152 - Using Electronic Health Data from the International Neonatal Transfusion Network to Explore the Impact of Red Blood Cell Transfusion on Cardiorespiratory Parameters in Preterm Neonates

Publication Number: 2146.152

Emöke Deschmann, Karolinska Institutet, Stockholm, Stockholms Lan, Sweden; Antoine Honoré, KTH, Stockholm, Stockholms Lan, Sweden; Till Rech, Charité - Universitätsmedizin Berlin, Berlin, Berlin, Germany; Alexandra Scrivens, Oxford university hospitals, Oxford, England, United Kingdom; Coen Zandvoort, University of Oxford, Oxford, England, United Kingdom; Ivan Binotto, University of Insubria, Calvene, Veneto, Italy; Sanja Zivanovic, IMPERIAL COLLEGE HEALTHCARE NHS TRUST, London, England, United Kingdom; Simon Stanworth, NHS Blood And Transplant, Oxford, England, United Kingdom; Christof Dame, Charité - Universitätsmedizin Berlin, berlin, Berlin, Germany; Caroline Hartley, University of Oxford, Oxford, England, United Kingdom

Background: Preterm neonates are commonly transfused, yet little is known about changes in cardiorespiratory stability of infants following transfusion. Most red blood cell (RBC) transfusions are given at a volume of 15 or 20 ml/kg, which is relatively high in relation to body weight compared to single-unit transfusions in adults.

Objective: We aimed to 1) describe the feasibility of electronic health data extraction to collect real-world information on RBC transfusions in preterm infants, 2) assess differences in infants' cardiorespiratory parameters in response to RBC transfusion.

Design/Methods: In this international multicenter study, selected neonatal intensive care units (NICUs) in Europe (Stockholm, Sweden; Berlin, Germany; Oxford and London, UK) participating in the Neonatal Transfusion Network were approached. In a cohort of preterm neonates with a gestational age < 32 weeks who received RBC transfusions, routine electronic health data from the NICU monitoring systems were collected retrospectively in Stockholm, London, and Berlin and prospectively in Oxford. Changes in heart rate, oxygen saturation, and respiratory rate were evaluated for 12 hours before and after the start of RBC transfusion.

Results: The feasibility of electronic health data extraction to collect real-world information on RBC transfusions was demonstrated in multiple NICUs, despite different electronic systems and capabilities. To date, a total of 223 transfusions in 99 infants were included from Stockholm, 87 transfusions in 24 infants from London, and 23 transfusions in 17 infants from Oxford. Data entry from Berlin was delayed because of administrative regulations required for access to the local data warehouse. Consistent with previous reports, changes in cardiorespiratory parameters were small across the whole cohort of preterm infants. However, strikingly some infants showed substantial changes in cardiorespiratory parameters following transfusion, while others showed no change (Figure).

Conclusion(s): This study demonstrates the ability to use real-world electronic health data to address questions relevant to neonatal transfusion practice. Preliminary analysis showed a broad variety of individual responses to RBC transfusion. Combining data from different centers across several countries provides a unique opportunity to address the causes of these variations (e.g., patient characteristics, transfusion factors), which could inform robust practice recommendations.

Variability in the cardiorespiratory response to transfusion
(A-F) Longitudinal changes in the mean and variability of HR, RR and SaO2 in combined subgroups of patients average across data where there is an increase in physiological parameter, no change or decrease. HR: heart rate in beats per minute (bpm). RR: respiratory rate in breaths per minute (bpm).