Neonatal Anemia

KEY POINTS

• 1.

  Anemia is an abnormal, and an unhealthy, reduction in the blood hemoglobin concentration or the hematocrit. There are limitations in our current definitions of anemia, which are based on “reference intervals” constructed from clinically obtained laboratory tests, not from tests performed on healthy volunteers.

• 2.

  Erythropoietin is the main physiologic regulator of red blood cell (RBC) production.

• 3.

  Neonatal RBCs frequently show morphologic features of immaturity such as anisocytosis, poikilocytosis, and macrocytosis. Some nucleated erythrocytes may also be seen.

• 4.

  The combination of shortened RBC survival, decreased production, and growth-related expansion of the blood volume is responsible for a progressive decrease of hemoglobin concentrations in early infancy.

• 5.

  Neonates with anemia should be evaluated for losses due to hemorrhage or hemolysis and for possible hyporegenerative etiologies. In growing premature infants, anemia of prematurity is an important cause of hyporegenerative anemia beyond 4 weeks after birth.

Introduction

Anemia is an abnormal, and an unhealthy, reduction in the blood hemoglobin concentration or the hematocrit. These two laboratory tests (hemoglobin and hematocrit) are somewhat similar in that each assesses the capacity of a subject’s blood to deliver oxygen to that subject’s tissues. It should be recognized, however, that neither measurement directly quantifies oxygen delivery or assesses whether the tissue’s oxygen needs are indeed being met. In fact, very often in neonatal medicine, poor oxygen delivery to tissues is not the result of anemia at all but is due to respiratory, or sometimes cardiac, disease. Consequently, a practical definition of anemia, in neonatal medicine, focuses on quantifying an erythrocyte number , not on the physiologic concept of inadequate oxygen delivery. Specifically, by convention, the number that defines anemia in a neonate is a blood hemoglobin concentration or a hematocrit that falls below the 5th percentile of the appropriate reference interval. The “appropriate” reference interval is one that accounts for the gestational age and the postnatal age of the neonate. This is because gestational age and postnatal age both profoundly affect where the hemoglobin concentration or hematocrit “should be.” This chapter reviews the reference interval–based definition and pathogenesis of neonatal anemia and provides practical approaches for diagnosing the exact cause of this condition when the cause is not obvious.

Fetal and Neonatal Erythropoiesis

Erythropoietin (Epo) is the main physiologic regulator of red blood cell production. However, it has additional biologic roles, some of which are particularly relevant during fetal and neonatal development. For instance, Epo is an important constituent of amniotic fluid, typically in concentrations of 25 to 40 mU/mL. A normal human fetus swallows 200 to 300 mL of amniotic fluid/kg/day and thus swallows 10 to 15 U of Epo/kg/day. ^, Epo does not cross the human placenta, and the source of the Epo in amniotic fluid is not the maternal circulation. In the second and third trimesters, amniotic fluid is largely derived from fetal urine, with minor constituents from fetal tracheal effluent and the placenta and fetal membranes. However, Epo in amniotic fluid is not derived from fetal urine. Fetal kidneys produce little Epo before delivery, and the first-voided urine of neonates generally has no detectable Epo. Studies using in situ hybridization and immunohistochemistry indicate that the source of Epo in amniotic fluid is largely maternal, from mesenchymal and endothelial cells in the deciduae and amnion.

Colostrum and breast milk contain biologically active Epo in concentrations of 10 to 20 mU/mL. Epo levels in milk do not correlate with Epo levels in the mother’s blood. In fact, during the first weeks of lactation, the mother’s serum Epo concentrations decrease, whereas her milk Epo concentrations increase, reaching the highest concentrations in women breastfeeding for a year or more. The source of Epo in breast milk is mammary gland epithelium. Epo in human amniotic fluid, colostrum, and breast milk is relatively protected from proteolytic digestion in the fetal and neonatal gastrointestinal tract. However, rather than being absorbed from the gastrointestinal tract into the blood, the Epo swallowed by the fetus and neonate binds to Epo receptors on the luminal surface of villous enterocytes, where it serves as an intestinal growth and development factor. Experimental animals artificially fed formulas devoid of Epo have retarded villous development, a condition that can be remedied by enteral recombinant Epo and blocked by anti-Epo antibody.

