Background: Abnormal complete blood count (CBC) and high plasma C-reactive protein (CRP) are associated with neonatal infections and could be helpful in the diagnosis of neonatal sepsis and to monitor the antibiotic treatment.
Objectives: The aim of this work is to evaluate and compare the performance of a bedside analyzer for blood count and C-reactive protein (CRP) with a conventional analyzer in a neonatal population.
Methods: 150 capillary or venous blood samples of term and preterm newborns were processed on an ABX-MicrosCRP200 analyzer and on a SysmexXE2100 (conventional hematology analyzer) for CBC, leukocyte differential, reticulocytes, and nucleated red blood cells (NRBC); high-sensitivity CRP (hs-CRP) was performed on a ModularPE. The differences between complete blood count and CRP were regressed against their means and assessed by means of intra-class-correlation.
Results: The intra-class-correlation for white blood cell (WBC) was 0.98, for hemoglobin 0.97, for hematocrit 0.96, for mean corpuscular volume 0.95, and for platelet 0.98. ABX-MicrosCRP200 overestimated the WBC (+1.27 x 10(3)/microL; p < 0.001), hematocrit (+1.80%; p < 0.001), and platelet (+13.55 x 10(3)/microL; p < 0.001). The intra-class-correlation for CRP was high (0.97), without systematic difference between the two values (p = 0.64).
Conclusions: The agreement between the two methods was high for both tests. However, the SD of the difference for WBC and platelet could be clinically important in leukopenic or thrombocytopenic newborns.