j o u r n a l o f s u r g i c a l r e s e a r c h  j a n u a r y 2 0 1 8 ( 2 2 1 ) 1 2 1 e1 2 7

Available online at www.sciencedirect.com

ScienceDirect

journal homepage: www.JournalofSurgicalResearch.com

Association for Academic Surgery

Outcomes of fetuses with primary hydrothorax

that undergo prenatal intervention (prenatal
intervention for hydrothorax)

Rodrigo A. Mon, MD,a,b Marjorie C. Treadwell, MD,b,c

Deborah R. Berman, MD,b,c Lori Day, MD,b,c
Jeannie Kreutzman, RN, MSN, CPNP,a,b George B. Mychaliska, MD,a,b
and Erin E. Perrone, MDa,b,*
a
Section of Pediatric Surgery, Department of Surgery, University of Michigan Health System, Ann Arbor, Michigan
b
University of Michigan Health System, Fetal Diagnosis and Treatment Center, Ann Arbor, Michigan
c
Division of Maternal-Fetal Medicine, Department of Obstetrics and Gynecology, University of Michigan Health
System, Ann Arbor, Michigan

article info abstract

Article history: Background: Primary hydrothorax is a congenital anomaly affecting 1 in 10,000-15,000

Received 6 February 2017 pregnancies. The natural history of this condition is variable with some fetuses having
Received in revised form spontaneous resolution and others showing progression. The associated pulmonary hy-
8 August 2017 poplasia leads to increased perinatal morbidity and mortality. Optimal prenatal interven-
Accepted 16 August 2017 tion remains controversial.
Available online 18 September 2017 Methods: After obtaining the Institutional Review Board approval, a retrospective review of all
patients evaluated for a fetal pleural effusion in the Fetal Diagnosis and Treatment Center at The
Keywords: University of Michigan, between 2006 and 2016 was performed. Cases with secondary etiologies
Hydrothorax for an effusion or when families decided to pursue elective termination were excluded.
Hydrops Results: Pleural effusions were identified in 175 patients. Primary hydrothorax was diagnosed
Thoracentesis in 15 patients (8%). The effusions were bilateral in 13/15 cases (86%) and 10/15 (66%) had
Thoracoamniotic shunt hydrops at presentation. All 15 patients with primary hydrothorax underwent prenatal
intervention. Thoracentesis was performed in 14/15 cases (93%). Shunt placement was per-
formed in 10/15 cases (66%). Shunt migration was seen in four patients (40%) and all of these
underwent prenatal shunt replacement. Overall survival was 76%. The rates of prematurity
and preterm premature rupture of membranes were 69% and 35%, respectively.
Conclusions: Fetal intervention for the treatment of primary hydrothorax is effective, and it
appears to confer a survival advantage. Both the fetuses and the mothers tolerated the
procedures well. Preterm labor and preterm premature rupture of membranes remain an
unsolved problem. Further studies are needed to understand the mechanisms behind the
development of fetal hydrothorax.
ª 2017 Elsevier Inc. All rights reserved.

* Corresponding author. C.S. Mott Children’s Hospital Pediatric Surgery, 1540 E. Hospital Dr, Ann Arbor, MI 48109-4211. Tel.: þ1 (734) 936
8464; fax: þ1 (734) 232 8667.
E-mail address: [email protected] (E.E. Perrone).
0022-4804/$ e see front matter ª 2017 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.jss.2017.08.034
122 j o u r n a l o f s u r g i c a l r e s e a r c h  j a n u a r y 2 0 1 8 ( 2 2 1 ) 1 2 1 e1 2 7

