Prognostic Factors and Clinical Features in Liveborn Neonates With Hydrops Fetalis
Prognostic Factors and Clinical Features in Liveborn Neonates With Hydrops Fetalis
ABSTRACT
The purpose of this study was to delineate the etiology and the clinical features of
liveborn neonates with hydrops fetalis, and to explore the prognostic factors for survival.
Medical records of 28 liveborn neonates with hydrops fetalis between April 1995 and
March 2005 were reviewed retrospectively. Demographic data, clinical manifestations,
laboratory findings, and outcomes were analyzed. Most patients presented with pleural
effusions (21 of 28) and ascites (22 of 28). The majority of patients had hydrops due to
cardiovascular diseases (seven of 28), hematologic disorders (six of 28), lymphatic
malformations (six of 28), and idiopathic origins (six of 28). The overall survival rate
was 50% and was highest (83%) in infants with lymphatic malformations. By univariate
analysis, risk factors for mortality are earlier ages at diagnosis and at birth, low Apgar
scores, need for resuscitation in the delivery room, low serum albumin level, and severe
acidemia. After using stepwise multiple logistic regression analysis, the most significant
factors associated with fatality were younger gestational age at birth and lower serum
albumin level. Hydrops fetalis remains a complex condition with a high mortality rate.
Hydrops resulting from lymphatic malformations has a favorable outcome. Preterm birth at
less than 34 weeks and serum albumin concentration lower than 2 g/dL are two poor
prognostic factors for survival.
Hydrops fetalis (HF) is an abnormal accumu- liveborn neonates.3 Most patients with HF diagnosed
lation of fluid in a fetus due to diverse etiologies. The in early fetal life have chromosomal anomalies, whereas
infants frequently present severely ill and the mortality those occurrences diagnosed after the second trimester are
rate is very high. HF is usually classified into two caused mainly by cardiovascular diseases.4–6 The prog-
categories: immune HF (IHF) and nonimmune HF nosis is dependent partly on the underlying disease, but
(NIHF). The former was caused mainly by Rh incom- with aggressive postnatal care, the survival rate is
patibility, but the occurrence has steadily declined because increased in selected cases. The purpose of this study
of maternal treatment with anti-D globulin. At present, is to identify the prognostic factors for liveborn neonates
NIHF appears more frequently, and it accounts for 90% with HF. In addition to underlying diseases, we also focus
of HF.1 The survival rate for fetuses with HF ranges from on the clinical features and biochemical data of these
12% to 24%,2 but is much higher (40% to 50%) in newborns. By doing so, we may provide additional
1
Division of Neonatology, Department of Pediatrics, Chang Gung Medicine 5, Fushing Street, Gueishan Shiang, Taoyuan, Taiwan 333,
Memorial Hospital, Chang Gung University College of Medicine; Republic of China.
2
Department of Public Health and Center of Biostatistics, Chang Am J Perinatol 2007;24:33–38. Copyright # 2007 by Thieme
Gung University College of Medicine, Taoyuan, Taiwan. Medical Publishers, Inc., 333 Seventh Avenue, New York, NY 10001,
Address for correspondence and reprint requests: Ming-Chou USA. Tel: +1(212) 584-4662.
Chiang, M.D., Division of Neonatology, Department of Pediatrics Accepted: August 11, 2006. Published online: December 27, 2006.
Chang Gung Memorial Hospital, Chang Gung University College of DOI 10.1055/s-2006-958158. ISSN 0735-1631.
33
34 AMERICAN JOURNAL OF PERINATOLOGY/VOLUME 24, NUMBER 1 2007
information for postnatal care to improve the survival rate massive (the entire lung with mediastinal shift).9 Ma-
and to identify significant prognostic factors to better ternal records were reviewed to determine the prenatal
counsel the parents. diagnosis and interventions.
