Hemoglobin increases with advancing gestational age: at term, cord blood hemoglobin is 16.8 g/dL (14-20 g/dL); hemoglobin levels in very low birthweight (VLBW) infants are 1-2 g/dL below those in term infants (Fig. 97-1). A hemoglobin value less than the normal range of hemoglobin for birthweight and postnatal age is defined as anemia (Table 97-1). A “physiologic” decrease in hemoglobin content is noticed at 8-12 wk in term infants (hemoglobin, 11 g/dL) and at about 6 wk in premature infants (7-10 g/dL).

Figure 97-1 Range (mean and 95% confidence limits) of hemoglobin concentration from 10 to 40 wk of gestational age in normal (zone I) fetuses obtained by cordocentesis (percutaneous umbilical blood sample). Solid circles depict maternal red blood cell isoimmunization; open circles indicate hemoglobin levels in fetuses with ultrasonographic evidence of hydrops (zone III).

(From Soothill PW: Cordocentesis: role in assessment of fetal condition, Clin Perinatol 16:755–770, 1989.)

Infants born by cesarean section may have a lower hematocrit (Hct) than those born vaginally. Anemia at birth is manifested as pallor, heart failure, or shock (Fig. 97-2). It may be due to acute or chronic fetal blood loss, hemolysis, or underproduction of erythrocytes. Specific causes include hemolytic disease of the newborn, tearing or cutting of the umbilical cord during delivery, abnormal cord insertion, communicating placental vessels, placenta previa or abruptio, nuchal cord, incision into the placenta, internal hemorrhage (liver, spleen, intracranial), α-thalassemia, congenital parvovirus infection or other hypoplastic anemias, and twin-twin transfusion in monozygotic twins with arteriovenous placental connections (Chapter 92).

Figure 97-2 Diagnostic approach to anemia in newborn infants. DIC, disseminated intravascular coagulation; G6PD, glucose-6-phosphate dehydrogenase; MCV, mean corpuscular volume.

(Modified from Blanchette VS, Zipursky A: Assessment of anemia in newborn infants, Clin Perinatol 11:489–510, 1984.)

Transplacental hemorrhage with bleeding from the fetal into the maternal circulation has been reported in 5-15% of pregnancies, but, unless severe, it is not usually sufficient to cause clinically apparent anemia at birth. The cause of transplacental hemorrhage is not clear, but its occurrence has been proven by demonstration of significant amounts of fetal hemoglobin and red blood cells (RBCs) in maternal blood on the day of delivery by the Kleihauer-Betke test or by flow cytometry methods to detect fetal cells in maternal blood. If the infant has severe anemia with heart failure, emergency exchange transfusion to restore Hct and oxygen-carrying capacity may be needed.

Acute blood loss usually results in severe distress at birth, initially with a normal hemoglobin level, no hepatosplenomegaly, and early onset of shock. In contrast, chronic blood loss in utero produces marked pallor, less distress, a low hemoglobin level with microcytic indices, and, if severe, heart failure.

Anemia appearing in the first few days after birth is also most frequently a result of hemolytic disease of the newborn. Other causes are hemorrhagic disease of the newborn, bleeding from an improperly tied or clamped umbilical cord, large cephalohematoma, intracranial hemorrhage, and subcapsular bleeding from rupture of the liver, spleen, adrenals, or kidneys. Rapid decreases in hemoglobin or Hct values during the first few days of life may be the initial clue to these conditions.

Later in the neonatal period, delayed anemia may develop as a result of hemolytic disease of the newborn, with or without exchange transfusion or phototherapy. Congenital hemolytic anemia (spherocytosis) occasionally appears during the 1st mo of life, and hereditary nonspherocytic hemolytic anemia has been described during the neonatal period secondary to deficiency of glucose-6-phosphate dehydrogenase (G6PD) and pyruvate kinase. Bleeding from hemangiomas of the upper gastrointestinal tract or from ulcers caused by aberrant gastric mucosa in a Meckel diverticulum or duplication is a rare source of anemia in newborns. Repeated blood sampling of infants requiring frequent monitoring of blood gas and chemistry parameters is a common cause of anemia among hospitalized infants. Deficiency of minerals such as copper may cause anemia in infants maintained on total parenteral nutrition.

Anemia of prematurity occurs in LBW infants 1-3 mo after birth, is associated with hemoglobin levels <7-10 g/dL, and is clinically manifested as pallor, poor weight gain, decreased activity, tachypnea, tachycardia, and feeding problems. Repeated phlebotomy for blood tests, shortened RBC survival, rapid growth, and the physiologic effects of the transition from fetal (low PaO₂ and hemoglobin saturation) to neonatal life (high PaO₂ and hemoglobin saturation) contribute to anemia of prematurity. The oxygen available to neonatal tissue is lower than that in adults, but a neonate’s erythropoietin response is attenuated for the degree of anemia, and as a result, hemoglobin and reticulocyte levels are low. In VLBW infants, delayed clamping of the umbilical cord with the infant held below the level of the placenta may enhance placental-infant transfusion and reduce postnatal transfusion needs. This maneuver should not delay any needed resuscitation and may lead to hyperviscosity.

