Evaluating the Role of C-Reactive Protein (CRP) in Screening for Neonatal Meningitis: A Comparison of Regional Practice

Background: Since adopting NICE CG149 guidance on neonatal sepsis, approximately 10% of babies in the UK are screened for sepsis at birth. Also the reported incidence of performing lumbar punctures (LPs) to screen for meningitis has increased by16%, without a corresponding increase in the detection rate. In our tertiary neonatal intensive care unit(NICU) we use an absolute CRP value of 20mg/l to screen for sepsis and 30mg/l for LP. The method is orthovitros while some other regional laboratories use immunoassay.

Objective:

• To quantify the rate of screening for meningitis at our NICU
• To compare our practice with other units in the Northwest Neonatal Network (NWN).
• To evaluate the role of a specific CRP threshold for performing LPs.

Methods: Retrospective analysis was performed of all babies undergoing an LP in our NICU over a 12 month period. Presence of risk factors for sepsis, clinical signs and peak CRP for all of were recorded along with CSF results.

We then circulated the audit throughout the hospitals in the NWN comparing their practice

Results: 291 LPs were performed on 273 patients over the period, at a frequency of 50/1000 live births (1 in 20). Only 6 were positive for meningitis (2%).
Of the LPs performed to screen for early onset sepsis, 2/256 were positive (0.8%) and of those for late onset sepsis, 4/35 were positive (11%). 1/3rd of babies screened did not have any identifiable risk factors for developing sepsis. Of all who got screened for sepsis but not meningitis, none presented later in the neonatal period with meningitis.

Throughout the NWN, the frequency was between 10-50/1000 livebirths and despite external standardisation interpretation of single CRP value varied.

Conclusion: Throughout the Network, differences exist in the analysis and interpretation of CRPs, resulting in a wide variation for screening for neonatal meningitis.

Absolute CRP levels did not correlate with the presence of meningitis.

By removing a CRP cut off, we could have reduced the number of LPs by 80%, without missing any cases of meningitis.