Pediatr Radiol (2008) 38:363–378

DOI 10.1007/s00247-007-0610-2

REVIEW

Malfunctioning central venous catheters in children:

a diagnostic approach
Alex Barnacle & Owen J. Arthurs & Derek Roebuck &
Melanie P. Hiorns

Received: 22 July 2007 / Accepted: 29 July 2007 / Published online: 12 October 2007
# Springer-Verlag 2007

Abstract Central venous access is increasingly becoming emergence of image-guided vascular access techniques has
the domain of the radiologist, both in terms of the insertion brought central venous access increasingly into the domain
of central venous catheters (CVCs) and in the subsequent of interventional radiology (IR), with perceived advantages
management of these lines. This article seeks to provide an of shorter operating times, an increased likelihood of achiev-
overview of the CVC types available for paediatric patients ing access in difficult cases, fewer procedural complications,
and a more detailed explanation of the spectrum of com- lower operating costs, and a probable improvement in long-
plications that may lead to catheter malfunction. A standard term venous patency rates [1–10]. This shift from general
catheter contrast study or ‘linogram’ technique is described. surgery to radiology has, however, brought with it an
The normal appearances of such a study and a detailed increased demand on radiology departments to maintain
pictorial review of abnormal catheter studies are provided, and manage indwelling central venous catheters (CVCs). In
together with a brief overview of how information from children, meticulous attention to catheter position is vital to
catheter investigations can guide the management of cath- ensure that CVCs are kept functioning for as long as pos-
eter complications. sible in order to minimize the number of catheter replace-
ments a child may have to undergo. In some practices,
Keywords Central venous catheter . Complications . malfunctioning catheters are removed and replaced without
Children investigation; in other centres, including our institution, it
is recognized that there are a number of interventions
available to salvage blocked or displaced catheters, allow-
Introduction ing alternative access sites to be preserved for future use.
Plain radiographs and fluoroscopic contrast examinations
Central venous access forms a vital part of the management are central to the radiological investigation of malfunction-
of many paediatric conditions. For many years, tunnelled ing catheters [11, 12]. Many radiologists, however, may be
central venous catheters were largely reserved for patients perplexed by the increasingly wide variety of venous access
on chemotherapy regimens or long-term total parenteral devices available and unaware of some of the potential
nutrition (TPN), but more recently, reliable central venous CVC-related complications that develop in children or the
access has proved central to the management of an extensive catheter-salvage techniques available.
range of conditions such as those requiring long-term This article gives a brief overview of the central venous
antibiotic therapy, chemotherapy or haemodialysis. The access devices available and the potential causes of catheter
malfunction. A standard catheter contrast study or ‘linogram’
technique is described. The normal appearances of such a
A. Barnacle : O. J. Arthurs : D. Roebuck : M. P. Hiorns (*) study and a detailed pictorial review of abnormal catheter
Radiology Department, Great Ormond Street Hospital,
Great Ormond Street,
studies are provided, together with a brief overview of how
London WC1N 3JH, UK information from catheter investigations can guide the
e-mail: hiornm@[Link] management of catheter complications.
364 Pediatr Radiol (2008) 38:363–378

Background tissues, exiting the skin at an appropriate position on the

chest wall. The tunnelled, extravascular portion of the
Central venous access is required for the safe delivery of a catheter is considered to provide a barrier to ascending
number of drugs and other intravenous therapies into large- infection and with time will adhere to the subcutaneous
calibre vessels, to allow adequate dilution of the drug and tissues, decreasing the likelihood of catheter displacement.
avoid vessel irritation. Here the term ‘central’ is taken to This adherence is encouraged by the presence of a short
encompass the brachiocephalic veins, the superior vena Dacron cuff attached to the catheter, which causes a focal
cava (SVC), the suprahepatic inferior vena cava (IVC) and region of fibrosis or ingrowth in the track and thus provides
the right atrium (RA). A centrally placed catheter with a further anchorage for the catheter (Fig. 1). Insertion of such
reasonable-size lumen also allows regular noninvasive catheters almost always requires general anaesthesia in
blood sampling, which has significant benefits for both children to allow safe venous puncture. The vein puncture
the clinical team and the patient. itself can be performed using a percutaneous, US-guided
There is much debate surrounding the issue of the ideal technique or a surgical open venotomy. Removal of a cuffed
CVC tip position and a detailed presentation of this ar- catheter that has been in situ for longer than a week or so
gument is beyond the remit of this article. Suffice to say generally also requires a general anaesthetic in children, as
that the argument that a catheter tip position at the level of removal or dissection of the cuff can be traumatic. Hence,
the RA is unsafe due to the risk of myocardial perforation tunnelled CVCs are best suited for medium- to long-term
and subsequent tamponade is increasingly unsupported in venous access requirements such as chemotherapy and TPN
modern practice with the advent of softer catheter tips, with regimens. Tunnelled catheters range in calibre from 2.7F
the probable exception of 2F catheters, traditionally inserted to 12F and have a single or double lumen. Haemodialysis
by the cot-side as neonatal long lines, which have a slightly catheters are very similar to other double-lumen catheter
stiffer tip. There is much in the literature regarding cardiac devices. The lumens are generally of a larger calibre,
perforation by 2F neonatal long lines in low-birth-weight however, to allow high flow rates for efficient dialysis and
infants and it seems generally accepted that these catheters the end holes or tips may be staggered (Fig. 2).
are safer when positioned outside the pericardial reflections An alternative to a traditional tunnelled CVC is a totally
[13–20]. This paper, however, deals specifically with radio- implanted CVC or implanted venous access device (iVAD),
logically or surgically inserted Silastic (silicone elastomer) commonly known as a portacath. This comprises a reservoir
catheters of 4F calibre and above. It is now recognized that attached to a standard Silastic catheter (Fig. 3). The catheter
these larger-calibre catheters are significantly less likely to is inserted using a standard tunnelled catheter technique.
develop tip thrombus or a fibrin sheath when left to move The reservoir is buried within the subcutaneous tissues of
freely within fast-flowing blood at the level of the right the chest wall, so that the device is completely implanted.
atrium and are likely to function for a significantly longer When required, the reservoir is accessed percutaneously
period of time in this position [8, 21–25]. The position of via a specialised noncoring needle. The reservoir can be
the RA is best estimated at a level 1.5 vertebral bodies accessed up to 1,000 times before replacement is needed.
below the carina on a chest radiograph [26, 27]. It is worth Such buried devices have a lower infection rate than ex-
remembering, however, that the position of a catheter tip ternalized catheters and allow greater freedom for social
will vary widely with patient positioning, particularly in activities such as swimming, but require patients to undergo
children, and should be evaluated with caution on any
image [8, 22, 28–30].

