Predictors, aetiology and outcome of early-onset neonatal bloodstream infection: a case-control study in a tertiary hospital in South Africa.

Difficulty identifying neonates at highest risk of early-onset bloodstream infection (EO-BSI) leads to high empiric antibiotic use. This retrospective case-control study analysed maternal and neonatal factors, pathogen profile, and outcomes of culture-confirmed EO-BSI (<72 hr of life) at a large neonatal unit in Cape Town, South Africa (1 January 2019-31 December 2021). Cases (neonates with culture-confirmed BSI) were matched 1:3 with randomly selected controls ('at risk' neonates with negative blood cultures, C-reactive protein < 10 mg/l and ≤4 days of antibiotics). Factors associated with EO-BSI were identified using multivariable logistic regression. Among 248 neonates included, 62 were cases and 186 were controls. Six factors independently predicted EO-BSI in 'at risk' neonates: ≥32 maternal risk factors; birth weight > 2500 g; hypo/hyperglycaemia; abnormal perfusion; seizures and invasive respiratory support. Group B streptococcus and Escherichia coli predominated on birth blood cultures (25/44; 56.8%), whereas Klebsiella. pneumoniae was dominant from 24 to 72 hr of life (13/20; 65%). Ampicillin plus gentamicin was the most frequently prescribed empiric regimen (82% in cases, 100% in controls), followed by regimens targeting healthcare-associated pathogens ≥ 24 hr of life. Cases were nearly 6 times more likely to demise than controls (RR 5.8, 95% CI = 2.7-12.5; case fatality rate 32.5%). Mortality was strongly associated with gram-negative pathogens, discordant antibiotic treatment and gestational age < 32 weeks. A clinical score to evaluate EO-BSI risk may reduce early-life antibiotic exposure. Beyond 24 hr of life, empiric antibiotics should provide coverage of healthcare-associated pathogens.