Pediatric Anesthesia ISSN 1155-5645

ORIGINAL ARTICLE

Upper body central venous catheters in pediatric cardiac

surgery
Jeffrey W. Miller1, Dien N. Vu1, Paul J. Chai2, Janet H. Kreutzer3, J. Blaine John4, David F. Vener5
& Jeffrey P. Jacobs6
1 Division of Pediatric Cardiac Anesthesiology, Congenital Heart Institute of Florida, Tampa, USA
2 Department of Pediatric Cardiac Surgery, Johns Hopkins Children’s Heart Surgery, Tampa, USA
3 Pediatric Cardiac Surgery Outcomes Research, St. Joseph’s Children’s Hospital, Tampa, USA
4 Division of Pediatric Cardiology, Congenital Heart Institute of Florida, St. Petersburg, USA
5 Department of Anesthesia, Division of Pediatric Cardiac Anesthesia, Texas Children’s Hospital/Baylor College of Medicine, Houston, TX, USA
6 Department of Pediatric Cardiac Surgery, Johns Hopkins Children’s Heart Surgery, St. Petersburg, USA

Keywords
pediatric cardiac surgery; invasive monitors;
morbidity; CVP lines; infant; congenital heart
disease
Correspondence
Jeff Miller, Division of Pediatric Cardiac
Anesthesia, Department of Anesthesiology,
Cincinnati Children’s Medical Center, MLC
2001, 3333 Burnet Avenue, Cincinnati, OH
45229, USA
Email: Jeff.Miller@CCHMC.org
Section Editor: Greg Hammer
Accepted 20 August 2013
doi:10.1111/pan.12261
Summary
Background: A central venous catheter located in the jugular or subclavian
vein provides rapid, reliable vascular access for pediatric heart surgery. How-
ever, intravascular catheters are associated with vessel injury. Stenosis or
thrombosis of central veins in the upper body can lead to ‘superior vena cava
syndrome’ with markedly elevated venous pressures in the head and neck,
causing facial swelling and headaches. This complication may be especially
serious for patients with superior cavopulmonary (Glenn) or total cavopul-
monary (Fontan) circulation. The authors hypothesized that upper body cen-
tral line placement would be associated with a low risk of venous thrombosis
or stenosis.
Methods: A three-year retrospective review of infant and univentricular car-
diac procedures at a single institution was performed. Two hundred and
thirty-five consecutive cardiac surgical patients <1 year of age or undergoing
palliation for univentricular cardiac anatomy up to five years of age during
January 2010 to December 2012 were included in this study. Upper body cen-
tral lines are routinely placed by the anesthesiologist after induction of anes-
thesia for pediatric cardiac surgery at the study institution. The major
exception is existing central venous access via an umbilical vein or femoral
vein. Patients <2 years of age received a 4.0-French, 5-cm double-lumen cen-
tral line [Cook Medical polyurethane, no antibiotic or heparin coating].
Those over two years of age received a 5.0-French, 8-cm triple lumen central
line [Cook Medical polyurethane, no antibiotic or heparin coating]. A retro-
spective review of charts, hospital reports, echocardiographic studies, and
cardiac catheterization studies was performed.
Results: The combined population of infants <1 year of age and patients
<5 years of age with functional univentricular hearts totaled 235 patients who
underwent 261 cardiac surgical operations. In this cohort of 261 cases, 171
size 4.0 or 5.0-French upper body central lines were inserted. A total of 158
right internal jugular vein catheters were placed. Two left internal jugular
lines, two left subclavian lines, and nine right subclavian lines were placed in
this population after failure to obtain right internal jugular access. Due to the
small sample size (N = 13), the central lines not placed in the right internal
jugular vein were excluded from further review. Two cases with right internal
jugular venous lines were excluded due to death (without known stenosis or

980 © 2013 John Wiley & Sons Ltd

Pediatric Anesthesia 23 (2013) 980–988
J.W. Miller et al. Central venous catheters in cardiac surgery

thrombosis) with the line in place. Twenty-three size 4.0- or 5.0-French right
internal jugular central venous lines were placed in patients over one year of
age (range 1.1–4.3 years) having modified Glenn- or Fontan-type surgery.
The central lines were removed with a median of 1.4 days after insertion
(range 0.7–8.2 days) for these older children, compared with a median of
4.2 days of age (range 0.3–19.3 days) for the 133 children <1 year of age. Ret-
rospective chart review of nursing notes, progress notes, cardiology notes, dis-
charge summaries, echocardiographic reports, and cardiac catheterization
reports for all patients who received an upper body central venous line (inter-
nal jugular or subclavian) showed no definitive diagnosis of an upper body
venous stenosis or thrombosis related to the central venous line. A further
targeted review of echocardiographic and cardiac catheterization studies for
univentricular cardiac patients failed to show stenosis or thrombosis of a ves-
sel associated with upper body central line placement.
Conclusions: This study describes one institution’s experience with routine
upper body central venous catheter placement for neonatal and infant cardiac
surgery as well as univentricular cardiac palliation (Glenn and Fontan proce-
dures) with minimal risk of clinically significant catheter-associated vessel
thrombosis or stenosis. No upper body central venous stenosis or thrombosis
was detected in association with perioperative catheter placement in the upper
body central venous system, primarily the right internal jugular vein in 156
cases. Right internal jugular central line placement for infant cardiac surgery
can be utilized with a low risk of direct venous thrombosis or stenosis.