Catheter types

There is a vast array of central venous access devices avail-

able for paediatric use; particular devices may be preferred
in different centres. In general, however, catheter subtypes
can be defined according to their site of insertion, whether
they are placed via a subcutaneous tunnel or, with particular
reference to haemodialysis catheters, according to function.
To achieve a central tip position, a catheter can be inserted
using a tunnelled or a nontunnelled technique. Tunnelled
Fig. 1 A 4F single-lumen Hickman (Bard Access Systems, Salt Lake
devices are usually inserted via access to a neck vein and are City, Utah) catheter for tunnelled central venous access. Inset
tunnelled to that site through the adjacent subcutaneous illustrates the Dacron ingrowth cuff
Pediatr Radiol (2008) 38:363–378 365

Fig. 2 A permanent (tunnelled) haemodialysis catheter with staggered

tip for optimal flow rates (12.5F Vas-Cath Soft Cell; Bard Access
Systems, Salt Lake City, Utah) Fig. 4 A triple-lumen 7F catheter (Arrow-Howes; Arrow Internation-
al, Reading, Pa.) for short-term nontunnelled central venous access

repeated needle access procedures. They are best suited for

intermittent therapies such as enzyme infusions in metabolic treatment in patients with osteomyelitis or infected neuro-
disorders, factor infusions in haemophilia or recurrent anti- surgical shunts and for repeated blood-sampling in needle-
biotic requirements for conditions such as cystic fibrosis. phobic patients, or children in whom peripheral access is
Nontunnelled CVCs are placed directly into a vein via a becoming technically difficult. PICCs are also increasingly
skin incision overlying the vein and are simply sutured to used for longer-term therapies such as chemotherapy,
the adjacent skin (Fig. 4). Access sites include the neck and pulmonary hypertension and TPN, and can function for
the extremities. Nontunnelled neck lines are commonly a year or more. Single-lumen peripherally inserted CVCs
placed for complex anaesthesia cases or in the intensive or PICCs are commonly 4F in calibre; this allows more
care setting. The neck veins can also be used for short-term reliable function than the 2F neonatal long lines, particu-
dialysis catheters in patients who do not require a longer- larly for blood sampling (aspiration via a catheter requires
term tunnelled device. For longer-term indications for a a faster flow rate in the vessel being sampled, and is
nontunnelled line, operators usually prefer to access venous therefore more sensitive to lumen calibre when compared
sites in the extremities. Small-calibre (4–5F) peripherally with infusion through the same catheter). Double-lumen 5F
inserted central catheters (PICCs) are usually placed via catheters are also widely used, although each lumen is
the veins of the upper arm (Fig. 5). In a cooperative child small and hence these catheters, although useful for
they can often be placed via a peripheral venous cannula, multidrug therapies, are less reliable for blood sampling.
without sedation or general anaesthesia, and are most useful Some types of PICC are now approved for CT contrast
for providing reliable access for medium-term therapies medium infusions via a pump injector.
in children. The catheter is advanced to the level of the
central vessels using fluoroscopic guidance in the radiology
department or a non-image-guided ‘best-estimate’ tech- Strategies in managing catheter malfunction
nique at the bedside. These catheters provide reliable
venous access for indications such as extended antibiotic CVC malfunction can be approached by the clinical team
and the radiologist in a number of ways. In some centres a

Fig. 3 A totally implantable port (Bard Access Systems, Salt Lake

City, Utah) with reservoir and 6.6F catheter. Inset illustrates the Fig. 5 A double-lumen 5F PICC (Bard Access Systems, Salt Lake
specialized noncoring port access needle City, Utah)
366 Pediatr Radiol (2008) 38:363–378

policy of removing all malfunctioning catheters without complication seems to occur most commonly in chubby
investigating the cause of the malfunction is maintained. In infants with redundant tissue in the neck (Fig. 11). Despite
most institutions, however, protocols are available for at- accurate initial placement, catheter tips can flick into an
tempted thrombolysis of blocked catheters and in some
units revision of displaced or incorrectly positioned lines is
attempted. In our opinion, an attempt should always be
made to diagnose the cause of a malfunctioning catheter,
with the aim of possible catheter salvage and, therefore,
a potential reduction in the number of both repeated an-
aesthesia and venous access attempts that a patient may
have to undergo. An institutional policy regarding the man-
agement of malfunctioning CVCs is helpful and should be
made with the facilities of that institution in mind. In
particular, contrast studies involving radiation exposure
should only be undertaken if the results of the study would
alter the subsequent management of the line. In general, an
IR department is more likely to be able to offer salvage pro-
cedures than a surgically managed service, due to differences
in both the imaging and interventional techniques available
within each speciality [3, 10, 31].

