Seminars in Fetal & Neonatal Medicine 16 (2011) 329e339

Contents lists available at SciVerse ScienceDirect

Seminars in Fetal & Neonatal Medicine

journal homepage: www.elsevier.com/locate/siny

Neonatal portal vein thrombosis: Diagnosis and management

Suzan Williams a, *, Anthony K.C. Chan b
a
b

Department of Hematology/Oncology, The Hospital for Sick Children, Toronto, Ontario, Canada
Department of Hematology/Oncology, McMaster University, Hamilton, Ontario, Canada

s u m m a r y
Keywords:
Anticoagulation
Liver lobe atrophy
Neonate
Portal hypertension
Portal vein thrombosis

Neonatal portal vein thrombosis (PVT) is an increasingly recognized event. Patients are generally
asymptomatic in the neonatal period. The diagnosis is made with Doppler ultrasound. Umbilical catheterization, exchange transfusion and sepsis are risk factors for neonatal PVT. Thrombophilia is possibly
a contributing risk factor. Although there are potential serious acute complications such as hepatic
necrosis, the outcome is good in the majority of cases, followed up to 8 years of age. Thrombus resolution
occurs in 30e70% in days to months. Liver lobe atrophy may occur following PVT, and does not appear to
be associated with any impairment of liver function. Non-occlusive thrombosis is more likely to resolve
than non-occlusive thrombosis. A subset of patients without resolution is at risk for developing portal
hypertension over the next decade of life. There are no current dening features present during the
neonatal period to enable identication of neonates at risk for portal hypertension. There is no evidence
that anticoagulation therapy improves time to resolution or decreases the likelihood of portal hypertension. Anticoagulation therapy may be considered. A management algorithm is proposed.
2011 Elsevier Ltd. All rights reserved.

1. Introduction
Neonatal portal vein thrombosis (PVT) has been described as
a rare event, but is becoming more commonly recognized. Estimates range from 1 in 100,000 live births1 to 36 per 1000 neonatal
intensive care unit admissions.2 Because portal venous thrombosis
rarely causes clinical problems during the neonatal period, historically the majority of cases remained unrecognized in the neonatal
period and were found later in childhood.3,4 This led to the
supposition that neonatal PVT was as described by Thompson and
Sherlock: exceedingly rare.5
By contrast with the reported rarity in the neonatal setting, PVT is
the major cause of extrahepatic portal hypertension and gastrointestinal bleeding in children.3 Prehepatic or extrahepatic portal
hypertension is most commonly the result of an organized
thrombus in the portal vein. Portal vein brosis, stenosis, or web
may also lead to elevations in portal venous pressure.6 Occlusion of
the portal vein associated with small periportal collaterals is known
as cavernous transformation of the portal vein (CTPV).7
The etiology of neonatal PVT is different from that in children and
adults. In adults PVT is most frequently secondary to cirrhosis.8 In
older children, PVT is related to liver transplantation, intra* Corresponding author. Department of Hematology/Oncology, The Hospital for
Sick Children, 555 University Ave., Toronto, Ontario M5G 1X8, Canada. Tel.: 1 416
813 8997; fax: 1 416 813 5327.
E-mail address: suzan.williams@sickkids.ca (S. Williams).
1744-165X/$ e see front matter 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.siny.2011.08.005

abdominal sepsis, splenectomy, sickle cell anemia, and antiphospholipid antibodies.8e14 In w50% of children with PVT, an
underlying etiology is not identied.10,15 PVT in neonates commonly
occurs secondary to the placement of an umbilical vein catheter
(UVC), with or without infection.16e18
2. Historical aspects
Historically, the etiology of portal vein occlusion in children has
been attributed to a variety of causes from congenital malformations
to thrombosis associated with infection and most recently to
thrombosis as a potential complication from trauma or vessel
catheterization. Prenatal onset of PVT and in association with
umbilical vein thrombosis has also been reported.19e21
Cavernomatous transformation of the portal vein was historically described as a congenital or tumour-associated formation. In
initial reports, a congenital malformation22 or hamartoma23 were
presumed.
In 1928, Klemperer reviewed previously reported cases and
suggested that thrombosis was responsible for the changes in the
portal vein.24 Wallgren in 1927 had postulated PVT as a complication of previous intra-abdominal infection, citing umbilical sepsis25
and appendicitis as possible causes. The possibility of umbilical
infection without sepsis having more disseminated complications
was described in a case of severe umbilical infection with cellulitis
of the anterior abdominal wall in 1957.26 Neonatal sepsis and

330

S. Williams, A.K.C. Chan / Seminars in Fetal & Neonatal Medicine 16 (2011) 329e339

umbilical sepsis leading to PVT was described in case series.27,28

Shaldon and Sherlock in 1962 reported 16 children with PVT; 11
had a history of sepsis.29 Acute appendicitis, omphalitis, and biliary
infections were then all cited as a source of infected thrombi
leading directly to portal thrombophlebitis. PVT, therefore, may be
either secondary to portal pyemia with the site of infection somewhere in the portal venous system, or due to local spread of
infection from a site adjacent to the portal vein.
Later, cavernomatous lesion in the portal vein was more strongly
connected as secondary to PVT and represented attempts at
a collateral circulation to carry portal blood to the liver.30,31 It is
now established that a PVT may completely resolve or it may
transform into a cavernoma with dilated pancreatico-duodenal and
prebiliary veins, secondary to portal hypertension.2,32,33
Despite umbilical infection and sepsis being common in the case
series initially describing PVT in children (Table 1), a signicant
proportion of cases did not have infection. Exchange transfusion via
the umbilical vein was postulated as an etiological factor in case
reports.34,35 In a case series, Oski et al. in 1963 reported four cases of
PVT; three having trauma and prolonged umbilical catheterization,
with three out of four (75%) having no evidence of infection.36
3. Umbilical vein catheterization
The transition from fetal to neonatal life causes major changes in
hepatic vascular anatomy. As blood ow from the umbilical vein to
the portal vein ends, the portal vein will take on its adult function of
carrying nutrients and toxins from the stomach and intestine to the
liver for processing in hepatic capillary-like vessel sinusoids. In
adults, the portal vein supplies 70e75% of liver blood ow, and
50e55% of the oxygen requirement.
In neonates, the umbilical vein is available as a site for central
venous access for the rst week of life. The umbilical vein joins the
left portal vein in the liver. There is a direct communication formed
between the umbilical vein and the ductus venosus, which
bypasses the liver and joins the inferior vena cava. The umbilicoportal conuence is the area at which the portal vein is traversed
by the junction of the umbilical vein and the ductus venosus
(Fig. 1).37 The normal umbilical venous catheter traverses through
the umbilical vein past the left branch of the portal vein, into the
ductus venosus, which lies at the cephalic aspect of the left portal
vein and ends in the right or middle hepatic vein close to their
entrance into the inferior vena cava.38 The ductus venosus is narrowed at its origin, and is not always perfectly aligned to the
umbilical vein. Therefore, the left branch of the portal vein may be
the recipient of the umbilical venous catheter tip during
placement.39,40
The catheterization of the umbilical vein has now long been
implicated as a cause for PVT presenting in childhood.3,4,36,41e43
There are several mechanisms by which umbilical venous
Table 1
Portal vein thrombosis (PVT) following neonatal umbilical catheterization, septicemia and umbilical sepsis.
Author

Year

Total
patients

Septicemia

Umbilical
sepsis

PVT

Thompson et al.5
Yadav et al.46
Schwartz et al.17
Guimaraes et al.47
Kim et al.37
Sakha et al.55
Gharehbaghi et al.54

1964
1993
1997
1998
2001
2007
2011

493 (470)a
47
173 (100)a
40
100
50
164

51 (10%)
7 (15%)
e
e
5 (5%)
11 (22%)
10 (6%)

37 (8%)b
4 (8.5%)
e
e
e
e
2 (1%)c

0
0
1 (1%)
0
43 (43%)
17 (34%)
5 (3%)

a
b
c

Patients followed.
Umbilical infections.
Peri-umbilical erythema.

