Assessing the need for transfusion of premature infants and role of hematocrit, clinical signs, and erythropoietin level.

There are no clear criteria for administration of blood to premature infants. In the past, indications for transfusion have included tachypnea, tachycardia, poor weight gain, apnea, bradycardia, pallor, lethargy, decreased activity, or poor feeding. Some have suggested that erythropoietin levels may also be useful in determining the need for transfusion. Data were studied from 11 premature infants with birth weights less than 1500 g collected throughout 469 hospital days. During that period the infants received a total of 37 blood transfusions. No overall relationship was found between hematocrit of 19% to 64% and heart rate, respiratory rate, or the occurrence of bradycardia; ie, these variables proved to be clinically unreliable as indicators of hematocrit. Furthermore, no predictable effect of transfusion could be identified on heart rate, respiratory rate, or on the incidence of apnea or bradycardia. It was anticipated that frequent episodes of apnea or bradycardia might increase serum erythropoietin concentration. To the contrary, more frequent bradycardia was associated with the low erythropoietin levels because those infants tended to receive transfusions for "symptomatic" anemia. The data are consistent with the concept that "anemia of prematurity" is not predictably associated with symptoms classically attributed to anemia. Possible reasons for this are that the premature infant has a different inherent response to anemia; that it is inappropriate to extrapolate symptoms of severe acute anemia to persons with mild or moderate chronic anemia; or, most likely, that other determinants of heart rate, respiratory rate, and apnea/bradycardia are of more importance than mild or moderate anemia.