CHAPTER 15: Fetal Disorders

Little was written of fetal disorders in the first edition of this textbook. General dropsy described above is today known as hydrops fetalis (Hydrops Fetalis). Hydrops is perhaps the quintessential fetal disorder, as it can be a manifestation of severe illness from a wide variety of etiologies. Fetal disorders may be acquired—such as alloimmunization, they may be genetic—congenital adrenal hyperplasia or α4-thalassemia, or they may be sporadic developmental abnormalities—like many structural malformations. In this chapter, fetal anemia and thrombocytopenia as well as immune and nonimmune fetal hydrops are reviewed. Fetal structural malformations are reviewed in Chapter 10, genetic abnormalities in Chapters 13 and 14, and conditions amenable to medical and surgical fetal therapies in Chapter 16. Because congenital infections arise as a result of maternal infection or colonization, they are considered in Chapters 64 and 65.

Of the many causes of fetal anemia, one of the most frequent is red cell alloimmunization, which results from transplacental passage of maternal antibodies that destroy fetal red cells. Alloimmunization leads to overproduction of immature fetal and neonatal red cells—erythroblastosis fetalis—a condition now referred to as hemolytic disease of the fetus and newborn (HDFN).

In addition, several congenital infections are also associated with fetal anemia, particularly parvovirus B19, discussed in Chapter 64 (Respiratory Viruses). In Southeast Asian populations, α4-thalassemia is a common cause of severe anemia and nonimmune hydrops. Fetomaternal hemorrhage occasionally creates severe fetal anemia and is discussed in Fetomaternal Hemorrhage. Rare causes of anemia include red cell production disorders—such as Blackfan-Diamond anemia and Fanconi anemia; red cell enzymopathies—glucose-6-phosphate dehydrogenase deficiency and pyruvate kinase deficiency; red cell structural abnormalities—hereditary spherocytosis and elliptocytosis; and myeloproliferative disorders—leukemias. Anemia may be identified through fetal blood sampling, described in Chapter 14 (Fetal Blood Sampling), or by Doppler evaluation of the fetal middle cerebral artery (MCA) peak systolic velocity, described in Management of the Alloimmunized Pregnancy.

Progressive fetal anemia from any cause leads to heart failure, hydrops fetalis, and ultimately death. Fortunately, the prevalence and the course of this otherwise devastating disorder have been dramatically changed by prevention and treatment. Prevention of D alloimmunization is with anti-D immune globulin. Identification and treatment of fetal anemia is with MCA Doppler studies and intrauterine transfusions, respectively. Severely anemic fetuses transfused in utero have survival rates exceeding 90 percent, and even in cases of hydrops fetalis, survival rates approach 80 percent (Lindenberg, 2013; Zwiers, 2017).

Red Cell Alloimmunization

Currently, 33 different blood group systems and 339 red cell antigens are recognized by the International Society of Blood Transfusion (Storry, 2014). Although some of these are immunologically and genetically important, many are so rare as to be of little clinical significance. Any individual who lacks a specific red cell antigen may produce an antibody when exposed to that antigen. Such antibodies can prove harmful to that individual if she receives an incompatible blood transfusion. Accordingly, blood banks routinely screen for erythrocyte antigens. These antibodies may also be harmful to a mother’s fetus during pregnancy. As noted, maternal antibodies formed against fetal erythrocyte antigens may cross the placenta to cause fetal red cell lysis and anemia.

Typically, a fetus inherits at least one red cell antigen from the father that is lacking in the mother. Thus, the mother may become sensitized if enough fetal erythrocytes reach her circulation to elicit an immune response. Even so, alloimmunization is uncommon for the following reasons: (1) low prevalence of incompatible red cell antigens; (2) insufficient transplacental passage of fetal antigens or maternal antibodies; (3) maternal-fetal ABO incompatibility, which leads to rapid clearance of fetal erythrocytes before they elicit an immune response; (4) variable antigenicity; and (5) variable maternal immune response to the antigen.

In population-based screening studies, the prevalence of red cell alloimmunization in pregnancy approximates 1 percent (Bollason, 2017; Koelewijn, 2008). Most cases of severe fetal anemia requiring antenatal transfusion are attributable to anti-D, anti-Kell, anti-c, or anti-E alloimmunization (de Haas, 2015).

Alloimmunization Detection

At the first prenatal visit, a blood type and antibody screen are routinely assessed, and unbound antibodies in maternal serum are detected by the indirect Coombs test (Chap. 9, Definitions). When the result is positive, the specific antibodies are identified, their immunoglobulin subtype is determined as either immunoglobulin G (IgG) or M (IgM), and the titer is quantified. Only IgG antibodies are a concern because IgM antibodies do not cross the placenta. Selected antibodies and their potential to cause fetal hemolytic anemia are listed in Table 15-1. The critical titer is the level at which significant fetal anemia could potentially develop. This may be different for each antibody, is determined individually by each laboratory, and usually ranges between 1:8 and 1:32. If the critical titer for anti-D antibodies is 1:16, a titer ≥1:16 indicates the possibility of severe hemolytic disease. An important exception is Kell sensitization, which is discussed in Alloimmunization to Minor Antigens.

CDE (Rh) Blood Group Incompatibility

The CDE system includes five red cell proteins or antigens: C, c, D, E, and e. There is no “d” antigen, and D-negativity is defined as the absence of the D antigen. Although most people are D positive or negative, more than 200 D antigen variants exist (Daniels, 2013). Rh was formerly termed rhesus because of a misconception that red cells from rhesus monkeys expressed human blood group antigen. In transfusion medicine, “rhesus” is no longer used (Sandler, 2017).

CDE antigens are clinically important. D-negative individuals may become sensitized after a single exposure to as little as 0.1 mL of fetal erythrocytes (Bowman, 1988). The two responsible genes—RHD and RHCE—are located on the short arm of chromosome 1 and are inherited together, independent of other blood group genes. The incidence of antigen positivity varies according to racial and ethnic origin. Nearly 85 percent of non-Hispanic white Americans are D-positive. The incidence approximates 90 percent for Native Americans, 93 percent for African Americans and Hispanic Americans, and at least 99 percent for Asian individuals (Garratty, 2004).