vena cava syndrome’). This complication is especially

Introduction
Some pediatric cardiovascular teams are reluctant to
place upper body central lines in the internal jugular
vein or subclavian vein in neonates, infants, and small
children with univentricular hearts due to a potential
risk of stenosis or thrombosis of the superior vena cava
(SVC). Percutaneous central line placement in neonates,
infants, and small children can be technically challeng-
ing. Multiple percutaneous attempts at central venous
cannulation may result in hematoma formation and
external compression or direct vascular wall injury. The
use of central venous catheters, including small percuta-
neously inserted central lines (PICCs), for prolonged
periods may result in complications including venous
stenosis, especially in neonates [1,2]. Risk factors include
longer duration, location in a subclavian vein, and
placement on the left side of the neck, possibly because
of a slower blood flow and a more tortuous venous
course with more direct vessel wall contact than the
right internal jugular vein [3]. The crossover point of
risk versus benefit for central lines in neonates, infants,
and small children undergoing cardiac surgery is
unknown.
Stenosis or thrombosis of central veins in the upper
body can lead to ‘postthrombotic syndrome’ [4] with
markedly elevated venous pressures in the head and
neck causing facial swelling and headaches (‘superior
serious for patients with superior cavopulmonary con-
nections (bidirectional Glenn anastomoses) or Fontan-
type circulation due to the obstruction of pulmonary
blood flow. If the azygous vein was ligated during the
creation of a superior cavopulmonary connection, only
small collateral venous channels remain to drain the
upper body in the event of superior vena cava obstruc-
tion. Cerebral edema, laryngeal edema, and chylothorax
can result. Postsurgical chylothorax may be associated
with innominate vein stenosis and thrombosis [5]. In
addition, in patients with biventricular circulation,
thrombosis of the SVC and innominate vein may com-
plicate the future implantation of transvenous pacing
leads [6].
Upper body venous stenosis or thrombosis is often
attributed to prior central venous catheter placement
without an adequate differential diagnosis. Manlhiot
noted an 11% incidence of arterial or venous vessel
thrombosis in pediatric cardiac surgery patients. Sixteen
percent of these intravascular thromboses were associ-
ated with an ‘indwelling access line in the vessel’ [7].
Stenosis of the upper body central veins can be congeni-
tal or acquired. Congenital stenosis of the SVC exists in
mild and severe forms [8]. Several anatomic and physio-
logic risk factors related to infant cardiac operations can
predispose patients to acquired obstruction of the SVC.
Mustard or Senning operations for transposition of the

