Original article
Neonatal percutaneous central venous lines: fit to
1
Neonatal Unit, Flinders Medical
Centre, Bedford Park, Australia;
2
Flinders Biomedical
Engineering, School of Medicine,
Flinders University of South
Australia, Flinders Medical
Centre, Bedford Park, Australia;
3
Centre for Perinatal Medicine,
Flinders Medical Centre, Bedford
Park, Australia
Correspondence to:
Dr Scott Morris, Centre for
Perinatal Medicine, Flinders
Medical Centre, Bedford Park,
South Australia 5042, Australia;
scott.morris@health.sa.gov.au
Accepted 21 December 2008
Published Online First
3 February 2009
F298
burst
C Smirk,1 T Soosay Raj,1 A-L Smith,2 S Morris3
ABSTRACT
Objective: To examine pressure changes in neonatal
percutaneous central venous catheters under varying
laboratory conditions and to quantify the risks of rupture in
clinical practice.
Design: We tested 27-gauge polyurethane Premicath and
24-gauge silicone ECC (both Vygon, Norristown, PA)
catheters. Burst pressures were determined by applying a
slowly ramped pressure to catheters that were occluded
at the tip. Flow–pressure relationships were defined by
increasing flow rates through patent catheters from 5 to
499 ml/h. Pressure changes during the manual flushing of
catheters were determined for patent and occluded
catheters and with different syringe sizes.
Results: The mean burst pressure for polyurethane
catheters (1730.8 kPa, 95% CI 1634.7 to 1826.8) was
higher than for silicone catheters (275.6 kPa, 95% CI
240.4 to 310.8). Polyurethane catheters demonstrated an
approximately fivefold greater margin of safety above
manufacturer recommended operating pressures before
burst compared to silicone catheters. Pressures remained
at safe levels in both catheters over the range of flows
generally used in neonatal practice. Hand-flushing of
obstructed silicone catheters caused rupture in 5/6
silicone catheters tested, in comparison to 0/16
polyurethane catheters.
Conclusions: Polyurethane central venous catheters have
a greater pressure tolerance than silicone catheters and are
less likely to rupture under experimental conditions.
Obstructed silicone catheters rupture easily when flushed.
Catheters were not tested in human infants.
Neonatal percutaneous central venous catheters
(CVCs) are frequently used for parenteral nutri-
tion. In circumstances where venous access is
difficult, CVCs are also used to give fluid boluses,
antibiotics or premedications for intubation. The
rupture of CVCs and the embolisation of frag-
ments have been described.1 2 However, the risk
of catheter rupture when flushing or administer-
ing rapid infusions via a CVC is unclear. There
are no published data that adequately examine
the burst pressures of neonatal CVCs. Although
Primhak et al described a burst pressure of
between 520 and 700 kPa in four silicone 24-
gauge CVCs in a letter to the editor, the authors
give no methodology and the validity of these
data is uncertain.3 We therefore conducted
experiments using new and used 27-gauge poly-
urethane and 24-gauge silicone CVCs to ascertain
burst pressures in obstructed catheters, to explore
the pressure tolerance of patent CVCs with
variable flow rates, and to determine pressure
changes during the manual flushing of patent and
obstructed CVCs.
Arch Dis Child Fetal Neonatal Ed 2009;94:F298–F300. doi:10.1136/adc.2008.147900
What is already known on this topic
c Case reports have documented the potential for
rupture of neonatal percutaneous central venous
catheters and embolism of fragments.
c Burst pressures and the risks of catheter rupture
with clinical use have not been clearly defined.
What this study adds
c Polyurethane Premicath central venous
catheters (CVCs) have a greater safety margin
between recommended operating pressures and
burst pressures when compared to silicone ECC
CVCs.
c Patent polyurethane CVCs and silicone CVCs are
unlikely to rupture with standard pump
pressures or when flushed to administer
medications.
c Blocked silicone ECC CVCs rupture easily if
flushing is attempted, and using a larger syringe
size does not prevent rupture.