Fig. 1. Anatomy of neonatal portal vein. RPV, right portal vein; RHV, right hepatic vein;
IVC, inferior vena cava; LHV, left hepatic vein; LPV, left portal vein; SV, splenic vein;
SMV, superior mesenteric vein; PV, portal vein (Published with permission from Kim et
al.110).

catheters may cause thrombosis; including damage to vessel walls,

disrupted blood ow, the infusion of substances such as total
parenteral nutrition that damage endothelial cells, and thrombogenic catheter materials.44,45
4. Variable incidence of PVT with umbilical vein
catheterization
There is a variable reported incidence of umbilical venous
catheter-related PVT in infants and children from zero to 43%
(Table 2).17,37,42,46e50 A number of explanations have been put forth.
Table 2
Portal vein thrombosis (PVT) following umbilical venous catheterization (UVC)
imaging.
Study

Year

UVC (n)

PVT N (%)

Time to image

Thompson et al.5
Yadav et al.46

1964
1993

493 (470)a
47

0b
0

Schwartz et al.17
Guimaraes et al.47
Kim et al.37

1997
1998
2001

173 (100)a
40
43

1 (0.6%)
0
43 (43%)

Sakha et al.55

2007

50

17 (34%)

Turebylu et al.56

2007

28

2 (7%)c

Gharehbaghi et al.54

2011

164

5 (3%)

School age
22 newborns:
within 4e8 weeks,
q3 months to age
12 months, 24 months.
25 children at age
1e5 years
3 days to 5 months
8e9 years
Within 1 week of UVC
q2e4 days till
UVC removal
q2e7 days till
resolution
1e2 weeks after
UVC removal
q1e2 months until
clot resolution
3, 5e7 days post UVC
Weekly while UVC
in place
3 days post UVC
24e48 h of UVC
48e72 h post
UVC removal
Weekly till discharge
or resolution

Patients followed.
Three children died and had suppurative thrombophlebitis in umbilical vein,
extending into ductus venosus, with no specic mention of portal vein on
postmortem.
c
Additional 4 (14%) had inferior vena cava clots.
b

S. Williams, A.K.C. Chan / Seminars in Fetal & Neonatal Medicine 16 (2011) 329e339

Differences between previous studies that contribute to the

difference in reported incidence of PVT with umbilical venous
catheter are: the prospective or retrospective nature of the study;
the diagnostic method used to study the thrombus; the time
schedule of examinations; and catheter variables such as size,
location, and duration.
Retrospective studies have shown that many infants and children
with extrahepatic portal hypertension have undergone catheterization of the umbilical vein during the newborn period.2e4,36,41,42
One of the more recent retrospective cohorts showed that ve out
of seven (71%) patients with portal hypertension or splenomegaly in
a cohort of patients with PVT had a history of umbilical venous
catheterization.32
Autopsy studies have estimated the incidence of umbilical
venous catheter-related thrombosis at 20e40% of neonates who die
with an umbilical venous catheter in place.42,48,50e53 Larroche
found that 40% of neonates having umbilical vein catheterization
developed portal venous thrombi after 25e48 h. Thrombi and
brin rings around the catheter or adherent to the intimae were
seen in 13.7% after a period of 12 h, and in 100% after three days.48
Thompson and Sherlock described 470 of 493 neonates, who
had umbilical venous catheters placed for exchange transfusion,
prospectively surveyed years later as children, as having no incidences of PVT or portal hypertension.5
Some more recent prospective studies of umbilical venous
catheters have also reported that properly inserted catheters are
not associated with PVT, with no reported incidences of PVT, when
ultrasound imaging was completed weeks to years after the
umbilical venous catheter placement.17,46,47 However, the absence
of portal hypertension by school age or portal vein occlusion on
imaging several weeks after the catheterization does not rule out
the possibility of a neonatal PVT.
Given the temporal evolution of thrombi, the timing and
frequency of ultrasound monitoring is relevant in determining the
incidence of PVT. More recent studies, with prospective ultrasound
examination starting hours to weeks after umbilical catheter
placement have described the incidence of PVT as 1.3e43% in neonates.17,37,54e56 Kim et al. showed a 43% rate of neonatal PVT
following umbilical venous catheters when ultrasound examination
was completed every 5e7 days while infants where in hospital. Most
of the small thrombi were attached to the catheter rather than the
wall of the portal vein, supportive of the idea that a catheter in the
lumen is closely related to the initiation of thrombosis. There was
clot resolution occurring in more than half of the cases within 23
days after catheter removal. Schwartz et al. commented that
thrombi which developed during the early neonatal period and
resolved before ultrasound imaging, which was performed at
a median of 2 months, would not have been detected, contributing
to a lower incidence of 1.3 %.17
Non-visualization of the portal vein is not considered as
representative of PVT in all reports. In the study by Yadav et al.,
there was an absence of splenoportal ow in at least seven (32%)
and possibly 12 (55%) of 22 neonates, which normalized by the age
of 2 years.46 The authors reported no PVT in this cohort. Guimaraes
et al. studied the late outcome of umbilical vein catheterization in
school-age patients (at age 8 years), and found intact portal
circulation by Doppler ultrasound in all 40 children except for two
with no portal vein visualized, one of which was associated with
small left hepatic lobe.47 Like Yadav et al., they reported no PVT in
their cohort of 40 patients with neonatal umbilical venous catheters. In the one patient where the ndings were persistent, the
authors postulated the non-visualization of the portal vein as
a normal variant,47 as has been previously described.47,57 However,
atrophy of the left hepatic lobe has been shown to be a potential
complication from neonatal PVT.2

331

The varying reported incidences of neonatal PVT following

umbilical vein catheterization reported by retrospective, prospective, and autopsy studies are compatible with a few assumptions.
The rst is that the development of PVT is common in the newborn
period. The second is that the majority of thrombi resolve spontaneously and therefore do not cause portal hypertension. Third, the
denition of PVT will inuence the incidence. When dened by
imaging, the incidence of PVT will be higher than when dened by
clinical symptoms of portal hypertension.
4.1. Catheter size, duration, location
Catheter-related variables such as size, duration of placement and
location also have an impact on the incidence of line-associated
thrombosis. In Schwartz et al.s study there was control of a smaller
(uniformly 3.0 F) catheter, with respect to catheter position, so that it
was not allowed to terminate in the liver,17 potentially contributing to
lower incidence than studies reporting on all patients with umbilical
venous catheters. Prolonged catheterization and transfusion through
the catheter were signicantly associated with PVT.37 The highest
rate of thrombosis (ve out of eight, 63%) was in neonates with
catheter placement in the portal vein, and the main site of thrombosis was the umbilicoportal conuence (the space of Rex).37
Outside of placement in the portal vein, a particular umbilical
venous catheter position does not appear to be signicantly associated with PVT. It has been suggested that PVT may occur less
frequently if the umbilical venous catheter tip is present in the right
atrium or in the inferior vena cava (high placement). There is
greater blood ow in the right atrium and inferior vena cava
compared to the portal sinus or umbilical vein (low placement). The
greater blood ow was thought to ensure adequate mixing of the
infusates with the neonates blood and to avoid direct exposure of
the liver to unphysiologic stimuli. In addition, the mechanical
endothelial damage and altered blood ow caused by the catheter
traversing the umbilicoportal conuence, serving as the initiating
thrombotic insult, would inuence the incidence of thrombosis.
There was no statistically signicant difference in rates of PVT
related to catheter position in Kim et al.s study. However, high
placement actually had a higher rate of thrombosis 27/68 (42%)
compared to low placement 4/15 (27%).37 In the Schwartz et al.s
study, the single infant with PVT had an umbilical vein catheter
which terminated in the right atrium.17 In Morag et al.s study, 45%
of infants with PVT had the umbilical venous catheter placed in an
appropriate high ow position.2
4.2. Patient factors
Apart from endothelial damage during catheter placement and
composition of the infusate, other factors associated with the
initiation and propagation of thrombosis include patient characteristics. Low birth weight, low ow state, hypercoagulability,
hypoxia, infection, sepsis, congenital malformations, and gestational diabetes mellitus can contribute to complications caused by
umbilical catheterization.42,53,58e61 The same variables that
contribute to complications necessitate the use of the catheter for
supportive care in critically ill infants.
In spite of the potential hazard of thrombus formation, the
umbilical venous catheter remains a practical and widely used
route for the monitoring and treatment of critically ill infants.62 As
a result, neonatal PVT will continue to be a relevant clinical entity.
5. Outcome
There are potentially signicant complications which can occur
following neonatal PVT (Table 3). Thrombosis of the umbilical

332

Table 3
Neonatal and childhood portal vein thrombosis: radiologic and clinical outcome.
Year

Diagnosis

Treatment

Alvares et al.3

1983

108 (1 months
to 15 years)

104 AG
4 US

3 VL
1 MS
76 PS

Schwartz et al.17
Kim et al.37

1997
2001

1
43 neo

US
US

2006

US

Morag et al.