© 2013 John Wiley & Sons Ltd 981

Pediatric Anesthesia 23 (2013) 980–988
Central venous catheters in cardiac surgery
great arteries have been associated with obstruction of
the SVC if the atrial baffle becomes stenotic [6]. The
anastomotic sites of modified Blalock–Taussig (BT)
shunts are in close proximity to the SVC where hema-
toma formation or scarring could externally compress
the upper body veins. Infant cardiac transplantation can
be associated with narrowing or even obstruction of the
SVC at the anastomosis [9]. Acute severe obstruction of
the SVC due to complications with SVC cannulation
has been reported at termination of cardiopulmonary
bypass [10]. Lesser degrees of unrecognized postsurgical
stenosis of the SVC after cannulation may exist, with
risk of delayed SVC obstruction. Physiologic risk factors
for venous occlusion include conditions with low venous
flow and hypercoagulable states [11] even in veins that
have not been catheterized [12]. Patients with superior
cavopulmonary connections or total cavopulmonary
(Fontan-type) circulations may be at increased risk due
to sluggish upper body venous flow combined with vas-
cular anastomoses and mediastinal scarring.
Vascular stenosis and thrombosis may be treated
medically with anticoagulation and thrombolytics, sur-
gically with operative thrombectomy and vascular
reconstruction, or with transcatheter thrombectomy and
implantation of dilatable stents [3]. However, prevention
by reducing any iatrogenic causes of upper body venous
stenosis is a significant priority. In the authors’ practice,
placement of upper body central venous catheters with a
strategy of rapid removal seems to be associated with a
low incidence of stenotic or thrombotic complications.
The authors hypothesized that upper body central line
placement would not be associated with a high risk of
clinically apparent venous thrombosis or stenosis.
Methods
This study included consecutive patients (235 patients,
261 operations) <1 year of age undergoing any cardiac
surgical procedure plus patients aged 1–5 years undergo-
ing bidirectional cavopulmonary anastomosis (BDCPA,
modified Glenn) or Fontan-type operations during Janu-
ary 2010 to December 2012 at a single institution. With
approval from the institutional review board, a retro-
spective review of charts, echocardiographic studies, car-
diac catheterization studies, and staff reports was
performed. A requirement for written informed consent
was waived by the institutional review board.
Method of management of central lines
At the study institution, upper body central lines are
routinely placed by the anesthesiologist after induction
of anesthesia for pediatric cardiac surgery. The major
982
J.W. Miller et al.
exception is existing central venous access via an umbili-
cal vein or femoral vein. After gowning, gloving, chlorh-
exidine skin preparation [13], and full draping, half of
the central lines were placed with continuous ultrasound
guidance (Dr. Miller) [14] and half by palpation and
landmark technique (Dr. Vu) according to physician
preference. Anesthesiologist assistant students partici-
pated in the majority of central line placements. Central
line placement was obtained using a standard Seldinger
technique with guidewire. If ultrasound was utilized,
wire placement was verified by ultrasound prior to dila-
tor insertion. The central lines were sutured in place and
covered with a bacteriostatic disk and a bio-occlusive
dressing. Patients <2 years of age received a 4.0-French,
5-cm double-lumen central line [Cook Medical polyure-
thane, no antibiotic or heparin coating]. Those over two
years of age received a 5.0-French, 8-cm triple lumen
central line [Cook Medical polyurethane, no antibiotic,
or heparin coating]. In patients with functionally uni-
ventricular hearts, a transthoracic line was placed at the
time of the cardiac operation and was left in place after
the upper body central line was removed. Placement of
both central lines allowed measurement of the upper
body venous pressure and calculation of the transpul-
monary pressure gradient after superior or total cavo-
pulmonary anastomosis.
Central lines were kept in place at the discretion of
the critical care team with a goal of rapid removal.
Patients with functionally univentricular hearts had
their percutaneous central lines removed before postop-
erative day two whenever possible. Maintenance intra-
venous fluids were administered through one port and
vasoactive medications through a second port. Unused
central line ports received a continuous flush of heparin-
ized saline (1 unit/ml) at 1 ml/hour [15]. Neonates and
infants with univentricular hearts received rectal aspirin
41 mg/day and the children aged 1–5 years (all univen-
tricular hearts) received aspirin 81 mg/day. After Bla-
lock–Taussig shunt placement, infants received a
postoperative intravenous heparin infusion of 10 units/
kg/hr. The central lines were removed in the pediatric
cardiovascular intensive care unit (PCVICU) by the bed-
side nurse, and manual pressure held on the site until
hemostasis was achieved. A sterile dressing was then
applied and left in place for two days.
Method of obtaining data
The study institution’s Congenital Cardiac Anesthesia
Society (CCAS) – Society of Thoracic Surgeons (STS)
Congenital Cardiac Anesthesia Database [16] was que-
ried for this patient population. Two senior intensive
care nurses reviewed the charts for central line location
© 2013 John Wiley & Sons Ltd
Pediatric Anesthesia 23 (2013) 980–988
J.W. Miller et al.
and removal time. Any notation of central line malfunc-
tion or complication was recorded.
Routine echocardiography was performed prior to
discharge from the PCVICU by experienced pediatric
cardiac ultrasonography technicians and reviewed by
pediatric cardiologists. The echocardiography reports of
all patients in this study were retrospectively reviewed
by the pediatric ultrasonography technicians and equiv-
ocal findings verified with a pediatric cardiologist. The
echocardiography reports were screened for findings of
upper body venous stenosis. Postoperative and dis-
charge ultrasound examinations of the upper body
venous system were reviewed. As this was not a prospec-
tive study, detailed upper body venous studies were only
available after superior cavopulmonary anastomosis or
Fontan procedures (42 cases). Ultrasound may not be
highly sensitive in detecting central line-associated
thrombus with venography being the gold standard [17].
The pre-Glenn and pre-Fontan catheterizations were
retrospectively reviewed by an experienced pediatric car-
diac interventional technician for evidence of stenosis of
the upper body central veins. Any pressure gradient
recorded between the SVC and the pulmonary artery or
Fontan conduit was noted. Upper body angiograms
were viewed for evidence of stenosis or thrombosis. The
full catheterization report was reviewed.
Method of analyzing data
Data were grouped according to the type of surgical
procedure performed. For each case, the time of percu-
taneous CVL placement and removal was used to calcu-
late the duration of the CVL placement. The duration of
CVL placement for each case was used to determine rel-
evant statistics for each age group, type of surgical pro-
cedure performed, and surgical severity score. Data
were analyzed using computerized spreadsheet software
(MICROSOFT EXCEL).
Results
The study population included consecutive patients
<1 year of age undergoing pediatric heart surgery at a
single institution between January 2010 and December
2012, excluding patients weighing <2.5 kg undergoing
ligation of the patent arterial duct (PDA) in the neonatal
intensive care unit. The population also included consec-
utive patients during the same period up to five years of
age undergoing Glenn- or Fontan-type procedures for
palliation of functionally univentricular hearts. These
patients were thought to be at highest risk of central
venous thrombosis or stenosis. The combined popula-
tion included 235 patients who underwent 261 cardiac
© 2013 John Wiley & Sons Ltd
Pediatric Anesthesia 23 (2013) 980–988
Central venous catheters in cardiac surgery
surgical operations. Eighteen patients had two proce-
dures, and four patients had three procedures on sepa-
rate hospital admissions during the three-year study
period.
In this cohort of 261 cases, we placed 171 size 4.0- or
5.0-French upper body central lines. Table 1 describes
the patient characteristics and median catheter duration
by procedure type. These upper body central lines were
removed a median of 4.0 days (range 0.3–19.2 days)
after insertion on the day of surgery. Patients undergo-
ing the Norwood procedure had their upper body cen-
tral lines removed a median of 5.2 days after insertion
(range 2.0–17.2 days). Patients having BDCPA proce-
dures had their upper body central lines removed a med-
ian of 1.3 days after insertion (range 0.3–5.2). Patients
having a Fontan procedure had their upper body central
lines removed at a median of 1.6 days after insertion
(range 0.7–8.2). A total of 158 right internal jugular vein
catheters were placed. The catheter duration for only
the right internal jugular central venous lines is
described in Figure 1.
Two left internal jugular lines, two left subclavian
lines, and nine right subclavian lines were placed in this
population after failure to obtain right internal jugular
access in infants <1 year of age. No evidence of upper
body stenosis or thrombosis was detected for these thir-
teen cases. However, due to the small sample size of cen-
tral lines placed in the left internal jugular or a
subclavian vein (N = 13), these cases were excluded
from further statistical analysis. Two patients died with
their right internal jugular venous catheters in place
(33.2 days ECMO after obstructed TAPVC repair,
7.0 days Fontan with a single lung), and their data were
excluded. Neither case was associated with a recognized
upper body venous stenosis or thrombosis.
During calendar years 2010–2012, twenty-four size
4.0- or 5.0-French upper body central lines (all RIJ)
were placed in patients over one year of age (range 1.1–
4.3 years) having BDCPA (modified Glenn surgery) or
TCPA (Fontan-type surgery). The central lines were
removed with a median of 1.4 days after insertion (range
0.7–8.2 days) for these 24 older children, compared with
a median of 4.2 days age (range 0.3–19.3 days) for the
133 right internal jugular lines placed in children <1 year
of age (Table 2).
The data were analyzed using the Society of Thoracic
Surgeons – European Association for Cardio-Thoracic
Surgery Congenital Heart Surgery (STS-EACTS) Mor-
tality Categories (‘STAT’ Mortality Categories) [18].
These are 5 categories that serve as the main complexity
adjustment tool for the STS Congenital Heart Surgery
Database. Category 1 has a lower risk and category 5
the higher risk of perioperative mortality. The 156
983
Central venous catheters in cardiac surgery J.W. Miller et al.

