Investigation of Pansusceptible Pseudomonas aeruginosa Meningitis Cases in Patients With External Ventricular Devices

Abstract

Background: During a 2-month period at an academic medical system, 4 cases of pansusceptible P. aeruginosa (PsA) meningitis were identified among neuroscience intensive care unit (NSICU) patients with an external ventricular device (EVD). Methods: We reviewed microbiology data for the previous 2 years to determine background PsA rates and to identify additional cases of PsA meningitis. A case was defined as the isolation of PsA from a CSF specimen. We convened a multidisciplinary group of stakeholders to review medical records of case patients and to conduct a series of observational rounds. Scalp swab specimens were collected from NSICU patients to detect possible skin colonization. Pulsed-field gel electrophoresis (PFGE) analyses were performed on PsA isolates from the 4 case patients and 5 patients with PsA isolates from other body sites. Results: There was no hospital-wide increase in PsA incidence, and no patient without an EVD had PsA cultured from CSF. Infections occurred, on average, 10 days (range, 6–15 days) after EVD insertion. Cases were geographically dispersed in the NSICU and did not share common staff. None of the PsA isolates were genetically related and all scalp cultures were negative. Observations included multiple opportunities for contact with water sources: sinks in proximity to the head of the bed, storage of supplies next to sinks, reuse of bath basins, and use of dilute peroxide to clean surgical wounds. Multiuse shampoos, conditioners and lotions, not approved for hospital use, were found on the unit. Furthermore, 3 of 4 patients received cefazolin >24 hours after 6 of their 7 neurosurgeries for an average of 4.7 days (range, 0.8–4 days). Care practices were changed to mitigate contact between EVD sites and environmental water sources, and extended cefazolin surgical prophylaxis was discontinued. EVD practices were revised, and clinical teams had their competency confirmed. No additional cases have been identified in the 16 months following these interventions. Conclusions: This cluster of EVD infections was likely caused by patient care practices that resulted in independent introductions of PsA from multiple nonsterile or contaminated water sources. Antibiotic selection of PsA by extended use of cefazolin perioperative prophylaxis may have also contributed. EVD care practices should be designed to limit contact between and EVD insertion sites and nonsterile water sources or potentially contaminated care supplies. To substantiate performance improvement efforts and ensure interinstitutional comparability, a practical, standardized EVD-associated infection surveillance definition is needed.

Funding: None

Disclosures: None