(e)UVC

Complete
resolution (CR)

100%
18/36 (50%)

2/36 (6%)

81/133 (61%)

16/133 (12%)

Turebylu et al.56
Sakha et al.55

2007
2007

Demirel et al.71

2009

Morag et al.32

2011

70d

US

14/37 (38%)
normal

Gharehbaghi et al.54

2011

5 occ

US

3/5 (60%)

US
US

US

None

ACT (LMWH
/UFH)

No change

Extension or
recurrence

Time to
CR/PR

LLA

PHTN

Death
2a

133
128/133 neo
5/133 inf
5
17 neo
16 nocc
1 occ
15 neo

ACT 59/133
(44%)

Partial
resolution (PR)

3/5 (60%)
13/13 (100%)

11/36
(30%)

5/36 (9%)

6 months
2e23 days
(e)UVC
2e626 days
mean: 63
median: 25
3e6 weeks
2e5 months

1/36 (3%)
30/133
(22.5%)

6/133 (4.5%)

22/133 (16%)b

NA

NA

2/5 (40%)
0

13/15 (87%)c

2/15 (13%)a
20/37 (54%)
2/20 (10%)
PHTN

3e6 weeks

2/37 (5%)
6/133 prior PHTN
2/6 (33%) CR
1/6 (17%) transplante
1/6 (17%) lost
2/6 (33%) PCS
2/5 (40%)

LLA, liver lobe atrophy; PHTN, portal hypertension; AG, angiogram; US, ultrasound; VL, variceal ligation or gastric devascularization; MS, mesocaval shunt; PS, portosystemic shunt surgery; neo, neonates; inf, infants; PCS,
portacaval shunt; lost, lost to follow-up; (e)UVC, umbilical venous catheter removal; nocc, non-occlusive; occ, occlusive.
a
Due to other causes.
b
25/133 (22.5%) lost to follow-up.
c
Described as recovered.
d
Follow-up at 2e8 years from prior cohort: 12/133 primary liver pathology, 22/133 deaths before 2 years, 29/99 (29%) lost to follow-up.
e
Patient with primary liver disease.

S. Williams, A.K.C. Chan / Seminars in Fetal & Neonatal Medicine 16 (2011) 329e339

Study

S. Williams, A.K.C. Chan / Seminars in Fetal & Neonatal Medicine 16 (2011) 329e339

portal sinus can lead to adjacent liver necrosis, found on postmortem examination, with risk factors for necrosis including
placement of the umbilical catheter in the portal vein, infusion of
alkaline solution, sodium bicarbonate via the catheter, and
thrombocytopenia.63 Cerebral infarct resulting from paradoxical
emboli, in which emboli from the PVT passing through the ductus
venosus and foramen ovale into systemic circulation, has been
reported.64 Hepatic hematoma with PVT related to malpositioned
umbilical venous catheter has been reported.49 Bleeding diathesis
with hemorrhagic ascites, massive intrapulmonary bleeding, and
subsequent death in diffuse intrahepatic PVT with marked hepatic
necrosis conrmed by postmortem has been documented.20
5.1. Resolution
The long term sequelae and clinical importance of thrombosis
detected on ultrasound in asymptomatic neonates are not fully
understood. The spontaneous regression of catheter-related
thrombi detected on ultrasound has been reported.17,65 In fact,
early spontaneous resolution has been postulated as a reason why
PVT was diagnosed only relatively rarely in the clinical setting of the
neonatal intensive care unit.3,17,37 The single detected PVT in the
Schwartz study had detection at 2 weeks and resolution at 6
months.17 In the ve PVT detected at 3e12 days in the Gharehbaghi
et al. study, three (60%) resolved by 3e6 weeks, whereas the two
other patients died before a second ultrasound.54 Recanalization
(partial or complete resolution) occurs more frequently in cases
with partial thrombi than in cases with occlusive thrombi. Kim et al.
found a statistically signicant relationship between the initial size
of the thrombus (occlusive versus non-occlusive) and clot resolution. Resolution occurred in 70% of neonates with non-occlusive
clots and 31% of neonates with occlusive clots. The time to detection of resolution after the removal of the catheter was at a mean of
10 days with a range of 2e23 days.37 There were similar ndings in
the Morag et al. study, in which there was a higher resolution rate in
non-occlusive versus occlusive clots with 77% resolution for
non-occlusive clots and 48% for non-occlusive clots; mean time
detected to resolution of the thrombus was 63 days (range: 2e626
days; median: 25 days). Although there was no association of
outcome with presence of an umbilical venous catheter, in the
subset of patients with umbilical venous catheters, there were
higher resolution rates in infants with appropriate placement of
umbilical venous catheters compared to low or intrahepatic
placements.2
5.2. Portal hypertension and cavernoma
If the PVT does not resolve, there may be long term implications.
It may transform into a cavernoma, with dilated pancreaticoduodenal and prebiliary veins, secondary to portal hypertension.33,37,46,47,66 Esophageal variceal bleeding, splenomegaly and
growth retardation are the most common clinical features in children with portal hypertension. Alveraz et al. found that portal
hypertension was diagnosed at a mean of 5.7 years after the acute
event. Eighty percent of patients with portal hypertension present
within the rst 3 years of life.3,33,41,67e69 Portal cavernomas and
portal hypertension were not seen in neonates with PVT followed
for 2e73 days (median: 8 days).37 Portal hypertension occurred in
(6/133) 4.5% followed for a median of 79 days (range: 1e1814 days).2
In the follow-up study of a subset of 70 of the neonatal cohort at
a median age of 5 years (range: 2e8 years), two children with portal
hypertension diagnosed during neonatal period required portacaval
shunting.32
However, portal hypertension appears to be uncommon
following neonatal PVT. This may in part be due to the predominant

333

left portal vein involvement. If the thrombus does not propagate

into the main or right portal vein, the effect on portal obstruction is
limited. Periportal collateral circulation would minimize portal
obstruction. The presumption is that even in unresolved PVT;
clinically signicant portal hypertension would not develop,
provided the clot remained in left portal vein only. Although
hepatic necrosis with PVT has been reported, even with extensive
PVT, signicant hepatic necrosis is not expected to occur, given the
compensatory blood ow from the hepatic artery.20
5.3. Liver lobe atrophy
Liver lobe atrophy is more common than portal hypertension
following neonatal PVT. Liver lobe atrophy occurred in 1 of 36 (3%)
infants with follow-up imaging in the Kim et al. study.37 Thirty out
of 133 (22.5%) of the infants in Morag et al.s study had lobar
atrophy, which was not demonstrated in early ultrasound examination but developed in later imaging.2 Rates of liver lobe atrophy
did not change in the subset of patients reported at age 2e8 years,
with left lobar atrophy noted in 20 (54%); in two associated with
splenomegaly, and in two associated with portal hypertension.32
Liver atrophy is more likely to occur with occlusive PVT than
non-occlusive PVT.2 Similar to the recent pediatric studies,
a correlation between hepatic lobar atrophy and ipsilateral portal
vein obstruction has been reported in adults.70
6. Thrombophilia
There have been reports of thrombophilia associated with PVT
in neonates and children (Table 4). Case series have reported factor
V Leiden, prothrombin gene mutation, protein C deciency, protein
S deciency, antithrombin deciency, increased factor VIII, methylene tetrahydrofolate reductase (MTHFR) variants with elevated
homocysteine, and elevated lipoprotein (a).1,2,32,71e77 One case
report describes inherited dysbroginemia presenting at age 15
years with PVT.78 In caseecontrol studies the odds ratio for children
with PVT having a thrombophilic defect compared to controls was
from 5.47 to 11.9.1,77
One of the difculties in interpreting the signicance of
thrombophilia ndings with respect to inhibitor protein deciencies (protein C, protein S and antithrombin) is teasing apart
genetic versus acquired deciencies. Testing is completed after the
thrombosis has occurred. Therefore, there is the possibility in the
absence of family studies with genetic testing, that the observed
deciency may be secondary to elevated consumption from portal
hypertension-associated portosystemic shunts or reduced
synthesis caused by a reduction of hepatic blood ow as a result of
thrombosis.
Revel-Vilk et al. found that inherited prothrombotic coagulation
proteins do not contribute signicantly to the pathogenesis of
venous thromboembolic disease in neonates, concluding that the
most signicant etiologic risk factors are the presence of a central
venous line and other medical conditions.79 In a prospective study
of 53 infants with umbilical vessel catheterization, in which 28 had
umbilical venous catheters, factor V Leiden, prothrombin, and
MTHFR mutations were screened for and serial ultrasounds
completed for detection of thrombosis. There was no statistically
signicant difference in rates of thrombosis in neonates with the
mutations compared to the neonates without mutations.80 Therefore, factor V Leiden, prothrombin gene mutations and MTHFR
mutations do not appear to add signicant risk for neonatal PVT.
The possibility remains that thrombophilia defects other than
factor V Leiden, prothrombin and MTHFR mutations may
contribute to the risk of neonatal PVT. As with other venous
thrombotic events in infants and children, thrombophilia may be