Table 1 Upper body central line placed on the day of surgery for 261 consecutive cases

Patient Patient Count of upper Upper body Upper body CVL

Total case weight — age — body CVL placed CVL duration— duration—
Procedure count median (kg) median (days) on day of surgery median (days) range (days)

Double-outlet right ventricle 7 4.0 112 7 (7 RIJ) 6.4 2.8–19.3

PDA closure 7 5.6 148 1 (0 RIJ) 1.3 –
Total anomalous pulmonary 7 4.7 63 6 (6 RIJ) 5.3a 3.3–14.1a
venous connectiona
Right ventricular outflow tract 11 6.1 208 6 (6 RIJ) 3.5 2.0–13.6
Pulmonary artery augmentation 5 3.0 7 2 (1 RIJ) 4.0 –
Modified Blalock–Taussig shunt 11 2.7 11 3 (3 RIJ) 9.2 3.3–12.1
Vascular ring repair 11 6 11 0 – –
Norwood-type procedures 14 3.2 6 8 (7 RIJ) 5.2 2.0–17.2
Aortic coarctation 16 4.0 35 1 (1 RIJ) 1.3 –
Other 17 5.0 180 6 (5 RIJ) 6.6 31. –11.1
Hypoplastic/interrupted aortic arch repair 19 3.3 11 11 (9 RIJ) 4.0 1.1–11.3
Tetralogy of Fallot 19 5.9 112 15 (15 RIJ) 5.3 1.7–13.4
Arterial switch procedures 20 3.3 6 15 (13 RIJ) 3.2 1.2–9.1
Total cavopulmonary anastomosis 22 13.1 984 21 (21 RIJ) 1.6a 0.7–8.2a
(Fontan-type procedures)
Atrioventricular septal defect repairs 23 4.8 144 19 (18 RIJ) 5.3 1.3–14.5
Cavopulmonary anastomosis 23 7.2 205 21 (19 RIJ) 1.3 0.3–5.2
(bidirectional Glenn procedure)
Ventricular septal defect repairs 34 5.7 169 29 (27 RIJ) 3.7 1.1–11.2
Total study population 261 4.8 117 171 (158 RIJ) 4.0 0.3–19.3
a
Excludes one infant after TAPVC repair that died with RIJ line in place at 33.2 days. Excludes one patient after Fontan procedure (hypoplastic
left heart syndrome, single lung) that died with RIJ line in place at 7 days. Neither patient had an apparent upper body venous thrombosis. Upper
body catheters placed in left internal jugular or subclavian vein are included (N = 13) in this table.