METHODS
Polyurethane 27-gauge CVCs (Premicath, Vygon,
Norristown, PA) with the introducing wire removed,
and 24-gauge silicone CVCs (ECC, Vygon) were
tested. Used CVCs were collected after removal from
babies when no longer required for total parenteral
nutrition or drug administration, wiped and flushed
with saline, and stored at 4uC until testing. For burst
pressure and infusion studies, a HPLC pump (P-500,
Pharmacia Chemicals, Uppsala, Sweden) was used
to generate controlled high flow and pressure
beyond the range of performance of pumps used in
clinical practice. CVC pressures were recorded
continuously during burst and infusion studies using
a digital oscilloscope (Virtual Instrument ADC-216,
PICO Technology, St Neots, UK) and a digital
manometer (DPM-II Universal Biometer, BioTek
Instruments, Winooski, VT) for pressures
,689.5 kPa. Pressure measurements during manual
flushing studies used the same digital manometer.
Studies were conducted over a 6-month period
from November 2007 to April 2008. The number of
new CVCs tested was determined by the con-
straints of our budget and the number of used
CVCs tested reflected availability within the
chosen time frame. Silicone CVCs are used less
commonly in our neonatal unit and therefore
fewer used CVCs were available for testing.

Burst pressures of occluded CVCs
Twenty three polyurethane (12 new and 11 used) and 14 silicone
(11 new and three used) CVCs were tested. Catheters were
obstructed at the tip with a short length of wire glued with
cyanacrylate for polyurethane CVCs and by a mechanical clamp
for silicone CVCs. CVCs were placed unrestrained in a normal
saline bath at 37uC. Water coloured with methylene blue was
infused via the pump at a set rate of 3 ml/h to achieve a slowly
ramped pressure increase. Rupture was indicated by methylene
blue efflux into the bath. Burst pressure was recorded, and the
sites and type of rupture were documented using a microscope.
Pressures during constant flow in patent CVCs
Twenty two polyurethane (11 new and 11 used) and 14 silicone
(11 new and three used) CVCs were tested without obstruction
with the same temperature and infusate as used above.
Incremental flow rates from 5 to 499 ml/h were applied and
pressures recorded after several seconds at each flow rate once
stabilised. Each catheter was tested once.
Pressure changes during manual flushing
Sixteen new polyurethane and six new silicone CVCs were tested
with a different operator (neonatal intensive care nurse or doctor)
flushing each CVC. A standard central line infusion set-up was
assembled on a bench top. A manometer was connected at the
proximal end of the CVC. The catheter and manometer were
hidden behind a drape to blind the operator to pressure readings
and the cause of obstruction. Two clinical scenarios were provided
to each operator. In the first ‘‘bolus’’ scenario, the operator was
asked to flush the CVC with normal saline to clear the line of
medication using 1, 2 and 10 ml syringes. They were instructed to
flush as they would in normal practice and to flush proximal to the
in-line bacterial filter. In the second ‘‘obstruction’’ scenario, the
same CVC was obstructed at its tip and the operator was asked to
clear the line flushing distally to the filter with the same choice of
syringes. Peak pressures were recorded for each flush episode. The
limit of measurement of the digital manometer was 689.5 kPa;
readings above this limit were recorded as 689.5 kPa for analysis.
Statistical methods
Flow–pressure relationships were analysed using a linear
regression model with random slopes in order to account for
the correlated structure of the data across flow rate.
Comparison of normally distributed data used an independent
samples t test or one way ANOVA with the Tukey post hoc
multiple comparison test as appropriate. Categorical data were
analysed using the x2 statistic. Statistical calculations used SPPS
15.0 for Windows 2006 (SPSS) and Stata v 10 (StataCorp).
RESULTS
Burst pressures of occluded CVCs
Time to rupture was 400–500 s for all CVCs. Figure 1 shows the
burst pressures for occluded catheters. The mean burst pressure
for polyurethane CVCs (1730.8 kPa, 95% CI 1634.7 to 1826.8)
was higher than for the silicone CVCs (275.6 kPa, 95% CI 240.4
to 310.8) (p,0.001). The mean and lowest burst pressures for
the polyurethane CVCs were 11.5 and 8.7 times, respectively,
above the safety limit of 150 kPa noted in the packaging insert.
The mean and lowest burst pressures for the silicone CVCs
were 2.6 and 1.6 times, respectively, above the working safety
limit of 100 kPa also noted in the packaging insert.
New and used polyurethane CVCs had similar mean burst
pressures (1777.3 kPa, 95% CI 1645.4 to 1909.1 vs 1680.1 kPa,
Arch Dis Child Fetal Neonatal Ed 2009;94:F298–F300. doi:10.1136/adc.2008.147900
Original article
95% CI 1521.4 to 1838.7; p = 0.51). New silicone CVCs had a
higher burst pressure (296.3 kPa, 95% CI 267.2 to 325.4) than used
silicone CVCs (199.7 kPa, 95% CI 45.5 to 353.9) (p = 0.008).
