Chylothorax in Children*

Guidelines for Diagnosis and Management

Vera Büttiker, MD; Sergio Fanconi, MD; and Réne Burger, MD

Study objective: To establish guidelines for the diagnosis and management of chylothorax in
children.
Design: Retrospective study.
Patients: Fifty-one patients with a diagnosis of chylothorax. Twelve patients were excluded
because of incomplete data or incorrect diagnosis. The following parameters were analyzed:
triglyceride level, total cell number, and lymphocyte percentage; amount of pleural effusion on
day of diagnosis, day 5, and day 14; and total time of pleural effusion. Prospectively, the same
parameters were analyzed in a control group of 10 patients with pleural drainage.
Intervention: Patients with chylothorax were treated primarily with fat-free oral nutrition; if chyle
did not stop, total parenteral nutrition with total enteric rest was started. If conservative therapy
was not successful, pleurodesis was performed.
Results: In children with chylothorax triglyceride, triglyceride content ranged from 0.56 to 26.6
mmol/L; all values except one were > 1.1 mmol/L. In 36 of 39 patients (92%), the cell count was
> 1,000 cells/mL. In 33 of 39 patients (85%), lymphocytes were > 90%. In patients without
chylothorax triglyceride, triglyceride levels ranged from 0.1 to 0.71 mmol/L (median, 0.38
mmol/L) and cell count was from 20 to 1400 cells/mL (median, 322 cells/mL), with a maximum of
60% lymphocytes. With fat-free nutrition, chyle disappeared in 29 of 39 patients. Five patients
died, and five required pleurodesis.
Conclusions: Pleural effusion in children is chyle when it contains > 1.1 mmol/L triglycerides
(with oral fat intake) and has a total cell count > 1,000 cells/mL, with a lymphocyte fraction
> 80%. Chylous effusions usually last long; however, after 6 weeks, the majority of the
effusions (29 of 39 patients) had ceased. Late surgical interventions reduce the number of
thoracotomies substantially, but can lead to very long hospitalization times. Early surgical
interventions (after < 3 weeks) lead to a high number of thoracotomies, but certainly reduce
hospitalization time. (CHEST 1999; 116:682– 687)

Key words: chyle; chylothorax; infants; newborn; pleural effusion; pleurodesis

Abbreviations: MCT 5 medium chain triglyceride; TPN 5 total parenteral nutrition

I ncomplication,
childhood, chylothorax is usually a postoperative
mainly occurring after cardiotho-
racic interventions or caused by thrombosis of the
left or right subclavian vein. It is rarely due to the
malformation of the pulmonary or thoracic lym-
phatic system that is associated with dysmorphic
syndromes.1–3 In adults, common causes of chylotho-
rax are thoracic or neck trauma or a malignancy at
the upper thoracic aperture.4 – 6
The definition of chyle in adults is well established
in the literature.5 For children, no clear definition
exists and very often adult values are applied.4,5,7
*From the ICU, University Children’s Hospital, Zürich, Switzer-
land.
Manuscript received April 14, 1999; revision accepted May 7,
1999.
Correspondence to: Vera Büttiker, MD, Steinwiesstr. 75, CH-
8032 Zürich, Switzerland; e-mail: BuettikerV@compuserve.com
One reason for this is the small number of chylotho-
races in this age group. Values derived from adult
patients are not necessarily applicable to children.
The aim of this study was to define the content and
distribution of cells and values of triglyceride con-
centration for chyle in pediatric patients. This should
help distinguish between chyle and pleural fluid of a
different etiology. Because therapy for chylothorax is
often long and difficult, a precise diagnosis is man-
datory.
Materials and Methods
We retrospectively analyzed all patients with the diagnosis of
chylothorax in our hospital database from the years 1985 to 1996.
There were a total of 51 patients having this diagnosis, but we had
to exclude 10 patients because of insufficient documentation. An
additional two patients were excluded because an analysis of the
available data dismissed the diagnosis of chylothorax.

