Content uploaded by Nancy L Moureau
Author content
All content in this area was uploaded by Nancy L Moureau on Jan 12, 2018
Content may be subject to copyright.
J VA
ISSN 1129-7298
J Vasc Access 2017; 00 (00): 000-000
© 2017 The Authors. This arcle is published by Wichg Publishing and licensed under Creave Commons Aribuon-NonCommercial-NoDerivaves 4.0
Internaonal (CC BY-NC-ND 4.0). Any commercial use is not permied and is subject to Publisher’s permissions. Full informaon is available at www.wichg.com
ORIGINAL RESEARCH ARTICLE
The use of “engineering controls” included the integraon of
NFCs into the IV access system both for needle safety and for
the prevenon of central-line associated bloodstream infec-
ons (CLABSI) (3-5). NFCs allow for the administraon of IV
uids, medicaons and blood to indwelling venous or arterial
catheters without the use of needles. NFCs are also used for
the withdrawal of blood samples and for aspiraon of blood
to check the catheter for patency. While the introducon
of NFCs greatly reduced the risk of needle-sck injuries for
healthcare workers, their use has been associated with other
complicaons such as an increase in catheter occlusions and
CLABSIs (6-11).
In response to the increase in CLABSI related to the use
of NFCs, medical device companies began designing and de-
veloping lower-risk devices (12). Over the past 20 years, 4
categories of NFC designs have emerged (2, 13). While there
is no regulatory body that recognizes the categories of NFC
as being indicave of funcon or performance (13), NFCs are
typically marketed as being “posive”, “negave”, “neutral”
(6, 12-16) or pressure-acvated an-reux (9, 13, 17). The
characteriscs of each category of NFC indicate the mecha-
nism and acon of the NFC upon connecon/disconnecon.
DOI: 10.5301/jva.5000781
Quantave assessment of reux in commercially
available needle-free IV connectors
Garret J. Hull1, Nancy L. Moureau2, Shramik Sengupta3
1 Cook Medical, Bloomington, Indiana - USA
2 PICC Excellence, Inc, Hartwell, GA; Greenville Memorial Hospital, Greenville, SC; Adjunct Associate Professor, Alliance for Vascular Access
Teaching and Research (AVATAR) Group, Centre for Health Pracce Innovaon, Menzies Health Instute Queensland, Grith University,
Brisbane - Australia
3 Department of Biomedical Engineering, University of Missouri, Columbia, MO - USA
Introducon and background
Prior to the advent of needle-free connectors (NFCs),
stainless-steel needles were used to access intravenous (IV)
y-sites, tubing ports and injecon sites. While this process
was eecve for IV pathway access, needle-sck injuries be-
came a substanal risk to healthcare workers, increasing the
potenal for occupaonally acquired blood-borne diseases
(1). In 1992, the Occupaonal Safety and Health Administra-
on (OSHA) recommended that healthcare facilies incorpo-
rate “engineering controls” to prevent such occurrences (2).
ABSTRACT
Introducon: Blood reux is caused by changes in pressure within intravascular catheters upon connecon or
disconnecon of a syringe or intravenous tubing from a needle-free connector (NFC). Changes in pressure, dif-
fering with each brand of NFC, may result in uid movement and blood reux that can contribute to intraluminal
catheter occlusions and increase the potenal for central-line associated bloodstream infecons (CLABSI).
Methods: In this study, 14 NFC brands represenng each of the four market-categories of NFCs were selected for
evaluaon of uid movement occurring during connecon and disconnecon of a syringe. Study objecves were to 1)
theorecally esmate amount of blood reux volume in microliters (μL) permied by each NFC based on exact
component measurements, and 2) experimentally measure NFC volume of uid movement for disconnecon reux
of negave, neutral and an-reux NFC and uid movement for connecon reux of posive displacement NFC.
Results: The results demonstrated uid movement/reux volumes of 9.73 μL to 50.34 μL for negave displace-
ment, 3.60 μL to 10.80 μL for neutral displacement, and 0.02 μL to 1.73 μL for pressure-acvated an-reux NFC.
Separate experiment was performed measuring connecon reux of 18.23 μL to 38.83 μL for posive displace-
ment NFC connectors.