patients in our study who received right internal jugular
central lines on the day of surgery had a median STAT
category of 2. The duration of RIJ upper body central
line placement by STAT category is provided in Table 3
and shown in Figure 2.
Retrospective chart review of nursing notes, progress
notes, cardiology notes, and discharge summaries for all
patients who received an upper body central venous line
showed no diagnosis of an upper body venous stenosis
or thrombosis related to the central venous line. All lines
were removed while still functioning except four which
were removed with the notation that they ‘did not flush’
or ‘did not aspirate’. Chart review and recorded ultra-
sound study review did not demonstrate any evidence of
SVC stenosis or SVC thrombosis in these four patients.
None of these four patients had a subsequent cardiac
catheterization. One patient in the study with hypoplas-
tic left heart syndrome (aortic atresia, mitral stenosis,
intact ventricular septum) was noted by ultrasound to
have no RIJ at the time of his Norwood procedure (fem-
oral line utilized), BDCPA procedure (LIJ central line
utilized), and intervening cardiac catheterizations. He
experienced a loss of his right subclavian vein after
cannulation for a cardiac catheterization.
Data from cardiac catheterization and venograms
after removal of central lines were available for seventeen
patients in the study population. The catheterizations
were performed between 3 and 854 days after the central
lines were placed. Three of the seventeen patients were
noted to have evidence of left-sided upper body central
venous narrowing. In all three of these patients, no evi-
dence of a prior left-sided central venous line was found
in the medical record. All three of these patients had
right internal jugular central venous lines at the time of
cardiac surgery. One four-month-old patient had innom-
inate vein stenosis after a modified Glenn procedure. He
never had a left-sided central venous line. The second
patient, a two-year-old patient with univentricular heart,
had left subclavian vein occlusion (Figure 3). He never
had a left-sided central venous line for surgery and was
not documented to have had a PICC line at any time.
The third patient, a three-year-old patient for a Fontan
procedure, was noted to have ‘innominate and left inter-
nal jugular venous thrombosis’. This was present from
the time of her initial Norwood palliation at another hos-
pital where she was documented to have experienced
multiple venous thromboses including the innominate
vein, left internal jugular vein, bilateral femoral veins,
and cerebral sinus. An extensive hypercoagulable evalua-
tion was nondiagnostic at that time. She only had an
umbilical central venous line for her Norwood procedure
at the other hospital prior to our study.

984 © 2013 John Wiley & Sons Ltd

Pediatric Anesthesia 23 (2013) 980–988
J.W. Miller et al. Central venous catheters in cardiac surgery

Figure 1 Duration of Right Internal Jugular CVL placement for all
patients by procedure. Patients with functional univentricular hearts
had their RIJ central lines removed on a median of 4.4 days for Nor-
wood procedures, a median of 1.3 days for BDCPA palliation (modi-
fied Glenn procedure), and a median of 1.6 days for Fontan
procedures. BDCPA = bidirectional cavopulmonary anastomosis
(Glenn), TCPC = total cavopulmonary connections (Fontan), RVOT =
right ventricular outflow tract, ASO = arterial switch operation,
VSD = ventricular septal defect repair, AoArch = hypoplastic aortic
arch repair, AVSD = atrioventricular septal defect repair, TOF =
tetralogy of Fallot repair, TAPVC = total anomalous pulmonary
venous connection repair, DORV = double-outlet right ventricle
repair, BT_Shunt = Blalock–Taussig shunt; box represents first/third
quartile, horizontal lines represent medians, and whiskers represent
the minimum and maximum values.