Initial examination of the catheter

It is important to examine both the catheter and its skin exit

site prior to performing any radiographic investigation, so
that a simple cause of catheter malfunction, such as a break
in the catheter itself or wear at the clamp site, can be
established. It is not uncommon to find that the skin suture of
a newly placed catheter is too tight and is restricting flow.

Fluoroscopic or plain radiographic views

Plain radiographic views of the catheter should be acquired

prior to any contrast-enhanced study. These can be acquired
as a chest radiograph or by using low-dose fluoroscopy at
the time of a planned contrast investigation. This allows
misplaced catheters (Fig. 6) and catheters of an incorrect
length to be identified (Fig. 7). The entire catheter must be
imaged. Studies limited to the intravascular portion of the
catheter will frequently miss abnormalities such as kinking
of the catheter within the skin, a finding that is particularly
common with portacaths (Fig. 8), and looping or twisting of
a catheter in the neck (Fig. 9). More than one view may be
required to confirm a kink in the catheter of a portacath;
angulation of the catheter most commonly occurs at the
origin of the subcutaneous tunnel, adjacent to the reservoir
(Fig. 10). Occasionally, a portacath reservoir can become
detached from its catheter, and children who are compul-
sive ‘port fiddlers’ can encourage a catheter to loop around
the port chamber so that over time the intravascular portion
Fig. 6 Misplaced catheter. a Fluoroscopic image of a 10-month-old
is shortened. A catheter may be accidentally pulled back or male with a portacath in situ, the tip of which has been placed into the
may migrate backwards into the soft tissues of the neck, coronary sinus. b The position of the tip is confirmed by a contrast
so that the tip lies outside the vein entirely. This second medium study
Pediatr Radiol (2008) 38:363–378 367

lying problem and obviate the need for access of the second
lumen. Water-soluble nonionic contrast medium should be
used and the operator should aim to minimize the dose of
contrast medium that the child receives.
The initial contrast medium injection should be per-
formed slowly while the entire catheter is screened so that
complications anywhere along the length of the line are
readily identified (Fig. 13). The operator should look for
signs of pooling of contrast medium in the soft tissues
surrounding the catheter. Extravasation and pooling of con-
trast medium in the neck are usually painful; a slow gentle
injection will minimize both the patient’s discomfort and
the risk of loss of the child’s confidence in the operator.
When examining a malfunctioning portacath device, tan-
gential and en face views of the port reservoir should be
obtained, to ensure that it fills uniformly and does not

Fig. 7 A chest radiograph in a 13-month-old male with a permanent

haemodialysis catheter, the tip of which lies at or beyond the level of
the tricuspid valve causing intermittently poor flow rates

incorrect position with changes in posture or intrathoracic

pressure; this occurs more commonly with catheter tips
placed at the level of the SVC rather than the RA. The
commonest site for a catheter tip to migrate to is the ipsi-
lateral internal jugular vein or the contralateral brachio-
cephalic vein (Fig. 12). Observing the catheter tip during
fluoroscopic screening, even before injecting contrast
medium, allows the operator to determine whether the tip
is moving freely or is wedged against a vessel or chamber
wall. Exaggerated movement of the catheter when in an
intracardiac position is strongly suggestive of malposition
of the catheter tip within the right ventricle; most operators
would agree that such a line should be shortened or replaced
in such a situation, even if the child is asymptomatic, due
to the risk of catheter-induced arrhythmias. Fluoroscopic
screening should be performed using a low pulse rate (three
pulses per second) and where available, the image-grab
facility should be used to minimize radiation exposure.

Performing a contrast study

If the initial examination of the line is unremarkable, a

contrast-enhanced study (‘linogram’) should be performed.
The catheter should be accessed using a sterile technique
and, where possible, any heparin solution dwelling in the
line should be aspirated and discarded. Inability to aspirate
from the catheter should be documented, as this narrows
the list of potential causes of catheter malfunction, as dis- Fig. 8 Kinking of the catheter. a Chest radiograph of a 6-year-old
male with a portacath in situ. The catheter is kinked just beyond the
cussed below. If only one lumen of a multilumen catheter is reservoir (arrow). b Fluoroscopic image of a femoral temporary
malfunctioning, imaging should first be performed through dialysis catheter in a 15-year-old male. The catheter is kinked at the
the functioning lumen as this may demonstrate the under- vessel entry site, causing poor flow rates
368 Pediatr Radiol (2008) 38:363–378

strate the cause of catheter malfunction. This may include a

significant change in tip position following a change in
posture or intermittent occlusion of the catheter lumen at the
level of the clavicle due to catheter compression between the
clavicle and the first rib (‘pinch-off’ syndrome).
If the child presents with symptoms of venous occlusion,
such as localized pain and swelling over the venous access
site or sudden swelling of the face or extremity, a formal
venogram via a peripheral cannula can be performed to
determine the presence of catheter-related venous thrombo-
sis (see below).