682 Clinical Investigations

 The diagnosis of chylothorax was made according to the
proposal of Staats et al4 and Straaten et al7: triglyceride levels in
pleural fluid had to be . 1.2 mmol/L, with a total cell number
. 1,000 cells/mL and a predominance of mononuclear cells. The
main diagnoses of the remaining 39 patients are presented in
Table 1.
The following parameters were recorded: triglyceride level,
total cell number, and content of lymphocytes in pleural effusion;
amount of fluid on the day of diagnosis, and at day 5 and 14; and
duration of pleural effusion after the initiation of therapy.
Additionally, after feeding five patients a formula containing fat,
we analyzed the change of triglyceride content. Therapy con-
sisted of a diet with a fat-free formula. Total parenteral nutrition
(TPN) with total enteric rest was begun when the diet was
unsuccessful (after 2 to 3 weeks); surgery was performed when
TPN was not effective. The length of each therapeutic modality
was recorded.
Results
There were 29 boys and 10 girls with the diagnosis
of chylothorax. Their ages ranged from 5 days to 10.5
years old (median, 2.1 years old). The 10 patients in
the control group were from 6 days to 7 years old
(median, 1.3 years old). All received regular formula.
Thirty-three patients (85%) developed chylothorax
after cardiac surgery. From the remaining six pa-
tients, two developed chylothorax after pneumonia.
One of these patients had Down’s syndrome; the
second patient was a newborn baby with bilateral
pneumonia and a consecutive bilateral pneumotho-
rax. Two girls had idiopathic chylothorax. In one, the
manifestation was at the age of 10 weeks. The other
girl presented at the age of 3 months with chylotho-
rax and then developed chyloascites and chyloperi-
card; she died at 6 months of age. From the remain-
ing two patients, one had chylothorax following an
operation for a diaphragmatic hernia, and the other
had esophageal atresia. Five patients died because of
underlying disease and not because of complications
resulting from chylothorax (Table 2): three had a
severe congenital cardiac malformation with no sat-
isfactory surgical correction; one developed a Budd-
Chiari syndrome and finally died; and one died of
acute respiratory distress after the aspiration of
stomach contents.
In 19 patients, chylothorax was on the right side; in
12 patients, it was left sided; and in 8 patients, it was
bilateral.
Table 1—Primary Diagnosis of Patients With
Chylothorax
Diagnosis No. of Patients
Congenital heart disease 33
Pneumonia 2 fat intake, especially in infants, consists mainly of
Esophageal atresia 1 triglycerides and almost no cholesterol, it is prefer-
Diaphragmatic hernia 1 able to measure triglyceride content in pleural effu-
Idiopathic 2
Triglyceride content ranged from 0.56 to 26.6
mmol/L (Fig 1). All values except one were . 1.1
mmol/L: the one patient with a triglyceride level of
0.56 mmol/L was never fed a formula containing fat.
The maximum value (26.6 mmol/L) was found in one
patient with idiopathic chylothorax. After the inges-
tion of only a small amount of milk (six 10- to 15-mL
formulas per day), the triglyceride concentration in
five patients with chylothorax increased to . 1.1
mmol/L (Fig 2).
Triglyceride levels in 10 patients without chylotho-
rax were measured for comparison; triglyceride con-
tent ranged from 0.18 to 0.71 mmol/L (median, 0.38
mmol/L; Fig 1).
In 36 of 39 patients (92%) with chylothorax, total
cell count was . 1,000 cells/mL (Fig 3). In only 3 of
39 patients (7%), total cell count was , 1,000 cells/
mL. In 33 of 39 patients (85%), lymphocytes were
. 90%; in the remaining 6 patients, lymphocytes
ranged from 57 to 89%. Total cell count of the
control group ranged from 20 to 1400 cells/mL
(median, 322 cells/mL), and lymphocytes ranged
from 10 to 60% (Fig 3).
The median amount of chyle at time of diagnosis
was 214 mL (range, 55 to 730 mL).
After thoracotomy, chylothorax developed be-
tween postoperative days 1 to 18. With a fat-free
formula, chyle disappeared within 14 days in 15
of 39 patients (38%). After 45 days of conservative
therapy, effusion had stopped in 30 of 39 patients
(77%). In two patients, chylous effusion ceased
only after 50 and 51 days, respectively; two
patients died after this time. In five patients
(13%), pleural effusion lasted . 10 weeks and a
pleurodesis was performed (Fig 4); this was
successful in all patients. Effusion ceased 3 to 5
days after surgery.
Discussion
Chyle is lymphatic fluid enriched with fat secreted
by intestinal cells. It is collected and then trans-
ported via the thoracic duct into the circulation.
Measuring fat content and demonstrating the pre-
dominance of lymphocytes in pleural effusion serves
to prove chyle. We used the definition of Staats et al4
to define chyle in our patients.
In our 39 patients, the diagnosis of chylothorax was
confirmed by pleural effusion with a triglyceride
level . 1.1 mmol/L and a cell count .1,000 cells/mL
with a predominance of lymphocytes. Because oral
sion when chyle is suspected.
CHEST / 116 / 3 / SEPTEMBER, 1999 683
 Table 2—Cause of Death in Five Patients