Table 2 Duration of right internal jugular central line placement by removed on postoperative (POD) day 13. On POD 22,
age he was noted on echocardiogram to have an ‘atrial sep-
Median CVL tal thrombus’ at the site of the patch repair. He subse-
Population Case Count Duration (days) Range (days) quently had a femoral venous line placed and developed
an inferior vena cava thrombus. He was diagnosed with
Age < 365 days 133 4.2 0.3–19.3
a hypercoagulation disorder. Second, at an estimated
(all procedures)
Age 1–5 years 23 1.4 0.7–8.2
gestational age (EGA) of 36 weeks, newborn with body
(univentricular heart) weight 2.5 kg was found to have chylothorax and severe
Study population 156 4.0 0.3–19.3 SVC and innominate vein occlusion after an arterial
switch procedure. This patient never had an upper com-
Upper body lines placed in left internal jugular or a subclavian vein
partment central venous line. She had a postoperative
excluded (N = 13) due to small sample size. No thrombosis or steno-
sis detected in these excluded patients. Two RIJ cases excluded
femoral venous line placed in the intensive care unit and
due to death with RIJ in place and without known thrombosis or developed multiple venous thromboses. The remaining
stenosis. three reports involved localized complications related to
femoral, transthoracic, or arterial catheters with no
Hospital reports were searched, and staff members upper body central venous stenosis or thrombosis.
were interviewed for central venous line thrombotic or
stenotic complications during our study period. Five
Discussion
incidents were reported. None were clearly related to
upper body central venous lines. First, a six-month-old No location for central venous catheterization is without
patient with Noonan’s syndrome had tetralogy of Fallot potential complications. While our institution utilizes
repair with a right internal jugular central line. He devel- the upper compartment vessels predominantly, even
oped chylothorax and had a prolonged postoperative in single-ventricle patients, other programs success-
course. His right internal jugular venous catheter was fully utilize either femoral, umbilical (in neonates), or