The prevalence of D alloimmunization complicating pregnancy ranges from 0.5 to 0.9 percent (Koelewijn, 2008; Martin, 2005). Without anti-D immune globulin prophylaxis, a D-negative woman delivered of a D-positive, ABO-compatible newborn has a 16-percent likelihood of developing alloimmunization. Two percent will become sensitized by the time of delivery, 7 percent by 6 months postpartum, and the remaining 7 percent will be “sensibilized”—producing detectable antibodies only in a subsequent pregnancy (Bowman, 1985). If there is ABO incompatibility, the D alloimmunization risk approximates 2 percent without prophylaxis (Bowman, 2006). The reason for the differing rates relative to ABO blood type results from erythrocyte destruction of ABO-incompatible cells, which thereby limits sensitizing opportunities. D sensitization also may occur following first-trimester pregnancy complications, prenatal diagnostic procedures, and maternal trauma (Table 15-2).

The C, c, E, and e antigens have lower immunogenicity than the D antigen but can cause hemolytic disease. Sensitization to E, c, and C antigens complicates approximately 0.3 percent of pregnancies in screening studies and accounts for about 30 percent of red cell alloimmunization cases (Howard, 1998; Koelewijn, 2008). Anti-E alloimmunization is the most common, but the need for fetal or neonatal transfusions is greater with anti-c alloimmunization than with anti-E or anti-C (de Haas, 2015; Hackney, 2004; Koelewijn, 2008).

The Grandmother Effect

In virtually all pregnancies, small amounts of maternal blood enter the fetal circulation. Real-time polymerase chain reaction (PCR) has been used to identify maternal D-positive DNA in peripheral blood from preterm and full-term D-negative newborns (Lazar, 2006). Thus, it is possible for a D-negative female fetus exposed to maternal D-positive red cells to develop sensitization. When such an individual reaches adulthood, she may produce anti-D antibodies even before or early in her first pregnancy. This mechanism is called the grandmother effect or theory because the fetus in the current pregnancy is jeopardized by maternal antibodies that were initially provoked by his or her grandmother’s erythrocytes.

Alloimmunization to Minor Antigens

Because routine administration of anti-D immunoglobulin prevents anti-D alloimmunization, proportionately more cases of hemolytic disease are caused by red cell antigens other than D (American College of Obstetricians and Gynecologists, 2016; Koelewijn, 2008). These are also known as minor antigens. Kell antigens are among the most frequent. Other antigens with potential to cause severe alloimmunization include Duffy group A—Fy^a, MNS, and Kidd—Jk^a (de Hass, 2015; Moise, 2008). Most cases of sensitization to minor antigens result from incompatible blood transfusions. However, if an IgG red cell antibody is detected and there is any doubt as to its significance, the clinician should err on the side of caution, and the pregnancy should be evaluated for hemolytic disease.

Only a few blood group antigens pose no fetal risk. Lewis antibodies—Le^a and Le^b, as well as I antibodies, are cold agglutinins. They are predominantly IgM and are not expressed on fetal red cells (American College of Obstetricians and Gynecologists, 2016). Another antibody that does not cause fetal hemolysis is Duffy group B—Fy^b.

Kell Alloimmunization

Approximately 90 percent of non-Hispanic white Americans and up to 98 percent of African Americans are Kell negative. Kell type is not routinely determined. Transfusion history is important, as nearly 90 percent of Kell sensitization cases result from transfusion with Kell-positive blood.

Kell sensitization may develop more rapidly and may be more severe than with sensitization to D and other blood group antigens. This is because Kell antibodies attach to erythrocyte precursors in the fetal bone marrow, thereby impairing the normal hemopoietic response to anemia. With fewer erythrocytes produced, there is less hemolysis, and severe anemia may not be predicted by the maternal Kell antibody titer. One option is to use a lower critical titer—1:8—for Kell sensitization (Moise, 2012). The American College of Obstetricians and Gynecologists (2016) has recommended that antibody titers not be used to monitor Kell-sensitized pregnancies.

ABO Blood Group Incompatibility

Incompatibility for the major blood group antigens A and B is the most common cause of hemolytic disease in newborns, but it does not cause appreciable hemolysis in the fetus. Approximately 20 percent of newborns have ABO blood group incompatibility, yet only 5 percent are affected clinically. And in such cases, the resulting anemia is typically mild.

The condition differs from CDE incompatibility in several respects. First, ABO incompatibility is often seen in firstborn neonates, whereas sensitization to other blood group antigens is not. This is because most group O women have developed anti-A and anti-B isoagglutinins before pregnancy from exposure to bacteria displaying similar antigens. Second, ABO alloimmunization rarely becomes more severe in successive pregnancies. Last, ABO incompatibility is considered a pediatric disease—rarely of obstetrical concern. This is because most anti-A and anti-B antibodies are IgM and do not cross the placenta. Fetal red cells also have fewer A and B antigenic sites than adult cells and are thus less immunogenic.

Consequently, fetal surveillance and early delivery are not indicated in pregnancies with prior ABO incompatibility. Careful neonatal observation is essential, however, because hyperbilirubinemia may require treatment with phototherapy or occasionally transfusion (Chap. 33, Polycythemia and Hyperviscosity).

Management of the Alloimmunized Pregnancy

An estimated 25 to 30 percent of fetuses from D-alloimmunized pregnancies will have mild-to-moderate hemolytic anemia. And without treatment, up to 25 percent will develop hydrops (Tannirandorn, 1990). If alloimmunization is detected and the titer is below the critical value, the titer is generally repeated every 4 weeks for the duration of the pregnancy (American College of Obstetricians and Gynecologists, 2016). Importantly, if a prior pregnancy was complicated by alloimmunization, serial titer assessment is not indicated, and the pregnancy is assumed to be at risk regardless of titer. Management of such pregnancies is discussed subsequently. In any pregnancy in which an antibody titer has reached a critical value, there is no benefit to repeating it. The pregnancy is at risk even if the titer drops, and further evaluation is still required.