Cells in the developing central nervous system produce Epo, which is present in relatively high concentrations in fetal cerebrospinal fluid (CSF). In fact, the highest concentrations of Epo in the CSF are in the most premature neonates, and by several years of age, CSF Epo concentrations are typically <1 mU/mL. Epo receptors are expressed on human fetal neurons, ^, and at least small quantities of recombinant Epo administered intravenously cross the blood-brain barrier and appear in the CSF. Epo is a natural neuroprotectant. Its production increases rapidly in the brain during hypoxia, and when Epo binds to receptors on neurons, antiapoptotic activity is induced. The clinical utility of recombinant Epo as a neuroprotectant is a topic of ongoing studies.

The liver is the primary site of fetal Epo production. The kidneys produce only about 5% of the total Epo during midgestation. The mechanisms regulating the switch in Epo production from the liver to the kidneys are not completely known but may involve developmental expression of transcription activators such as hypoxia inducible factor and hepatic nuclear factor 4 ^, or developmental methylation of promoter and enhancer regions. The switch might involve the GATA transcription factors, particularly GATA-2 and GATA-3, which are negative regulators of Epo gene transcription.

Identifying Anemia Using Reference Intervals

“Reference intervals” are generally used to interpret laboratory tests in neonatology in place of “normal ranges.” The difference is that reference intervals are constructed from clinically obtained laboratory tests, not from tests performed on healthy volunteers. In order to approximate a normal range, the laboratory tests included in a reference interval data set are only those from neonatal patients who have minimal pathology or pathology not thought to be related to the test under consideration. For instance, reference intervals for the hematocrit of neonates exclude data from neonates with clinical issues known to affect the hematocrit, such as erythrocyte transfusions, hemolytic disease, hemorrhage, or reduction transfusion. Reference intervals for hematocrit on the day of birth, according to gestational age of the neonate, are shown in Fig. 43.1A . Before 28 weeks’ gestation, anemia is defined by a hematocrit below 30%. At term, anemia is defined by a hematocrit below 42%. Clearly, the hematocrit should increase gradually during the period from 23 weeks to term. Thus the definition of anemia, at birth, requires knowledge of the gestational age. Fig. 43.1B gives the same information for blood hemoglobin concentration on the day of birth. The same basic pattern is seen as in Fig. 43.1A .

In the days and weeks after birth, the hematocrit and hemoglobin gradually decrease. Fig. 43.2 demonstrates the reference interval for decreasing hematocrit (A) and hemoglobin (B) of term neonates (≥35 weeks) and the hematocrit (C) and hemoglobin (D) of preterm neonates (<35 weeks).

Circulating erythrocytes in the fetus have features reminiscent of “stress erythropoiesis” in adults. These features include anisocytosis, poikilocytosis, macrocytosis, and the presence of nucleated erythrocytes. Marrow cellularity in the fetus is relatively high. Erythroid precursors account for 30% to 65% and myeloid cells for 45% to 75% of nucleated marrow cells at birth. The myeloid to erythroid ratio at birth is approximately 1.5:1. Marrow cellularity decreases after birth, attaining a density that is normal for adults by 1 to 3 months. Initially, this decrease in cellularity results from a rapid decline in red cell production. At 1 week of age, erythroid elements account for only 8% to 12% of nucleated cells, and the myeloid to erythroid ratio exceeds 6:1. The normal adult proportion of myeloid to erythroid precursors is not established until the third month. Both the percentage and absolute number of lymphocytes increase during the first 2 months, so that by 3 months of age, they constitute nearly 50% of marrow nucleated cells. Differential counts of bone marrow aspirates from preterm infants are similar to those of term infants.