Introduction Data collected included estimated gestational age (EGA) at

diagnosis, the presence or absence of hydrops, the type, side
Primary hydrothorax is a congenital anomaly affecting 1 in and number of prenatal interventions, the percentage of
10,000-15,000 pregnancies.1,2 Fetuses can present with either lymphocytes in the pleural fluid, complications associated
unilateral of bilateral effusions and males are affected more with intervention, the EGA at birth, and mortality. Preterm
frequently than females (2:1). This condition is also commonly birth was defined as delivery before 36 wk of gestation.
referred to as “primary chylothorax”. It has been argued that Postnatal records were reviewed and interventions were
because the normal mean percentage of lymphocytes in the documented. For some patients born at an outside hospital,
fetal blood is >80%, this parameter cannot be used prenatally we had access to electronic records; the rest had telephone or
to characterize an effusion as chyle. Therefore, the term email follow-up. Descriptive statistics were performed using
“hydrothorax” is preferred.3 Microsoft Excel 2016 (Microsoft Corporation, Redmond, WA).
The natural history of this condition is variable with some All prenatal interventions were performed by maternal-
fetuses having spontaneous resolution and others showing fetal medicine physicians. Thoracentesis and thor-
progression. A better understanding of this natural history is acoamniotic shunt placement were done under ultrasound
fundamental, as we try to determine prognosis and develop guidance, with the mother receiving intravenous sedation
criteria for fetal intervention. Perinatal mortality in fetuses (see Figs. 14). For shunt placement, a double pigtail silastic
with pleural effusions has been reported to be between 22% catheter (Rocket Medical, Hingham, MA) was used in all cases.
and 53%.2,4,5 The effusion can behave like a space-occupying The mothers were discharged home after the procedure and
lesion with resultant lung growth restriction. In the most se- follow-up ultrasound was performed a week later.
vere cases, fetuses may develop large effusions and progress
to hydrops. The associated pulmonary hypoplasia leads to
increased perinatal morbidity and mortality.6,7 The prognosis Results
and the effectiveness of prenatal intervention can also vary
widely in the presence of secondary causes of hydrothorax. Pleural effusions were identified in 175 patients. We excluded
These include chromosomal anomalies, infections, cardiac patients with cardiac or structural anomalies, chromosomal
malformations, and other structural anomalies.6,8 Therefore, anomalies, or infections. The specific associated anomalies
a comprehensive fetal and maternal evaluation is necessary are described in Table 1. In this study, 80% of the patients had
before prenatal intervention is considered. structural or chromosomal abnormalities. We also excluded
Since the introduction of prenatal therapy, fetuses with patients who died before birth with either elective pregnancy
large pleural effusions seem to have an improved outcome.9,10 termination (n ¼ 11) or had intrauterine fetal demise before
The majority of fetuses are treated with drainage or shunting. intervention (n ¼ 7).
Shunt placements are associated with shunt dysfunction, Primary hydrothorax was therefore diagnosed in 15/175
prematurity, or other complications. The incidence of these pleural effusion patients (8%). The effusions were bilateral in
complications in current practice and the outcomes associated 13/15 cases (86%). In this series 10 patients (66%) had hydrops
with treatment need to be better defined.11,12 There are re- at presentation.
ported cases of chemical pleurodesis being used as first-line All 15 patients with primary hydrothorax underwent pre-
therapy for this condition, but the role of this therapy is not natal intervention. A modified version of the algorithm pro-
clear.13,14 The aim of this study is to review our experience with posed by Yinon et al.15 is used at our institution to determine
fetuses diagnosed with primary hydrothorax, which under- the timing and type of prenatal intervention (Fig. 5).
went prenatal intervention and describe the clinical outcomes. Mean EGA at the time of the first intervention was 27w1d
(standard deviation [SD]  4.1). Thoracentesis was performed
in 14/15 patients (93%). One patient had primary shunt
Methods placement after the baby showed rapid progression from a
small effusion to a large effusion with mediastinal shift and
A retrospective review of all patients evaluated for a fetal hydrops. Fetal pleural fluid was analyzed in all 15 patients,
hydrothorax at the Fetal Diagnosis and Treatment Center at and the analysis revealed a lymphocyte fraction >80% in 11
The University of Michigan, between 2006 and 2016 was per- patients. The median lymphocyte fraction was 92%.
formed. Institutional Review Board approval was obtained Shunt placement was performed in 10/13 patients (76%).
(HUM00091706). Due to the retrospective nature of this The mean gestational age at the time of initial shunt
approved study, the Institutional Review Board granted a placement was 28w1d (SD  3.82) (see Table 2). Two patients
waiver for informed consent of patients included. Cases with were excluded from this part of the analysis. The first pa-
other possible etiologies for an effusion such as chromosomal tient moved out of state after the initial thoracentesis and
anomalies, infections, cardiac malformations, and other was lost to follow-up. The second patient had a thoracent-
structural anomalies were defined as secondary hydrothorax esis with subsequent reaccumulation of the effusion. Shunt
and were excluded. Cases where the family decided to pursue placement was recommended but the family declined any
elective termination of pregnancy were also excluded. further intervention. The fetus eventually had intrauterine
Hydrops was defined as the presence of fluid in at least one demise.
other compartment in addition to the thorax (peritoneum, When looking only at the patients that presented with
pericardium, or integument).4,6 hydrops, 7/10 (70%) underwent shunt placement. Three
 mon et al  prenatal intervention for hydrothorax 123