Determining Fetal Risk

Up to 40 percent of D-negative pregnant women carry a D-negative fetus. The presence of anti-D antibodies reflects maternal sensitization but does not indicate whether the fetus is D-positive. If a woman became sensitized in a prior pregnancy, her antibody titer might rise to high levels during the current pregnancy even if the current fetus is D-negative, due to an amnestic response. In a non-Hispanic white couple in which the woman is D-negative, there is an 85-percent chance that the man is D-positive. But, in 60 percent of these cases, he will be heterozygous at the D-locus. And, if he is heterozygous, then half of his children will be at risk for hemolytic disease. Transfusion history is relevant. Alloimmunization to a red cell antigen other than D may have occurred following a blood transfusion in the past, and if that antigen is not present on paternal erythrocytes, the pregnancy is not at risk.

Initial evaluation of alloimmunization begins with determining the paternal erythrocyte antigen status. Provided that paternity is certain, if the father is negative for the red cell antigen to which the mother is sensitized, the pregnancy is not at risk. In a D-alloimmunized pregnancy in which the father is D-positive, it is helpful to determine paternal zygosity for the D antigen using DNA-based analysis. If the father is heterozygous—or if paternity is not known—the woman should be offered assessment of fetal genotype. Traditionally, this was done with amniocentesis and PCR testing of uncultured amniocytes, which has a positive-predictive value of 100 percent and negative-predictive value of approximately 97 percent (American College of Obstetricians and Gynecologists, 2016; Van den Veyver, 1996). Fetal testing for other antigens—such as E/e, C/c, Duffy, Kell, Kidd, and M/N—is also available with this method. Chorionic villus sampling is not recommended because of greater risk for fetomaternal hemorrhage and subsequent worsening of alloimmunization.

Noninvasive fetal D genotyping has been performed using cell-free DNA (cfDNA) from maternal plasma (Chap. 13, Fetal DNA in the Maternal Circulation). The reported sensitivity exceeds 99 percent, the specificity exceeds 95 percent, and positive- or negative-predictive values are similarly very high (de Haas, 2016; Johnson, 2017; Moise, 2016; Vivanti, 2016). Fetal D genotyping with cfDNA is routinely used in parts of Europe. There are two potential indications in D-negative pregnant women: (1) in women with D alloimmunization, testing can identify fetuses that are also D-negative and do not require anemia surveillance, and (2) in women without D alloimmunization, anti-D immune globulin might be withheld if the fetus is D negative. In the case of the latter, the American College of Obstetricians and Gynecologists (2017) does not recommend routine cfDNA screening in D-negative pregnancies until it becomes cost-effective.

Management of the alloimmunized pregnancy is individualized and may consist of maternal antibody titer surveillance, sonographic monitoring of the fetal MCA peak systolic velocity, amnionic fluid bilirubin studies, or fetal blood sampling. Accurate pregnancy dating is critical. The gestational age at which fetal anemia developed in prior pregnancies is important because anemia tends to occur earlier and be sequentially more severe.

Middle Cerebral Artery Doppler Velocimetry

Serial measurement of the peak systolic velocity of the fetal MCA is the recommended test for detection of fetal anemia (Society for Maternal–Fetal Medicine, 2015a). The anemic fetus shunts blood preferentially to the brain to maintain adequate oxygenation. The velocity rises because of increased cardiac output and decreased blood viscosity. The technique is discussed in Chapter 10 (Ductus Arteriosus) and requires training and experience (American College of Obstetricians and Gynecologists, 2016).

In a landmark study, Mari and coworkers (2000) measured the MCA peak systolic velocity serially in 111 fetuses at risk for anemia and in 265 normal control fetuses. The threshold value of 1.5 multiples of the median (MoM) for gestational age correctly identified all fetuses with moderate or severe anemia. This provided a sensitivity of 100 percent, with a false-positive rate of 12 percent.

The MCA peak systolic velocity is followed serially, and values are plotted on a curve like the one shown in Figure 15-1. If the velocity is between 1.0 and 1.5 MoM and the slope is rising—such that the value is approaching 1.5 MoM—surveillance is generally increased to weekly Doppler interrogation. If the MCA peak systolic velocity exceeds 1.5 MoM and the gestational age is younger than 34 or 35 weeks, fetal blood sampling should be considered and followed by fetal transfusion if needed (Society for Maternal–Fetal Medicine, 2015a). The false-positive rate of MCA peak systolic velocity increases significantly beyond 34 weeks, due to the normal augmentation in cardiac output that develops at this gestational age (Moise, 2008; Zimmerman, 2002).

Amnionic Fluid Spectral Analysis

This test is included for historical interest. More than 50 years ago, Liley (1961) demonstrated the utility of amnionic fluid spectral analysis to measure bilirubin concentration and to thereby estimate hemolysis severity. Amnionic fluid bilirubin concentration was measured by a spectrophotometer and was represented as the change in optical density absorbance at 450 nm—ΔOD₄₅₀. The likelihood of fetal anemia was determined by plotting the ΔOD₄₅₀ value on a graph that was divided into zones. These zones roughly correlated with fetal hemoglobin concentration, and thus with anemia severity. The original Liley graph was valid from 27 to 42 weeks’ gestation and was subsequently modified by Queenan (1993) to include gestational ages as early as 14 weeks. However, the amnionic fluid bilirubin level is normally high in midpregnancy, limiting the reliability of this technique.

Middle cerebral artery velocimetry is more accurate than ΔOD₄₅₀ assessment and does not confer risks for increased alloimmunization associated with amniocentesis. It has replaced ΔOD₄₅₀ assessment for this purpose.

Fetal Blood Transfusion

If there is evidence of severe fetal anemia, because of either elevated MCA peak systolic velocity or development of fetal hydrops, management is strongly influenced by gestational age. Fetal blood sampling and intrauterine transfusion are generally performed prior to 34 to 35 weeks (Society for Maternal-Fetal Medicine, 2015a). Intravascular transfusion into the umbilical vein under sonographic guidance is the preferred method of fetal transfusion. Transfusion into the fetal peritoneal cavity may be necessary with severe, early-onset hemolytic disease in the early second trimester, a time when the umbilical vein is too narrow to readily permit needle entry. With hydrops, although peritoneal absorption is impaired, some prefer to transfuse into both the fetal peritoneal cavity and the umbilical vein.

