A pragmatic evaluation of serum IL-6, PCT and CRP compared to routine CSF biomarkers and CDC/IDSA criteria in healthcare-associated Ventriculitis.

Background: Healthcare-associated ventriculitis (HAV) is a serious complication following neurosurgical procedures, especially in patients with external ventricular drains. Diagnosis remains challenging due to nonspecific clinical features, limited physical examination options, and the low sensitivity of microbiological methods. While cerebrospinal fluid (CSF)-derived biomarkers are linked to a higher infection rate, serum biomarkers have a low predictive value. However, they are still part of our daily routine monitoring for infections. In addition, the predefined criteria catalog (Centers for Disease Control [CDC]/Infectious Diseases Society of America [IDSA]) may be promising, but also underestimate the burden of the disease. This study aims to explore, with a matched case-control design, whether systemic markers (interleukin-6 [IL-6], C-reactive protein [CRP], and procalcitonin [PCT]) and their albumin ratios are associated with the clinical decision to initiate antibiotic therapy for HAV and how their performance compares to that of routine CSF markers. Furthermore, differences between microbiologically confirmed and unconfirmed HAV are explored. Finally, we reflect on our treatment decision with reference to the CDC and IDSA criteria. Methods: Adult neurosurgical intensive care patients with and without HAV treated between 2020 and 2023 were included. Patients with microbiologically confirmed and suspected ventriculitis (mcHAV and suspHAV) were both characterized by the initiation of targeted empirical antibiotic therapy and had positive and negative CSF cultures, respectively. Serum and CSF parameters were analyzed from day -4 to day 0 (initiation of therapy). Group comparisons, logistic regression, and area under the receiver operating characteristic curve (AUROC) analyses were performed to evaluate diagnostic accuracy. Results: Of the 1,817 screened neurosurgical Intensive Care Unit patients, 196 met the inclusion criteria, and after matching, 33 treated patients (16 mcHAV and 17 suspHAV) and 33 matched controls were included. Neither serum levels of IL-6, CRP, and PCT nor their albumin ratios differed between the mcHAV/suspHAV group and the control group and were not associated with treatment initiation. In contrast, CSF markers showed strong discriminatory ability: CSF white blood cell count (WBC) (AUROC 0.828), the CSF/blood WBC ratio (0.797), CSF glucose (0.754), and the CSF/blood glucose ratio (0.768) were significantly associated with ventriculitis. No significant correlations were detected between CSF and serum biomarkers. According to the CDC and IDSA criteria, 50% of treated patients met the definition for definitive HAV and 97% for probable HAV, whereas only 2.9% of controls were culture positive. Conclusion: In this exploratory matched case-control study, in contrast to serum infection markers, higher routine CSF markers were associated with a greater likelihood of treatment despite imperfect microbiological confirmation, highlighting the need for improved bedside techniques and innovative CSF sampling methods to reduce diagnostic uncertainty in HAV.