Fig. 9 Fluoroscopic image of a tunnelled catheter in a 2-year-old

child. The catheter has twisted in the soft tissues of the neck, causing
the catheter tip to be partially withdrawn (arrow)

contain thrombus (Fig. 14). If the initial slow injection of

contrast medium appears normal, a controlled but forceful
hand injection should then be performed to study in more
detail the flow of contrast medium away from the catheter
tip. Digital subtraction angiography (DSA) should be used
during image acquisition where possible, to achieve a diag-
nostic study with one injection, minimizing both contrast
agent and radiation dose. Cooperative patients should be
encouraged to breath-hold. If the study is performed in an
imaging suite without DSA capability, screening should be
focused on the catheter tip and the fluoroscopic pulse rate
may have to be increased to either 15 pulses per second or
‘continuous’ screening to obtain adequate images. Diag-
nostic imaging of the contrast medium jet is hardest in
small patients with rapid breathing and significant cardiac
motion.
The contrast medium jet should be of the same diameter
as the lumen of the catheter and should exit the catheter in a
linear forward direction, rapidly moving away from the
catheter tip rather than pooling at the tip. The jet should fan
out appropriately to fill the lumen of the vessel or chamber
(Fig. 15). The operator should be aware of the configura-
tion of the catheter, so that a catheter with a side hole, such
as the Groshong PICC (Bard Access Systems, Salt Lake
City, Utah) is not misinterpreted as a fractured or partially
occluded line (Fig. 16).
It is important to remember to study both lumens of a
double-lumen catheter if injection of the first lumen is normal.
In patients with intermittent catheter malfunction, attempts
Fig. 10 Kinking of the catheter. a, b AP (a) and lateral (b) views of a
should be made to recreate the situation in which problems
portacath in a 2-year-old female demonstrating sharp angulation of the
occur; often, catheter function is position-dependent and proximal catheter, limiting infusion flow rates. The kink is not clearly
placing the child in the troublesome position may demon- seen on the AP view
Pediatr Radiol (2008) 38:363–378 369

thrombus (see below). It is vital to identify this problem

during an initial slow instillation of contrast medium as a
forceful injection may damage the vessel wall.
A fibrin sheath or sleeve thrombus is caused by build-up
of fibrin material around the intravascular portion of the
catheter. This phenomenon may occur in the majority of
catheters, as early as 24 h after catheter insertion [8, 32,
33]. Fibrin sheaths cause malfunction only when the sheath
extends around or over the tip of the catheter. A complete
sheath prevents any contrast medium from flowing away

Fig. 11 Chest radiograph of a 9-month-old infant with a tunnelled

CVC in situ. The catheter has migrated back into the redundant soft
tissues of the neck and chest wall

Once the study is completed, the catheter should always

be flushed and, where required, ‘locked’ with heparin so-
lution of an appropriate concentration. This should be
documented in the clinical notes as part of the examination.
If the catheter is deemed unsafe for use, this too should be
clearly documented.

Interpreting an abnormal contrast study

Once contrast medium has been instilled, a careful exam-

ination of the entire length of the catheter should be
performed. Abnormal findings may be subtle, such as
narrowing of the lumen at the site of the skin suture or
clamp (Fig. 17). Contrast medium will leak into the soft
tissues at the site of any break in the catheter wall (Fig. 18).
In such cases the point of extravasation can often be clin-
ically predicted by the patient’s clinical symptoms (focal
swelling within the subcutaneous tunnel or leak of infusate
from the catheter exit site) and the contrast agent dose
therefore minimized.
Inability to aspirate from a catheter that otherwise
flushes well indicates one of three possible causes of
catheter malfunction: the presence of a fibrin sheath,
catheter tip thrombus, or position of the catheter tip against
a vessel or chamber wall. All three causes create a valve
effect at the catheter tip, so that the obstruction is tem-
porarily alleviated by forward pressure during infusion
but recurs with negative pressure on attempted aspiration
(Fig. 19).
Gentle contrast medium injection through a catheter that Fig. 12 Migration of the catheter tip. a Chest radiograph of a 4-year-
is resting against a vessel or chamber wall will demonstrate old male with a portacath in situ. The catheter tip lies in the right
internal jugular vein (short arrow). Note that the proximal catheter is
rapid but asymmetric or oblique flow of contrast agent also kinked (long arrow). b Chest radiograph of a 2-year-old female
away from the catheter tip (Fig. 20). This appearance with a tunnelled CVC in situ. The catheter tip lies in the left
should be distinguished from an irregular jet due to tip brachiocephalic vein
370 Pediatr Radiol (2008) 38:363–378

Fig. 21. Thrombus within the catheter lumen itself is seen

as a filling defect in the catheter during injection (Fig. 23).
Tip thrombus causes irregularity of the jet of contrast
medium as it exits the catheter and may deflect the jet away

Fig. 13 Fluoroscopic imaging of a tunnelled catheter in a 4-month-

old female. The entire catheter is imaged during slow instillation of
contrast medium

from the catheter tip; instead it tracks retrogradely along the

intravascular portion of the catheter, between the catheter
and the fibrin sheath, and then spills into the soft tissues
at the point of venous access (Fig. 21). In such cases,
the patient will present with symptomatic swelling at the
venous access site, usually the neck, during attempted
catheter flushes. It is essential to identify initial contrast
medium pooling at the catheter tip in the early phase of the
injection followed by tracking of contrast medium back
along the length of the catheter on later images, so that this
is not misinterpreted as extravasation due to catheter frac-
ture. To ensure a correct diagnosis in such cases, sequential
images should be acquired or ‘grabbed’ during a very slow
injection. In some cases, the fibrin sheath may be very
thin and it is broken or fenestrated during flushing of the
catheter. A circumferential but incomplete fibrin sheath will
cause narrowing of the contrast jet as it exits the catheter,
with delayed fanning out of contrast medium. This is a
subtle, but classic, sign of a fibrin sheath (Fig. 22). The
operator must ensure that the contrast agent jet is of a
similar diameter to that of the catheter lumen, so that a
circumferential thrombus or fibrin sheath causing partial
occlusion of the end hole is not missed.
Catheter thrombus is most likely to occur in catheters
Fig. 14 Normal portacath. a, b Tangential (a) and en face (b) views
that are inadequately or infrequently flushed, or that are of a portacath reservoir in a 7-year-old male. Note that in both views
sited in small vessels with a low volume and rate of blood there is normal opacification of the reservoir during contrast medium
flow around the tip, as has occurred in the case shown in instillation
Pediatr Radiol (2008) 38:363–378 371