Transfusion is generally recommended only if the fetal hematocrit is <30 percent (Society for Maternal-Fetal Medicine, 2015a). Once hydrops has developed, the hematocrit is generally 15 percent or lower. The red cells transfused are type O, D-negative, cytomegalovirus-negative, packed to a hematocrit of approximately 80 percent to prevent volume overload, irradiated to prevent fetal graft-versus-host reaction, and leukocyte-poor. The fetal–placental volume allows rapid infusion of a relatively large quantity of blood. Before transfusion, a paralytic agent such as vecuronium may be given to the fetus to minimize movement. In a nonhydropic fetus, the target hematocrit is generally 40 to 50 percent. The volume transfused may be estimated by multiplying the estimated fetal weight in grams by 0.02 for each 10-percent rise in hematocrit needed (Giannina, 1998). In the severely anemic fetus at 18 to 24 weeks’ gestation, less blood is transfused initially, and another transfusion may be planned for approximately 2 days later. Subsequent transfusions usually take place every 2 to 4 weeks, depending on the hematocrit.

The MCA peak systolic velocity threshold for severe anemia is higher following an initial transfusion—1.70 MoM rather than 1.50 MoM (Society for Maternal-Fetal Medicine, 2015a). It is hypothesized that the change in threshold compensates for the contribution of donor cells in the initial transfusion, because donor cells (from adults) have a smaller mean corpuscular volume. Alternately, the timing of subsequent transfusions is based on anemia severity and posttransfusion hematocrit. Following transfusion, the fetal hematocrit generally drops by approximately 1 percent per day. A more rapid initial decline may be encountered in the setting of fetal hydrops.

Outcomes

Procedure-related complications have declined significantly at experienced centers in recent years, with overall survival rates exceeding 95 percent (Zwiers, 2017). Complications include fetal death in approximately 2 percent, need for emergent cesarean delivery in 1 percent, and infection and preterm rupture of membranes in 0.3 percent each, respectively. The stillbirth rate exceeds 15 percent if transfusion is required before 20 weeks (Lindenberg, 2013; Zwiers, 2017). Considering that fetal transfusion is potentially lifesaving in severely compromised fetuses, these risks should not dissuade therapy.

Van Kamp (2001) reported that if hydrops had developed, the survival rate approached 75 to 80 percent. However, of the nearly two thirds with resolution of hydrops following transfusion, more than 95 percent survived. The survival rate was <40 percent if hydrops persisted.

Lindenberg (2012) reviewed long-term outcomes following intrauterine transfusion in a cohort of more than 450 alloimmunized pregnancies. Alloimmunization was secondary to anti-D in 80 percent, anti-Kell in 12 percent, and anti-c in 5 percent. Approximately a fourth of affected fetuses had hydrops, and more than half also required exchange transfusion in the neonatal period. Among nearly 300 children aged 2 to 17 years who participated in neurodevelopmental testing, fewer than 5 percent had severe impairments. These included severe developmental delay in 3 percent, cerebral palsy in 2 percent, and deafness in 1 percent.

Prevention of Anti-D Alloimmunization

Anti-D immune globulin is one of the success stories of modern obstetrics. It has been used for nearly five decades to prevent D alloimmunization. In countries without access to anti-D immune globulin, up to 10 percent of D-negative pregnancies are complicated by hemolytic disease of the fetus and newborn (Zipursky, 2015). With immunoprophylaxis, however, the alloimmunization risk is reduced to <0.2 percent. Despite long-standing and widespread use, its mechanism of action is not completely understood.

As many as 90 percent of alloimmunization cases occur from fetomaternal hemorrhage at delivery. Routine postpartum administration of anti-D immune globulin to at-risk pregnancies within 72 hours of delivery lowers the alloimmunization rate by 90 percent (Bowman, 1985). Additionally, provision of anti-D immune globulin at 28 weeks’ gestation reduces the third-trimester alloimmunization rate from approximately 2 percent to 0.1 percent (Bowman, 1988). Whenever there is doubt whether to give anti–D immunoglobulin, it should be given. If not needed, it will not cause harm, but failure to provide it when needed can have severe consequences.

Current preparations of anti-D immune globulin are derived from human plasma donated by individuals with high-titer anti-D immunoglobulin D antibodies. Formulations prepared by cold ethanol fractionation and ultrafiltration must be administered intramuscularly because they contain plasma proteins that could result in anaphylaxis if given intravenously. However, formulations prepared using ion exchange chromatography may be administered either intramuscularly or intravenously. This is important for treatment of significant fetomaternal hemorrhage, which is discussed subsequently. Both preparation methods effectively remove viral particles, including hepatitis and human immunodeficiency viruses. Depending on the preparation, the half-life of anti-D immune globulin ranges from 16 to 24 days, which is why it is given both in the third trimester and following delivery. The standard intramuscular dose of anti-D immune globulin—300 μg or 1500 IU—will protect the average-sized mother from a fetal hemorrhage of up to 30 mL of fetal whole blood or 15 mL of fetal red cells.

In the United States, anti-D immune globulin is given prophylactically to all D-negative, unsensitized women at approximately 28 weeks’ gestation, and a second dose is given after delivery if the newborn is D-positive (American College of Obstetricians and Gynecologists, 2017). Before the 28-week dose of anti-D immune globulin, repeat antibody screening is recommended to identify individuals who have become alloimmunized (American Academy of Pediatrics, 2017). Following delivery, anti-D immune globulin should be given within 72 hours. Recognizing that 40 percent of neonates born to D-negative women are also D negative, administration of immune globulin is recommended only after the newborn is confirmed to be D positive (American College of Obstetricians and Gynecologists, 2017). If immune globulin is inadvertently not administered following delivery, it should be given as soon as the omission is recognized, because there may be some protection up to 28 days postpartum (Bowman, 2006). Anti-D immune globulin is also administered after pregnancy-related events that could result in fetomaternal hemorrhage (see Table 15-2).

