Background Early onset sepsis (EOS) is a leading cause of morbidity and mortality in newborns. The incidence of culture-positive EOS in UK neonates is 0.9/1000 live births. We aimed to compare EOS investigation practice with The National Institute of Clinical Excellence (NICE) guidelines (2012) for investigation, and determine factors that influence LP acquisition.

Method We conducted a retrospective observational study in a regional neonatal unit, using data from 213 babies investigated for EOS (< 48 h). We reviewed patient notes, electronic records, and pathology results in 2 epochs (September–October 2012, April–May 2013). Clinical concern of infection was judged retrospectively using NICE criteria (any ‘Red Flag’, >2 risk factors, or other clinical indicators).

Results Data was collected from 213 term babies (228 infection episodes). Clinical concern for infection was identified in 52/228. These infants had more respiratory distress and shock than those without clinical concern. (p < 0.05). In total 31/228 babies had an LP; 14/31 infants had clinical concern of infection. Clinical concern for infection was similar in the ‘LP group’ and ‘non-LP group’. The ‘LP group’ had a larger proportion of maternal pyrexia (18/31 vs. 60/137, p < 0.05) and median maximum CRP (19 mg/l vs. 1 mg/l, p < 0.05) than the ‘non-LP group’. There were no positive blood/CSF cultures.

Conclusion Despite its low incidence EOS causes substantial clinical concern. LPs were not carried out in a number of infants despite meeting NICE criteria. Maternal risk factors and maximum CRP appear to determine the acquisition of LP more than physician ‘concern’ of infection.