Objective To study whether the incidence of intraventricular haemorrhage (IVH) in very premature infants (<32 weeks gestation) with thrombocytopaenia is lower when using a liberal platelet-transfusion guideline compared with a restrictive guideline.

Study design A retrospective cohort study comparing the incidence of IVH in very premature infants with thrombocytopaenia (platelet count <150×10⁹/l) admitted between 2007 and 2008 to two neonatal intensive care unit in The Netherlands. The restrictive platelet-transfusion unit (N=353 infants <32 weeks gestation) transfused only in case of active haemorrhage and a platelet count <50×10⁹/l. The liberal-transfusion unit (N=326 infants <32 weeks gestation) transfused according to predefined platelet count thresholds. Primary outcome was the incidence and severity of IVH in infants with thrombocytopaenia in both units.

Results The number of infants with thrombocytopaenia that received a platelet transfusion was significantly lower in the restrictive-transfusion unit compared with the liberal-transfusion unit, 15% (21/145) versus 31% (41/141), (p<0.001). The incidence of IVH in infants with thrombocytopaenia in the restrictive-transfusion and liberal-transfusion units was 30% (44/145) and 29% (41/141), respectively (p=0.81). The incidence of severe IVH (grade 3 or 4) in the restrictive-transfusion and liberal-transfusion units was 8% (12/145) and 11% (16/141), respectively (p=0.38).

Conclusion In the restrictive-transfusion unit, the rate of platelet transfusions was significantly lower, but the incidence and severity of IVH was similar to the liberal-transfusion unit.