Introduction. World Health Organization (WHO) recorded 5 millions neonatal mortality each year due to sepsis and 98% were in developing countries. Diagnosis of neonatal sepsis needs to be confirmed with positive culture from normally sterile sites. On the other hand, posponing treatment will worsen the disease and increase mortality.

Aim. This study conducted to evaluate bacterial pattern of neonatal sepsis and to compare laboratory parameter differences between suspected and confirmed sepsis.

Methods. This was retrospective descriptive analytic study on 94 neonates in Perinatology Division, Department of Child Health, Faculty of Medicine University of Sumatera Utara, from November 2016 until January 2017. Blood cultures were taken to confirm the diagnosis. Laboratory parameters were collected from medical records. Variables with significant results were analyzed for their accuracy. P value <0.05 were considered statistically significant with 95% confidence interval.

Results. Out of 94 neonates, bacteria were found in 55.3% neonates, with most common etiology was Klebsiella pneumonia (22.6%). There were significant neutrophil/lymphocyte ratio and procalctonin differences between suspected and confirmed sepsis (p 0.025 and 0.008 respectively). With diagnostic threshold of 9.4, sensitivity and specificity of neutrophil/lymphocyte ratio were 61.5% and 66.7%, respectively. Procalcitonin sensitivity and specificity were 84.6% and 71.4%, respectively, with a diagnostic threshold of 3.6 mg/L.

Conclusion. Bacteria found in 55.3% suspected sepsis neonates. The most common etiology was Klebsiella pneumonia. Neutrophil/lymphocyte ratio and Procalcitonin were significantly higher in confirmed sepsis. Procalcitonin can be a reliable diagnostic marker of neonatal sepsis compared with neutrophil/lymphocyte ratio.

Keywords. Neonatal sepsis. Procalcitonin. Neutrophil/lymphocyte ratio