Conclusions: This study revealed signicant dierences in reux volumes for uid displacement based on NFC
design. While more research is needed on eects of blood reux in catheters and NFCs, results highlight the need
to consider NFCs based on performance of individual connector designs, rather than manufacturer designaon of
posive, negave and neutral markeng categories for NFCs without an-reux mechanisms.
Keywords: Intravenous catheters, Needleless connectors, Occlusion, Reux
Accepted: June 13, 2017
Published online: December 4, 2017
Corresponding author:
Shramik Sengupta, PhD
Department of Bioengineering
University of Missouri
1406 E Rollins St.
MO 65211 Columbia, USA
senguptas@missouri.edu
Assessment of reux in needle-free connectors
2
© 2017 The Authors. Published by Wichg Publishing
NFCs are designed as end caps to lock/luer onto the hub of
catheters. Eciency may be measured on the ability of the
NFC design to prevent movement of uids and inadvertent
reux of blood into the catheter (18, 19). The key features
and dening characteriscs of each NFC design along with
manufacturer recommended clamping sequences are sum-
marized in Table I. The specics of each type of NFC are fur-
ther explored in the Discussion secon of this paper.
Unfortunately, the large variety of NFCs, the dierent de-
signs, performance and instrucons for use with each device
is a source of confusion among clinicians. In a 2011 survey of
4000 healthcare workers in which 554 responded (9):
• 21.9% did not know which brand of needle-free IV con-
nector was used with their CVCs (2)
• 25.4% did not know if their connector was ‘positive’,
‘negative’ or ‘neutral’ (2)
• 47.2% did not understand the correct way to flush and
clamp a catheter with the NFC used by their institution (9).
It is not uncommon for several types or brands of NFCs to be
used on peripheral and central venous access devices within
a single hospital, such as one category for peripheral cath-
eters and a dierent category for central catheters (2, 6, 9).
Evidence suggests that lack of training for device usage cor-
relates with an increase in CLABSI (20). This was demon-
strated in a 2009 study that followed ve hospitals adopng
new NFCs (12). When switching NFC brands from ‘negave’
to ‘posive’ connectors, which may have dierent instruc-
ons for use, some hospitals reported an increase in blood-
stream infecons (8). However, hospitals that went back
to their original NFC subsequently reported infecon rates
had returned to previous levels (8, 15, 21). This evidence
suggests that NFC design dierences, variaons in instruc-
ons for use and inconsistencies in asepc technique may
contribute to increased incidence of CLABSI associated with
NFCs (2, 8, 9, 12).
Central venous access devices (CVAD) have undergone
design changes; some of these changes included the elimina-
on of clamps. Despite the instrucons for use of many NFC
manufactures to use clamps on catheters aer disconnecon,
the opon to perform this acon is oen not present (22).
Given the inconsistencies noted with NFC usage, the aim
of the research was to clarify NFC funcon by quanfying
uid movement and volume of reux occurring within each
NFC aer disconnecon or connecon.
Materials and methods
For this study, 14 commercially available NFCs were se-
lected to quantavely study the uid reux volume in mi-
croliters (µL or mm3) of each NFC. Study objecves were to (i)
theorecally esmate amount of blood reux volume in mi-
croliters (μL) permied by each NFC based on exact compo-
nent measurements, and (ii) experimentally measure amount
of uid movement or reux. The 14 NFCs are represented in
each of the four current markeng categories of NFC (13).
The NFCs selected for evaluaon in this study included:
Negave displacement
• BD Carefusion Smartsite
• BD Q-Syte
• Baxter Interlink
• ICU Medical Clave*
Neutral displacement
• ICU Medical Microclave Clear
• Baxter One-Link
• RyMed Invision
• Nexus NIS-6P
Pressure acvated an-reux
• ICU Medical Neutron
• Nexus TKO-5
• Nexus TKO-6P
*ICU Medical Clave is included in the negave displacement
NFC group based on terminology and grouping established in
prior publicaon (13).