The nature of rupture was different for the two CVCs.
Polyurethane CVCs showed coiling with increasing pressure
followed by a localised longitudinal 1–3 mm split that occurred
in various places along the catheter length. Silicone CVCs
developed a localised bleb either at the hub or within 2.5 cm of
the obstructed tip before longitudinal splitting at this site of focal
thinning. No catheters split transversely or fragmented at bursting.
Pressures during constant flow in patent CVCs
Figure 2 shows the flow–pressure relationships for the two
catheter types. Intraluminal pressure was linearly related to
flow for both CVCs (p,0.001) with a statistically greater
increase in pressure per unit increase in flow for the
polyurethane CVCs compared to the silicone CVCs (1.64
(95% CI 1.57 to 1.71) vs 0.37 (95% CI 0.29 to 0.44); p,0.001).
There was no statistically detectable effect of new versus used
CVCs when this variable was included in the model.
For polyurethane CVCs, the package insert recommended
safe pressure (150 kPa) was exceeded at 100 ml/h but remained
below our measured burst pressures at 499 ml/h. For silicone
CVCs, pressure exceeded the recommended safe pressure limit
(100 kPa) at 300 ml/h and reached our range of measured burst
pressures at 400 ml/h.
Manual flushing
The peak pressures generated by flushing obstructed polyur-
ethane CVCs (348.9 kPa, 95% CI 295.1 to 403.3) were greater
than for patent polyurethane CVCs (169.6 kPa, 95% CI 142.7 to
196.5) (p,0.001). Higher pressures were generated flushing the
obstructed polyurethane CVCs with a 1 ml syringe in comparison
to a 10 ml syringe, with similar pressures otherwise noted
between syringe sizes (1 ml syringe 451.6 kPa, 95% CI 365.4 to
537.8; 2 ml syringe 333.7 kPa, 95% CI 242.0 to 425.4; 10 ml
syringe 239.2 kPa, 95% CI 157.2 to 321.3; p = 0.001 1 ml vs 10 ml
using one-way ANOVA). In all cases pressures were below our
observed range of burst pressures and no polyurethane CVCs
ruptured. No statistical differences in pressures generated were
noted for different syringe sizes when flushing patent lines.
For silicone CVCs, peak pressures generated were higher in
the obstructed group (245.5 kPa, 95% CI 146.9 to 343.4) as
compared to the patent group (36.5 kPa, 95% CI 18.6 to 55.2)
(p = 0.002) and similar to the pressures measured in obstructed
polyurethane CVCs (p = 0.057). No differences in pressures
were noted between syringe sizes for flushing silicone CVCs for
either patent or obstructed catheters. Rupture of obstructed
silicone CVCs was frequent during flushing (5/6 silicone CVCs
ruptured compared to 0/16 polyurethane CVCs; p,0.001).
Rupture occurred with both 1 ml and 10 ml syringe sizes. No
ruptures occurred when flushing patent silicone catheters.
DISCUSSION
These data document the pressures generated under different
circumstances in two commonly used CVCs and help to quantify
the risks of rupture. Product information gives a similar safe
pressure limit for both CVCs (150 kPa for the 27-gauge
polyurethane Premicath, and 100 kPa working pressure or
120 kPa bolus pressure for the 24-gauge silicone ECC).
However, our data show an approximately fivefold greater safety
margin before burst for the polyurethane CVC in comparison to
the silicone CVC. Silicone CVCs showed a narrow safety margin
F299
 Original article
Figure 1 Burst pressures for occluded polyurethane Premicath and
silicone ECC central venous catheters (CVCs). Burst pressures for
individual CVCs are shown, with box plots summarising the data. New
CVCs are indicated by open symbols, and used CVCs by symbols with
cross hairs. Solid lines indicate manufacturer safety limits for Premicath
(upper line) and ECC (lower line).
to burst above the manufacturer’s recommended pressures. The
comparative resilience of the polyurethane CVC was not affected
by routine clinical use. In contrast, measured burst pressures of
used silicone CVCs tended to be lower than those of new CVCs.
Higher pressures were generated in the polyurethane CVCs
with constant flow when compared to silicone CVCs. This
reflects the less distensible polyurethane catheter wall and
smaller catheter diameter. The data indicate that flow rates
used in neonatal practice would be unlikely to cause rupture.