Fig. 1 e Fetal heart, fetal lungs, prenatal ultrasound of a normal fetus (four-chamber view). (Color version of figure is
available online.)

patients (30%) with hydrops at the time of initial evaluation placement was not felt appropriate. There was reac-
did not have a shunt placed. One was in labor at presentation cumulation of the hemothorax after initial drainage, and the
and delivered at 29w4d of gestation. The fetus survived for patient underwent thoracentesis again immediately before
89 d and died because of cardiac arrest. The second patient delivery to facilitate neonatal resuscitation. Delivery timing
had a hemothorax diagnosed at thoracentesis; shunt was dictated by the development of severe preeclampsia in

Fig. 2 e Insertion of the shunt introducer into the fetal chest, under ultrasound guidance. (Color version of figure is available
online.)
124 j o u r n a l o f s u r g i c a l r e s e a r c h  j a n u a r y 2 0 1 8 ( 2 2 1 ) 1 2 1 e1 2 7

Fig. 3 e Ultrasound image showing the shunt deployed across the fetal chest wall, with almost complete resolution of the
effusion. (Color version of figure is available online.)

the mother. The baby did well and was discharged on day of Another complication of prenatal intervention was shunt
life #20. The third patient’s family declined shunt placement migration. Migration of the shunt was seen in 4/10 patients
with eventual fetal demise. (40%). It was noticed in the first follow-up ultrasound in all
No immediate maternal or fetal complications from the four cases, an average of 8 d later (range 5-16 d). It was decided
prenatal procedures were identified. There was a high rate of that these patients would benefit from a replacement shunt.
prematurity with mean EGA at delivery of 34 wk (SD  4.2) All four patients underwent another intervention to replace
(range 25w4d-40w0d). In addition, 35% of patients had docu- the shunt. All patients requiring multiple shunts survived to
mented preterm premature rupture of membranes. Survival discharge. One patient had inward migration of the shunt. The
among premature patients was 6/9 (66%). As expected, patient underwent a thoracotomy with pleurectomy in the
advanced gestational age at delivery was associated with neonatal period for persistent effusion. The migrated shunt
better outcomes with 100% survival in the four patients that was identified along the chest wall, but it was completely
were delivered at term. embedded in scar tissue, so it was left in situ.

Fig. 4 e Ultrasounds of bilateral hydrothoraxes before treatment (left) and resolution of effusions after shunt placement
(right) in the same patient. (Color version of figure is available online.)
 mon et al  prenatal intervention for hydrothorax 125