In newborn infants the hemoglobin (Hb) concentration and hematocrit of capillary blood are 5% to 10% higher than those of venous blood. The difference between capillary and venous values is greatest at birth but disappears by 3 months of age. The discrepancy is greatest in preterm infants and in those with hypotension, hypovolemia, and acidosis. ^,

Reticulocytes at birth are approximately 5% of erythrocytes, with a range of 4% to 7%. ^, Reticulocytes remain elevated for the first 1 to 3 days, typically dropping abruptly to 0% to 1% by day 7. Nucleated red cells are seen regularly on blood smears during the first day of life, constituting about 0.1% of the red cell population (500 normoblasts/mm ³ ) but are not common in the circulation after the first 3 days unless intermittent or chronic hypoxia is present.

Red cell morphology is characterized by macrocytosis and poikilocytosis. Target cells and stomatocytes are prominent. Similarly, a high proportion of siderocytes (3.2% vs. the normal adult mean of 0.1%) is seen.

Measuring the circulating red blood cell (RBC) volume in a fetus or neonate is difficult. Mock et al. used a nonradioactive method, based on in vivo dilution of biotinylated RBCs enumerated by flow cytometry, to estimate the correlation between hematocrit and circulating RBC volume in infants <1300 g and found that venous hematocrit values correlated highly with the circulating erythrocyte volume ( r , 0.907; P <.0001). ^,

Neonates have a shorter red cell survival than do children and adults. The life span of red cells from term infants is estimated to be 60 to 80 days using the ⁵¹ Cr method and 45 to 70 days using methods involving ⁵⁹ Fe. Fetal studies using [ ¹⁴ C] cyanate-labeled red cells in sheep revealed an average red cell life span of 64±6 days. The mean red cell life span increases linearly from 35 to 107 days as the fetal age increases from 97 days (midgestation) to 136 days (term).

Neonatal red cells transfused into adults have a short survival, indicating that factors intrinsic to the newborn red cell are responsible. This conclusion gains further support from a demonstration that adult red cells survive normally in newborn recipients. The life span frequency function is not parametrically distributed, in that most cells are destroyed before the mean survival is reached. Shortened red cell survival corresponds with erythropoietic rates at birth that are three to five times greater than those of normal adults.

The abrupt transition from the relative hypoxia of the uterus to an oxygen-rich environment profoundly alters erythropoiesis. During the first 2 months of life, the infant experiences both the highest and lowest Hb concentrations occurring at any time in development. Epo levels at birth are usually well above the normal adult range and fall markedly in the immediate postnatal period. By 24 hours, the Epo value is below the normal adult range, where it remains throughout the first month of life. The decrease in Epo level is followed by a decline in the number of bone marrow precursors and a decrease in the reticulocyte count.

The combination of shortened RBC survival, decreased production, and growth-related expansion of the blood volume is responsible for a progressive fall of the Hb concentration to a mean of approximately 11 g/dL at 2 months of age. The lower reference interval for infants of this age is approximately 9 g/dL. This nadir is called physiologic anemia , in that it is not associated with apparent distress and is not prevented with nutritional supplements. Stabilization of the Hb concentration is heralded by an increase in reticulocytes at 4 to 8 weeks. Thereafter, the Hb concentration rises to a mean level of 12.5 g/dL, where it remains throughout infancy and early childhood.

At term, the placenta and umbilical cord contain 75 to 125 mL of blood (30–40 mL/kg), or approximately one-fourth to one-third of the fetal blood volume. Linderkamp et al. compared postnatal alterations in blood viscosity, hematocrit, plasma viscosity, red cell aggregation, and red cell deformability in the first 5 days of postnatal life in full-term neonates with early (less than 10 seconds) and late (3 minutes) cord clamping. The residual placental blood volume decreased from 52±8 mL/kg of neonatal body weight after early cord clamping to 15±4 mL/kg after clamping. The neonatal blood volume was 50% higher in the late cord-clamped infants.

Additional placental transfer of blood to preterm infants occurs by delayed clamping of the umbilical cord. Transfer of about 10 to 15 mL/k body weight can be expected by delaying clamping for 30 to 60 seconds and has been claimed to reduce intraventricular hemorrhage and late-onset sepsis.

