Optimizing haemoglobin measurements in VLBW newborns: Insights from a comparative retrospective study

Introduction: Haemoglobin levels assessment is a crucial part of neonatal intensive care practice, the painful experience of repeated heel pricks and venepunctures blood sampling may negatively affect neonatal clinical course. To date the reliability of haemoglobin levels obtained by point-of-care testing (POCT) analysis if compared to standard blood cell count remains controversial. Materials and methods: Retrospective study conducted on all inborn premature infants (gestational age < 32 weeks) admitted to NICU of the IRCCS Giannina Gaslini Institute during the period May 2021-April 2023. We considered blood samplings occurred within the first 28 days of life recording the laboratory haemoglobin levels (Hblab) (reference method), the point-of-care haemoglobin levels (HbPOCT) (alternative method) and the type of puncture (arterial, venous and capillary). A Bland-Altman analysis was performed to evaluate the Hb agreement, it determines the bias (mean difference between the reference and alternative methods) and limits of agreement (LOA; lower, l-LOA; upper, u-LOA) of measures. An acceptable limit of agreement was 1 g/dl according to the existing literature. Results: We considered 845 blood samplings from 189 enrolled patients. The comparison between the reference and the alternative method showed a good agreement for the capillary sampling technique with l-LOA of -0.717 (-0.776; -0.659) and u-LOA of 0.549 (0.490; 0.607), these results were not achievable with the other techniques, with LOAs over +/- 1 g/dl threshold (venous