Liam O'Connell (Ireland)

Cork University Maternity Hospital Department of Neonatology

Author Of 1 Presentation

A REVIEW OF LUMBAR PUNCTURES PERFORMED FOR EARLY ONSET NEONATAL SEPSIS IN NEONATES OF 35 WEEKS GESTATION OR GREATER IN A TERTIARY NEONATAL UNIT BETWEEN 2016 AND 2020

Date
Wed, 11.05.2022
Session Time
10:00 - 11:00
Session Type
Oral Presentations Session
Room
DIMITRIS MITROPOULOS HALL
Lecture Time
10:32 - 10:42

Abstract

Backgrounds:

Early onset sepsis (EOS) is a diagnostically challenging clinical concern in neonatology. We sought to ascertain the incidence of EOS in our unit, and to examine the role of lumbar puncture (LP) in EOS evaluation, including the use of CRP as an indication.

Methods

A retrospective electronic chart review of neonates who were born in CUMH at greater than 35 weeks gestational age from 01/01/2016-31/12/2020 who underwent a LP due to concern for EOS. Neonates were identified based on electronic chart pharmacy records of receiving 3+ days of IV benzylpenicilin or cefotaxime, correlated with laboratory reports for LP.

Results:

676 neonates were treated for greater than 3 days with intravenous antibiotics due to concern for EOS (23/1000). 205 LPs were undertaken (7/1000). In 41 cases (20%) CSF was either not obtained, or was too bloodstained for microscopy. 18 neonates had proven bacteraemia (0.62/1000). There were 6 cases of culture/PCR negative CSF with a high cell count concerning for ventriculitis and 1 case of culture positive group B streptococcus meningitis in the setting of bacteraemia with the same organism. The incidence of confirmed or suspected bacterial meningitis was 0.24/1000.

Conclusions/Learning Points:

In neonates without bacteraemia or signs of meningism, none had suspected or confirmed bacterial meningitis with a pre-LP CRP less than 45 or a maximum CRP less than 50. The relative rarity of neonatal meningitis precludes definitive threshold definition by this study. However using higher CRP thresholds of 40 versus 20 in neonates without other indications for LP in this cohort would have prevented 34 LPs without missing a case of ventriculitis/meningitis. A sizable cohort (20%) in whom LP was felt indicated had failed or entirely bloodstained CSF collection, challenging diagnosis.

Hide