Delayed cord clamping (≈1-2 min or after cessation of cord pulsation) may be beneficial in otherwise well newborns in preventing anemia in full-term infants, with effects extending beyond the neonatal period. The benefits of delayed cord clamping persist for 2-6 mo as improved hematocrit, iron status as measured by ferritin concentration and stored iron, and a clinically important reduction in the risk of anemia in infancy. Late clamping may result in delivery of an extra 20-40 mL of blood and 30-35 mg of iron to the newborn. Polycythemia is a risk with delayed clamping but is often asymptomatic.

Treatment of neonatal anemia by blood transfusion depends on the severity of symptoms, the hemoglobin level, and the presence of co-morbid diseases (bronchopulmonary dysplasia, cyanotic congenital heart disease, respiratory distress syndrome) that interfere with oxygen delivery. The need for treatment with blood should be balanced against the risks of transfusion, including hemolytic transfusion reactions, exposure to blood product preservatives and other potential toxins, volume overload, possible increased risk of retinopathy of prematurity and necrotizing enterocolitis, graft-versus-host (GVH) reaction, and transfusion-acquired infection (cytomegalovirus [CMV], HIV, parvovirus, hepatitis B and C) (Chapter 468). The risk of CMV infection can be almost eliminated by the use of leukoreduced blood. In the infant <1,500 g, CMV antibody-negative leukoreduced blood should be used. The risk of acquiring HIV and hepatitis B and C viruses is reduced but not eliminated by antibody screening of donated blood. Blood banking techniques that limit multiple donor exposure should be encouraged.

Although transfusion guidelines for preterm infants have been proposed (Table 97-2), they have not been subjected to rigorous clinical study. Nonetheless, these guidelines have led to a decline in the number of unnecessary transfusions. The use of restrictive vs more liberal transfusion guidelines has been examined in two randomized trials, one conducted at University of Iowa and a second multicentric trial known as the PINT (Premature Infants in Need of Transfusion) study. The restrictive guidelines in the two groups were generally similar. In the Iowa trial, the transfusion thresholds in the liberal- and restrictive-transfusion groups were <46% and <34%, respectively, in tracheally intubated infants receiving assisted ventilation; <38% and <28%, respectively, in infants receiving nasal continuous positive airway pressure or supplemental oxygen; and <30% and <22%, respectively, in infants breathing room air. The transfusion thresholds for the liberal groups were higher in the Iowa trial than in the PINT study. In both trials, the use of restrictive thresholds resulted in fewer transfusions and also increased the number of infants who received no transfusions at all. However, in the Iowa trial (but not in the PINT study), restrictive transfusion thresholds were associated with increases in major cranial ultrasonographic abnormalities and in the frequency of apneic spells. Although these findings need further evaluation in clinical studies, the issue of finding an appropriate transfusion threshold in premature infants remains unresolved.

Asymptomatic full-term infants with a hemoglobin level of 10 g/dL may be monitored, whereas symptomatic neonates born after abruptio placentae or with severe hemolytic disease of the newborn need immediate transfusion. Preterm infants who have repeated episodes of apnea and bradycardia despite theophylline therapy and a hemoglobin level ≤8 g/dL may benefit from RBC transfusion. In addition, infants with respiratory distress syndrome or severe bronchopulmonary dysplasia may need hemoglobin levels of 12-14 g/dL to improve oxygen delivery. No transfusion is needed to replace blood removed for testing or for mild asymptomatic anemia. Asymptomatic neonates with reticulocytopenia and hemoglobin levels ≤7 g/dL may require transfusion; if a transfusion is not provided, close observation is essential. Packed RBC transfusion (10-20 mL/kg) is given at a rate of 2-3 mL/kg/hr to raise the hemoglobin concentration; 2 mL/kg raises the hemoglobin level 0.5-1 g/dL. Hemorrhage should be treated with whole blood if available; alternatively, fluid resuscitation is initiated, followed by packed RBC transfusion.

Recombinant human erythropoietin (r-HuEPO) may be considered in the treatment of chronic or anticipated anemia in an attempt to decrease or eliminate transfusions when families, for religious reasons, request all possible measures to avoid transfusions. Therapy with r-HuEPO must be supplemented with oral iron. Doses and regimens vary. In anemia of prematurity, r-HuEPO does not provide a major reduction in transfusion requirements or the number of donors; therefore, routine use of erythropoietin in VLBW infants is not recommended. Early initiation of r-HuEPO therapy may produce a small reduction in the total transfusion volume per infant. There were concerns about an increased risk of severe retinopathy of prematurity in the r-HuEPO group. The effects of late initiation of r-HuEPO (≥8 days) have also been associated with small reductions in the total blood volume transfused per infant and the number of transfusions per infant.

Erythroblastosis fetalis is caused by the transplacental passage of maternal antibody active against paternal RBC antigens of the infant and is characterized by an increased rate of RBC destruction. It is an important cause of anemia and jaundice in newborn infants despite the development of a method of preventing maternal isoimmunization by Rh antigens. Although more than 60 different RBC antigens are capable of eliciting an antibody response, significant disease is associated primarily with the D antigen of the Rh group and with incompatibility of ABO factors. Rarely, hemolytic disease may be caused by C or E antigens or by other RBC antigens, such as C^W, C^X, D^U, K (Kell), M, Duffy, S, P, MNS, Xg, Lutheran, Diego, and Kidd. Anti-Lewis antibodies do not cause disease.