associated chromosomal abnormalities have an even higher

Table 1 e Description of associated malformations in
fetuses with pleural effusions (excluded from the rate of fetal demise and neonatal mortality.4
analysis). Prenatal intervention is usually reserved for those patients
with a diagnosis of primary hydrothorax, and it has been re-
Associated abnormalities Number of cases
ported to improve the chances of survival, particularly in
Structural abnormalities 54 (30%) those fetuses that develop hydrops.2,3 In those series, the
Congenital heart defects 35 (20%) survival advantage was seen in patients in whom the hydrops
Cystic hygroma 22 (13%) resolved or improved significantly after prenatal interven-
Infection 1 (0.5%) tion.6 When left untreated, pleural effusions associated with
Turner’s syndrome 15 (8%) hydrops have a reported perinatal mortality of 69%-75%.17 In
our series, overall survival for patients treated prenatally for
Down’s syndrome 10 (5.5%)
primary hydrothorax, both hydropic and nonhydropic, was
Other chromosomal anomalies 5 (3%)
consistent with previously published studies.6,11,15
Several forms of intervention have been proposed. These
include thoracentesis, placement of a thoracoamniotic shunt,
Overall survival for primary hydrothorax was 76% (10/13). and less frequently, in utero pleurodesis.9,11,13 The indications
Complications associated with pulmonary hypoplasia for each intervention may vary, depending on the progression
accounted for the three deaths. When analyzing the perinatal of the effusion, the state of the pregnancy, and the overall
mortality, all patients who died had hydrops at the time of condition of the fetus.15
initial diagnosis, with a mean EGA at diagnosis of 28w0d It has been proposed that shunting could be the first-line
(SD  3). treatment because the fluid tends to reaccumulate rapidly
The mean follow-up duration was 11 mo (0-36 mo). As after thoracentesis, and the bulk of primary pleural effusions
previously stated, one patient was lost to follow-up. One are associated with lymphatic fluid that will ultimately benefit
patient was diagnosed with pulmonary lymphangiectasia at from shunting.3 We believe, there is still a role for thor-
2 y of age. Another survivor has a diagnosis of cerebral palsy acentesis as it confirms the diagnosis of a hydrothorax. We
and hydrocephalus, requiring a ventriculoperitoneal shunt noted one fetus that was found to have a hemothorax on
and home oxygen. The rest of the survivors appear to be thoracentesis and a shunt was not indicated due to risk of
doing well, without significant health problems. complications and blood loss. In addition, in a few cases, the
effusion may resolve or stabilize after thoracentesis and
further intervention is not required.
In our series, 10 patients underwent shunt placement. The
Discussion median gestational age at the time of shunt placement was
29w4d. The rate of immediate fetal demise following shunt
Primary hydrothorax in the fetus is a condition associated placement has been reported at 11%.9 There was no immediate
with significant perinatal morbidity and mortality. The fetal mortality associated with shunt placement in this series.
mechanisms behind the prenatal development of a hydro- This rate of shunt migration in our cohort of 40% is
thorax are not well understood. In some cases, there is consistent with the reported incidence in the literature of
spontaneous resolution of the effusion.5 On the other hand, 20%-65%.3,12 Derderian et al.6 noticed that even a short period
some patients progress to hydrops rapidly and this can be life of drainage might help resolve the chyle leak; therefore, sug-
threatening.7 gesting repeat shunt placement is not always needed. The
Overall mortality associated with prenatal pleural effu- risks of an additional procedure and possible development of
sions has been reported to be between 22% and 53%.5,8,16 Pa- preterm labor should be taken into account when managing
tients with secondary pleural effusions and those with these patients. At earlier gestational ages, the concerns for
pulmonary hypoplasia as well as the risk of hydrops have to be
balanced against the risk of prematurity. Prematurity and
Small to Large effusion with
Moderate
mediastinal shift, premature rupture of membranes remains a common
hydrops or
effusion complication of prenatal intervention.11,12 Our data support
polyhydramnios
this, as survival was only 66% with premature birth and 100%
with full-term delivery.
Chest-wall deformity has been described as a complication
Effusion Persistent Progression
Regresses without or of thoracoamniotic shunt placement. The mechanisms sug-
regression development gested for this problem include altered rib growth, secondary to
of hydrops
expansion, followed by decompression of the chest wall and
direct mechanical chest-wall trauma during shunt insertion. It
No Shunt if < Shunt or
intervention 36 weeks Drain appears to be more common if the shunt is placed before 20 wk
of gestation.18 In our series, there were no documented cases of
chest-wall deformity after shunt placement.
Weekly follow up
The study has several limitations inherent to retrospective
Fig. 5 e Clinical approach to primary hydrothorax. Modified reviews from a single institution. The number of cases with
from Yinon et al.15 primary hydrothorax is small compared to the total sample,
126 j o u r n a l o f s u r g i c a l r e s e a r c h  j a n u a r y 2 0 1 8 ( 2 2 1 ) 1 2 1 e1 2 7

Table 2 e Intervention s and outcomes of each patient identified with hydrothorax.

