A 6-week-old girl was admitted for worsening generalized edema of unknown etiology. She was the 2,880-g product of a 32-week gestation to a 32-year-old G1P1 woman. Pregnancy was complicated by ultrasound discovery of ascites at 31 weeks that had not been present at 21 weeks. TORCH (toxoplasmosis, other infections, rubella, cytomegalovirus, and herpes simplex) virus and parvovirus titers of the mother were negative. On the day of delivery, scalp edema was noted by ultrasound and delivery was induced vaginally. Apgar scores were 8 at 1 minute and 9 at 5 minutes.
Physical examination at birth included appropriate-for-gestational-age growth parameters, no dysmorphic features, normal cardiac exam, three-vessel umbilical cord, ascites, and facial and scalp edema.
Pertinent laboratory tests in the newborn nursery included: hemoglobin 14.3 g/dL with normal white blood cell differential and platelet count, hemoglobin 9.6 g/dL on day 3 of life with 6% reticulocytes and 9 nucleated red blood cells/high-powered field, peak bilirubin (total/direct) of 9.8/0.3 mg/dL on day 3 of life, albumen 2.4 mg/dL, normal serum transaminases, negative cytomegalovirus DNA by polymerase chain reaction, normal karyotype and whole genome microarray, small patent ductus arteriosus on echocardiogram, normal newborn metabolic screen, and normal brain MRI. Abdominal ultrasound revealed a large amount of ascites without masses or organomegaly.
She remained in the outside neonatal intensive care unit (NICU) for 4 weeks for treatment of edema. She was given Epogen (Amgen, Inc.) to treat the anemia. At discharge, she had minimal abdominal distention and mild peripheral edema. Three days prior to this admission at 6 weeks of age, her parents noticed that the hand and foot swelling recurred. Her eyelids had always remained puffy. They had been measuring abdominal circumferences at home, which waxed and waned depending upon stool output. Her father felt that she had been “breathing heavier” for the past few days.
Family history was remarkable in that both parents are of Albanian descent but not consanguineous. She had been taking a combination of breast milk by breast and cow’s milk formula by bottle. Her development at 6 weeks was normal.
On physical exam at 6 weeks of age, her weight and height were in the 55th percentile and her head circumference was in the 25th percentile. She was not dysmorphic. Lungs were clear. Cardiac exam was normal. Abdomen was soft but distended. Liver was palpable 2 cm below the right costal margin. She had 1+ pitting edema to just above the knees. Neurologic exam was normal.
The following tests were normal during this admission: Complete blood count, urinalysis, serum chemistries and liver functions tests, and thyroid function tests. Serum albumen was 3.1 mg/dL. Abdominal ultrasound revealed a mildly thickened, echogenic gall bladder of questionable significance.
Robert Listernick, MD, moderator: I’d like to use correct terminology. This baby had fetal hydrops?
Kathryn Farrow, MD, PhD, neonatologist: Fetal hydrops is defined as abnormal fluid accumulation in two or more body compartments, such as the abdomen, pleural or pericardial spaces, and the skin. This child fits the definition having had ascites and scalp edema.
Dr. Listernick: Is there in utero treatment for hydrops?
Dr. Farrow: That depends upon the etiology. Immune hydrops leading to severe anemia can be treated with transfusions, whereas hydrops secondary to arrhythmias may respond to maternal treatment with digoxin. Premature delivery may be an important step depending on the biophysical profile of the fetus; it’s not clear why this infant was delivered at 31 weeks gestation. I suspect you’ll be getting into more specific etiologies in a bit.
Dr. Listernick: I will indeed. Before we do, I noticed that the physicians performed both a karyotype as well as a whole genome microarray. Were both tests necessary?
Joel Charrow, MD, pediatric geneticist: Almost certainly not. The microarray is likely to be more informative and will detect any deletion or duplication. The microarray will not detect a balanced translocation, which is not a concern in this clinical situation. If you are looking for aneuploidy, whole genome microarray is the best test.
Dr. Listernick: Let’s begin to parse out the various causes of fetal hydrops. We can start with the world of immune-mediated hydrops.