Posive displacement*
• B. Braun Ultrasite
• BD Carefusion MaxPlus
• B. Braun Caresite
TABLE I - Fluid displacement characteriscs of negave, neutral displacement, pressure acvated an-reux, and posive, needle-free
connectors
Types of NFCs Negave displacement
NFC
Neutral displacement
NFC
An-reux
NFC
Posive displacement
NFC
Fluid movement upon
disconnecon
Blood reuxes into
catheter
Blood reuxes into
catheter
Fluid restricted by
diaphragm
Fluid moves
toward paent
Fluid movement
upon connecon
Fluid moves toward
paent
Fluids moves
toward paent
Fluid restricted by dia-
phragm
Blood reuxes in
to catheter
Manufacturer
recommended
clamping sequence
Clamp before
disconnecon
No specied clamping No specied clamping Clamp aer disconnec on
of male luer
Hull et al 3
© 2017 The Authors. Published by Wichg Publishing
*As dened in publicaons as posive pressure mechanical
valve with reux occurring on connecon with a nal uid
push at disconnecon clearing blood from catheter p, de-
scribed as a compression/decompression mechanism creang
posive (disconnecon) and negave pressure (on acvaon)
resulng in uid displacement uctuaons (2, 16).
To meet the goals of this study, two independent
experiments were performed, with experiment 2 separated
into 2A and 2B to clearly separate dierences of posive
displacement devices from negave, neutral and an-reux
NFC.
Experiment 1: theorecal esmates of NFC reux
volume (Fig. 1)
Purpose
To calculate the amount of blood reux created by sucon
pressure which occurs in NFC designs.
Materials
• Computer with SolidWorks® Professional 2015 software
• OGP SmartScope Flash 200 Optical Comparator Mea-
surement System
• SolidWorks CAD models of each NFC
Method
To esmate the amount of blood reux caused by sucon
pressure during periods of compression and subsequent non-
compression from syringe connecon and disconnecon, di-
mensions of each component of all NFCs were precisely mea-
sured using a computerized opcal measuring system (OGP
SmartScope Flash 200) by Simplicated Innovaon LLC. Each
component of the NFC was precisely dimensioned (±0.001”)
and built into 3-dimensional (3-D) model using Computer
Aided Design (CAD) 3-D soware (SolidWorks™ Professional
2015). These dimensions provided the necessary input for
the computaonal study and the elastomeric compression
in Experiment 1. Geometrical and mass property calculaons
were used to generate data measuring the mechanical reux
created during compression of the so septum during the
connecon and disconnecon of a male luer locking syringe
to each NFC septum.
Two pictorial models were created for each NFC using
SolidWorks modeling tools: the rst was a ‘un-accessed’ or un-
compressed “at rest” representaon in which the NFC is not
connected to a male luer from the syringe or IV tubing set. The
second picture demonstrates what happens when the same
NFC is ‘accessed’ by a male luer connector, showing compres-
sion of the so septum, the corresponding movement inside
the NFC, and the resulng sucon pressure and uid reux or
ow. For this experiment, a 10-mL BD syringe or a blunt can-
nula was used (Fig. 1). The volume of the “accessed” and “un-
accessed” 3-D models were calculated. Subtracng the volume
of the “un-accessed” or uncompressed “at rest” area from the
“accessed” or compressed area of the NFCs produces the theo-
recal amount of uid displacement available to reux into an
IV catheter upon connecon or disconnecon of the male luer.
Theorecal calculaons of reux values obtained for each
of the 14 connectors are listed in Table II.
Experiment 2A: Actual venous simulaon of negave,
neutral and an-reux NFC reux volume (Tab. II)
Purpose
Quanfy the amount of uid movement or reux asso-
ciated with the disconnecon (negave, neutral and an-
reux NFC) of a male luer lock to each of the 11 (of 14)
NFCs.
To measure blood reux or uid movement associated
with each NFC upon disconnecon of a male luer, an in
vitro venous model was created in the laboratory using the
following:
Materials (Fig. 2)
• An industry standard 10-mL syringe (BD Luer-Lok™ tip
syringe)
• Needle-free connectors (NFCs)
• Clear PVC tubing and stopcock
• Glass capillary rod (6 mm OD × 1.2 mm ID × 12”)
Fig. 1 - Experiment 1 - Pictorial model of a needle-free connector
(Microcla ve®) sh owing its inter na l me ch anism when (A) unaccessed
and (B) accessed using a 10-mL BD Syringe. Internal uid pathway
is displayed in red, moving silicone par ts in yellow, and the outer
housing as translucent. Below the images are the corresponding
volumes of the uid pathway and the dierences between them,
which yield the expected volume of reux (reux volumes for the
other NFCs were calculated in a similar manner).