# Hydrops Thoracentesis Shunt Shunt EGA at delivery Postnatal Alive
(%lymph) placement revision treatment @ D/
C
1 Yes 93% Bilateral 2 (R at 32w4d Chest tube then Yes
(30w0d) 31w0d and pleurectomy
32w0d)
2 Yes 78% Right (28w3d) 1 (R at 36w0d None Yes
30w5d)
3 Yes 95% Bilateral (R at n/a 34w4d Chest tube then No
30w2d; L at pleurectomy
30w6d)
4 No 89% Left 2 (30w6d) n/a 34w1d None Yes
5 Yes 73% None n/a 31w0d None Yes
(hemothorax)
6 Yes 96% None n/a Declined further intervention n/a No
resulting in intrauterine fetal
demise
7 No 67% None n/a Lost to F/U. None Lost
to F/U.
8 No 16% None n/a 40w0d None Yes
9 Yes 92% None n/a 29w4d Chest tube then No
(presented in pleuro-peritoneal
labor) shunt
10 No 95% Right (26w5d) n/a 39w4d None Yes
11 No 96% Bilateral (L at n/a 35w1d Pleurectomy Yes
31w6d;
R at 33w3d)
12 Yes 95% Bilateral (R at n/a 39w6d Chest tube Yes
25w0d;
L at 26w5d)
13 Yes 89% Bilateral (R at 2 (R at 25w4d No data Yes
19w0d; L at 20w0d and
20w5d) 21w4d)
14 Yes 94% Bilateral (L at 2 (R at 33w4d Chest tube Yes
28w2d; R at 30w0d; L at
29w0d) 30w4d)
15 Yes 98% Bilateral (R at n/a 31w5d Chest tube No
30w3d; L at
31w3d)

suggesting that most patients have an alternative etiology, may require multiple interventions as the shunt can migrate
including structural or chromosomal abnormality as the or become clogged. Thoracentesis immediately before de-
cause of the effusion. Mallmann et al.9 described a better livery may facilitate the neonatal resuscitation in cases not
survival rate in patients with trisomy 21 compared to euploid drained before that time.
patients with primary hydrothorax. We did not include pa- Both the fetuses and the mothers appeared to tolerate the
tients with chromosomal abnormalities in our analysis. This procedures well. Preterm labor and preterm premature
underlines the fact that the mechanisms at play in the rupture of membranes remain the unsolved problem that can
development of hydrothorax in the fetus are not completely add significant morbidity. Other early complications of pre-
understood and may help direct future studies. natal intervention include direct trauma to the chest wall,
shunt dislodgement, shunt obstruction, and fetal demise.
Chest-wall deformity has been reported as a late complication
Conclusions and, although we did not see, must be recognized as a less
frequent complication that should be included in the coun-
Fetal intervention for the treatment of primary hydrothorax is seling. Further studies are needed to understand the mecha-
effective. It appears to confer a survival advantage, particu- nisms behind the development of fetal hydrothorax and to
larly in the setting of hydrops. The incidence of complications better define the best timing for intervention and the role for
associated with shunt placement is low. However, patients repeat shunting after shunt malfunction or dislodgement.
 mon et al  prenatal intervention for hydrothorax 127