© 2013 John Wiley & Sons Ltd 985

Pediatric Anesthesia 23 (2013) 980–988
Central venous catheters in cardiac surgery
Table 3 Duration of RIJ central line placement by STAT category
STAT Category 1 2
Median (range) in days 3.8 (1.1–11.2) 2.1 (0.3–13.6)
Case count 36 56
Upper body lines placed in left internal jugular or a subclavian vein excluded (N = 13) due to small sample size. No thrombosis or stenosis
detected in these excluded patients. Two RIJ cases excluded due to death with RIJ in place and without known thrombosis or stenosis.
Figure 2 Duration of Right Internal Jugular CVL Placement by STAT
Category. Most RIJ central lines were removed rapidly in patients
with functional univentricular hearts after BDCPA (Glenn) and TCPA
(Fontan), which are category 2 procedures. Society of Thoracic Sur-
geons – European Association for Cardio-Thoracic Surgery Congenital
Heart Surgery (STS-EACTS) Mortality Categories (‘STAT’ Mortality
Categories). Category 1 is least complex. In this patient popula-
tion, category 1 consisted of mostly VSD repairs and TOF without
surgically placed catheters for perioperative monitoring
and infusions in these patients. [19] Many programs pre-
fer utilizing infradiaphragmatic venous catheters, espe-
cially in patients under one month of age. The most
recent data from the CCAS-STS Congenital Heart Sur-
gery dataset from Spring 2013 shows that of the 2597
lines recorded in children undergoing either CPB or no-
CPB cardiovascular cases in neonates (0–30 day),
38.8% were placed in either the internal jugular or sub-
clavian vessels, while 27.2% were femoral and 32.9%
were umbilical/central. In infants (31 day–1 year,
N = 4421), 70.5% were placed in the internal jugular or
subclavian vein, while 22% were femorally placed.
[David Vener. Personal Communication. July 2013].
A central venous catheter located in the jugular or
subclavian vein provides rapid, reliable vascular access
for pediatric heart surgery. This study describes a single-
institutional experience with primarily right internal jug-
ular venous central lines placed in the operating room
986
J.W. Miller et al.
3 4 5 Total
5.0 (1.3–14.5) 5.2 (1.1–19.3) 4.4 (2.0–17.2) 4.0 (0.3–19.3)
26 31 7 156
ventriculotomy; category 2 BDCPA Glenn procedures, Fontan proce-
dures, and TOF repairs with transannular patching; category 3 com-
plete AVSD repair and arterial switch operations; category 4 aortic
arch repair, DORV, ASO with VSD, and TAPVC repairs; and category
5 Norwood procedures. Box represents first/third quartile, horizontal
lines represent medians, and whiskers represent the minimum and
maximum values.
by senior pediatric cardiac anesthesiologists for pediatric
cardiac surgery in neonates, infants, and univentricular
children with a very low incidence of upper body venous
stenosis or thrombosis. Catheter removal was planned
for 2–4 days postoperatively.
Central venous access may be obtained by internal jug-
ular, subclavian, transthoracic atrial, percutaneous intra-
venous central catheter (PICC), umbilical venous, or
femoral techniques. Each has advantages and disadvan-
tages related to placement, number of lumens, maximum
flow rate, infiltration risk, air bubble detection, vessel
injury, thrombosis, and infection. For the pediatric car-
diac anesthesiologist, an upper body central line offers
many advantages over the other venous access sites. An
upper body central venous line is easily accessible to the
anesthesiologist at the head of the operating table. Bolus
injections and infusions can be delivered with direct visu-
alization of the central line, including brief negative pres-
sure to see a ‘flash back’ of blood confirming the
© 2013 John Wiley & Sons Ltd
Pediatric Anesthesia 23 (2013) 980–988
J.W. Miller et al.
Figure 3 An occluded left subclavian vein in a patient who did not
receive a left-sided upper body PICC line or central line at any time.
Central venous occlusion may occur in patients who undergo pediat-
ric cardiac procedures with no upper body central venous access.
intravascular location of the line. The skin site of punc-
ture and soft tissues can be examined during surgery
(under the surgical drapes) for evidence of infiltration or
line dislodgement. The SVC and right atrium can be
observed by transesophageal echocardiography for cath-
eter location or turbulence and cavitation as a saline flush
is administered to verify intravascular placement.
Many risk factors exist for venous stenosis or throm-
bosis. A percutaneously placed, femoral venous, multil-
umen catheter in an infant with hyperalimentation fluid
infused is possibly the highest risk scenario [2,3,12].
These risk factors are potentially cumulative. It is
important to weigh the risks versus the benefits of cen-
tral line placement as the occurrence of upper body
venous thrombosis is a major complication, especially in
patients with functionally univentricular hearts. Knowl-
edge of the venous anatomy of the patient from prior
catheterizations and echocardiograms can be very use-
ful. Preplacement ultrasound examination may suggest
which cannulation sites are most accessible.
Rapid removal of upper body central lines may reduce
the risk of stenotic complications. Neonates, infants,
and older patients with functionally univentricular
hearts often have transthoracic catheters placed by the
surgeon directly into the right atrium or common atrium
if more than a few days of central venous access are
expected. This line tends to be maintained in the inten-
sive care unit until the patients are feeding postopera-
tively, while the upper body central line is removed as
© 2013 John Wiley & Sons Ltd
Pediatric Anesthesia 23 (2013) 980–988
Central venous catheters in cardiac surgery
soon as any immediate perioperative instability resolves.
These results show that STAT category 2 patients (pri-
marily Glenn- and Fontan-type procedures in the study
population) had their upper body central lines removed
earlier than the lower risk STAT category 1 patients
(primarily VSD repairs), consistent with the strategy for
rapid removal in the univentricular patient population.
Central venous thromboses and stenoses did occur in
this patient population in the lower body venous system.
However, upon review, upper body central venous lines
were not directly associated with these complications in
this study. One case of atrial septal patch thrombus, one
left subclavian vein occlusion (in a patient with no prior
history of left upper body central venous lines), and one
case of complete superior vena cava thrombosis (in a
patient with no prior history of upper body central
venous lines) were found. It is possible that placement of
a central venous catheter in any location may lead to
thrombosis of other vessels that are not directly cannu-
lated due to alterations in flow patterns or coagulation
factors.[11] Using the statistical ‘rule of three’ [20],
under the conditions of this study, the upper 95% confi-
dence interval of the postoperative risk of a thrombosis
or major stenosis occurring in the internal jugular vein
of cardiac infants based on the zero incidence of throm-
bosis in 156 children with right internal jugular venous
catheters in this setting is 3 of 156 or 1.9%.
The Joint Congenital Cardiac Anesthesia Society
(CCAS) - Society of Thoracic Surgeons (STS) Congeni-
tal Cardiac Anesthesia Database will allow for multiin-
stitutional analysis of upper body central venous line
complications in neonates, infants, and children after
congenital heart surgery. The next proposed version of
the STS-CCAS Database includes the complication
‘systemic vein obstruction’ to better track thrombosis or
stenosis related to central lines. However, this may not
capture incidents of venous obstruction at sites unre-
lated to catheter placement.
Limitations of this study
This study was limited by being retrospective and
restricted to one institution. No attempt to determine
whether subclinical, low-grade stenosis of the central
veins occurred after central line removal. Retrospective
review of ultrasound reports, selected ultrasound exam-
inations, and cardiac catheterization studies of the cen-
tral venous system was graded on a simple yes/no scale
for patency. This grading of low-grade stenoses could be
performed in a future prospective study using focused
ultrasound examinations before central line placement
and after central line removal. The sample size of left
internal jugular and subclavian central lines was small
987
Central venous catheters in cardiac surgery J.W. Miller et al.

(N = 13) compared with the sample size for right inter- data, Susan Wilde RN and Coretta McAlmont RN for
nal jugular central venous lines (N = 156), so these performing the chart reviews, Allie Williams RDCS and
results are best applied to right internal jugular vein Stacey Lopez RDCS for reviewing the echocardiogra-
catheterizations. phy studies, and Christopher Q. Rogers, Ph.D., for his
assistance with statistical analysis. This research was
carried out without funding.
Conclusions
Placement of right internal jugular central venous lines
Disclosures
for infant cardiac surgery can be successfully accom-
plished in the majority of patients. In this study, no Jeffrey Phillip Jacobs is Chair of the Society of Thoracic
upper body central venous stenosis or thrombosis was Surgeons (STS) Congenital Heart Surgery Database
detected in direct association with perioperative catheter Task Force. David Vener is Chair of the Database Com-
placement in the upper body central venous system, mittee Congenital Cardiac Anesthesia Society.
primarily the right internal jugular vein. In the study
institution with a consistent comprehensive management
Ethical approval
strategy, right internal jugular venous catheterization
for infant congenital heart surgery is associated with a St. Joseph’s Hospital, Tampa, FL IRB File #3305.
low risk of direct, clinically significant stenotic or throm-
botic complications.
Conflict of interest
No conflicts of interest declared.
Acknowledgements
The authors gratefully acknowledge Samuel C. Weiss
RCIS for his collection of the cardiac catheterization