Anti-D immune globulin may produce a weakly positive—1:1 to 1:4—indirect Coombs titer in the mother. This is harmless and should not be confused with development of alloimmunization. Additionally, as the body mass index increases above 27 to 40 kg/m², serum antibody levels decrease by 30 to 60 percent and may be less protective (MacKenzie, 2006; Woelfer, 2004). D-negative women who receive other types of blood products—including platelet transfusions and plasmapheresis—are also at risk of becoming sensitized, and this can be prevented with anti-D immune globulin. Rarely, a small amount of antibody crosses the placenta and results in a weakly positive direct Coombs test in cord and infant blood. Despite this, passive immunization does not cause significant fetal or neonatal hemolysis.

It is estimated that in 2 to 3 per 1000 pregnancies, the volume of fetomaternal hemorrhage exceeds 30 mL of whole blood (American College of Obstetricians and Gynecologists, 2017). A single dose of anti-D immune globulin would be insufficient in such situations. If additional anti-D immune globulin is considered only for women with risk factors such as those shown in Table 15-2, then half of those who require additional immune globulin may be missed. For this reason, all D-negative women should be screened at delivery, typically with a rosette test, followed by quantitative testing if indicated (American College of Obstetricians and Gynecologists, 2017).

The rosette test is a qualitative test that identifies whether fetal D-positive cells are present in the circulation of a D-negative woman. A sample of maternal blood is mixed with anti-D antibodies that coat any D-positive fetal cells present in the sample. Indicator red cells bearing the D-antigen are then added, and rosettes form around the fetal cells as the indicator cells attach to them by the antibodies. Thus, if rosettes are visualized, there are fetal D-positive cells in that sample. In the setting of D incompatibility, or any time a large fetomaternal hemorrhage is suspected—regardless of antigen status, a Kleihauer-Betke test or flow cytometry test are used. These are discussed in Fetal Thrombocytopenia.

The dosage of anti-D immune globulin is calculated from the estimated volume of the fetal-to-maternal hemorrhage, as described in Fetal Thrombocytopenia. One 300-μg dose is given for each 15 mL of fetal red cells or 30 mL of fetal whole blood to be neutralized. If using an intramuscular preparation of anti-D immune globulin, no more than five doses may be given in a 24-hour period. If using an intravenous preparation, two ampules—totaling 600 μg—may be given every 8 hours. To determine if the administered dose was adequate, the indirect Coombs test may be performed. A positive result indicates that there is excess anti–D immunoglobulin in maternal serum, thus demonstrating that the dose was sufficient. Alternatively, a rosette test may be performed to assess whether circulating fetal cells remain.

Serological Weak D Phenotypes

Formerly called D^u, these are the most common antigenic D variants in the United States and Europe. Serological weak D phenotypes have been further refined into two general categories using molecular analysis—RHD genotyping. Molecular weak D phenotypes carry reduced numbers of intact D antigens on the red cell surface. Those designated partial D types have protein deletions associated with abnormal D antigens that lack epitopes (Sandler, 2017). When this distinction is known, it can have clinical consequences in terms of sensitization risk and need for anti-D immune globulin.

Traditionally, serological weak D individuals have been considered to be D-positive or -negative depending on the clinical situation. For the purposes of blood donation, they are categorized as D-positive, whereas transfusion recipients with weak D are considered D-negative. In pregnancy, weak D has also been considered D-negative, so that individuals receive immune globulin and avoid potential sensitization (American College of Obstetricians and Gynecologists, 2017; Sandler, 2015).

Many non-Hispanic white Americans who test positive for weak D have weak D phenotypes 1, 2, or 3. Individuals with these phenotypes may be managed as though they are D-positive. Because they are not at risk for alloimmunization, anti-D immune globulin is not needed (Sandler 2015, 2017). In contrast, individuals with partial D antigens may be at risk for D-sensitization and do require immune globulin. Molecular RHD genotyping has been suggested for pregnant women with weak D phenotype, but cost-benefit analysis of this strategy is presently lacking (American College of Obstetricians and Gynecologists, 2017). If molecular genetic testing has not been performed in those with serologic weak D phenotype, D immunoprophylaxis should be administered to those with weak D phenotype.

A small amount of fetomaternal bleeding likely occurs in all pregnancies, and in two thirds, this may be sufficient to provoke an antigen-antibody reaction. As shown in Figure 15-2, the incidence increases with advancing gestation and the volume of fetal blood in the maternal circulation. Fortunately, a large blood loss—true fetomaternal hemorrhage—is rare. In one series of more than 30,000 pregnancies, fetomaternal hemorrhage ≥150 mL occurred in 1 per 2800 births (de Almeida, 1994). The prevalence of fetomaternal hemorrhage of at least 30 mL—the volume of fetal blood covered by a standard 300-μg dose of anti-D immune globulin—is estimated to be 3 per 1000 pregnancies (Wylie, 2010).

Selected causes of fetomaternal hemorrhage are shown in Table 15-2. It also may occur with placenta previa, placental chorioangioma, or vasa previa (Giacoia 1997; Rubod, 2007). In each of these circumstances, however, fetomaternal hemorrhage is extremely uncommon if not rare. And, in more than 80 percent of cases, no cause is identified. With significant hemorrhage, the most common presenting complaint is decreased fetal movement (Bellussi, 2017; Wylie, 2010). A sinusoidal fetal heart rate pattern is infrequently seen but warrants immediate evaluation (Chap. 24, Periodic Fetal Heart Rate Changes). Sonography may demonstrate elevated MCA peak systolic velocity, and indeed this is reported to be the most accurate predictor (Bellusi, 2017; Wylie, 2010). Hydrops is an ominous finding. If fetomaternal hemorrhage is suspected, an elevated MCA peak systolic velocity or sonographic evidence of hydrops prompts consideration of urgent fetal transfusion or delivery.