This information is all helpful for optimizing counseling and 6. Derderian SC, Trivedi S, Farrell J, et al. Outcomes of fetal
care of these patients. intervention for primary hydrothorax. J Pediatr Surg.
2014;49:900e903. discussion 3-4.
7. Klam S, Bigras JL, Hudon L. Predicting outcome in
primary fetal hydrothorax. Fetal Diagn Ther.
2005;20:366e370.
Acknowledgment 8. Petersen S, Kaur R, Thomas JT, Cincotta R, Gardener G. The
outcome of isolated primary fetal hydrothorax: a 10-year
Authors’ contribution: R.A.M., E.E.P., G.B.M., D.R.B., and M.C.T. review from a tertiary center. Fetal Diagn Ther. 2013;34:69e76.
were responsible for either conception or design of the work. 9. Mallmann MR, Graham V, Rosing B, et al. Thoracoamniotic
R.A.M., J.K., and E.E.P. were responsible for data collection. shunting for fetal hydrothorax: predictors of intrauterine
course and postnatal outcome. Fetal Diagn Ther.
R.A.M., E.E.P., G.B.M., M.C.T., D.R.B., and L.D. performed the
2017;41:58e65.
data analysis and interpretation. R.A.M. and E.E.P. were 10. Wilson RD, Baxter JK, Johnson MP, et al. Thoracoamniotic
responsible for drafting the article. Critical revision of the shunts: fetal treatment of pleural effusions and congenital
article was given by R.A.M., J.K., E.E.P., G.B.M., and M.C.T. The cystic adenomatoid malformations. Fetal Diagn Ther.
final version to be published was approved by R.A.M., J. K., 2004;19:413e420.
E.E.P., G.B.M., D.R.B., L.D., and M.C.T. 11. Picone O, Benachi A, Mandelbrot L, Ruano R, Dumez Y,
Dommergues M. Thoracoamniotic shunting for fetal pleural
effusions with hydrops. Am J Obstet Gynecol.
2004;191:2047e2050.
Disclosure 12. Peranteau WH, Adzick NS, Boelig MM, et al.
Thoracoamniotic shunts for the management of fetal lung
The authors have no personal or financial relationships to lesions and pleural effusions: a single-institution review and
disclose. predictors of survival in 75 cases. J Pediatr Surg.
2015;50:301e305.
13. Nygaard U, Sundberg K, Nielsen HS, Hertel S, Jorgensen C.
New treatment of early fetal chylothorax. Obstet Gynecol.
references 2007;109:1088e1092.
14. O’Brien B, Kesby G, Ogle R, Rieger I, Hyett JA. Treatment of
primary fetal hydrothorax with OK-432 (Picibanil): outcome in
1. Downie L, Sasi A, Malhotra A. Congenital chylothorax: 14 fetuses and a review of the literature. Fetal Diagn Ther.
associations and neonatal outcomes. J Paediatr Child Health. 2015;37:259e266.
2014;50:234e238. 15. Yinon Y, Kelly E, Ryan G. Fetal pleural effusions. Best Pract Res
2. Longaker MT, Laberge JM, Dansereau J, et al. Primary fetal Clin Obstet Gynaecol. 2008;22:77e96.
hydrothorax: natural history and management. J Pediatr Surg. 16. Estoff JA, Parad RB, Frigoletto Jr FD, Benacerraf BR. The
1989;24:573e576. natural history of isolated fetal hydrothorax. Ultrasound Obstet
3. Rustico MA, Lanna M, Coviello D, Smoleniec J, Nicolini U. Fetal Gynecol. 1992;2:162e165.
pleural effusion. Prenat Diagn. 2007;27:793e799. 17. Weber AM, Philipson EH. Fetal pleural effusion: a review and
4. Ruano R, Ramalho AS, Cardoso AK, Moise Jr K, Zugaib M. meta-analysis for prognostic indicators. Obstet Gynecol.
Prenatal diagnosis and natural history of fetuses presenting 1992;79:281e286.
with pleural effusion. Prenat Diagn. 2011;31:496e499. 18. Merchant AM, Peranteau W, Wilson RD, et al. Postnatal
5. Aubard Y, Derouineau I, Aubard V, Chalifour V, Preux PM. chest wall deformities after fetal thoracoamniotic shunting
Primary fetal hydrothorax: a literature review and proposed for congenital cystic adenomatoid malformation. Fetal Diagn
antenatal clinical strategy. Fetal Diagn Ther. 1998;13:325e333. Ther. 2007;22:435e439.