from the long axis of the catheter (Fig. 24). Thrombus may
extend into the vessel itself, which will cause irregular
filling of the vessel beyond the catheter tip (Fig. 25).
Catheter-induced venous thrombosis is influenced by
catheter dwell time, the size of the catheter relative to the
vessel, the access vein used, infusate type, and the presence
of any prior venous interventions or systemic comorbidities
in the patient [34–38]. It is likely to be a relatively common
occurrence [39], although evidence for this in the literature
is skewed by variations in the diagnostic techniques and

Fig. 15 Fluoroscopic image of contrast medium flowing freely from

the catheter tip in an 8-year-old male. The contrast medium
immediately fans out to fill the right atrium

Fig. 17 Catheter narrowing. a Image acquired during contrast

injection through a double-lumen catheter in a 7-year-old male. The
catheter is damaged at the site of previous kinks or catheter clamps
(arrows). b Focal narrowing of the opacified lumen of a PICC (arrow)
Fig. 16 Normal appearance of the contrast medium jet exiting the in a 3-year-old male where the skin suture device has been attached to
side hole of a PICC the catheter too tightly
372 Pediatr Radiol (2008) 38:363–378

pathological definitions used in various studies. Many cases

of pericatheter venous thrombosis are asymptomatic and the
catheter continues to function, so this complication is likely
to be under-reported. Where they occur, symptoms include
localized pain and swelling overlying the accessed vein and
sudden facial or extremity swelling. It is important that the
clinical team rule out septic thrombophlebitis in symptom-
atic patients. Isolated facial or extremity swelling is, however,
strong evidence for acute venous thrombosis [34]. Contrast-

Fig. 19 Line drawing demonstrating the effect of a fibrin sheath. a

Flushing the catheter displaces the sheath or thrombus from the
catheter tip and allows saline to exit the lumen. b Attempted aspiration
sucks the sheath back onto the catheter tip, causing withdrawal
occlusion

enhanced catheter studies will demonstrate irregular occlu-

sion of the involved vein (see Fig. 25). In well-established
cases, drainage may occur only via tortuous collateral
vessels. It may be appropriate to perform a venogram of

Fig. 20 Fluoroscopic contrast study of a haemodialysis catheter in a

Fig. 18 Extravasation. a Extravasation of contrast medium into the 9-year-old female. The lumen with a more distal tip has been
soft tissues at the site of a catheter break. b Contrast medium is seen opacified. The catheter is too long and the distal tip lies up against
leaking from a break in the connection between the portacath reservoir the inferior wall of the right atrium, causing the contrast jet to be
and the catheter deflected
Pediatr Radiol (2008) 38:363–378 373

the affected vessel via a peripheral cannula rather than the the services of a proactive IR department. Most institutions
catheter at this stage, to assess drainage of the extremity at least have a ward-based thrombolysis protocol for occluded
(Fig. 26). Some operators may prefer to perform a venogram CVCs and these techniques play a significant role in
immediately prior to placement of a new catheter, to allow extending the life of a malfunctioning catheter [10, 40–45].
time for possible recanalization of the vessel or formation of Catheters that are found to be too long can be shortened
substantial collaterals in the interim period (Fig. 27). by a variety of techniques. The simplest involves with-
drawal of the intravascular portion of the catheter at the
venous access site while the tunnelled portion is left in situ.
Implications for catheter salvage techniques The catheter is then re-measured and shortened, prior to re-
puncture of the venous access site and replacement of the
Diagnostic catheter studies are central to the management intravascular portion of the catheter [46, 47]. If there is
of malfunctioning catheters. A detailed discussion of cath- sufficient excess length, the catheter can be cut when it is
eter salvage techniques is beyond the remit of this review; a partially withdrawn, and the tip wired to maintain access,
number of the techniques outlined below are dependent on avoiding the need for re-puncture.

Fig. 21 Fibrin sheaths. a, b

Contrast medium is instilled
into a tunnelled catheter in a 4-
month-old child. The catheter
has been partially pulled back
and the tip lies within the left
brachiocephalic vein. A fibrin
sheath occludes the catheter.
Instead of flowing away from
the catheter tip, contrast medium
tracks back along the side of the
catheter (a arrow) and extrava-
sates into the tissues of the neck
(b). c Contrast study performed
in a 2-year-old child with a
catheter occluded by a fibrin
sheath. Contrast medium tracks
back along the catheter and into
the subcutaneous tunnel, outlin-
ing the cuff, as shown diagram-
matically in the line drawing
374 Pediatr Radiol (2008) 38:363–378