One limitation of quantitative tests for fetal cells in the maternal circulation is that they do not provide information regarding hemorrhage timing or chronicity (Wylie, 2010). In general, anemia developing gradually or chronically, as in alloimmunization, is better tolerated by the fetus than acute anemia. Chronic anemia may not produce fetal heart rate abnormalities until the fetus is moribund. In contrast, significant acute hemorrhage is poorly tolerated by the fetus and may cause profound fetal neurological impairment from cerebral hypoperfusion, ischemia, and infarction. In some cases, fetomaternal hemorrhage is identified during stillbirth evaluation (Chap. 35, Risk Factors).

Laboratory Tests

Once fetomaternal hemorrhage is recognized, the volume of fetal blood loss should be estimated. The volume is essential to calculate the appropriate dose of anti D-immune globulin if the woman is D-negative, and it may influence obstetrical management.

The most commonly used quantitative test for fetal red cells in the maternal circulation is the acid elution or Kleihauer-Betke (KB) test (Kleihauer, 1957). Fetal erythrocytes contain hemoglobin F, which is more resistant to acid elution than hemoglobin A. After exposure to acid, only fetal hemoglobin remains, such that after staining, the fetal erythrocytes appear red and adult erythrocytes appear as “ghosts” (Fig. 15-3). The fetal cells are then counted and expressed as a percentage of adult cells. The KB test is labor intensive. Importantly, there are two scenarios in which it may not be accurate: (1) maternal hemoglobinopathies such as β-thalassemia in which the fetal hemoglobin level is elevated and (2) pregnancies at or near term, when the fetus has already started to produce hemoglobin A.

Hemorrhage Quantification

The volume of fetomaternal hemorrhage is calculated from the KB test result using the following formula:

One method is to estimate the maternal blood volume (MBV) as 5000 mL for a normal-size, normotensive women at term. Thus, for 1.7-percent positive KB-stained cells in a woman of average size with a hematocrit of 35 percent and whose fetus has a hematocrit of 50 percent:

The fetal-placental blood volume at term approximates 125 mL/kg. For a 3000-g fetus, that would equate to 375 mL. Thus, this fetus lost approximately 15 percent (60 ÷ 375 mL) of the fetal-placental volume. Because the hematocrit is 50 percent in a term fetus, this 60 mL of whole blood represents 30 mL of red cells lost into the maternal circulation. This loss should be well tolerated hemodynamically but would require two 300-μg doses of anti-D immunoglobulin to prevent alloimmunization. A more precise method to estimate the maternal blood volume includes a calculation based on the maternal height, weight, and anticipated physiological maternal blood volume accrual (Table 41-1).

Fetomaternal hemorrhage can also be quantified using flow cytometry, which uses monoclonal antibodies to hemoglobin F or to the D antigen, followed by quantification of fluorescence (Chambers, 2012; Welsh, 2016). Flow cytometry is an automated test that can analyze a greater number of cells than the KB test. Further, it is unaffected by maternal levels of fetal hemoglobin or by fetal levels of hemoglobin A. Flow cytometry has been reported to be more sensitive and accurate than the KB test, however, it uses specialized technology not routinely available in many hospitals (Chambers, 2012; Corcoran, 2014; Fernandes, 2007).

Alloimmune Thrombocytopenia

This condition is also referred to as neonatal alloimmune thrombocytopenia (NAIT) or fetal and neonatal alloimmune thrombocytopenia (FNAIT). Alloimmune thrombocytopenia (AIT) is the most common cause of severe thrombocytopenia among term newborns, with a frequency of 1 to 2 per 1000 births (Kamphuis, 2010; Pacheco, 2013; Risson, 2012). FNAIT is caused by maternal alloimmunization to paternally inherited fetal platelet antigens. The resulting maternal antiplatelet antibodies cross the placenta in a manner similar to red cell alloimmunization (Red Cell Alloimmunization). Unlike immune thrombocytopenia, the maternal platelet count is normal with FNAIT. And, unlike anti-D alloimmunization, severe sequelae may affect the initial at-risk pregnancy.

Maternal platelet alloimmunization is most often against human platelet antigen-1a (HPA-1a). It accounts for 80 to 90 percent of cases and is associated with the greatest severity (Bussel, 1997; Knight, 2011; Tiller, 2013). This is followed in order of frequency by HPA-5b, HPA-1b, and HPA-3a. Alloimmunization to other antigens accounts for only 1 percent of reported cases.

Approximately 85 percent of non-Hispanic white individuals are HPA-1a positive. Two percent are homozygous for HPA-1b and thus at risk for alloimmunization. Importantly, however, only 10 percent of homozygous HPA-1b mothers who carry an HPA-1a fetus will produce anti-platelet antibodies. Approximately a third of affected fetuses or neonates will develop severe thrombocytopenia, and 10 to 20 percent of those with severe thrombocytopenia sustain an intracranial hemorrhage (ICH) (Kamphuis, 2010). As a result, population-based screening studies have identified FNAIT-associated ICH in 1 per 25,000 to 60,000 pregnancies (Kamphuis, 2010; Knight, 2011).

FNAIT may present in various ways. In some cases, neonatal thrombocytopenia may be an incidental finding or the newborn may manifest petechiae. In the other extreme, a fetus or neonate may develop devastating ICH—often before birth. Of 600 pregnancies with FNAIT identified through a large international registry, fetal or neonatal ICH complicated 7 percent of cases (Tiller, 2013). Hemorrhage affected the first-born child in 60 percent and occurred before 28 weeks’ gestation in half. A third of affected children died soon after birth, and 50 percent of survivors had severe neurological disabilities. Bussel and coworkers (1997) evaluated fetal platelet counts before therapy in 107 fetuses with FNAIT. Thrombocytopenia severity was predicted by a prior sibling with perinatal ICH, and 98 percent of cases were identified this way. The initial platelet count was <20,000/μL in 50 percent. In cases in which the platelet count was initially >80,000/μL, they noted that it dropped by more than 10,000/μL each week in the absence of therapy.