Statistical analysis was performed using SPSS for
Windows, version 10.0 (SPSS, Inc., Chicago, IL). Dis-
MATERIALS AND METHODS crete variables were analyzed by x2 test. Continuous
Medical records of liveborn neonates with HF admitted variables were analyzed by Student’s t-test. Data are
to the neonatal intensive care unit at Chang Gung expressed as mean standard deviation. Statistical sig-
Children’s Hospital were reviewed retrospectively for nificance was defined as p < 0.05. After univariate anal-
the period between April 1995 and March 2005. HF ysis to screen for statistically significant variables, a
was defined as the presence of general anasarca (sub- stepwise multiple logistic regression analysis was per-
cutaneous thickening > 5 mm) revealed on real-time formed to adjust confounding variables simultaneously.
fetal ultrasonography plus at least two of the following
conditions: pleural effusion, ascites, pericardial effusion,
placental edema, or polyhydramnios. NIHF was defined RESULTS
as infants with HF without evidence of isoimmuniza-
tion. In our patient population, we divided the diagnosis Demographic Data
into the following categories according to the classifica- Twenty-eight liveborn neonates were eligible for our
tion by Forouzan7: hematologic disorders, cardiovas- study with a sex distribution of 11 boys and 17 girls.
cular diseases, lymphatic malformations, chromosomal During 1995 to 2005, there were 18,976 newborns
anomalies, thoracic abnormalities, and idiopathic ori- admitted to the neonatal intensive care unit in this
gins. Hematologic disorders were evaluated by complete tertiary-center population, giving an approximate prev-
blood count, blood smear, Coombs test, blood typing, alence of HF of 1.5 per 1000 admissions. Fourteen
hemoglobin electrophoresis, and Kleihauer-Betke stain. patients survived, yielding an overall survival rate of
Cardiovascular diseases and thoracic abnormalities were 50%. Among the 28 neonates, the maternal age was
investigated by radiography and ultrasonography. In 28.9 6.0 years. GA at diagnosis and at birth was
addition to a predominance of lymphocytes presented 32.5 3.7 and 33.9 3.3 weeks, respectively, and the
in visceral effusions, lymphatic malformations were es- BW was 2600 844 g. Apgar scores were 3 2 at 1
tablished by computed tomography (CT) or magnetic minute and 5 2 at 5 minutes, respectively. Seventy-five
resonance imaging (MRI).8 Karyotyping was offered to percent of patients (21 of 28) were delivered via cesarean
identify the chromosomal anomalies. Furthermore, section.
screening for congenital infection (including toxoplas-
mosis, rubella, cytomegalovirus, herpes simplex virus,
syphilis, and parvovirus B19) was performed in some Clinical Features
patients who lacked a definitive diagnosis. If there were Twenty-one of 28 neonates (75%) presented with pleural
no results of the prenatal and postnatal workup that effusions, 22 newborns (79%) presented with ascites, two
could explain the cause of hydrops, it was regarded as newborns (7%) presented with pericardial effusions, and
having an idiopathic origin. seven newborns (25%) presented with cardiomegaly. A
Demographic data, such as gender, maternal age, total of 16 neonates (57%) had two or more serous cavity
gestational ages (GA) at diagnosis and at birth, mode of effusions. Seven patients (25%) received intrauterine
delivery, birthweight (BW), and Apgar scores were therapy. Fetal pleurocentesis was performed in three
recorded. In addition, the following information was cases, a pleuroamniotic shunt was placed in three cases,
collected: numbers and locations of body cavities with and digoxin was prescribed for a fetus with supraven-
fluid accumulation, whether the excess fluid was tapped tricular tachycardia. Among those with prenatal therapy,
in the delivery room (DR), need for resuscitation in the three patients (43%), including the one with fetal ar-
DR, administration of inotropic agents, ventilator mode rhythmia, survived. Forty-three percent of the patients
and settings if used, vital signs, and underlying etiology. required resuscitation in the DR after birth. Ninety-
Hemogram, biochemical data, and blood gas analysis three percent of all patients underwent ventilator support
from the first blood sampling were obtained. The com- for respiratory distress and 79% received inotropic agents
position of effusions, hematologic and immunologic for unstable vital signs.
studies, bacterial and viral cultures, and chromosomal The clinical features are summarized in Table 1.
