A retrospective, multicenter study of 972 very-low-birth-weight (VLBW) infants treated in 6 US neonatal intensive care units (NICUs) has shown that platelet transfusions do not significantly affect the incidence of intraventricular hemorrhage (IVH).
Thrombocytopenia is a risk factor for IVH, but investigators found no correlation between its severity and risk for IVH. Nor did they find platelet transfusions to have a significant effect on the incidence of IVH.
To describe platelet transfusion practices in US NICUs, senior author Martha Sola-Visner, MD, of Boston Children’s Hospital in Massachusetts, and colleagues studied NICU admissions from January 1, 2006, to December 31, 2007. They collected the last data on December 4, 2008.
Of the 972 VLBW infants, 231 (23.8%) received at least 1 platelet transfusion. And more males received transfusions (61%) than females.
Infants who received transfusions were more premature at 26.3 weeks’ gestation age compared with 28.8 weeks for those who did not receive transfusions, P<0.001.
Transfused infants were also smaller, with a mean birth weight of 805 g compared with 1113 g in the group that did not receive a transfusion, P<0.001.
Platelet transfusions
The 231 transfused infants received a total of 1002 platelet transfusions, with a mean of 4.3 per infant (range 1 to 63 transfusions).
Forty-one percent of infants had transfusions during the first 7 days of life only, amounting to 281 transfusions; 32.9% had transfusions after the first 7 days only, and 26.4% had transfusions during both periods. Seven hundred twenty-one transfusions were administered after day 7.
Almost two thirds of the transfusions, 65.4% or 653 of 998 transfusions, were given to infants who had a pre-transfusion platelet count of at least 50,000/μL.
The investigators poined out that this finding “was in contrast to UK NICUs,” where transfusions are administered at a median platelet count of 27,000/μL.
Illness severity
The investigators found significant differences among NICU sites in terms of clinical markers for transfusions.
Overall, 189 VLBW infants had platelet counts less than 100,000/μL in the first 7 days of life for a total of 402 days. And at least 1 platelet transfusion was given on 212 of those days. Of these, 198 transfusions (93.4%) had a marker of severe illness or bleeding.
On the other hand, of the 190 patient days without a transfusion, 113 (59.5%) had at least 1 of these markers (P<0.001).
Thrombocytopenia and IVH risk
The investigators evaluated the risk for IVH based on the lowest platelet count before the diagnosis of IVH was made.
They found that infants with thrombocytopenia were at higher risk for IVH, with a hazard ratio of 2.17 for any platelet count less than 150,000/μL (P<0.001).
Nevertheless, for the 314 infants with at least 1 platelet count less than 150,000/μL during the first 7 days of life, they found no association between severity of thrombocytopenia and the risk for subsequent IVH (P=0.70).
Transfusion and IVH risk
To determine whether platelet transfusions protected VLBW infants from IVH during their first 7 days of life, the investigators performed a Cox regression analysis in 756 infants.
They found that 134 infants (17.7%) had an IVH, including 62 (8.2%) with grade III or IV. So in the unadjusted model, they found a significant association between platelet transfusion and subsequent IVH, P=0.004.
However, when they adjusted the model for clinical covariates, only infants with grade III or IV IVH had a significantly greater risk with platelet transfusion, P=0.01.
Clinical covariates included sex, gestational age less than 28 weeks, 5-minute Apgar score less than 7, antenatal corticosteroid treatment, and pregnancy-induced hypertension as an indication for delivery.
The investigators also adjusted the model for clinical covariates and nadir platelet count of less than 15,000/μL. In this model, platelet transfusion became nonsignificant, even for IVH of grade III or IV.
The investigators noted that the degree to which their results are generalizable to infants with more severe thrombocytopenia is unclear, since infants in this analysis often had transfusions at platelet levels between 50,000/μL and 150,000/μL. They also collected the data approximately 8 years ago, and transfusion practices may have changed since then.
The 6 NICU study sites included Boston Children’s Hospital, Boston, Massachusetts; University of Iowa Children’s Hospital, Iowa City, Iowa; and 4 NICUs affiliated with Intermountain Health Care in Utah.
The investigators published their findings in JAMA Pediatrics.