Diagnosis and Management

Alloimmune thrombocytopenia is typically diagnosed following delivery of a neonate with severe and unexplained thrombocytopenia to a woman whose platelet count is normal. Rarely, the diagnosis is ascertained after identifying fetal ICH. The condition recurs in 70 to 90 percent of subsequent pregnancies, is often severe, and usually develops earlier with each successive pregnancy. Traditionally, fetal blood sampling was performed to detect fetal thrombocytopenia and to tailor therapy, with transfusion of platelets if the fetal platelet count was <50,000/μL. Because of procedure-related complications, however, experts recommend abandoning routine fetal platelet sampling in favor of empirical treatment with intravenous immune globulin (IVIG) and prednisone (Berkowitz, 2006; Pacheco, 2011).

Therapy is stratified according to whether a prior affected pregnancy was complicated by perinatal ICH, and if so, at what gestational age (Table 15-3). Pioneering work by Bussel (1996) and Berkowitz (2006) and their colleagues demonstrated the efficacy of such treatment. In one series of 50 pregnancies with fetal thrombocytopenia secondary to FNAIT, IVIG raised the platelet count by approximately 50,000/μL, and no fetus developed ICH (Bussel, 1996). Among pregnancies at particularly high risk—based on a platelet count <20,000/μL or sibling with FNAIT-associated ICH—the addition of corticosteroids to IVIG increased the platelet count in 80 percent of cases (Berkowitz, 2006). Cesarean delivery has been recommended at or near term. A noninstrumental vaginal delivery is generally considered only if fetal blood sampling has demonstrated a platelet count >100,000/μL (Pacheco, 2011).

Additional considerations include risks and costs associated with therapy. Side effects of IVIG may include fever, headache, nausea/vomiting, myalgia, and rash. Maternal hemolysis also has been described (Rink, 2013). Costs for IVIG may exceed $70 per gram or nearly $10,000 for each weekly 2-g/kg infusion for an average-size pregnant woman (Pacheco, 2011).

Immune Thrombocytopenia

Also known as immune or idiopathic thrombocytopenic purpura (ITP), this autoimmune disorder is characterized by antiplatelet IgG antibodies that attack platelet glycoproteins. In pregnancy, these antibodies may cross the placenta and cause fetal thrombocytopenia. Maternal ITP is discussed in Chapter 56 (Platelet Disorders). Fetal thrombocytopenia is usually mild. However, neonatal platelet levels may fall rapidly after birth, with a nadir at 48 to 72 hours of life. Neither the maternal platelet count, identification of antiplatelet antibodies, nor treatment with corticosteroids effectively predicts fetal or neonatal platelet counts (Hachisuga, 2014). Importantly, fetal platelet counts are usually adequate to allow vaginal delivery without an increased risk of ICH. In a recent review of more than 400 pregnancies with ITP, there was no case of fetal or neonatal ICH and no infant with any central nervous system abnormality (Wyszynski, 2016). Fetal bleeding complications are considered rare, and fetal blood sampling is not recommended (Neunert, 2011). Delivery mode is based on standard obstetrical indications.

This term refers to excessive accumulation of serous fluid. Strictly defined, hydrops fetalis is edema of the fetus. Traditionally, the diagnosis was made after delivery of a massively edematous neonate, often stillborn (Fig. 15-4). With sonography, hydrops has become a prenatal diagnosis. It is defined as two or more fetal effusions—pleural, pericardial, or ascites—or one effusion plus anasarca. As hydrops progresses in severity, edema is invariably a component, and is usually accompanied by placentomegaly and hydramnios. Clinically significant edema is defined sonographically as skin thickness >5 mm, and placentomegaly if the placenta thickness is at least 4 cm in the second trimester or 6 cm in the third trimester (Bellini, 2009; Society for Maternal–Fetal Medicine, 2015b). Hydrops may result from a wide range of conditions with varying pathophysiologies, each with the potential to make the fetus severely ill. It is divided into two categories. If found in association with red cell alloimmunization, it is termed immune, otherwise, it is nonimmune.

Immune Hydrops

The incidence of immune hydrops has dramatically declined with the advent of anti-D immune globulin, MCA Doppler studies for detection of severe anemia, and prompt fetal transfusion when needed (Fetal Blood Transfusion). However, fewer than 10 percent of hydrops cases are caused by red cell alloimmunization (Bellini, 2012; Santolaya, 1992).

The pathophysiology underlying hydrops remains unknown. Immune hydrops is postulated to share several physiological abnormalities with nonimmune hydrops. As shown in Figure 15-5, these include decreased colloid oncotic pressure, increased hydrostatic (or central venous) pressure, and enhanced vascular permeability. Immune hydrops results from transplacental passage of maternal antibodies that destroy fetal red cells. The resultant anemia stimulates marrow erythroid hyperplasia and extramedullary hematopoiesis in the spleen and liver. The latter likely causes portal hypertension and impaired hepatic protein synthesis, which lowers plasma oncotic pressure (Nicolaides, 1985). Fetal anemia also may raise central venous pressure (Weiner, 1989). Finally, tissue hypoxia from anemia may increase capillary permeability, such that fluid collects in the fetal thorax, abdominal cavity, and/or subcutaneous tissue.

The degree of anemia in immune hydrops is typically severe. In a series of 70 pregnancies with fetal anemia from red cell alloimmunization, Mari and coworkers (2000) found that all those with immune hydrops had hemoglobin values <5 g/dL. As discussed in Fetal Blood Transfusion, immune hydrops is treated with fetal blood transfusions.

Nonimmune Hydrops

At least 90 percent of cases of hydrops are nonimmune (Bellini, 2012; Santolaya, 1992). The prevalence estimate is 1 per 1500 second-trimester pregnancies (Heinonen, 2000). The number of specific disorders that can lead to nonimmune hydrops is extensive. Etiologies and the proportion of births within each hydrops category from a review of more than 6700 affected pregnancies are summarized in Table 15-4. A cause is identified in at least 60 percent prenatally and in more than 80 percent postnatally (Bellini, 2009; Santo, 2011). Currently, approximately 20 percent of cases remain idiopathic (Bellini, 2015). As shown in Figure 15-5, several different pathophysiological processes are proposed to account for the final common pathway of hydrops fetalis.