334

S. Williams, A.K.C. Chan / Seminars in Fetal & Neonatal Medicine 16 (2011) 329e339

Table 4
Thrombophilia and paediatric portal vein thrombosis.
Study

FVL

PTG

Y PC

Y PS

Y AT

MTHFR/[
Homocys

Combined/other

OR (95% CI)

Heller et al.1
Demirel et al.71

24
16

4 (17%)
0

0
0

1 (4%)
0

0
0

1 (4%)
0

1 (4%)
0

5.47 (1.7  17.6)

Morag et al.2
Morag et al.32
El-Karaksy et al.72

40
25
40

1 (2.5%)
2 (8%)
11 (27.5%)

0
0
5 (12.5%)

0
NP
11 (27.5%)

0
3 (12%)
0

1 (2.5%)
1 (4%)
0

NP
NP
NP

ACLA 2 (8%)
PTG MTHFR 1 (7%)
[ FVIII 0 (0%)
[ FVIII 11 (27.5%)

6 for FVL

Pietrobattist et al.77

5 (12.5%)
Y PC Y AT def 1 (2.5%)
Y PC FVL APCR 1 (2.5%)
Y PC PTG 1 (2.5%)
FVL APCR PTG 2 (5%)

31

2 (7%)

3 (10%)

4 (13%)

4 (13%)

NP

Pinto et al.73

14

1 (7%)

6 (43%)

3 (21%)

1 (7%)

UttenreutherFischer et al.75
Dubuisson et al.74
Seixas et al.108
Schobess et al.109

23

2 (9%)

NP

NP

NP

NP

16 (68%)
(13% homo)
4 hetero (28%)
3 homo (21%)
NP

20
20
9

NP
0
2 (22%)

NP
NP
NP

9 (45%)
NP
NP

13 (65%)
NP
NP

10 (50%)
NP
NP

NP
NP
NP

11.91 (1.41  100.77)

No thrombophilia
in control patients

1/64 control patient FVL

FVL, factor V Leiden; PTG, prothrombin gene mutation; PC, protein C; PS, protein S; AT, antithrombin deciency; FVIII, factor 8; MTHFR, methylene tetrahydrofolate reductase
variant mutation; Homocys, homocysteine; APCR, activated protein C resistance; ACLA, anticardiolipin antibody; NP, not performed or not provided; homo, homozygous
mutation; hetero, heterozygous mutation.

one of multiple factors contributing to thrombotic risk. There is no

available data on how the presence of a prothrombotic state
impacts on the care of a neonate with PVT.
7. Diagnosis
There may be an absence of clinical and laboratory signs with
PVT in the neonate. Thrombocytopenia may be seen at the time of
diagnosis, but is not specic for PVT. Thrombocytopenia
(<100  109/L) has been reported in 26/133 infants (19.5%). In 13 of
these infants, the thrombocytopenia may have been secondary to
other conditions, including necrotizing enterocolitis (in two
infants) or sepsis (in 11 infants).2 Consumption from the thrombosis or the concomitant clinical risk factors for thrombosis, such as
sepsis, may explain the low platelet count in the acute phase. In the
late chronic stage, a low platelet count would be most likely
secondary to hypersplenism with portal hypertension (Table 5).36
In contrast to adults, liver function is usually normal in children
presenting with PVT.8 There can be mild liver biochemical abnormalities in children with PVT. The extent to which similar abnormalities occur in neonates is less clear. Nine out of 133 (7%)
neonates had abnormal liver enzymes as the indication for ultrasound which identied the PVT.2 In the follow-up study in 2011 on
70 of the 133 neonates diagnosed with PVT, 25 had liver functions
assessed and mild abnormalities were detected in nine (36%)
children with a mean age of testing of 5.5 years (range: 1.5e8
years). There were mild elevations of aspartate aminotransferase

(AST) in three (12%) and alanine aminotransferase (ALT) in six (24%)

children, with normal albumin, direct bilirubin, and gammaglutamyl transferase levels.32 In a study on ultrasound imaging in
16 infants with abdominal thrombosis, the presence of main left
intrahepatic PVT in six infants was clinically unsuspected. In four of
the six infants (67%), there was an association with transitory
abnormalities of the liver function tests, with no other cause for the
elevation of AST and ALT found. Levels returned to normal with
thrombus resolution (Table 6).65
Investigational radiographic studies in the diagnosis of PVT
include abdominal ultrasound, computed tomography (CT), angiography, and magnetic resonance imaging (MRI). Accurate interpretation of the ndings identied by each of these modalities can
be difcult, especially in the neonates. A CT scan accurately identies PVT and the presence of collateral vessels but requires exposure to radiation and the use of intravenous contrast materials.
Similarly, angiogram requires exposure to radiation and intravenous contrast material. MRI performs as least as well or better than
CT in diagnosing PVT, without the issues of ionizing radiation
exposure and intravenous contrast.81 In comparing the efcacy of
imaging techniques in identifying portal vein patency, Weinreb
et al. found MRI superior to CT in visualizing hepatic architecture
and vascular anatomy and patency in 27 children undergoing
evaluation of suspected liver disease.82 Unfortunately MRI, like CT,
cannot be performed at the bedside of an ill neonate. Therefore,
each of these modalities may have limited utility in the neonatal
population.

Table 5
Clinical presentation of neonatal portal vein thrombosis.
Study

Year

Signs

Symptoms

Kim et al.37
Morag et al.2

2001
2006

None

None

55

2007

43 neonates
128 neonates
5 infants
17 neonates
16 nocc
1 occ
15 neonates

Sakha et al.

Demirel et al.71

2009

Gharehbaghi et al.54

2011

Laboratory tests
Thrombocytopenia
26/133 (19.5%)

Sepsis 11 (65%)

Hepatosplenomegaly
2 (13%)

5 neonates
5 occ

nocc, non-occlusive; occ, occlusive; AST, aspartate aminotransferase; ALT, alanine aminotransferase.

2/5 (40%) [ AST, ALT

S. Williams, A.K.C. Chan / Seminars in Fetal & Neonatal Medicine 16 (2011) 329e339

335

Table 6
Clinical presentation of childhood portal vein thrombosis.
Study

Year

Patients

Symptoms

Signs

Laboratory tests

Alvares et al.3

1983

108 children
Mean age:
5 years 7 months
Range: 1 month
to 15 years

86 GI bleeds (80%)
Hematemesis 58 (53%)
Melena 19 (17.5%)
Hemorrhoid
bleeding 5 (4.6%)
Diarrhea 5 (4.6%)
Abdominal
distension 5 (4.6%)
Splenectomy 5 (4.6%)
Abdominal pain 2 (2%)

Splenomegaly 101 (94%)

Hepatomegaly 13 (12%)
Ascites 9 (8.3%)

El-Karaksy et al.72

2004

40 children
PHTN 2
to PVT

Splenomegaly 35 (87.5%)
Splenectomy 5 (12.5%)
Incidental splenomegaly
3 (7.5%)

Pietrobattist et al.77

2010

31 children

Peter et al.107

2003

86 children
EHPVO

Ascites 2 (2.3%)
Splenomegaly 11 (12.8%)

Normal liver serum biochemistry

81 (94.2%)

Morag et al.32

2011

70 children
Median age:
55 months
Range:
24e96 months

Hematemesis melena
15 (37.5%)
Hematemesis 7 (17.5%)
Melena 1 (2.5%)
Abdominal distension
11 (27.5%)
Abdominal pain 3 (7.5%)
Neonatal sepsis 2 (10%)
Umbilical sepsis 3 (5%)
Umbilical sepsis (6%)
Gastroenteritis and
Dehydration (6%)
Family history of
VTE (3%)
Parental consanguinity
(1%)
UGI bleeding
53 (61.6%)
Jaundice
Splenomegaly 3/37 (8%)

Liver function tests (106 tested)

[ serum transaminase 11 (10%)
(11 previous blood transfusions, 6
hepatitis B surface antigen positive)
Hypoalbuminemia (2.5 to 3.5 g/dL) 12 (11%)
[ serum unconjugated bilirubin 5 (4.7%)
Thrombocytopenia (40 000e150 000) 64 (63%)
Lymphopenia (<1000 lymphocytes) in 25/96 (26%)
Neutropenia (<1500 neutrophils) 8/98 (8%)
Anemia 4 (4%)
Y factor II 7 (7%)
Y factor I 5/104 (5%)
[ high brin split product 2 (2%)
Y factor V
Endoscopy ndings:
Esophageal varices 40 (100%)
Gastric varices 10 (25%)
Congestive gastropathy 3 (7.5%)

GI, gastrointestinal; PHTN, portal hypertension; PVT, portal vein thrombosis; UGI, upper gastrointestinal; VTE, venous thromboembolism; EHPVO, extrahepatic portal vein
occlusion.