Fig. 22 Fibrin sheaths. a Cath-

eter contrast study in a 5-year-
old male. Contrast medium exits
the catheter tip in a linear man-
ner, but the contrast jet is sig-
nificantly narrower than the
lumen of the catheter (arrow),
implying the presence of a cir-
cumferential, nonocclusive fi-
brin sheath. The effect of the
fibrin sheath is shown diagram-
matically. b Contrast medium
exits the tip of a portacath in an
11-year-old male. The jet fails to
fan away from the tip (arrow) of
the catheter, instead simply fill-
ing the lumen of the sheath. c A
similar appearance is seen dur-
ing a contrast study of a PICC in
another patient. An arrow marks
the catheter tip. Note the appar-
ent irregular calibre of the cath-
eter itself, a sign that contrast
medium is tracking back along
the catheter as shown in Fig. 21c

Kinks in the subcutaneous tunnel of a CVC can often be alternative technique involves venous access via a femoral
straightened by accessing either the venous access site or approach, allowing the displaced catheter to be snared and
the skin exit site of the catheter. Intervention at the venous pulled back into the appropriate vessel [10, 49–51].
access site may result in an increase in the intravascular Catheter tip thrombus can be dislodged by inserting a tip
length of the catheter, a result that is beneficial in catheters deflecting guidewire through the catheter and rotating the
that have migrated back into the tunnel or the neck. wire once it is just beyond the tip of the catheter [9, 10].
Catheter tips that have flicked out of position can occa- Such techniques, although effective, are rarely long lasting.
sionally be repositioned using one of several techniques. Fibrin sheath stripping techniques can be attempted in
Often a forceful injection of saline through the displaced patients in whom replacement of a malfunctioning catheter
catheter will encourage the tip to return to its original is undesirable, such as those with a high risk for general
position, in the direction of flowing blood [10]. Obstinate anaesthesia or limited alternative access sites. Again, fem-
catheters may respond to placement of a guidewire through oral venous access is obtained and the catheter is ensnared
the catheter lumen, which stiffens the catheter sufficiently using a standard vascular snare device (Fig. 28). Gentle
to encourage repositioning of the catheter [10, 48]. An manipulation of the snare along the length of the catheter is
Pediatr Radiol (2008) 38:363–378 375

Fig. 25 Contrast study of a left femoral catheter in a 7-month-old

Fig. 23 Image acquired during contrast medium injection through one child. The short white arrow shows the tip of the catheter. Contrast
lumen of a double-lumen catheter in a 2-year-old female. There is a medium outlines the common iliac veins and IVC. There are several
filling defect within the catheter (arrow) at the site of a focal filling defects in the IVC (long white arrows) consistent with venous
intraluminal thrombus thrombus beyond the catheter tip. Note that contrast medium also
refluxes into ascending lumbar veins on the left (black arrows)

performed in an attempt to dislodge and retrieve the fibrin

sheath [9, 10, 51–55]. Symptomatic pulmonary embolism is
unlikely, although the procedure is contraindicated in sub-
jects with a right to left shunt or severe cardiopulmonary
disease that will not tolerate a pulmonary embolus [32, 55].

Fig. 24 Thrombus. a Fluoroscopic image acquired during a catheter

contrast study in a 3-year-old female. The contrast medium jet is
deflected away from the long axis of the catheter, implying the
presence of partially occlusive thrombus at the catheter tip. The effect Fig. 26 Left arm venography in a 1-year-old female with a
of the tip thrombus is shown diagrammatically. b Thrombus is malfunctioning tunnelled catheter in situ. Contrast medium fails to
adherent to the tip of the distal lumen of a haemodialysis catheter in a fill the occluded SVC and there are irregular filling defects within the
7-year-old male. Contrast medium exits the lumen eccentrically and left brachiocephalic vein, consistent with mural thrombi. Small
outlines the thrombus collaterals are forming in the left supraclavicular region
376 Pediatr Radiol (2008) 38:363–378

Fig. 27 Right arm venogram in an 8-year-old boy performed prior to

placement of a new catheter. The study demonstrates multiple
Fig. 29 US image of the right internal jugular vein in a 12-year-old
irregular venous collaterals draining the arm following occlusion of
female. A linear filling defect is attached to the vessel wall (arrow),
the subclavian and brachiocephalic veins
likely to represent a previous fibrin sheath in this patient with a history
of multiple previous access catheters

Implications for catheter replacement techniques

of a documented fibrin sheath or venous occlusion. Resiting
The results of catheter contrast medium studies are often
a catheter into the lumen of a fibrin sheath is highly likely to
valuable in planning catheter replacement. Operators are
result in a rapid reestablishment of the sheath and subsequent
likely to alter their approach to venous access in the presence
catheter occlusion. In cases where reaccess of a vessel con-
taining a fibrin sheath is necessary, it is sensible to attempt an
US-guided puncture parallel to the fibrin sheath, rather than
into it (Fig. 29). Alternatively, the sheath can be deliberately
cannulated so that a guidewire is placed through the sheath
followed by a standard angioplasty balloon that, when
inflated, encourages break-up of the sheath or incorporation
of the sheath into the vessel wall [9, 53].
Documentation of complete venous occlusion is vital prior
to replacement of a CVC. This allows a more considered ap-
proach to catheter placement and appropriate discussion with
the family prior to a difficult procedure. Recanalization and
venous stenting techniques or use of unusual access sites
should be considered in children who would otherwise require
open surgical procedures to maintain reliable venous access.

