There are two recently completed large randomized clinical trials of blood transfusions in the preterm infants most at risk of requiring them. Liberal and restrictive strategies were compared with composite primary outcome measures of death and neurodevelopmental impairment. Infants managed under restrictive guidelines fared no worse in regard to mortality and neurodevelopment in early life. The studies had remarkably similar demographics and used similar transfusion guidelines. In both, there were fewer transfusions in the restrictive arm. Nevertheless, there were large differences between the studies in regard to transfusion exposure with almost 3 times the number of transfusions per participant in the transfusion of prematures (TOP) study. Associated with this, there were differences between the studies in various outcomes. For example, the combined primary outcome of death or neurodevelopmental impairment was more likely to occur in the TOP study and the mortality rate itself was considerably higher. Whilst the reasons for these differences are likely multifactorial, it does raise the question as to whether they could be related to the transfusions themselves? Clearly, every effort should be made to reduce exposure to transfusions and this was more successful in the Effects of Transfusion Thresholds on Neurocognitive Outcomes (ETTNO) study. In this review, we look at factors which may explain these transfusion differences and the differences in outcomes, in particular neurodevelopment at age 2 years. In choosing which guidelines to follow, centers using liberal guidelines should be encouraged to adopt more restrictive ones. However, should centers with more restrictive guidelines change to ones similar to those in the studies? The evidence for this is less compelling, particularly given the wide range of transfusion exposure between studies. Individual centers already using restrictive guidelines should assess the validity of the findings in light of their own transfusion experience. In addition, it should be remembered that the study guidelines were pragmatic and acceptable to a large number of centers. The major focus in these guidelines was on hemoglobin levels which do not necessarily reflect tissue oxygenation. Other factors such as the level of erythropoiesis should also be taken into account before deciding whether to transfuse.

Objective. To assess the efficacy of recombinant human erythropoietin (rHuEpo) in the treatment of the anemia of prematurity. Methodology. A double-blind, placebo-controlled study was conducted on 80 preterm infants (≤32 weeks; postnatal age, 2 to 8 weeks; central hematocrit ≤35%). Patients were randomly assigned to receive subcutaneous rHuEpo (Eprex, 600 U/kg per week) or an equivalent volume of placebo, for up to 6 weeks. All patients received supplements of vitamin E (25 IU) and iron (3 mg/kg per day). The iron supplement was increased if declining serum ferritin measurements were noted. Results. Treatment and placebo groups did not differ significantly with respect to mean gestational age, birth weight hematocrit, or reticulocyte count at study entry. Fewer transfusions were administered to those receiving erythropoietin (7 compared with 21; P = .002). Compared with the placebo group, the infants receiving rHuEpo had a higher mean hematocrit (32.3 ± 4% vs 29.3 ± 6.2%; P = .014) and absolute reticulocyte count (223 ± 73 vs 124.9 ± 73 x 109/L; P < .001) at the end of the study. The mean neutrophil count was not significantly reduced at study exit (P = .8), nor at any other period during the trial in the rHuEpo group. Intercurrent events (mostly infections) were not increased in the treatment group, although there was one case of sudden infant death syndrome at age 4 months. Conclusions. Using a dose of rHuEpo of 600 U/kg per week, this study has shown a clear reduction in the requirement for blood transfusion in preterm infants.

No abstract