From a practical perspective, neonatal PVT is diagnosed via

Doppler ultrasound. Abdominal ultrasound with colour Doppler is
less invasive, less traumatic to the patient, and more economical
than CT or MRI.
Ultrasound is an operator-dependent modality and technical
factors (including the skill level of the sonographer, patient cooperation, and the presence of abdominal gas and anatomic variations)
may prevent the identication of vascular structures. Despite this,
portal vein obstruction is well evaluated even by grey-scale sonography. The features of portal blood ow distinguishable from
systemic venous ow are: (i) continuous ow, with little or no cardiac
or respiratory variation; (ii) relatively low mean and peak frequencies compared to systemic veins, which generally have a window
due to concentration of ow velocities in a narrow band near the
peak; and (iii) mild turbulence, visible as roughness of the Doppler
spectrum envelope and audible as a distinct low-pitched rumble.83
In a study comparing portal vein patency by Raby et al., the portal
vein was assessed in 115 patients (age range: 6 months to 75 years)
by ultrasound scanning and arterial portography over a 2-year
period.84 Ultrasound scans correctly assessed portal vein patency
in 31/34 (90%) patients with diffuse liver disease, 47/52 (90%) with
liver tumours and 6/6 (100%) with primary PVT. Ultrasound
assessment of portal vein patency or occlusion had an overall
accuracy of 87.5%, and was more reliable in demonstrating patency
(90% accuracy) than occlusion (68% accuracy). Ultrasound assessment was conrmed as accurate by arterial portography in a subset
of 21 cases that underwent subsequent liver transplantation.

Therefore, the accuracy of ultrasound is comparable to arterial

portography. Technical factors, however, may preclude the visualization of the portal vein and its collaterals and prevent establishment of their patency.
Ultrasonography has been supplanted by Doppler ultrasound,
which improves assessment by providing information on portal
vein patency and blood ow characteristics.85 Doppler ultrasound
and especially color ow Doppler is useful for conrming the
changes in ow patterns around the thrombus and the resumption
of normal ow pattern in follow-up imaging as the thrombosis
resolves.65,83,86
A grading system for PVT based on ultrasound ndings has been
suggested. Grade 1 PVT was dened as non-occlusive PVT with
normal liver parenchyma; grade 2 as occlusive PVT with normal
liver parenchyma; and grade 3 as occlusive PVT with ultrasonographic abnormalities of the liver parenchyma.2
The authors initially postulated an association between grade 3
PVT (occlusive PVT with parenchymal abnormalities) and poor
outcome; dened as portal hypertension or left liver lobe atrophy.
In the follow-up paper on a cohort of children, representing
a subset of the neonates in the original paper, this hypothesis was
contradicted. There did not appear to be an association between the
grade of thrombus and the risk of progressive splenomegaly. Of the
ve children found to have splenomegaly, three (60%) had grade 1
PVT on neonatal imaging.32 Although the grading system may not
be useful for prognosis, it may be of value in data collection for
prospective studies.

336

S. Williams, A.K.C. Chan / Seminars in Fetal & Neonatal Medicine 16 (2011) 329e339

8. Treatment
8.1. Impact of anticoagulation
The role of anticoagulation in PVT management is unclear. Given
that there is often rapid resolution e and even in the absence of
resolution, there may be no clinical sequelae from the liver lobe
atrophy e the question of whether anticoagulation therapy would
lead to a decrease in the rates of portal hypertension in childhood
remains unanswered.
There is an absence of prospective data on anticoagulation in the
literature. Multiple strategies have been reported in the literature
from observation, anticoagulation, and thrombolysis. In the retrospective Morag et al. study with reported resolution rates of 81/133
(61%) with complete or partial resolution and 16/133 (12%) with no
change, and no infants with extension described, anticoagulation
treatment was given to 59 infants. The authors describe no association of outcome with the use of anticoagulation. The indications
for anticoagulation treatment included the presence of a second,
occlusive thrombus with liver parenchymal changes or involving
two branches of the portal vein, post cardiac surgery. Dosages and
lengths of treatment varied signicantly.2
8.2. Absence of guidelines
There are no evidence-based guidelines on treatment of
neonatal PVT. The American College of Chest Physicians have been
unable to develop specic guidelines due to insufcient data. The
literature on neonatal PVT consists largely of case reports, and the
infrequency of the diagnosis hinders obtaining evidence that would
permit informed pediatric guidelines for therapy. Differences in
etiology and pathophysiology reduce the usefulness of the
extrapolation of therapeutic guidelines for adults with PVT to
children, much less neonates.87
Although there are insufcient data to make strong recommendations regarding anticoagulation therapy for neonatal PVT,
the options include conventional anticoagulation therapy in ageappropriate doses, short term anticoagulation therapy, or close
monitoring of the thrombus with objective tests and the use of
anticoagulation therapy if thrombus extension occurs. The treatment in each neonate should be individualized with consideration
of the risk:benet ratio. Importantly, there is no current evidence
that anticoagulation therapy would lead to improved outcome
following neonatal PVT, namely reduction in the incidence of portal
hypertension.
8.3. Current recommendations for anticoagulation in neonates
There are several Grade 2C or weak treatment recommendations based on low quality evidence.88 The guiding principles in
light of the lack of evidence would be to avoid anticoagulation if
possible and treat for the minimum duration possible. Anticoagulation treatment with unfractionated heparin (UFH) or low
molecular weight heparin (LMWH) can be initiated. Alternatively,
radiographic monitoring with anticoagulation therapy if extension
occurs can be considered. Initial anticoagulation would be appropriate in the absence of contraindications such as bleeding, coagulopathy, intracranial hemorrhage or cerebral infarct. If
anticoagulation is selected, initial administration would be in the
form of UFH or LMWH, and subsequent LMWH for 10 days to 3
months.87
If no therapy is chosen, interval imaging to rule out an extending
thrombosis would be appropriate. Imaging at the end of the initial
phase of therapy would be prudent to decide on need for further

therapy. Imaging should be repeated after completion of anticoagulation therapy to rule out extension when off therapy.
Treatment of symptomatic acute PVT, extending to the main
portal vein, by means of regional streptokinase infusion has been
reported.89 However, aggressive therapy is not likely warranted in
the majority of cases. Thrombolytic therapy should be reserved for
major vessel occlusion with organ compromise. In contrast to the
ischemic symptoms caused by thromboembolism of the arteries,
PVT seldom causes clinical problems during the neonatal period,
may resolve without any intervention, and even unresolved
thrombosis is expected to remain silent if it does not propagate into
other branches. Therefore unless the PVT extends into the adjacent
vessels e inferior vena cava, renal veins or right atrium e with
clinical symptoms, the risk of thrombolysis likely outweighs the
benet. In order to ensure adequate plasminogen to maximize
efcacy of thrombolytic therapy, supplementation with plasminogen via the use of fresh frozen plasma should occur immediately
prior to thrombolysis.
8.4. Considerations for anticoagulation
After initial treatment with UFH or LMWH, vitamin K antagonists (VKA) with a target INR of 2e3 could theoretically be used to
continue anticoagulation therapy. However, the use of warfarin in
the neonatal period is problematic. Formula-fed infants will receive
large amounts of vitamin K, and will be warfarin resistant. Dosing
becomes difcult as there is no commercially available liquid
formulation of warfarin.90 Given the anticipated difculties, anticoagulation with warfarin would not be recommended in the
neonatal period.
LMWH may be preferable to UFH if anticoagulation therapy is
used to treat neonatal PVT, given the predictable pharmacokinetics,
reduced monitoring requirements, and possibly decreased rate of
major bleeding. In a randomized, controlled trial of therapeutic
anticoagulation in children, LMWH therapy was compared with
UFH and VKA for the treatment of venous thromboembolic events
in children. There was a major bleeding rate of 12.5% in the UFH/
VKA versus 5.6% in the LMWH arm.91 A major bleeding rate of 0.7%
of children treated with the low molecular weight heparin, enoxaparin, has been reported.92 The rate of major bleeding in children
aged <3 months receiving enoxaparin has been reported as 3%.93
Neonates have an increased dose requirement compared to
older infants and children. Increased dose requirement in neonates
is multifactorial. Anticoagulant proteins are decreased in the
neonate when compared to adult levels.94,95 The levels of natural
anticoagulants such as antithrombin and protein C are decreased to
35e50% of adult levels during the neonatal period.95,96 Neonates
also have an increased volume of distribution which will accentuate
the volume of distribution for hydrophilic drugs such as the
commonly used low molecular weight heparin, enoxaparin.97
Preterm infants tend also to have higher dose requirements than
term infants.98,99 Therefore, a higher low molecular weight heparin
dosing than in older children should be considered to facilitate
faster time to therapeutic levels, and to decrease the number of
venipunctures required for monitoring.100
Apart from bleeding, there are other potential side-effects of
anticoagulation. Heparin-induced thrombocytopenia (HIT) is an
immune-mediated thrombocytopenia accompanied by a paradoxical prothrombotic state, occurring after heparin exposure.101 The
incidence of HIT in pediatric intensive care units is 1.5e2.3%.102 The
incidence of HIT in neonates is reportedly lower. In 108 neonates
receiving heparin for more than 5 days, none developed HIT.103 HIT
occurs less frequently following LMWH exposure than UFH exposure.104 However, HIT has been reported in children following the
use of enoxaparin.105 Heparin use has also been associated with