Robert Liem, MD, pediatric hematologist: Since the 1960s and the introduction of RhoGAM (Kedrion Biopharma, Inc.), the incidence of Rh isoimmunization and hemolytic disease of the newborn has plummeted. However, there are still cases of severe ABO isoimmunization that occur. In addition, Rh isoimmunization may occur, even in first pregnancies, due to previous sensitization during prior maternal transfusions, unrecognized miscarriages, or exposure to red cell antigens in Rh-negative drug abusers. We also have seen cases of isoimmunization with the other Rh antigens, C or E, despite the use of RhoGAM, which is specific for the D antigen.
Dr. Farrow: In the past, the standard way we determined if the fetus was in need of transfusion due to severe anemia secondary to isoimmunization was by measurement of amniotic fluid bilirubin via repeated amniocenteses. We can now predict the need for transfusion by noninvasively measuring fetal middle cerebral artery peak systolic velocity using Doppler imaging.
Dr. Liem: There are other non-immune hematologic causes of fetal hydrops. These include enzyme defects such as G6PD (glucose-6-phosphate dehydrogenase) deficiency or pyruvate kinase deficiency or alpha thalassemia due to four gene deletions in the two pairs of alpha globin genes, leading to the formation of Bart’s hemoglobin (nonfunctional tetramers of unpaired gamma hemoglobin chains). Maternal infection with parvovirus B19 can lead to severe hypoproliferative fetal anemia and hydrops. Finally, in utero bleeding that results in severe anemia may lead to fetal hydrops. This may occur from maternal placental abruption, maternal-fetal or twin-twin transfusions, or even intracranial hemorrhage. This infant was definitely anemic at birth, but not to the degree where we would usually see fetal hydrops.
Dr. Listernick: Let’s discuss the cardiac causes of fetal hydrops next.
Angira Patel, MD, pediatric cardiologist: Congestive heart failure leads to fetal hydrops. Structural lesions such as atrioventricular valve incompetence, Ebstein anomaly, or hypoplastic left heart may lead to hydrops. In utero myocarditis may lead to significantly decreased cardiac function and heart failure. Finally, one should consider the presence of arrhythmias, either supraventricular tachycardia or congenital heart block secondary to such maternal autoimmune conditions as systemic lupus erythematosus or Sjögren’s syndrome.
Dr. Listernick: How soon can heart failure from these congenital anomalies be detected?
Luciana Young, MD, pediatric cardiologist: Fetal echocardiograms are typically requested around 18 to 20 weeks of gestation, at which point the diagnosis of a structural heart lesion can be made by abdominal ultrasound. Arrhythmias may be intermittent, so they may be missed in an individual patient at any given point in time. However, typically if heart failure is going to develop, the arrhythmia is present more than 50% of the time in utero.
Dr. Listernick: Treatment?
Dr. Young: Structural abnormalities are difficult to treat in utero. Fetal interventions typically include balloon valvuloplasty for critical pulmonary or aortic stenosis or relief of a restrictive atrial septum in hypoplastic left heart syndrome. For fetal supraventricular tachycardia, maternal administration of digoxin is often the first line of treatment. However, it may be difficult for a hydropic fetus to absorb the digoxin transplacentally. In this case, the digoxin may need to be administered intramuscularly to the fetus. Additional agents may include flecainide or amiodarone. If the supraventricular tachycardia is difficult to treat, dual antiarrhythmic therapy may be indicated. Maternal sotalol may be considered for treatment of atrial flutter. Congenital heart block due to maternal autoimmune antibodies may be treated with corticosteroids, although recent studies question their efficacy. Maternal theophylline has also been administered in some cases to increase the underlying ventricular rate; however, the atria and ventricles will remain dysynchronous.
Dr. Listernick: Finishing up this great triumvirate of causes of fetal hydrops would be congenital infections.
