Platelet transfusions in infants with necrotizing enterocolitis do not lower mortality but may increase morbidity

Necrotizing enterocolitis (NEC), a serious multisystemic inflammatory disease most commonly seen in premature neonates, is often associated with thrombocytopenia. Infants with severe forms of NEC commonly have platelet counts of less than 50,000/mm(3), occasionally less than 10,000/mm(3). Despite an absence of data to support the practice, platelet transfusions are commonly used to maintain a certain arbitrary platelet count in an effort to prevent bleeding. As platelet transfusions contain a variety of bioactive factors including pro-inflammatory cytokines, we hypothesized that a higher number and volume of platelet transfusions would not be associated with an improvement in mortality or morbidity.

A retrospective cohort analysis was conducted of the medical records of all infants between 1997 and 2001 with Bell's Stage 2 or 3 NEC associated with platelet counts of <100,000/mm(3). The medical records were evaluated for the following variables: platelet counts, number and volume of platelet transfusions, symptoms of bleeding, and hospital course. Mortality and development of short bowel syndrome and/or cholestasis were correlated to the total number and volume (total ml and ml/kg) of platelet transfusions. Differences between the outcome groups were compared using the independent t-test, Fisher's exact test and Mann-Whitney tests.

A total of 46 infants met the study criteria (gestational age 28+/-4 weeks and birth weight 1166+/-756 g, mean+/-SD). There were a total of 406 platelet transfusions administered to the study population. Of these, 151 (37.2%) were given in the presence of active bleeding, with 62% of these resulting in the cessation of bleeding within 24 hours. Other listed indications for platelet transfusions were hypovolemia and severe thrombocytopenia. On analysis of the entire cohort, there was no statistical improvement in either mortality or morbidity (short bowel syndrome and cholestasis) with greater number and/or volume of platelet transfusions. Furthermore, we found that infants who developed short bowel syndrome and/or cholestasis had been given a significantly higher number and volume of platelet transfusions when compared to those who did not have these adverse outcomes [median (minimum - maximum) - number of transfusions : 9 (0 to 33) vs 1.5 (0 to 20), p=0.010; volume of transfusions (ml/kg): 121.5 (0 to 476.6) vs 33.2 (0 to 224.3), p=0.013].

This retrospective analysis suggests that greater number and volume of platelet transfusions in infants with necrotizing enterocolitis are associated with greater morbidity in the form of short bowel syndrome and/or cholestasis without the benefit of lower mortality.