S. Williams, A.K.C. Chan / Seminars in Fetal & Neonatal Medicine 16 (2011) 329e339

osteopenia and osteoporosis. There is evidence that prolonged

heparin therapy is associated with osteopenia and therefore prolonged therapy should be avoided.106
8.5. Prevention of neonatal PVT
Reducing the risk factors for PVT in an attempt to prevent PVT
may not be feasible in a neonate with a need for central venous
access. Nevertheless, prevention strategies would include placement of the umbilical venous catheter in initial correct position,
removal of the umbilical venous catheter within one week, and
avoidance of transfusion through the umbilical venous catheter.
Given the lack of evidence for anticoagulation, preventive strategies
would be prudent.
9. Summary
Neonatal PVT is an increasingly recognized event. There may be
no symptoms or laboratory abnormalities on bloodwork. Some

337

patients may have accompanying thrombocytopenia or liver serum

biochemical abnormalities. As a result of the lack of symptoms, the
diagnosis may not be suspected. Diagnosis is via Doppler ultrasound. Umbilical catheterization and sepsis are risk factors for
neonatal PVT. Thrombophilia is possibly a contributing risk factor.
However, neonatal PVT may still occur in the absence of risk factors.
From the available literature, there appears to be a good outcome in
the majority of cases followed up to 8 years of age. Non-occlusive
thrombosis is more likely to resolve than occlusive thrombosis.
Most patients likely do not require treatment, as resolution occurs
without specic treatment. There is a mean time to resolution of
days to months, dependent on how frequently imaging is
completed. Approximately 40% of patients with PVT in the neonatal
period do not have rapid resolution, but the vast majority will not
have clinically signicant sequelae. Liver atrophy is common
following PVT, and appears to be of little clinical signicance in the
childhood years. However, there is a subset of patients (w5%) that
will go on to develop the more serious complication of portal
hypertension in childhood. Unfortunately, there are no current

PVT

Remove UVC,
HUS, INR,
aPTT, FBN,
PLT, LFTs

Coagulopathy

No Coagulopathy

No ACT
ACT UFH

US @ 5-7 days

Extension

ACT LMWH

US @ 10 days

Resolution
Stable/smaller
on
3 images

Resolution

Stable/Smaller

Stop ACT

ACT X 3 mos
LMWH

US @ 5-7 days

No PVT

Extension
IVC/RA/RV
with
symptoms

Consider tPA

Extension

Ensure adequate ACT

ACT

PVT

Thrombophilia work-up 6-12 mos

US to R/O PHTN yearly X 5
Fig. 2. Neonatal portal vein thrombosis management. PVT, portal vein thrombosis; ACT, anticoagulation therapy; aPTT, activated partial thromboplastin time; FBN, brinogen; HUS,
head ultrasound to rule out hemorrhage; INR, international normalized ratio; IVC/RA/RV, inferior vena cava/right atrial/right ventricular; LFT, liver function tests; LMWH, low
molecular weight heparin; PHTN, portal hypertension; PLT, platelets; PVT, portal vein thrombosis; tPA, tissue plasminogen activator; US, ultrasound; UFH, unfractionated heparin;
UVC, umbilical vein catheter; mos, months.

338

S. Williams, A.K.C. Chan / Seminars in Fetal & Neonatal Medicine 16 (2011) 329e339

dening features present during the neonatal period to identify and

dene neonates at risk for portal hypertension. There is no current
evidence that anticoagulation therapy makes a difference to resolution rates or incidence of portal hypertension. However, given the
potential morbidity of portal hypertension, it may be prudent to
consider anticoagulation therapy in the absence of coagulopathy
(Fig. 2). Neonates should be followed for at least 5 years after PVT to
monitor for the development of portal hypertension in an attempt
to avoid presentation with gastrointestinal hemorrhage in
childhood.

Practice points
 Neonatal PVT is increasingly recognized.
 Neonatal PVT may have no signs or symptoms.
 Diagnosis of neonatal PVT is via colour Doppler
ultrasound.
 Umbilical venous catheters and sepsis are risk factors
for neonatal PVT.
 60e70% of neonatal PVT may have spontaneous resolution in the neonatal period.
 Non-occlusive PVT are more likely to resolve than
occlusive PVT.
 Non-resolving neonatal PVT are associated with liver
lobe atrophy, which may be of no clinical significance.
 A small subset of neonates with PVT will develop portal
hypertension in childhood.
 The role of anticoagulation in the management of
neonatal PVT is unclear.
 The role of thrombophilia in neonatal PVT is unclear.
Conict of interest statement
None declared.
Funding sources
None.
References
1. Heller C, Schobess R, Kurnik K, et al. Abdominal venous thrombosis in
neonates and infants: role of prothrombotic risk factors e a multicentre
caseecontrol study. For the Childhood Thrombophilia Study Group. Br J
Haematol 2000;111:534e9.
2. Morag I, Epelman M, Daneman A, et al. Portal vein thrombosis in the neonate:
risk factors, course, and outcome. J Pediatr 2006;148:735e9.
3. Alvarez F, Bernard O, Brunelle F, Hadchouel P, Odievre M, Alagille D. Portal
obstruction in children. I. Clinical investigation and hemorrhage risk. J Pediatr
1983;103:696e702.
4. ODonnell B, Moloney MA. Development and course of extrahepatic portal
obstruction in children. Lancet 1968;1:789e91.
5. Thompson EN, Sherlock S. The aetiology of portal vein thrombosis with
particular reference to the role of infection and exchange transfusion. Q J Med
1964;33:465e80.
6. Glassman MS, Klein SA, Spivak W. Evaluation of cavernous transformation of
the portal vein by magnetic resonance imaging. Clin Pediatr (Phila)
1993;32:77e80.
7. Fisher MR, Wall SD, Hricak H, McCarthy S, Kerlan RK. Hepatic vascular
anatomy on magnetic resonance imaging. Am J Roentgenol 1985;144:739e46.
8. Wilson KW, Robinson DC, Hacking PM. Portal hypertension in childhood. Br J
Surg 1969;56:13e22.
9. Brady L, Magilavy D, Black DD. Portal vein thrombosis associated with antiphospholipid antibodies in a child. J Pediatr Gastroenterol Nutr
1996;23:470e3.
10. Arav-Boger R, Reif S, Bujanover Y. Portal vein thrombosis caused by protein C
and protein S deciency associated with cytomegalovirus infection. J Pediatr
1995;126:586e8.
11. Skarsgard E, Doski J, Jaksic T, et al. Thrombosis of the portal venous system
after splenectomy for pediatric hematologic disease. J Pediatr Surg
1993;28:1109e12.
12. Laishram H, Cramer B, Kennedy R. Idiopathic acute portal vein thrombosis:
a case report. J Pediatr Surg 1993;28:1106e8.
13. Arnold KE, Char G, Serjeant GR. Portal vein thrombosis in a child with
homozygous sickle-cell disease. West Indian Med J 1993;42:27e8.

