this significance was lost when adjusted for confounding factors in multivariate analyses, suggesting a strong interdependence between the variables. This highlights the difficulty of isolating individual risk factors in a cohort with a complex array of clinical presentations and underlying comorbidities.