results were collated. The volume of pleural effusions Laboratory findings disclosed hypoalbuminemia
was identified according to the appearance of effusions (2.1 0.6 g/dL) and acidemia (pH, 7.005 0.25). Pro-
on a chest radiograph: small (not more than one fourth found anemia was noted in two patients with isoimmune
of the lung field), moderate (one fourth to one half of the hemolytic anemia in which hemoglobin (Hb) 4.1 g/dL
lung field), large (more than half of the lung field), or was found in one with Rh type (E, c) incompatibility,
LIVEBORN NEONATES WITH HYDROPS FETALIS/HUANG ET AL 35
Table 1 Summary of the Clinical Features of Liveborn Table 2 Etiology and Outcome of Liveborn Neonates
Neonates with Hydrops Fetalis with Hydrops Fetalis
Surviving Fatal Cases Total Patients
Clinical Features Cases (n ¼ 14) (n ¼ 14) (n ¼ 28)
Survived Died Total
Intrauterine fetal 7 (25%) Etiology and Outcome (n ¼ 14) (n ¼ 14) (n ¼ 28)
therapy Idiopathic origins 3 3 6
Pleurocentesis 1 2 Hematologic disorders 6
Pleuroamniotic 1 2 Rh type (E, c) 1
shunt
incompatibility
Digoxin 1
Anti-M alloantibodies 1
Resuscitation in 2 10 12 (43%) Homozygous 1
the DR a-thalassemia
Mechanical 12 14 26 (93%) Fetomaternal 2
ventilation hemorrhage
IMV 9 9 Twin–twin transfusion 1
HFOV 3 5 syndrome
Inotropic agents 9 13 22 (79%)
Cardiovascular diseases 7
used
Structural anomalies
Pleural effusions 11 10 21 (75%) Ebstein anomaly 1 1
Small 4 1 Atrioventricular 1 2
Moderate 5 3 septal defect
Large 2 3 Arrhythmias, PSVT 1
Massive 3 Hemangioma of 1
Ascites 10 12 22 (79%) iliac artery
2 serous cavity 8 8 16 (57%)
Lymphatic malformations 5 1 6
effusions
Chromosomal anomalies 2
HFOV, high-frequency oscillatory ventilator; IMV, intermittent Trisomy 21 1
mandatory ventilation.
Chromosome 1
22p deletion
and Hb 1.8 g/dL was found in one with anti-M alloanti-
Thoracic abnormalities, 1 1
bodies. Three infants presented with nonimmune ane-
CCAM
mia: one due to thalassemia (Hb, 6.3 g/dL), and two
infants with fetomaternal hemorrhage (Hb, 3.3 and 7.7 PSVT, paroxysmal supraventricular tachycardia; CCAM, congenital
cystic adenomatoid malformation.
g/dL, respectively). White blood cell counts, platelet
counts, and C-reactive protein levels were within normal
limits. from those with hematologic disorders, there would be
no surviving infants with NIHF secondary to hemato-
logic disorders. Eleven of the 14 infants who died in the
Etiology and Outcome neonatal period did so within 2 weeks.
There were two patients with IHF and 26 patients with
NIHF. The majority of patients had hydrops due to
cardiovascular diseases (seven of 28), hematologic dis- Prognostic Factors
orders (six of 28), lymphatic malformations (six of 28), By univariate analysis, the risk factors for mortality are
and idiopathic origins (six of 28). Table 2 summarizes younger GAs at diagnosis (30.9 4.0 versus 33.9 3.1
the etiology and outcome of liveborn neonates with HF. weeks; p ¼ 0.036) and at birth (32.0 3.4 versus
Infants with lymphatic malformations had a favorable 35.9 1.7 weeks; p ¼ 0.001), low Apgar scores (2 2
survival rate (83%) in this study. The survival rate was versus 4 2 at 1 minute; p ¼ 0.015; and 4 2 versus
33% in infants with hematologic disorders because only 6 2 at 5 minutes; p ¼ 0.030), need for resuscitation in
two infants with isoimmune hemolytic anemia survived. the DR (p ¼ 0.006), low serum albumin level (1.7 0.3
No fetal transfusion was provided but exchange trans- versus 2.4 0.6 g/dL; p ¼ 0.001), and severe acidemia
fusion therapy was done in the patient with Rh type (6.862 0.21 versus 7.136 0.20; p ¼ 0.001; Table 3).