14. Kowal-Vern A, Radhakrishnan J, Goldman J, Hutchins W, Blank J. Mesenteric

and portal vein thrombosis after splenectomy for autoimmune hemolytic
anemia. J Clin Gastroenterol 1988;10:108e10.
15. Pinkerton JA, Holcomb Jr GW, Foster JH. Portal hypertension in childhood. Ann
Surg 1972;175:870e86.
16. Rehan V, Seshia MM. Complications of umbilical vein catheter. Eur J Pediatr
1994;153:141.
17. Schwartz DS, Gettner PA, Konstantino MM, et al. Umbilical venous catheterization and the risk of portal vein thrombosis. J Pediatr 1997;131:760e2.
18. Kooiman AM, Kootstra G, Zwierstra RP. Portal hypertension in children due to
thrombosis of the portal vein. Neth J Surg 1982;34:97e103.
19. Stiller RJ, Neale D, Schwartz D, Laifer S, Kleinman G. Prenatal diagnosis of
portal vein thrombosis by ultrasound. Ultrasound Obstet Gynecol
2003;22:295e8.
20. Ries M, Zenker M, Kandler C, Rauch R, Fischer E. Severe bleeding diathesis in
a premature baby with extensive hepatic necrosis due to portal vein thrombosis of prenatal onset. Ann Hematol 1999;78:339e40.
21. Rubabaza P, Persadie RJ. Two cases of umbilical vein thrombosis, one with
associated portal vein thrombosis. J Obstet Gynaecol Can 2008;30:338e43.
22. Balfour QW, Stewart TG. Case of enlarged spleen complicated with ascites
both depending upon varicose dilatation and thrombosis of the portal vein.
Edinb Med J 1869;14:589e98.
23. Pick L. Total hemangiomatous obliteration of the portal vein. Virchows Arch;
1909:197.
24. Klemperer P. Cavernous transformation of the portal vein. Arch Pathol
1928;6:353e77.
25. Wallgren A. Contribution to the study of the splenomegalies of childhood.
Acta Paediatr 1927;6:1.
26. Somers K. Cavernous transformation of the portal vein following umbilical
sepsis. Br Med J 1957;2:335e6.
27. Hsia DY, Gellis SS. Portal hypertension in infants and children. Am J Dis Child
1955;90:290e1.
28. Clatworthy Jr HW, Boles Jr ET. Extrahepatic portal bed block in children:
pathogenesis and treatment. Ann Surg 1959;150:371e83.
29. Shaldon S. The investigation of portal hypertension in childhood. Proc R Soc
Med 1962;55:769e70.
30. Gibson JB, Richards RL. Cavernous transformation of the portal vein. J Pathol
Bacteriol 1955;70:81e96.
31. Parker RA, Seal RM. Cavernous transformation of the portal vein. J Pathol
Bacteriol 1955;70:97e103.
32. Morag I, Shah PS, Epelman M, Daneman A, Strauss T, Moore AM. Childhood
outcomes of neonates diagnosed with portal vein thrombosis. J Paediatr Child
Health; 2011.
33. Sarin SK, Agarwal SR. Extrahepatic portal vein obstruction. Semin Liver Dis
2002;22:43e58.
34. Habif DV. Treatment of esophageal varices by partial esophagogastrectomy
and interposed jejunal segment. Surgery 1959;46:212e37.
35. Tizard JP. Portal hypertension following exchange transfusion through the
umbilical vein. Proc R Soc Med 1962;55:772.
36. Oski FA, Allen DM, Diamond LK. Portal hypertension e a complication of
umbilical vein catheterization. Pediatrics 1963;31:297e302.
37. Kim JH, Lee YS, Kim SH, Lee SK, Lim MK, Kim HS. Does umbilical vein catheterization lead to portal venous thrombosis? Prospective US evaluation in
100 neonates. Radiology 2001;219:645e50.
38. Ades A, Sable C, Cummings S, Cross R, Markle B, Martin G. Echocardiographic evaluation of umbilical venous catheter placement. J Perinatol
2003;23:24e8.
39. Kunad T. On difculties and dangers in catheterization of the umbilical vein.
Kinderarztl Prax 1967;35:293e300.
40. Rosen MS, Reich SB. Umbilical venous catheterization in the newborn: identication of correct positioning. Radiology 1970;95:335e40.
41. Junker P, Egeblad M, Nielsen O, Kamper J. Umbilical vein catheterization and
portal hypertension. Acta Paediatr Scand 1976;65:499e504.
42. Scott JM. Iatrogenic lesions in babies following umbilical vein catheterization.
Arch Dis Child 1965;40:426e9.
43. Alvarez F. Risk of portal obstruction in newborns. J Pediatr 2006;148:715e6.
44. Glimelius B, Busch C, Hook M. Binding of heparin on the surface of cultured
human endothelial cells. Thromb Res 1978;12:773e82.
45. Pottecher T, Forrler M, Picardat P, Krause D, Bellocq JP, Otteni JC. Thrombogenicity of central venous catheters: prospective study of polyethylene, silicone and polyurethane catheters with phlebography or post-mortem
examination. Eur J Anaesthesiol 1984;1:361e5.
46. Yadav S, Dutta AK, Sarin SK. Do umbilical vein catheterization and sepsis lead
to portal vein thrombosis? A prospective, clinical, and sonographic evaluation.
J Pediatr Gastroenterol Nutr 1993;17:392e6.
47. Guimaraes H, Castelo L, Guimaraes J, et al. Does umbilical vein catheterization
to exchange transfusion lead to portal vein thrombosis? Eur J Pediatr
1998;157:461e3.
48. Larroche JC. Umbilical catheterization: its complications. Anatomical study.
Biol Neonate 1970;16:101e16.
49. Image interpretation session: 1995. Case 7. Hepatic hematoma and left portal
vein thrombosis due to an inappropriately placed umbilical vein catheter.
Radiographics 1996;16:230e2.
50. Tanke RB, van Megen R, Daniels O. Thrombus detection on central venous
catheters in the neonatal intensive care unit. Angiology 1994;45:477e80.

S. Williams, A.K.C. Chan / Seminars in Fetal & Neonatal Medicine 16 (2011) 329e339
51. Khilnani P, Goldstein B, Todres ID. Double lumen umbilical venous catheters
in critically ill neonates: a randomized prospective study. Crit Care Med
1991;19:1348e51.
52. Schmidt B, Zipursky A. Thrombotic disease in newborn infants. Clin Perinatol
1984;11:461e88.
53. Wigger HJ, Bransilver BR, Blanc WA. Thromboses due to catheterization in
infants and children. J Pediatr 1970;76:1e11.
54. Gharehbaghi MM, Nemati M, Hosseinpour SS, Taei R, Ghargharechi R.
Umbilical vascular catheter associated portal vein thrombosis detected by
ultrasound. Indian J Pediatr 2011;78:161e4.
55. Sakha SH, Rafeey M, Tarzamani MK. Portal venous thrombosis after umbilical
vein catheterization. Indian J Gastroenterol 2007;26:283e4.
56. Tchuenbou J, Bacq Y, Fimbel B, Metman EH. [Portal vein thrombosis associated
with factor V Leiden mutation in a woman who underwent exchange transfusion at birth]. Gastroenterol Clin Biol 2003;27:645e7.
57. Fraser-Hill MA, Atri M, Bret PM, Aldis AE, Illescas FF, Herschorn SD. Intrahepatic portal venous system: variations demonstrated with duplex and color
Doppler US. Radiology 1990;177:523e6.
58. Seibert JJ, Taylor BJ, Williamson SL, Williams BJ, Szabo JS, Corbitt SL. Sonographic detection of neonatal umbilical-artery thrombosis: clinical correlation. Am J Roentgenol 1987;148:965e8.
59. Colburn MD, Gelabert HA, Quinones-Baldrich W. Neonatal aortic thrombosis.
Surgery 1992;111:21e8.
60. Tooley WH. What is the risk of an umbilical artery catheter? Pediatrics
1972;50:1e2.
61. Krafte-Jacobs B, Sivit CJ, Mejia R, Pollack MM. Catheter-related thrombosis in
critically ill children: comparison of catheters with and without heparin
bonding. J Pediatr 1995;126:50e4.
62. Butler-OHara M, Buzzard CJ, Reubens L, McDermott MP, DiGrazio W,
DAngio CT. A randomized trial comparing long-term and short-term use of
umbilical venous catheters in premature infants with birth weights of less
than 1251 grams. Pediatrics 2006;118:e25e35.
63. Devlieger H, Snoeys R, Wyndaele L, et al. Liver necrosis in the new-born
infant: analysis of some precipitating factors in neonatal care. Eur J Pediatr
1982;138:113e9.
64. Parker MJ, Joubert GI, Levin SD. Portal vein thrombosis causing neonatal
cerebral infarction. Arch Dis Child Fetal Neonatal Ed 2002;87:F125e7.
65. Stringer DA, Krysl J, Manson D, Babiak C, Daneman A, Liu P. The value of
Doppler sonography in the detection of major vessel thrombosis in the
neonatal abdomen. Pediatr Radiol 1990;21:30e3.
66. Mitra SK, Kumar V, Datta DV, et al. Extrahepatic portal hypertension: a review
of 70 cases. J Pediatr Surg 1978;13:51e7.
67. Schettino GC, Fagundes ED, Roquete ML, Ferreira AR, Penna FJ. Portal vein
thrombosis in children and adolescents. J Pediatr (Rio J) 2006;82:171e8.
68. Webb LJ, Sherlock S. The aetiology, presentation and natural history of extrahepatic portal venous obstruction. Q J Med 1979;48:627e39.
69. Lykavieris P, Gauthier F, Hadchouel P, Duche M, Bernard O. Risk of gastrointestinal bleeding during adolescence and early adulthood in children with
portal vein obstruction. J Pediatr 2000;136:805e8.
70. Hann LE, Getrajdman GI, Brown KT, et al. Hepatic lobar atrophy: association
with ipsilateral portal vein obstruction. Am J Roentgenol 1996;167:1017e21.
71. Demirel N, Aydin M, Zenciroglu A, et al. Neonatal thrombo-embolism: risk
factors, clinical features and outcome. Ann Trop Paediatr 2009;29:271e9.
72. El-Karaksy H, El-Koofy N, El-Hawary M, et al. Prevalence of factor V Leiden
mutation and other hereditary thrombophilic factors in Egyptian children
with portal vein thrombosis: results of a single-center caseecontrol study.
Ann Hematol 2004;83:712e5.
73. Pinto RB, Silveira TR, Rosling L, Bandinelli E. [Thrombophilic disorders in
children and adolescents with portal vein thrombosis.]. J Pediatr (Rio J)
2003;79:165e72.
74. Dubuisson C, Boyer-Neumann C, Wolf M, Meyer D, Bernard O. Protein C, protein S
and antithrombin III in children with portal vein obstruction. J Hepatol
1997;27:132e5.
75. Uttenreuther-Fischer MM, Vetter B, Hellmann C, et al. Paediatric thromboembolism: the inuence of non-genetic factors and the role of activated
protein C resistance and protein C deciency. Eur J Pediatr 1997;156:277e81.
76. Uttenreuther-Fischer MM, Ziemer S, Gaedicke G. Resistance to activated
protein C (APCR): reference values of APC-ratios for children. Thromb Haemost
1996;76:813e4.
77. Pietrobattista A, Luciani M, Abraldes JG, et al. Extrahepatic portal vein
thrombosis in children and adolescents: inuence of genetic thrombophilic
disorders. World J Gastroenterol 2010;16:6123e7.
78. Bandyopadhyay R, Bandyopadhyay SK, Chatterjee U. Inherited dysbrinogenaemia presenting with portal vein thrombosis. J Indian Med Assoc
2010;108:180e2.
79. Revel-Vilk S, Chan A, Bauman M, Massicotte P. Prothrombotic conditions in an
unselected cohort of children with venous thromboembolic disease. J Thromb
Haemost 2003;1:915e21.
80. Turebylu R, Salis R, Erbe R, Martin D, Lakshminrusimha S, Ryan RM. Genetic
prothrombotic mutations are common in neonates but are not associated
with umbilical catheter-associated thrombosis. J Perinatol 2007;27:490e5.
81. Zirinsky K, Markisz JA, Rubenstein WA, et al. MR imaging of portal venous
thrombosis: correlation with CT and sonography. Am J Roentgenol
1988;150:283e8.