Importantly, the etiology of nonimmune hydrops varies according to when in gestation it is identified. Of those diagnosed prenatally, aneuploidy accounts for approximately 20 percent, cardiovascular abnormalities for 15 percent, and infections for 14 percent—the most common of these being parvovirus B19 (Santo, 2011). Overall, only 40 percent of pregnancies with nonimmune hydrops result in a liveborn neonate, and of these, the neonatal survival rate is only about 50 percent. Sohan and colleagues (2001) reviewed 87 pregnancies with hydrops and found that 45 percent of those diagnosed before 24 weeks’ gestation had a chromosomal abnormality. The most frequent aneuploidy was 45,X—Turner syndrome, and in such cases, the survival rate was <5 percent (Chap. 13, Polyploidy). If hydrops is detected in the first trimester, the aneuploidy risk is nearly 50 percent, and most have cystic hygromas (Fig. 10-22).

Although the prognosis of nonimmune hydrops is guarded, it is heavily dependent on etiology. In large series from Thailand and Southern China, α4-thalassemia is the predominant cause of nonimmune hydrops, accounting for 30 to 50 percent of cases and conferring an extremely poor prognosis (Liao, 2007; Ratanasiri, 2009; Suwanrath-Kengpol, 2005). In contrast, treatable etiologies such as parvovirus, chylothorax, and tachyarrhythmias, which each comprise about 10 percent of cases, can result in survival in two thirds of cases with fetal therapy (Sohan, 2001).

Diagnostic Evaluation

Hydrops is readily detected sonographically. As noted, two effusions or one effusion plus anasarca are required for diagnosis. Edema may be particularly prominent around the scalp, or equally obvious around the trunk and extremities. Effusions are visible as fluid outlining the lungs, heart, or abdominal viscera (Fig. 15-6).

In many cases, targeted sonographic and laboratory evaluation will identify the underlying cause of fetal hydrops. These include cases due to fetal anemia, arrhythmia, structural abnormality, aneuploidy, placental abnormality, or complications of monochorionic twinning. Depending on the circumstances, initial evaluation includes the following:

1. Indirect Coombs test for alloimmunization

2. Targeted sonographic fetal and placental examination, including:

   • A detailed anatomical survey to assess for the structural abnormalities listed in Table 15-4

   • MCA Doppler peak systolic velocity to assess for fetal anemia

   • Fetal echocardiography with M-mode evaluation

3. Amniocentesis for fetal karyotype and for parvovirus B19, cytomegalovirus, and toxoplasmosis testing as discussed in Chapter 64. Consideration of chromosomal microarray analysis if fetal anomalies are present

4. Kleihauer-Betke test for fetomaternal hemorrhage if anemia is suspected, depending on findings and test results

5. Consideration of testing for alpha-thalassemia and/or inborn errors of metabolism.

Isolated Effusion or Edema

Although one effusion or anasarca alone is not diagnostic for hydrops, the above evaluation should be considered if these are encountered, as hydrops may develop. For example, an isolated pericardial effusion may be the initial finding in fetal parvovirus B19 infection (Chap. 64, West Nile Virus). An isolated pleural effusion may represent a chylothorax, which is amenable to prenatal diagnosis, and for which fetal therapy may be lifesaving if hydrops develops (Chap. 16, Percutaneous Procedures). Isolated ascites also may be the initial finding in fetal parvovirus B19 infection, or it may result from a gastrointestinal abnormality such as meconium peritonitis. Finally, isolated edema, particularly involving the upper torso or the dorsum of the hands and feet, may be found in Turner or Noonan syndrome or may represent congenital lymphedema syndrome (Chap. 13, Polyploidy).

Mirror Syndrome

An association between fetal hydrops and development of maternal edema in which the fetus mirrors the mother is attributed to Ballantyne. He called the condition triple edema because the fetus, mother, and placenta all became edematous. The etiology of the hydrops is not related to development of mirror syndrome. It has been associated with hydrops from D alloimmunization, twin-twin transfusion syndrome, placental chorioangioma, and with fetal cystic hygroma, Ebstein anomaly, sacrococcygeal teratoma, chylothorax, bladder outlet obstruction, supraventricular tachycardia, vein of Galen aneurysm, and various congenital infections (Braun, 2010).

In a review of more than 50 cases of mirror syndrome, Braun (2010) found that approximately 90 percent of women had edema, 60 percent had hypertension, 40 percent had proteinuria, 20 percent had liver enzyme elevation, and nearly 15 percent had headache and visual disturbances. Based on these findings, it is reasonable to consider mirror syndrome a form of severe preeclampsia (Espinoza, 2006; Midgley, 2000). Others, however, have suggested that it is a separate disease process with hemodilution rather than hemoconcentration (Carbillon, 1997; Livingston, 2007).

Some reports describe the same imbalance of angiogenic and antiangiogenic factors that is observed with preeclampsia, and this suggests a common pathophysiology (Espinoza, 2006; Goa, 2013; Llurba, 2012). These findings, which include elevated concentrations of soluble fms-like tyrosine kinase-1 (sFlt-1), decreased placental growth factor (PlGF) levels, and elevation of soluble vascular endothelial growth factor receptor-1 (sVEGFR-1) concentrations, are discussed further Chapter 40 (Endothelial Cell Injury).

In most cases with mirror syndrome, prompt delivery is indicated and followed by resolution of maternal edema and other findings (Braun, 2010). However, in isolated cases of fetal anemia, supraventricular tachycardia, hydrothorax, and bladder outlet obstruction, successful fetal treatment resulted in resolution of both fetal hydrops and maternal mirror syndrome (Goa, 2013; Livingston, 2007; Llurba, 2012; Midgley, 2000). Normalization of the angiogenic imbalance has also been described following fetal transfusion for parvovirus B19 infection. Fetal therapy for these conditions is reviewed in Chapter 16. Given the parallels to severe preeclampsia, delaying delivery to effect fetal therapy should be considered only with caution. If the maternal condition deteriorates, delivery is recommended.