Diagnosis Cause of Death Time of Death, d

Fallot’s tetralogy, diaphragmatic hernia Heart failure after cardiothoracic procedure 35

Transposition of great arteries, prematurity of Heart failure with postoperative sepsis (normal 19
35 wk immunoglobulins)
Ventricular septal defect, anomalous pulmonary Heart failure after cardiopulmonary resuscitation 21
venous return
Trisomy 21, atrioventricular canal, and Fallot’s Heart failure, aspiration syndrome 75
tetralogy
Persistent pleural effusion and chyloascites of No operation possible because of cardial instability, heart 90
unknown etiology failure, Budd-Chiari syndrome

In the literature, it is reported1,6,8 –11 that chylo-
thorax usually occurs in newborn and pediatric pa-
tients as a complication of thoracic and cardiac
surgery. This was also the case in 33 of our 39
patients. One patient had congenital chylothorax,
probably as a result of birth trauma or caused by
lymphatic malformation. This form of chylothorax is
rarely described in the literature.7,12,13 Chylothorax is
also described in children with trisomy 21 and
Noonan’s syndrome; in both instances, vascular and
lymphatic malformations are found.3 Only one pa-
tient in our series had trisomy 21 and chylothorax
develop after pneumonia. In older children and
adults, nontraumatic chylothoraces that are caused
by an obstruction of the thoracic duct secondary to
fibrosis or tumors are described.6 In these cases, the
spontaneous resolution of chylous leaks rarely oc-
curs. Postoperative chylothoraces obviously have a
traumatic origin; in our patients, they developed
between day 1 and day 18 after thoracotomy. This
has been also widely reported in the litera-
ture.1,6,9,10,14 There is a temporal relation between
the time of diagnosis and the initiation of oral fat
intake.
With one exception, all of our patients had a
triglyceride content . 1.1 mmol/L. This is similar
to the findings of Staats et al,4 who reported
triglyceride levels $ 1.2 mmol/L in adult patients
with chylous effusion. Our patient, who had a
triglyceride level of 0.56 mmol/L and a diagnosis
of chylothorax, had no oral fat intake at time of
diagnosis; this child suffered massive effusion after
repair of a congenital diaphragmatic hernia with
intraoperative injury of the thoracic duct. In this
situation, feeding the patient a formula containing
fat was regarded as unnecessary. After cardiac
surgery, especially in newborn infants, oral feeding
is often delayed, but small amounts of formula
were usually sufficient to allow the diagnosis of
chylothorax. Without oral fat intake, the distinc-
tion between chylous and nonchylous effusion is
difficult or even impossible to determine. On the
other hand, in 13 patients, triglyceride content was
. 1.1 mmol/L without oral fat intake; these chil-
dren all had surgery at $ 3 months of age and had
normal oral fat intake preoperatively. Obviously,
their bowel still contained enough fat to allow the
diagnosis. It seems, especially in newborn infants,
that a minimal oral fat intake is mandatory to allow
the diagnosis of chylothorax.
Staats et al4 and others5 also noted that a milky

Figure 1. Triglyceride concentration in pleural fluid. Each point Figure 2. Triglyceride concentration in pleural fluid in five
represents one patient. There is a clear difference between infants without enteral feeding and suspected chylothorax. Feed-
patients with chylous effusion and those with nonchylous effu- ing small amounts of a formula increased triglyceride content and
sion. confirmed the diagnosis.

684 Clinical Investigations

 Figure 3. Total number of cells and lymphocytes. In chylous effusion, the fraction of lymphocytes is
. 80%. Note the logarithmic scale.