339

82. Weinreb JC, Cohen JM, Armstrong E, Smith T. Imaging the pediatric liver: MRI
and CT. Am J Roentgenol 1986;147:785e90.
83. Weltin G, Taylor KJ, Carter AR, Taylor CR. Duplex Doppler: identication of
cavernous transformation of the portal vein. Am J Roentgenol
1985;144:999e1001.
84. Raby N, Karani J, Powell-Jackson P, Meire H, Williams R. Assessment of portal
vein patency: comparison of arterial portography and ultrasound scanning.
Clin Radiol 1988;39:381e5.
85. Niederau C, Sonnenberg A, Muller JE, Erckenbrecht JF, Scholten T, Fritsch WP.
Sonographic measurements of the normal liver, spleen, pancreas, and portal
vein. Radiology 1983;149:537e40.
86. Berland LL, Lawson TL, Foley WD. Porta hepatis: sonographic discrimination of
bile ducts from arteries with pulsed Doppler with new anatomic criteria. Am J
Roentgenol 1982;138:833e40.
87. Monagle P, Chalmers E, Chan A, et al. Antithrombotic therapy in neonates and
children: American College of Chest Physicians Evidence-Based Clinical
Practice Guidelines (8th edition). Chest 2008;133:887Se968S.
88. Guyatt GH, Cook DJ, Jaeschke R, Pauker SG, Schunemann HJ. Grades of
recommendation for antithrombotic agents: American College of Chest
Physicians Evidence-Based Clinical Practice Guidelines (8th Edition). Chest
2008;133:123Se31S.
89. Rehan VK, Cronin CM, Bowman JM. Neonatal portal vein thrombosis
successfully treated by regional streptokinase infusion. Eur J Pediatr
1994;153:456e9.
90. Veldman A, Nold MF, Michel-Behnke I. Thrombosis in the critically ill
neonate: incidence, diagnosis, and management. Vasc Health Risk Manag
2008;4:1337e48.
91. Massicotte P, Julian JA, Gent M, et al. An open-label randomized controlled
trial of low molecular weight heparin compared to heparin and coumadin for
the treatment of venous thromboembolic events in children: the REVIVE trial.
Thromb Res 2003;109:85e92.
92. Ignjatovic V, Najid S, Newall F, Summerhayes R, Monagle P. Dosing and
monitoring of enoxaparin (low molecular weight heparin) therapy in children. Br J Haematol 2010;149:734e8.
93. Bauman ME, Belletrutti MJ, Bajzar L, et al. Evaluation of enoxaparin dosing
requirements in infants and children. Better dosing to achieve therapeutic
levels. Thromb Haemost 2009;101:86e92.
94. Andrew M, Paes B, Milner R, et al. Development of the human coagulation
system in the healthy premature infant. Blood 1988;72:1651e7.
95. Kuhle SMM. Maturation of the coagulation system during childhood. Prog
Pediatr Cardiol 2005;21:3e7.
96. Kuhle S, Male C, Mitchell L. Developmental hemostasis: pro- and anticoagulant systems during childhood. Semin Thromb Hemost 2003;29:329e38.
97. Reiter P. Neonatal pharmacology and pharmacokinetics. Neoreviews
2002;3:e229e36.
98. Malowany JI, Knoppert DC, Chan AK, Pepelassis D, Lee DS. Enoxaparin use in
the neonatal intensive care unit: experience over 8 years. Pharmacotherapy
2007;27:1263e71.
99. Streif W, Goebel G, Chan AK, Massicotte MP. Use of low molecular mass
heparin (enoxaparin) in newborn infants: a prospective cohort study of 62
patients. Arch Dis Child Fetal Neonatal Ed 2003;88:F365e70.
100. Belletrutti M, Mitchell L, Massicotte MP. Does increasing the starting dose
of enoxaparin result in better treatment of children requiring thromboprophylaxis? A retrospective cohort study [abstract]. E-PAS 2007;61.
6316e11.
101. Dager WE, White RH. Pharmacotherapy of heparin-induced thrombocytopenia. Expert Opin Pharmacother 2003;4:919e40.
102. Schmugge M, Risch L, Huber AR, Benn A, Fischer JE. Heparin-induced
thrombocytopenia-associated thrombosis in pediatric intensive care patients.
Pediatrics 2002;109:E10.
103. Klenner AF, Fusch C, Rakow A, et al. Benet and risk of heparin for maintaining peripheral venous catheters in neonates: a placebo-controlled trial.
J Pediatr 2003;143:741e5.
104. Warkentin TE, Levine MN, Hirsh J, et al. Heparin-induced thrombocytopenia
in patients treated with low-molecular-weight heparin or unfractionated
heparin. N Engl J Med 1995;332:1330e5.
105. Dager WE, White RH. Low-molecular-weight heparin-induced thrombocytopenia in a child. Ann Pharmacother 2004;38:247e50.
106. Newall F, Johnston L, Ignjatovic V, Monagle P. Unfractionated heparin therapy
in infants and children. Pediatrics 2009;123:e510e8.
107. Peter L, Dadhich SK, Yachha SK. Clinical and laboratory differentiation of
cirrhosis and extrahepatic portal venous obstruction in children.
J Gastroenterol Hepatol 2003;18:185e9.
108. Seixas CA, Hessel G, Ribeiro CC, Arruda VR, Annichino-Bizzacchi JM. Factor V
Leiden is not common in children with portal vein thrombosis. Thromb
Haemost 1997;77:258e61.
109. Schobess R, Junker R, Auberger K, Munchow N, Burdach S, Nowak-Gttl U.
Factor V G1691A and prothrombin G20210A in childhood spontaneous
venous thrombosis e evidence of an age-dependent thrombotic onset in
carriers of factor V G1691A and prothrombin G20210A mutation. Eur J Pediatr
1999;158(Suppl. 3):S105e8.
110. Kim JH, Lee YS, Kim SH, et al. Does umbilical vein catheterization lead to
portal venous thrombosis? Prospective US evaluation in 100 neonates. Radiology 2001;219:645e50.