Ben Z. Katz, MD, pediatric infectious disease physician: First, let me say that most infectious disease physicians have a pet peeve about the use of the term “TORCH infections.” The term connotes a mish-mash of different infections that really don’t have much to do with one another. For instance, congenital rubella is a congenital malformation syndrome that develops when an infant is exposed to rubella virus in the first trimester, whereas congenital herpes infection is a true infection due to peripartum acquisition of herpes virus. That being said, in utero parvovirus infection would be the classic infectious cause of fetal hydrops due to the development of severe anemia. In addition, myocarditis due to a number of in utero viral infections can also lead to hydrops, as was said earlier.
Dr. Listernick: What about the concept of ordering TORCH titers?
Dr. Katz: Ridiculous. I try to teach clinicians to test for what is clinically suspected rather than ordering a meaningless panel. In addition, the test that’s ordered depends on the agent being considered. A positive urine culture for cytomegalovirus in the first 3 weeks is diagnostic of a congenital infection; after 3 weeks, a positive urine culture might be the result of post-natal infection. The diagnosis of rubella is always made on a serologic basis.
Dr. Farrow: The strange thing about this infant is that her edema is waxing and waning. In my experience, these infants have a fairly prolonged hospital course, but by the time they go home the edema has resolved.
Dr. Listernick: From looking at the notes, there seemed to be a lot of concern about the possibility of liver disease when the infant was readmitted.
Estella Alonso, MD, pediatric hepatologist: A perfect example of in utero liver disease is gestational alloimmune liver disease, previously known as neonatal hemochromatosis. These babies often are born with cirrhosis, and usually have small, contracted livers that may be associated with ascites. Hydrops is not a common manifestation of neonatal liver disease unless there is associated cardiomyopathy or multiorgan failure. With that said, her admission serum albumin of 3.1 mg/dL should not be associated with significant edema.
Dr. Listernick: So when this child was admitted here, all the consultants basically said “not my disease!” Dr. Charrow, in the genetics department, was more interested in following her and keeping an open mind.
Dr. Charrow: It took us a while to get all the medical records so as to understand what had or had not been done in the way of testing. In the back of my mind the whole time was the possibility of a lysosomal storage disease, which has been known to cause non-immune hydrops. As I followed her over two subsequent outpatient visits, the edema waxed and waned but she definitely began to develop hepatomegaly without splenomegaly, and I felt her facial features were becoming a little coarse.
Dr. Listernick: Can you define coarse?
Dr. Charrow: It’s difficult. In general, it’s the absence of sharply defined facial features. The skin or the subcutaneous tissue becomes thickened around the eyebrows, lips, nose, and mouth. The supraorbital ridges become prominent. It’s very much a “gestalt.” Often, it is not recognizable until 6 months to 1 year of age, but I have seen it as early as 6 weeks in an infant with mucolipidosis II.
Dr. Listernick: How did you proceed with testing?
Dr. Charrow: We started testing for activity of the various lysosomal enzymes. Normally, I prefer to do targeted testing based on signs and symptoms. However, I wasn’t quite sure how to be surgically precise in my testing in this instance, so we performed a test I don’t often do — a lysosomal enzyme panel that tests levels of the major enzymes. She was found to be deficient in beta-galactosidase, leading to a diagnosis of GM1 gangliosidosis.
Dr. Listernick: What are the genetics of GM1 gangliosidosis?
Dr. Charrow: It is inherited in an autosomal recessive fashion, but it has a very interesting phenotypic spectrum. Some mutations produce a skeletal dysplasia with normal intelligence and no neurologic disease (Morquio syndrome). In contrast, infantile GM1 gangliosidosis is a progressive neurologic disease that leads to developmental regression, seizures, corneal clouding, retinal cherry red spots, cardiomyopathy, and death in the first several years of life. There are milder forms of the disease that may present throughout childhood, even in the adult years. The severity of GM1 gangliosidosis is related to the level of beta-galactosidase activity. This child essentially has no enzyme activity, indicating the severe infantile phenotype.
Dr. Listernick: Treatment?
Dr. Charrow: There is no treatment for GM1 gangliosidosis. Stem cell transplantation has been ineffective, and there is no enzyme replacement therapy available.
Dr. Listernick: Tragic. Thank you, everyone.