Utility of haematological parameters and C-reactive protein in the detection of neonatal sepsis

Abstract

Objective: To evaluate various haematological parameters, individually and in combination, to formulate a haematological scoring system (HSS, defined by Rodwell et al.), which can then be used to screen for sepsis in neonates who are clinically suspected of infection.1

Methods: The study cohort consisted of 150 neonates (from birth to 3 days old) with clinically suspected infection. Blood was collected by peripheral venepuncture in all neonates. A complete blood count, differential leucocyte count, total leucocyte count (TLC), total neutrophil count (TNC), immature neutrophil count, band form count and platelet count were performed. Immature total neutrophil count (I/T) and immature/mature neutrophil count (I/M) ratios were then obtained. C-reactive protein (CRP) was measured semiquantitatively and blood culture and antibiotic sensitivity were performed in each case. The haematological parameters were compared individually and in combination (by HSS) with CRP.

Results: Twenty-one (14%) neonates had blood culture proven sepsis. On evaluation of various haematological parameters, TLC < 10 x 109/L, TNC < 8 x 109/L, I/M > 0.25, I/T > 0.14, band count > 15% and platelet count < 150 x 109 were found to have optimal sensitivities and negative predictive values (NPV). Using these values, an HSS was formulated. A haematological score > or = 3 had a sensitivity of 86% and NPV of 96%. C-reactive protein as a single test had a sensitivity of 76% and NPV of 96%. A combination of CRP with haematological parameters decreased the sensitivity and NPV of the HSS.

Conclusions: A haematological score can be obtained by a complete blood count and examination of peripheral blood smear, thus permitting an objective assessment of haematological changes that occur in a neonate suspected of sepsis. C-reactive protein does not have any advantage over HSS, either as a single test or in combination.