aspect, first described by Wallis and Schölberg15 in Pieterson and Jakobson18 demonstrated substantial
1911, can be helpful in diagnosis. In fact, chylous increases of triglyceride content in pleural effusion
effusion often has a turbid aspect; however, in our when MCT oil was given.
opinion, this can also be misleading: 2 of our 10 The time of surgery usually is not defined as
control patients without chylothorax had turbid ef- uniformly as is the initial therapy. Some authors1,8
fusions, but triglyceride content was only 0.12 recommend surgery if effusion persists for . 2
mmol/L and 0.50 mmol/L, respectively. The milky weeks; others1,8,19 regard an amount . 100 mL per
aspect can, therefore, only serve as a hint for further year of age in children as an indication for surgery.
investigations; it is not proof that the effusion is Most authors, however, recommend an extended
chylous. period of conservative management and do not
The aspect of effusion also depends on total cell proceed to surgical treatment until . 4 weeks of
pleural effusion.10,11,14 Operative procedures include
content. In almost all of our patients with chyle
pleurodesis: the ligation of the main duct with
(92%), total cell count was . 1,000 cells/mL. In two
adjacent leaking lymphatics.2,10,14,19 The right chest
thirds of our patients with chyle (69%), the lympho-
approach for duct ligation is not always successful
cyte fraction was . 90%; in the remaining patients, it
because variations in ductal anatomy allow the lymph
was . 80%. Three patients had a total cell count
to bypass the area of ligation. Identification of
, 1,000 cells/mL, but a high triglyceride level de-
spite low oral fat intake. This low fat intake was the
only common feature we could find as a possible
explanation for the low cell number. Patients with
lymphocytes between 80% and 90% of the total
count had an increased number of polymorphonu-
clear granulocytes because of a possible irritation of
the pleura by the operation or the drainage. In
patients without chylothorax, lymphocytes were al-
ways , 60% of the total cell count.
Usually, initial therapy for postoperative chylotho-
rax has been pleural space drainage, use of medium
chain triglyceride (MCT) oil, fat-free oral alimenta-
tion, or enteric rest with TPN.2,9,14,16,17 In our
patients, a fat-free formula consisting of proteins and Figure 4. Duration of pleural drainage and time of pleurodesis.
starch was given; if effusion did not abate after 2 to Squares indicate the number of days until the pleural effusions
stopped. Plus signs indicate the number of days until the patient
3 weeks, TPN was started together with total enteric died. Circles indicate the number of days until pleurodesis was
rest. MCT oil was not used in our institution, as performed.