Conclusion

Reliable central venous access is key to the management of

many paediatric conditions. The insertion of central venous
access devices is increasingly becoming the remit of the inter-
ventional radiologist, a paradigm shift that will require general
radiologists to play some role in the subsequent management
of these catheters. In institutions where the specific cause of
catheter malfunction may alter subsequent management,
Fig. 28 The distal portion of a malfunctioning haemodialysis catheter
has been snared from a femoral approach in an attempt to strip the radiologists need to be familiar with the catheter types used
fibrin sheath from the catheter and their imaging appearances, both on plain radiography
Pediatr Radiol (2008) 38:363–378 377

and on contrast investigations. Accurate diagnosis of CVC 19. Nowlen TT, Rosenthal GL, Johnson GL et al (2002) Pericardial
malfunction can significantly alter outcome for patients in effusion and tamponade in infants with central catheters. Pediatrics
110:137–142
whom repeated venous access procedures can often become a 20. Beardsall K, White DK, Pinto EM et al (2003) Pericardial effusion
source of greater morbidity than their primary condition. and cardiac tamponade as complications of neonatal long lines:
are they really a problem? Arch Dis Child Fetal Neonatal Ed 88:
F292–F295
21. Cadman A, Lawrence JA, Fitzsimmons L et al (2004) To clot or
not to clot? That is the question in central venous catheters. Clin
References Radiol 59:349–355
22. Schutz JC, Patel AA, Clark TW et al (2004) Relationship between
1. Trerotola SO, Johnson MS, Harris VJ et al (1997) Outcome of chest port catheter tip position and port malfunction after interven-
tunneled hemodialysis catheters placed via the right internal jugular tional radiologic placement. J Vasc Interv Radiol 15:581–587
vein by interventional radiologists. Radiology 203:489–495 23. Cohn DE, Mutch DG, Rader JS et al (2001) Factors predicting
2. Mauro MA, Jaques PF (1997) Radiologic placement of long-term subcutaneous implanted central venous port function: the rela-
central venous catheters: a review. J Vasc Interv Radiol 4:127–137 tionship between catheter tip location and port failure in patients
3. Rush J, Baskin KM, Cahill AM et al (2005) Central venous access with gynecologic malignancies. Gynecol Oncol 83:533–536
by different service providers: does it matter who puts in the port? 24. Petersen J, Delaney JH, Brakstad MT et al (1999) Silicone venous
(abstract). Pediatr Radiol 35 [Suppl 1]:S76 access devices positioned with their tips high in the superior vena
4. Roebuck DJ, Kleidon TM, McLaren CA et al (2005) Internal jugular cava are more likely to malfunction. Am J Surg 178:38–41
vein (IJV) patency after central venous (CV) access (abstract). Pediatr 25. Lucas H, Attard-Montalto SP, Saha V et al (1996) Central venous
Radiol 35 [Suppl 1]:S75 catheter tip position and malfunction in a paediatric oncology unit.
5. Blum AS (1999) The role of the interventional radiologist in Pediatr Surg Int 11:159–163
central venous access. J Intraven Nurs 22 [Suppl 6]:S32–S39 26. Jimenez M, Baskin KM, Cahill AM et al (2005) The elusive
6. McBride KD, Fisher R, Warnock N et al (1997) A comparative cavoatrial junction: anatomic and biometric study of the superior
analysis of radiological and surgical placement of central venous vena cava (abstract). Pediatr Radiol 35 [Suppl 1]:S59
catheters. Cardiovasc Intervent Radiol 20:17–22 27. Yoon SZ, Shin JH, Hahn S (2005) Usefulness of the carina as a
7. Lamaris JS, Post PJ, Zonderland HM et al (1990) Percutaneous radiographic landmark for central venous catheter placement in
placement of Hickman catheters: comparison of sonographically paediatric patients. Br J Anaesth 95:514–517
guided and blind techniques. AJR 155:1097–1099 28. Nazarian GK, Bjarnason H, Dietz CA et al (1997) Changes in
8. Lund GB, Trerotola SO, Scheel PF et al (1996) Outcome of tunneled catheter tip position when a patient is upright. J Vasc
tunneled hemodialysis catheters placed by radiologists. Radiology Intervent Radiol 8:437–441
198:467–472 29. Kowalski CM, Kaufman JA, Rivitz SM et al (1997) Migration of
9. Namyslowski J, Trerotola SO (2000) Interventional radiologic central venous catheters: implications for initial catheter tip
placement and management of infusion catheters. In: Savada SJ, positioning. J Vasc Intervent Radiol 8:443–447
Trerotola SO (eds) Venous interventional radiology with clinical 30. Chang TC, Funaki B, Szymski GX (1998) Are routine chest
perspectives. Thieme, New York, pp 325–346 radiographs necessary after image-guided placement of internal
10. Kidney DD, Nguyen DT, Deutsch LS (1988) Radiologic evalu- jugular central venous access devices? AJR 170:335–337
ation and management of malfunctioning long-term central vein 31. Garofalo RS, Zaleski GX, Lorenz JM et al (1999) Exchange of
catheters. AJR 171:1251–1257 poorly functioning tunneled permanent hemodialysis catheters.
11. Kim FM, Burrows PE, Hoffer FA et al (1996) Interpreting the AJR 173:155–158
results of pediatric central venous catheter studies. Radiographics 32. Weschler RJ, Spirn PW Conant EF et al (1993) Thrombosis and
16:747–754 infection caused by thoracic venous catheters: pathogenesis and
12. Boardman P, Hughes JP (1998) Radiological evaluation and imaging findings. AJR 160:467–471
management of malfunctioning central venous catheters. Clin 33. Hoshal VL, Ause RG, Hoskins PA (1971) Fibrin sleeve formation
Radiol 53:10–16 on indwelling subclavian central venous catheters. Arch Surg
13. Garg M, Chang CC, Merritt RJ (1989) An unusual case 102:353–358
presentation: pericardial tamponade complicating central venous 34. Namyslowski J (2002) Management of catheter-induced throm-
catheter. J Perinatol 9:456–457 bosis. Tech Vasc Interv Radiol 5:85–88
14. Rogers BB, Berns SD, Maynard EC et al (1990) Pericardial 35. Allen AW, Megargell JL, Brown DB et al (2000) Venous
tamponade secondary to central venous catheterisation and thrombosis associated with the placement of peripherally inserted
hyperalimentation in a very low birthweight infant. Pediatr Pathol central catheters. J Vasc Interv Radiol 11:1309–1314
10:819–823 36. Gonsalves GF, Eschelman DJ, Sullivan KL et al (2003)
15. Sasidharan P, Billman D, Heimler R et al (1996) Cardiac arrest in an Incidence of central vein stenosis and occlusion following upper
extremely low birth weight infant: complication of percutaneous extremity PICC and port placement. Cardiovasc Intervent Radiol
central venous catheter hyperalimentation. J Perinatol 16:123–126 26:123–127
16. Makwana N, Lander A, Buick R et al (2003) Unusual complica- 37. Trerotola SO, Kuhn-Fulton J, Johnson MS et al (2000) Tunneled
tion of a central venous catheter in a neonate. Arch Dis Child infusion catheters: increased incidence of symptomatic venous
Fetal Neonatal Ed 88:F440 thrombosis after subclavian versus internal jugular access.
17. Leipala JA, Petaja J, Fellman V (2001) Perforation complications Radiology 217:89–93
of percutaneous central venous catheters in very low birth weight 38. Oguzkurt L, Tercan F, Torun D et al (2004) Impact of short-term
infants. J Paediatr Child Health 37:168–171 hemodialysis catheters on the central veins: a catheter venographic
18. Liossis G, Bardin C, Papageorgiou A (2003) Comparison of risks study. Eur J Radiol 52:293–299
from percutaneous central venous catheters and peripheral lines in 39. Chait P, Verbeeteb K, Amaral J et al (2005) Central venous
infants of extremely low birth weight: a cohort controlled study of catheter (CVC) related thrombosis in haematology-oncology
infants <1000 g. J Matern Fetal Neonatal Med 13:171–174 pediatric population (abstract). Pediatr Radiol 35 [Suppl 1]:S79
378 Pediatr Radiol (2008) 38:363–378

