Is peripartum maternal fever alone a reliable predictor of neonatal sepsis? A single-centre, retrospective cohort study.

AIM

To determine the incidence rate of early-onset neonatal sepsis (EONS) among term neonates (gestation greater than 37 weeks) admitted to the neonatal intensive care unit for suspected sepsis and the association of EONS with maternal fever (temperature greater than 38°C).

METHODS

A single-centre retrospective cohort study of all term neonates (gestation >37 weeks) admitted to and treated in the neonatal unit at the Townsville University Hospital between March 2015 and March 2020. Neonatal sepsis was confirmed with positive neonatal blood culture. Data on neonatal birth/stay and maternal pregnancy were collected from the electronic medical records and neonatal database.

RESULTS

Data from 737 neonates who were admitted for treatment of EONS were analysed. Sixty % (426) reported maternal intrapartum fever, with 1.1% (5) of neonates developing blood culture-proven sepsis. Forty % did not report intrapartum fever (311), with 3% (9) of neonates developing sepsis. As such, the sensitivity and specificity of maternal fever are 1.14% and 97%, respectively. The positive predictive value was 35.7%, and the negative predictive value was 40.1%. Fourteen neonates developed EONS, and all of them were symptomatic. Seventy-eight % (334/426) of the women in the febrile group received epidural analgesia compared to 5% (16/311) in the afebrile group. Of the 95 neonates born to women with chorioamnionitis, one (1.0%) of the neonates born to women with chorioamnionitis developed sepsis.

CONCLUSIONS

Intrapartum maternal fever is an unreliable predictor for EONS and leads to unnecessary antibiotic treatment. Symptoms in the neonate are a more reliable indicator of an ill neonate with blood culture-proven sepsis.