References
1 Dhande V, Kattwinkel J, Alford B. Recur-
rent bilateral pleural effusions secondary to
superior vena cava obstruction as a compli-
cation of central venous catheterization.
Pediatrics 1983; 72: 109–113.
2 Sellilto M, Messina F. Central venous cathe-
terization and thrombosis in newborns:
update on diagnosis and management. J
Matern Fetal Med 2012; 25: 26–28.
3 Agarwal A, Patel BM. Central vein stenosis:
a nephrologist’s perspective. Semin Dial
2007; 20: 53–62.
4 Goldenberg NA, Donadini MP, Kahn SR
et al. Post-thrombotic syndrome in children:
a systematic review of frequency of occur-
rence, validity of outcome measures, and
prognostic factors. Haematologica 2010; 95:
1952–1959.
5 Zuluaga MT. Chylothorax after surgery for
congenital heart disease. Curr Opin Pediatr
2012; 24: 291–294.
6 Ing FF, Mullins CE, Grifka RG et al. Stent
dilation of superior vena cava and innomi-
nate vein obstructions permits transvenous
pacing lead implantation. Pacing Clin Elec-
trophysiol 1998; 21: 1517–1530.
7 Manlhiot C, Menjak IB, Brand~ ao LR et al.
Risk, clinical features, and outcomes of
thrombosis associated with pediatric
cardiac surgery. Circulation 2011; 124:
1511–1519.
8 Ro PS, Hill SL, Cheatham JP. Congenital
superior vena cava obstruction causing ana-
sarca and respiratory failure in a newborn:
successful transcatheter therapy. Catheter
Cardiovasc Interv 2005; 65: 60–65.
9 Jayakumar A, Hsu DT, Hellenbrand WE
et al. Endovascular stent placement for
venous obstruction after cardiac
transplantation in children and young adults.
Catheter Cardiovasc Interv 2002; 56: 383–386.
10 Masayuki S, Yasufumi N, Naoko I et al.
Acute development of superior vena cava
syndrome after pediatric cardiac surgery.
Pediatr Anesth 2008; 18: 997–998.
11 Gokce M, Altan I, Unal S et al. Recurrent
pediatric thrombosis: the effect of underlying
and/or coexisting factors. Blood Coagul Fibri-
nolysis 2012; 23: 434–439.
12 Gray BW, Gonzalez R, Warrier KS et al.
Characterization of central venous catheter–
associated deep venous thrombosis in
infants. J Ped Surg 2012; 47: 1159–1166.
13 Tamma PD, Aucott SW, Milstone AM.
Chlorhexidine use in the neonatal intensive
care unit: results from a national survey.
Infect Control Hosp Epidemiol 2010; 31:
846–849.
14 Wu SY, Ling Q, Cao LH et al. Real-time
two-dimensional ultrasound guidance for
central venous cannulation: a Meta-analysis.
Anesthesiology 2013; 118: 361–375.
15 Shah PS, Shah VS. Continuous heparin infu-
sion to prevent thrombosis and catheter
occlusion in neonates with peripherally
placed percutaneous central venous cathe-
ters. Cochrane Database Syst Rev 2008; 2:
CD002772.
16 Vener DF, Guzzetta N, Jacobs JP et al.
Development and implementation of a new
data registry in congenital cardiac anesthesia.
Ann Thorac Surg 2012; 94: 2159–2165.
17 Roy M, Turner-Gomes S, Gill G et al. Accu-
racy of Doppler echocardiography for the
diagnosis of thrombosis associated with
umbilical venous catheters. J Pediatr 2002;
140: 131–134.
18 O’Brien SM, Clarke DR, Jacobs JP et al. An
empirically based tool for analyzing mortal-
ity associated with congenital heart surgery.
J Thorac Cardiovasc Surg 2009; 138:
1139–1153.
19 Aiyagari R, Song JY, Donohue JE et al.
Central venous catheter-associated complica-
tions in infants with single ventricle: compar-
ison of umbilical and femoral venous access
routes. Pediatr Crit Care Med 2012; 13:
549–553.
20 Hanley JA, Lippman-Hand A. If nothing
goes wrong, is everything all right? Interpret-
ing zero numerators JAMA 1983; 249:
1743–1745.

988 © 2013 John Wiley & Sons Ltd

Pediatric Anesthesia 23 (2013) 980–988