40. Chesler L, Feusner JH (2002) Use of tissue plasminogen activator 49. Thalhammer A, Jacobi V, Balzer J et al (2002) Repositioning of
(rt-PA) in young children with cancer and dysfunctional central malpositioned or flipped central venous catheters. Eur Radiol
venous catheters. J Pediatr Hematol 24:653–656 12:698–700
41. Whigham CJ, Lindsey JI, Goodman CJ et al (2002) Venous port 50. Hartnell GG, Gates J, Suojanen JN et al (1996) Transfemoral re-
salvage utilizing low dose tPA. Cardiovasc Intervent Radiol positioning of malpositioned central venous catheters. Cardiovasc
25:513–516 Interv Radiol 19:329–331
42. Santilli J (2002) Fibrin sheaths and central venous catheter occlusions: 51. Bessoud B, de Baere T, Kuoch Y et al (2003) Experience at a
diagnosis and management. Tech Vasc Interv Radiol 5:89–94 single institution with endovascular treatment of mechanical
43. Savader SJ, Ehrman KO, Porter DJ et al (2001) Treatment of complications caused by implanted central venous access devices
hemodialysis catheter-associated fibrin sheaths by rt-PA infusion: in pediatric and adult patients. AJR 180:527–532
critical analysis of 124 procedures. J Vasc Interv Radiol 12:711–715 52. Angle JF, Shilling AT, Schenk WG et al (2002) Utility of
44. Semba CP, Deitcher SR, Li X et al (2002) Treatment of occluded percutaneous intervention in the management of tunnelled
central venous catheters with alteplase: results in 1,064 patients. J hemodialysis catheters. Cardiovasc Intervent Radiol 26:9–18
Vasc Interv Radiol 13:1199–1205 53. Janne d’Othee B, Tham JC, Sheiman RG (2006) Restoration of
45. Liu CY, Jain V, Shields AF et al (2004) Efficacy and safety of patency in failing tunnelled haemodialysis catheters: a comparison
reteplase for central venous catheter occlusion in patients with of catheter exchange, exchange and balloon disruption of
cancer. J Vasc Interv Radiol 15:39–44 the fibrin sheath, and femoral stripping. J Vasc Interv Radiol
46. Kaufman JA, Fan CM, Geller SC et al (1997) Percutaneous 17:1011–1015
revision of excess length from an implanted long-term central 54. Grey RJ, Levitin A, Buck D et al (2000) Percutaneous sheath
venous access device. AJR 169:1732–1734 stripping versus transcatheter urokinase infusion for malfunc-
47. Murthy R, Arbabzadeh M, Richard H et al (2004) Endovascular tioning well-positioned tunneled central venous dialysis catheters:
technique for revision of excess catheter length in subcutaneous a prospective, randomized trial. J Vasc Interv Radiol 11:1121–
implanted venous access devices. Cardiovasc Intervent Radiol 1129
27:259–261 55. Crain MR, Mewissen MW, Ostrowski GJ et al (1996) Fibrin
48. Lois JF, Gomes AS, Pusey E (1987) Non-surgical repositioning of sleeve stripping for salvage of failing hemodialysis catheters:
central venous catheters. Radiology 156:329–333 technique and initial results. Radiology 198:41–44