When the Cause of Neonatal Anemia Is Not Obvious

Once anemia has been recognized in a neonate, using appropriate reference intervals, it is important to determine the cause of the anemia. In many instances, the explanation for anemia in a neonate is obvious. However, occasionally the cause is unclear. Diagnosing the cause of the anemia, not just the fact that anemia exists, is important. This is because diagnosing the cause may reveal something about the propensity for future anemia, such as with genetic hemolytic anemia. Moreover, the cause can be important to families or the obstetric management team or to quality-improvement initiatives that aim to reduce the incidence of neonatal anemia.

Thus every time anemia is diagnosed in a neonate, the cause of the anemia should be sought, and when appropriate, the cause should be documented in the medical record. An effective approach to finding the cause of the anemia in a neonate when the cause is not obvious involves careful consideration of each of the “three H’s”: namely, (1) hyporegenerative, (2) hemorrhagic, and (3) hemolytic ( Fig. 43.3 ).

It can be helpful to classify whether the anemic neonate’s RBCs are normal in size or are larger or smaller than normal. Similarly, it can be helpful to know whether the anemic neonate’s RBC content of hemoglobin is normal or is greater or less than normal. Toward this end, reference intervals for erythrocyte indices are shown in Fig. 43.4 . Reference intervals for the mean corpuscular volume (MCV, measured in fL) are shown in Fig. 43.4A and for the mean corpuscular hemoglobin (MCH, measured in pg) in Fig. 43.4B . Microcytic anemia is diagnosed when the MCV is below the 5th percentile for gestational age. Thus an anemic extremely low gestational age neonate with an MCV less than about 104 fL has microcytic anemia. Likewise, an anemic term neonate with an MCV less than about 98 fL has microcytic anemia. In the way the MCV informs on red cell size, the MCH informs on erythrocyte “paleness” or hypochromia, due to an amount of hemoglobin in erythrocytes that is below normal (below the 5th percentile reference interval for gestational age). Thus an anemic extremely low gestational age neonate with an MCH less than about 35 pg would have hypochromic anemia, whereas an anemic term neonate with an MCH less than about 33 pg would have hypochromic anemia.

These reference intervals are valid for the first day of life, and probably for the first week, assuming no erythrocyte transfusion is given. However, rigorous reference intervals based on postnatal age are not yet available, and thus it is not clear precisely how the erythrocyte indices change over the weeks and months after birth. In adults, the reference interval for MCV is 88±8 fL (thus an MCV less than 80fL defines microcytosis), and the reference interval for MCH is 30±3 pg (thus an MCH less than 30 pg defines red cell hypochromia).

Microcytic and hypochromic erythrocytes in anemic neonates can be recognized by examining the blood smear unless they are so mild as to be unrecognizable from normal. New complete blood cell count (CBC) parameters, besides the erythrocyte indices, give additional credence to the diagnosis of microcytosis and hypochromasia. These are the % Micro R (the percentage of RBCs with an MCV less than 60 fL) and the % HYPO-HE (the percentage of RBCs with an MCH less than 17 pg). In healthy adults these parameters are both typically less than 1%, meaning that fewer than 1% of the erythrocytes are extremely microcytic or hypochromic. Precise reference intervals, based on gestational age and postnatal age, have not yet been published for neonates.

The other RBC index that can sometimes help identify the cause of an unknown variety of neonatal anemia is the mean corpuscular hemoglobin concentration (MCHC). This is a measure of the concentration of hemoglobin in red cells. Spherical erythrocytes typically have a high MCHC. Unlike the reference intervals for MCV and MCH, the reference interval for MCHC does not change with gestational or postnatal age. It should remain in the range 34±1 g/dL throughout life. Fig. 43.5 illustrates MCHC histograms of three groups of neonates: (1) Coombs (direct antiglobulin test [DAT]) negative, (2) Coombs positive, and (3) neonates with a confirmed diagnosis of hereditary spherocytosis (HS). An elevated MCHC in an anemic neonate is a means, although an imperfect means, of suggesting the diagnosis of HS. As seen in the Fig. 43.5 histogram, MCHC measurements of normal neonates overlap with those with proven HS. However, anemic neonates with an MCHC greater than about 36.5 or 37 g/dL are quite likely to have HS and should be further evaluated with that diagnosis in mind.