Factory default pressure settings on standard neonatal infusion
pumps are less than 100 kPa and with these limits high flows
are achievable in both CVC types.
Hand flushing of medication in patent CVCs appears safe.
However, despite the use of experienced neonatal nurses and
doctors, blocked silicone CVCs usually ruptured with hand
flushing. This practice should therefore be abandoned. The use of
larger syringe sizes to minimise risk of rupture of blocked silicone
CVCs, as suggested by Primhak et al and product information, is
not supported by our data.3 Flushing a blocked Premicath also
generated widely variable and high pressures, and although no
catheters ruptured and pressures generated were well below our
range of measured burst pressures, given the limited number of
CVCs tested this practice is also not recommended.
Salis et al described burst pressure tolerance of larger calibre 16-
and 18-gauge CVCs under conditions of power injection for
radiographic studies.4 Although their experimental conditions and
pressure data are not immediately relevant to neonatal practice,
lower burst pressure was documented for silicone CVCs in
comparison to polyurethane CVCs. These and our current data
are consistent with the observation that case reports of neonatal
CVC rupture and embolism are largely restricted to silicone CVCs.1 2
We found that rupture of CVCs did not result in fragmentation
but rather a short longitudinal split. The immediate effect of this
in practice would be extravasation of fluid. A distal rupture would
remain unsuspected and could predispose to later embolism.
The external validity of our absolute pressure data is difficult to
ascertain in the absence of other comparable published data.
Furthermore, we emphasise that the data obtained under the
F300
Figure 2 Flow–pressure relationships for patent polyurethane
Premicath and silicone ECC central venous catheters (CVCs). Comparison
is made to observed burst pressure in occluded CVCs.
particular conditions of this laboratory study cannot be directly
extrapolated to clinical practice. CVCs were not tested in human
infants. Burst pressures for silicone CVCs in this study are
considerably less than those recorded in four CVCs by Primhak et
al, although these authors’ methodology is unclear.3 The finding
that 24-gauge silicone CVCs may become more susceptible to
burst after normal clinical use requires further study, and no firm
conclusion can be drawn from the data due to small numbers.
The greater pressure tolerance of the 27-gauge polyurethane
CVC has implications for neonatal practice. In our neonatal
unit this CVC is used more frequently than silicone CVCs in
babies of all sizes because of ease of insertion and a capacity to
tolerate high infusion rates with acceptable pressure. The
exception is where central placement of the tip is essential
and the polyurethane CVC is too short to allow this. Any CVC
where pump pressure is exceeded is considered blocked and
electively replaced rather than hand-flushed. It should however
be emphasised that a lower risk of rupture does not indicate
that the polyurethane CVCs are safer than silicone CVCs in
other respects. This is highlighted by data suggesting a greater
incidence of pericardial and pleural effusion with polyurethane
CVCs in comparison to silicone CVCs.5
Acknowledgements: The authors acknowledge Jeanette Birtles for assistance with
equipment, and Richard Woodman for statistical advice. We purchased 10 ECC and 10
Premicath CVCs at cost price from Vygon Australia for this study.
Funding: Purchase of central venous catheters was funded within the Flinders
Medical Centre Neonatal Unit equipment budget.
Competing interests: None.
Ethics approval: Ethics approval was obtained from the Flinders Medical Centre
Human Research Ethics Committee.
REFERENCES
1. Neubauer A-P. Percutaneous central iv access in the neonate: experience with 535
silastic catheters Acta Paediatr 1995;84:756–60.
2. Andrews RE, Tulloh RMR, Rigby ML. Percutaneous retrieval of endovascular catheter
fragments. Arch Dis Child 2002;87:149–50.
3. Primhak RH, Gathercole N, Reiter H. Pressures used to flush central venous
catheters. Arch Dis Child Fetal Neonatal Ed 1998;78:F234.
4. Salis AI, Eclavea A, Johnson MS, et al. Maximal flow rates possible during power
injection through currently available PICCs: an in vitro study. J Vasc Interv Radiol
2004;15:275–81.
5. Pezzati M, Filippi L, Chiti G, et al. Central venous catheters and cardiac tamponade in
preterm infants. Intensive Care Med 2004;30:2253–6.
Arch Dis Child Fetal Neonatal Ed 2009;94:F298–F300. doi:10.1136/adc.2008.147900