Background and aims Sepsis is a major cause of morbidity and mortality in newborns. We aimed to assess the efficacy of the current empiric regimens commonly used in neonatal intensive care units (NICU) and to describe the characteristics of newborns with sepsis.

Methods Infants admitted to the NICU with a sepsis diagnosis during January 2012 and December 2013 were assessed in a retrospective manner. While infants with early onset neonatal sepsis (EO-NS) diagnosis had received an empiric regimen including ampicillin and gentamicin, those with a late onset neonatal sepsis (LO-NS) diagnosis received carbapenem, vancomycin and fluconazole. Culture antibiogram results were compared with empirical treatment choices. Mortality and recovery rates with empiric treatment was documented. We used the MedCalc statistical software for statistical analysis.

Results We identified 124 infants (66 males and 58 females). Of these 43 (34.7%) were diagnosed with EO-NS and 81 (65.3%) with LO-NS. While the rate of positive blood culture results was 36% in the EO-NS group it was 84% in the LO-NS group. Treatment success rates were 93,5% in the EO-NS group, and 81,7% in the LO-NS group. Carbapenem resistance was identified in 10,5% of the newborns in the LO-NS group. Mortality rates were 4.6% in the EO-NS and 8,9% in the LO-NS group.

Conclusions The success rates of empiric antibiotic treatment was high in both the EO-NS and LO-NS group. The microbial flora of the NICU should be taken into account when deciding for the empirical antibiotic treatment regimens in infants with hospital acquired sepsis.