Fig. 43.6 is a composite timeline showing typical changes in erythrocytes during the first 90 days following birth. As with the other reference interval diagrams in this chapter, the lower line indicates the 5th percentile lower limit of the reference group and the upper line indicates the 95th percentile upper limit. Blood hemoglobin and reticulocyte concentrations typically decrease during the first days after birth, as does the immature reticulocyte percentage. When these values remain elevated in an anemic neonate, they suggest hemorrhage or hemolysis, reflecting an increase in erythropoietic activity of the marrow in an attempt to compensate for the RBC loss. The reticulocyte hemoglobin content shown in the lower-most panel of Fig. 43.6 reflects the iron content of reticulocytes. Anemic neonates with reticulocyte hemoglobin below the 5th-percentile lower limit are likely to have iron deficiency.

The red cell distribution width (RDW) is another way to characterize the erythrocytes of an anemic neonate. The RDW describes the variation in RBC size within a blood sample. Thus it is a standard way to numerically express erythrocyte anisocytosis, meaning variance in MCV or RBC size. Fig. 43.7A shows the reference interval for RDW at birth according to gestational age. The 95th percentile upper reference interval is about 22% to 23% in extremely low gestational age neonates and 20% in those of older gestation at birth. Fig. 43.7B shows the reference interval during the first 2 weeks after birth, according to gestational age grouping. Anemic neonates with an elevated RDW (above the 95th percentile) typically have reticulocytosis because reticulocytes are larger than mature erythrocytes; thus there is more variation in RBC size. Reticulocytes are not typically measured in neonates as part of the CBC but are ordered as a separate laboratory test. However, the RDW is part of each CBC; therefore if an elevated RDW is noted in an anemic neonate, one can expect reticulocytosis, suggesting hemorrhage or hemolysis and arguing against hyporegenerative anemia.

Another parameter of interest in considering the cause of neonatal anemia is the nucleated RBC count. Fig. 43.8 shows reference intervals for nucleated red blood cells (NRBCs) shown two ways: as an absolute number of NRBCs per microliter blood and with reference to the number of NRBCs per 100 white blood cells. Elevated NRBCs at birth suggest hypoxia in utero occurring 36 hours or so prior to birth. Low to normal values in anemic neonates are typically seen in anemia due to erythrocyte hypoproduction.

Fragmented RBCs occur in microangiopathic conditions. In neonates these conditions are typically disseminated intravascular coagulation (DIC), necrotizing enterocolitis, and sepsis. When erythrocytes circulate past intraluminal fibrin strands, they can be caught, tethered, and torn. After incurring this damage, the erythrocyte membrane can reseal and the damaged cell can circulate as a red cell fragment. When seen on a blood film the damaged red cells are identified as “schistocytes.” Although it is not yet approved by the Food and Drug Administration in the United States, the fragmented red cells (FRCs) parameter is quantified by electronic cell counters as a routine part of the CBC. Fig. 43.9 shows reference intervals for FRCs per microliter of blood over the first 90 days after birth. Values greater than about 900,000 FRCs/μL are abnormal (above the 95th-percentile upper reference interval). Anemic neonates with an elevated FRC have hemolytic anemia from a microangiopathic condition and likely have DIC, necrotizing enterocolitis, or sepsis.

Fig. 43.3 lists clinical and laboratory elements that indicate whether a neonate who has anemia with unknown cause is likely to have anemia due to hypoproduction, hemorrhage, or hemolysis. In the next sections, exact diagnoses under these three categories are listed and detailed.

Hyporegenerative Anemia

Impaired erythrocyte production can occur in a fetus or neonate for a variety of reasons. Lack of an appropriate marrow environment (as seen in osteopetrosis), lack of specific substrates or their carriers (e.g., iron, folate, vitamin B ₁₂ , or transcobalamin II deficiency), and lack of specific growth factors (e.g., decreased Epo production or abnormalities in Epo receptors) can be causative. The most common hyporegenerative anemia in neonatal intensive care unit patients is “anemia of prematurity,” discussed below. A rare collection of hyporegenerative neonatal anemias on a genetic basis are discussed in the section following.