(E, c) incompatibility. The remaining four infants, in- Using stepwise multiple logistic regression analysis, the
cluding one with homozygous a-thalassemia, two with most significant factors associated with fatality are
fetomaternal hemorrhage, and one with twin–twin younger GA at birth and lower serum albumin level.
transfusion syndrome, all died as a result of severe By using receiver operating characteristics analysis, the
underlying disease. If infants with IHF were separated area under the curve is 0.964. The cutoff points of GA at
36 AMERICAN JOURNAL OF PERINATOLOGY/VOLUME 24, NUMBER 1 2007
Table 3 Comparison of Clinical Characteristics between Surviving and Fatal Cases in Liveborn Neonates with
Hydrops Fetalis
Characteristic Surviving Cases (n ¼ 14) Fatal Cases (n ¼ 14) p
birth and of serum albumin concentration for survival are clinicians should be aware that hemangiomas and arte-
34 weeks (sensitivity, 1.00; specificity, 0.71) and 2.0 g/ riovenous malformations can be a rare cause of HF.13
dL (sensitivity, 0.86; specificity, 0.64), respectively. Two patients had chromosomal anomalies. In our
series, all hydropic patients were diagnosed during the
second or third trimesters; we speculate that the true
DISCUSSION incidence of abnormal fetal karyotypes associated with
NIHF accounted for the majority of HF. Cardiovascular HF is probably higher, because if a chromosomal anom-
diseases are one of the main causes for NIHF. Seventy- aly is confirmed in early pregnancy, the parents select
one percent of infants with cardiovascular diseases had termination or there is an increased tendency for sponta-
structural anomalies including Ebstein anomaly and neous abortion or delivery as a stillbirth. In our clinical
atrioventricular septal defects (AVSDs). Although observation, there was no definitive case resulting from
many cardiac malformations have been reported, the an infectious disease. Rodrı́guez et al14,15 reported that
most common were AVSDs.10–12 Intrauterine conges- cardiovascular diseases are more common in the live-
tive heart failure with high hydrostatic pressure secon- born, whereas congenital infections are the most com-
dary to inadequate cardiac output with right atrial mon causes for stillborn fetuses with NIHF. Our results
overload leads to fetal hydrops.4–6 Prenatal intervention concur with these authors for liveborn series.
has not been offered to fetuses with structural anomalies In our study, 7% of HF was IHF, which is similar
in our institution to date, but fetal arrhythmia can be to a previous report by Heinonen et al.2 In contrast with
treated medically. In one of our patients, transplacental other observations, minor group incompatibility, instead
digitalization successfully converted paroxysmal supra- of Rh incompatibility, is the major cause of IHF. Wu
ventricular tachycardia to normal sinus rhythm. Large et al16 reported that anti-E and anti-E þ c antibodies
series have shown improved survival rates following play an important role in maternal irregular antibodies to
appropriate maternal drug administration in hydropic result in neonatal hemolytic diseases in Taiwan. How-
fetuses with arrhythmia.4,5 One hydropic fetus was ever, anti-M alloantibody is seldom the cause of neonatal
diagnosed with a bulky hemangioma over the external hemolytic disease.17 Exchange transfusion therapy and
and internal iliac arteries after birth. We speculate that intensive neonatal care can provide a better outcome in
the hemangioma functions as a huge arteriovenous this situation.
shunt, leading to cardiac overload and high-output HF caused by a lymphatic malformation present-
cardiac failure. It is difficult to diagnose prenatally, and ing as a chylothorax had favorable prognosis, which is in
LIVEBORN NEONATES WITH HYDROPS FETALIS/HUANG ET AL 37
agreement with other investigators.5,18,19 In addition to function, and could support the cardiovascular circula-
CT and MRI, lymphatic malformations can be also tion in hydropic infants with hypotension.
