Thrombocytopenia affects 20–35% of patients admitted to Neonatal Intensive Care Units (NICUs). Platelet transfusions are usually administered to thrombocytopenic neonates at higher thresholds than those used for older children or adults, although there is a paucity of evidence to guide these decisions.
In this study, we used a web-based survey to investigate the platelet transfusion thresholds used in level 1 NICUs (equivalent to level 3 in the US) in German-speaking European countries (Austria, Germany and Switzerland, AUT/GER/SUI). This survey was identical to the one previously used to investigate the transfusion practices of US neonatologists, thus allowing for a direct comparison of both populations. Eleven common clinical case scenarios of thrombocytopenia in preterm or term neonates were described and neonatologists were asked at which platelet count they would order a transfusion. For each case scenario, the median and the most frequently selected thresholds were determined. The Mann-Whitney-U-test was applied to compare the distribution of platelet transfusion thresholds between AUT/GER/SUI and US neonatologists. Univariate cumulative logit models (proportional odds model) were used to evaluate the differences between AUT/GER/SUI and US practices. In this analysis, an odds ratio >1 indicates an increased odd to select a higher threshold.
At least one neonatologist from 100 of the 171 (58%) eligible level 1 NICUs (AUT n=2; GER n=92; SUI n=6) participated in the survey, for a total of 144 neonatologists. Their answers were compared with those of 1006 U.S. neonatologists previously surveyed. In 9 of the 11 scenarios, AUT/GER/SUI neonatologists selected substantially lower platelet transfusion thresholds than US neonatologists (P<.0001). For example, in a preterm infant (27 weeks of gestation and 950g) who was clinically stable and not bleeding at two days of life, the median platelet transfusion threshold was 30×109/L in AUT/GER/SUI vs. 50×109/L in the US (P<.0001). If the same infant had an intracranial hemorrhage, the most frequent threshold among AUT/GER/SUI neonatologists increased to 50×109/L, while most US neonatologists rose their trigger to 100×109/L (P<.0001). To quantify the differences in transfusion practices between the two groups of physicians, we then calculated (for each case scenario) the odds of US neonatologists selecting a higher transfusion category than their European colleagues. In nine of the eleven scenarios the odds ratios were between 3.25 and 5.05, reflecting the high likelihood that a US neonatologist would choose a higher transfusion threshold than an AUT/GER/SUI neonatologist facing the same patient. Only in two case scenarios (term infant with alloimmune thrombocytopenia and premature infant prior to lumbar puncture) the odds ratios were <2, reflecting more similar (although still significantly different) responses. Finally, in order to estimate the clinical impact of the observed differences, we recorded the platelet counts of every neonate admitted to our NICU over a six month period, and determined whether they would receive a platelet transfusion by extrapolating the answers of AUT/GER/SUI vs. US neonatologists to the first clinical vignette. Using this approach, we estimated that 18 out of 1000 neonates would be transfused if AUT/GER/SUI patterns were applied, compared to 33 transfused infants if US patterns were applied (1.8-fold increase).
This first international comparative survey on platelet transfusion practices in neonates revealed substantially higher transfusion thresholds in the US than in AUT/GER/SUI. These differences might have substantial clinical implications, since platelet transfusions are associated with higher neonatal mortality and morbidity. In neonates with necrotizing enterocolitis and thrombocytopenia, receiving a higher number of platelet transfusions has been associated with a higher incidence of short bowel syndrome and cholestasis, and there is evidence to support the hypothesis that platelet transfusions can worsen bowel injury. Thus, well-designed clinical studies are needed to address the risks/benefits of these different approaches.
Josephson:Immucor: Speakers Bureau.
Asterisk with author names denotes non-ASH members.