CHEST / 116 / 3 / SEPTEMBER, 1999 685

leaking sites on the side of effusion and ligation of all
tributaries together with the main duct is successful
in as many as 90% of patients.20 However, this
requires an extensive thoracotomy. Usually the deci-
sion to perform surgery is delayed, mainly with the
intention of preventing an additional thoracotomy. A
possible solution to this problem was presented by
Murphy et al19: as early as 8 days after diagnosis, they
inserted a pleuroperitoneal shunt with good results
in 75% of the patients. Especially in patients with
chylothorax after thoracic procedures, pleuroperito-
neal shunts were successful, less so in patients with
caval thrombosis or high right atrial pressures. An
advantage of this procedure is that it only requires
the insertion of a pleural catheter. If successful, this
technique avoids long parenteral nutrition and hos-
pitalization time.
It is not clear from the small series of patients with
chylothorax reported in the literature how often
nonoperative management is successful. In our se-
ries, surgical interventions were performed very late
in comparison. So, our patients practically represent
the spontaneous course of this disease with a mini-
mal number of surgical procedures (in 5 of 34
patients) that were performed for recalcitrant chylo-
thoraces that were not abating spontaneously. This is
in contrast to the report of Milsom et al,21 who
performed a surgical procedure in 19 of 20 patients
with chylothorax. Nevertheless, the rate of compli-
cations in our patients is not higher than those
usually reported in the literature. Consequences of
late surgical interventions are very long hospitaliza-
tion times and imminent nutritional and infectious
complications.22 Only one of our patients had septi-
cemia, the preceding Ig level, however, was normal.
If a surgical intervention in our series had been done
after 3 weeks, then 20 of 39 patients would have
undergone surgery. This explains why some authors
report a high rate of surgical intervention in their
patients. Opinions and recommendations about the
length of conservative management vary consider-
ably and depend also on the experience of the
surgical team in performing a certain procedure.
Advocates of operative interventions before 3 weeks
have reported disappointing results with conserva-
tive management. We think that it can be worthwhile
to wait . 3 weeks, provided that the nonoperative
management is aggressive, ie, TPN with total enteric
rest. To improve overall management, those patients
with pertinacious chylothoraces should be identified
early. In patients with high central venous pressures,
thrombosis of the superior vena cava, or after caval-
pulmonary anastomosis, conservative management
usually fails.
From our patients, a considerable number (11 of
39 patients) had effusion lasting . 4 weeks; seven
686
effusions were . 6 weeks. In five of the seven
patients mentioned, a pleurodesis was performed
successfully. In the remaining two patients, the
cardiovascular situation was too unstable for a tho-
racotomy and they died. Retrospectively, we think
that waiting 6 weeks or even longer as was done in
two patients (105 days and 150 days, respectively)
brings no advantage and is too long.
In conclusion, the diagnosis of chylothorax in
newborn and pediatric patients can be made when
analysis of the pleural fluid shows a triglyceride level
of . 1.1 mmol/L and an absolute cell count . 1,000
cells/mL, with a lymphocyte fraction . 80%. This
definition is simple and it allows, in most cases, a
definite diagnosis, provided there is minimal oral fat
intake. Surgery performed as early as 7 to 10 days
after diagnosis of chylothorax certainly shortens hos-
pitalization time. Waiting 2 to 4 weeks, however,
reduces the need for surgical intervention substan-
tially. Delaying surgery in a patient with pertinacious
chylothorax . 4 weeks is not recommendable.
References
1 Stringel G, Stanley M, Bass J. Surgical management of
persistent postoperative chylothorax in children. Can J Surg
1984; 27:543–546
2 Bond JS, Guzzetta PC, Snyder ML, et al. Management of
pediatric postoperative chylothorax. Ann Thorac Surg 1993;
56:469 – 473
3 Chernick V, Reed MH. Pneumothorax and chylothorax in the
neonatal period. J Pediatr 1970; 76:625– 632
4 Staats BA, Ellefson RD, Budahn LL, et al. The lipoprotein
profile of chylous and nonchylous pleural effusions. Mayo
Clin Proc 1980; 55:700 –704
5 Ingram RH. Diseases of the pleura, mediastinum and dia-
phragm. In: Wilson JD, et al, eds. Harrison’s principles of
internal medicine. 12th ed. 1991; 1112
6 Fairfax AJ, McNabb WR, Spiro SG. Chylothorax: a review of
18 cases. Thorax 1986; 41:880 – 885
7 Straaten van HL, Gerards LJ, Krediet TG. Chylothorax in the
neonatal period. Eur J Pediatr 1993; 152:2–5
8 Hargus EP, Stanley D, Carson SD, et al. Chylothorax and
chylopericardial tamponade following Blalock Taussig anas-
tomosis. J Thorac Cardiovasc Surg 1978; 75:642– 645
9 Feteih W, Syamasundar R, Whisennand HH, et al. Chylo-
pericardium: new complication of Blalock-Taussig anastomo-
sis. J Thorac Cardiovasc Surg 1983; 85:791–798
10 Le Coultre C, Oberhänsli I, Mossaz A, et al. Postoperative
chylothorax in children: differences between vascular and
traumatic origin. J Pediatr Surg 1991; 26:519 –523
11 Allen EM, van Heeckeren DW, Spector ML, et al. Manage-
ment of nutritional and infectious complications of postoper-
ative chylothorax in children. J Pediatr Surg 1991; 26:1169 –
1174
12 Koffler H, Papile LA, Burstein RL. Congenital chylothorax:
two cases associated with maternal polyhydramnios. Am J Dis
Child 1978; 132:638
13 Brodman RF, Zavelson TM, Schiebler GL. Treatment of
congenital chylothorax. NY State J Med 1975; 75:553–557
14 Rheuban KS, Kron IL, Carpenter MA, et al. Pleuroperitoneal
shunts for refractory chylothorax after operation for congen-
Clinical Investigations
 ital heart disease. Ann Thorac Surg 1992; 53:85– 87
15 Wallis RLM, Schölberg HA. On chylous and pseudo-chylous
ascitis. Q J Med 1911; 4:153–204
16 Gershanik JJ, Jonsson HT, Riopel DA. Dietary management
of neonatal chylothorax. Pediatrics 1994; 53:400 – 403
17 Joyce LD, Lindsay WG, Nicoloff DM. Chylothorax after
median sternotomy for intrapericardial cardiac surgery.
J Thorac Cardiovasc Surg 1976; 71:475– 480
18 Pieterson B, Jakobson B. Medium chain triglycerides for
treatment of spontaneous neonatal chylothorax. Acta Paediatr
Scand 1977; 66:121–125
19 Murphy MC, Newman BM, Rogers BM. Pleuroperitoneal
shunts in the management of persistent chylothorax. Ann
Thorac Surg 1989; 48:195–200
20 Robinson CLN. The management of chylothorax. Ann Thorac
Surg 1985; 39:90 –95
21 Milsom JW, Kron IL, Rheuban KS. Chylothorax: an assess-
ment of current surgical management. J Thorac Cardiovasc
Surg 1985; 18:221–227
22 McWilliams CB, Fan LL, Murphy SA. Transient T-cell
depression in postoperative chylothorax. J Pediatr 1981;
99:595–597
CHEST / 116 / 3 / SEPTEMBER, 1999 687