confirmed by lymphangiography, lymphoscintigraphy, In conclusion, hydrops fetalis remains a complex
and lung biopsy. These diagnostic methods are accurate condition with high mortality and morbidity. The out-
to distinguish the congenital errors of lymphatic devel- come of HF is dependant mainly on the underlying
opment (such as lymphangiomas, lymphangiectasis, or etiology, GA at birth, and serum albumin level. Our
lymphatic dysplasia syndrome).20 We adopted conserva- observation suggests that hydrops resulting from lym-
tive management in most patients, including pleural phatic malformations has a favorable outcome. Preterm
drainage and a high-protein, medium-chain triglyceride birth at < 34 weeks and a serum albumin concen-
diet, and symptoms resolved spontaneously in surviving tration < 2 g/dL are two poor prognostic factors for
cases. In addition, prenatal interventions by either in- survival.
termittent thoracocentesis or an in-dwelling pleuro-
amniotic shunt are warranted in severe progressive
cases to minimize the risk of pulmonary hypoplasia REFERENCES
resulting from intrathoracic compression.21
Various poor prognostic factors for mortality have 1. Santolaya J, Alley D, Jaffe R, Warsof SL. Antenatal
classification of hydrops fetalis. Obstet Gynecol 1992;79:
been described, including the presence of a pleural
256–259
effusion,22 two serous cavity effusions, a need for chest 2. Heinonen S, Ryynanen M, Kirkinen P. Etiology and
compressions at birth,6 and lower Apgar scores at 1 and outcome of second trimester non-immunologic fetal hydrops.
5 minutes.23 By univariate analysis, the fatal cases had a Acta Obstet Gynecol Scand 2000;79:15–18
tendency to lower Apgar score, need for resuscitation in 3. Wy CA, Sajous CH, Loberiza F, Weiss MG. Outcome of
the DR, and more severe acidemia than the surviving infants with a diagnosis of hydrops fetalis in the 1990s. Am J
group. These parameters suggesting a poor condition at Perinatol 1999;16:561–567
4. McCoy MC, Katz VL, Gould N, Kuller JA. Non-immune
birth reflect the severity of hydrops, and strongly influ-
hydrops after 20 weeks’ gestation: review of 10 years’
ence outcome. Moreover, using stepwise multiple logis- experience with suggestions for management. Obstet Gyne-
tic regression analysis, we conclude that the most col 1995;85:578–582
significant factors for fatality are GA at birth younger 5. Sohan K, Carroll SG, De La Fuente S, Soothill P, Kyle P.
than 34 weeks and a serum albumin level < 2 g/dL. A Analysis of outcome in hydrops fetalis in relation to
hydropic fetus who is delivered prematurely is subject to gestational age at diagnosis, cause and treatment. Acta
the additional complications of prematurity. Poor re- Obstet Gynecol Scand 2001;80:726–730
6. Wafelman LS, Pollock BH, Kreutzer J, Richards DS,
sponse to neonatal resuscitation in premature infants
Hutchison AA. Nonimmune hydrops fetalis: fetal and
with hydrops is attributed to the decreased lung com- neonatal outcome during 1983–1992. Biol Neonate 1999;
pliance (secondary either to hyaline membrane disease or 75:73–81
extrapulmonary compression), cardiovascular compro- 7. Forouzan I. Hydrops fetalis: recent advances. Obstet Gynecol
mise, and even intracranial hemorrhage.24 Prenatal in- Surv 1997;52:130–138
tervention to prolong pregnancy in selected patients 8. Barker PM, Esther CR Jr, Fordham LA, Maygarden SJ,
might improve the survival rate. Funkhouser WK. Primary pulmonary lymphangiectasia
in infancy and childhood. Eur Respir J 2004;24:413–
Phibbs et al25 have demonstrated a close correla-
419
tion between serum albumin concentration and severity 9. Rocha G, Fernandes P, Rocha P, Quintas C, Martins T,
of HF. They proposed that anemia or cardiac disorders Proenca E. Pleural effusions in the neonate. Acta Paediatr
lead to congestive heart failure with elevated central 2006;95:791–798
venous pressure, which results in increased capillary 10. Allan LD, Sharland GK, Chita SK, Lockhart S, Maxwell DJ.
filtration pressure and restriction of lymphatic return. Chromosomal anomalies in fetal congenital heart disease.
Heart failure leads to congestion of the liver, causing a Ultrasound Obstet Gynecol 1991;1:8–11
11. Iskaros J, Jauniaux E, Rodeck C. Outcome of nonimmune
decrease in albumin synthesis. In addition, hypoxia
hydrops fetalis diagnosed during the first half of pregnancy.
increases capillary permeability to albumin in asphyxi- Obstet Gynecol 1997;90:321–325
ated hydropic infants. Finally, hypoalbuminemia causes 12. Jauniaux E. Diagnosis and management of early non-immune
a significant decrease in plasma colloid osmotic pressure, hydrops fetalis. Prenat Diagn 1997;17:1261–1268
allowing edema formation and serous effusion.25 Except 13. Hartung J, Chaoui R, Kalache K, Tennstedt C, Bollmann R.
for albumin transfusion after birth, Shimokawa et al26 Prenatal diagnosis of intrahepatic communications of the
reported that intrauterine albumin infusion is effective in umbilical vein with atypical arteries (A-V fistulae) in two
cases of trisomy 21 using color Doppler ultrasound. Ultra-
treating idiopathic HF. In our observation, there is no
sound Obstet Gynecol 2000;16:271–274
evidence that transfusing albumin aggressively in the 14. Rodrı́guez MM, Chaves F, Romaguera RL, Ferrer PL, de la
postnatal period is effective. However, albumin trans- Guardia C, Bruce JH. Value of autopsy in nonimmune
fusion, in selective cases, might be beneficial in removing hydrops fetalis: series of 51 stillborn fetuses. Pediatr Dev
fluid from lungs and subsequently to improve pulmonary Pathol 2002;5:365–374
38 AMERICAN JOURNAL OF PERINATOLOGY/VOLUME 24, NUMBER 1 2007
15. Rodrı́guez MM, Bruce JH, Jimenez XF, et al. Nonimmune following antenatal diagnosis. Acta Paediatr 1995;84:749–
hydrops fetalis in the liveborn: series of 32 autopsies. Pediatr 755
Dev Pathol 2005;8:369–378 22. Nakayama H, Kukita J, Hikino S, Nakano H, Hara T. Long-
16. Wu KH, Chu SL, Chang JG, Shin MC, Peng CT. term outcome of 51 liveborn neonates with non-immune
Hemolytic disease of the newborn due to maternal irregular hydrops fetalis. Acta Paediatr 1999;88:24–28
antibodies in the Chinese population in Taiwan. Transfus 23. Simpson JH, McDevitt H, Young D, Cameron AD. Severity
Med 2003;13:311–314 of non-immune hydrops fetalis at birth continues to predict
17. De Young-Owens A, Kennedy M, Rose RL, Boyle J, survival despite advances in perinatal care. Fetal Diagn Ther
O’Shaughnessy R. Anti-M isoimmunization: management 2006;21:380–382
and outcome at the Ohio State University from 1969 to1995. 24. Phibbs RH, Johnson P, Kitterman JA, Gregory GA, Tooley
Obstet Gynecol 1997;90:962–966 WH. Cardiorespiratory status of erythroblastotic infants. I.
18. Klam S, Bigras JL, Hudon L. Predicting outcome in primary Relationship of gestational age, severity of hemolytic diseases,
fetal hydrothorax. Fetal Diagn Ther 2005;20:366–370 and birth asphyxia to idiopathic respiratory distress syndrome
19. Zenker M, Ries M. Hydrops fetalis caused by chylothorax: an and survival. Pediatrics 1972;49:5–14
exception to the rule. J Perinat Med 1997;25:388 25. Phibbs RH, Johnson P, Tooley WH. Cardiorespiratory status
20. Faul JL, Berry GJ, Colby TV, et al. Thoracic lymphangio- of erythroblastotic newborn infants. II. Blood volume,
mas, lymphangiectasis, lymphangiomatosis, and lymphatic hematocrit and serum albumin concentration in relation to
dysplasia syndrome. Am J Respir Crit Care Med 2000;161: hydrops fetalis. Pediatrics 1974;53:13–23
1037–1046 26. Shimokawa H, Hara K, Maeda H, Miyamoto S, Koyanagi T,
21. Mussat P, Dommergues M, Parat S, et al. Congenital Nakano H. Intrauterine treatment of idiopathic hydrops
chylothorax with hydrops: postnatal care and outcome fetalis. J Perinat Med 1988;16:133–138