Assessment of reux in needle-free connectors
4
© 2017 The Authors. Published by Wichg Publishing
TABLE II - Results of the theorecal calculaons and actual in vitro
venous values
NFC
category
Brand Experiment 1:
theorecal
calculaons
(μL)
Experiment
2: actual in
vitro venous
values (μL)
SD
Negave Carefusion
Smartsite®
27. 92 50.37 1.069
BD Q-Syte® 23.20 38.34 0.721
Baxter Interlink® 11.98 13.18 0.1 34
ICU CLAVE® 8.02 9.73 0.265
Neutral ICU MicroClave® 7.7 7 10.80 0.458
Baxter One-Link® 15.87 8.05 0.058
Rymed InVision 2.93 6.54 0.375
Nexus NIS®-6P 5.21 3.60 1.487
An-reux ICU NeutronTM 5.21 1.73 2.656
Nexus TKO®-5 4.03 0.34 0.661
Nexus TKO®-6P 5.26 0.02 0.029
Posive B. Braun
Ultrasite®
59.69 38.83 2.619
Carefusion
MaxPlus®
75.81 23 .73 1.872
B. Braun Caresite® 10.65 18.23 4.60 4
For Experiment 1, “theorecal” values are in microliter (μL) volumes math-
emacally calculated using the method shown in Figure 1. For Experiment 2,
actual in vitro venous values in microliter (μL) volumes are the results from
uid reux obtained upon disconnecon of negave, neutral displacement
and pressure acvated an-reux NFCs and upon connecon for posive dis-
placement NFC.
NFC = needle-free connector; SD = standard deviaon.
• Metric ruler
• Collection bag with water (green food coloring added).
Method
The in vitro venous model was designed to replicate the
condions which cause blood reux into an IV catheter dur-
ing disconnecon of a syringe from an NFC. This venous sim-
ulaon apparatus was designed to replicate the peripheral
venous pressure found in the human vasculature (2, 23). For
the purposes of this study, an average venous pressure of
8 mmHg was used. A glass capillary rod was used to allow
for visualizaon as well as accurate measurement of uid
movement.
For the categories of negave displacement, neutral dis-
placement and pressure-acvated an-reux NFCs (Fig. 2),
each of the NFCs was connected to the PVC tubing, which
was aached to a stopcock and the vercally posioned glass
capillary rod on the model. The following procedure was per-
formed a total of 30 mes per NFC type and brand:
1. A 10-mL syringe was filled with water and attached di-
rectly to the NFC.
2. The stopcock on the model was turned “OFF” to the glass
capillary tube.
3. The 10-mL syringe plunger was depressed and all air was
purged from the NFC, PVC tubing and stopcock attached
to the model.
4. The stopcock was turned “ON” to the glass capillary
rod.
5. The syringe plunger was slowly depressed allowing fluid
to fill the glass capillary rod until the fluid level reached
108 mm (equal to 8 mmHg of simulated venous pres-
sure) on the metric ruler.
6. The syringe was disconnected from the NFC and the amount
of fluid reflux in the glass capillary tube was recorded.
Three sterile samples of each NFC were tested; steps 1-6 were
repeated 10 mes for each sample for a total of 30 venous
simulaons per NFC. All 30 tests were totaled and averaged
to obtain a stascally signicant total uid reux distance
into the glass capillary rod. In this study, all experiments were
conducted by one person, minimizing user variances, result-
ing in relavely low standard deviaons.
The inside diameter of the glass capillary rod is 0.60 mm.
The average distance of uid reux volume of each of the 14
NFCs was used to calculate the total reux volume in micro-
liters (mm3 or µL)
Fig. 2 - In vitro venous model used to measure actual reux. Con-
sists of the following: (1) Standard 10-mL BD Luer Lock syringe,
(2) Needle-free connector, (3) Clear PVC tubing , (4) Glass capillary
rod, (5) Metric ruler, (6) 3-way stopcock, (7) Collecon bag with
dyed water.
Hull et al 5
© 2017 The Authors. Published by Wichg Publishing
Fig. 3 - Visual representaon of
the consequences of reux into a
20-gauge catheter lumen using the
Experiment 1, theorecal/mathe-
macal calculaons and Experiment
2, actual in vitro venous values. The
individual pictures of each NFC illus-
trate the distance in microliters (μL)
the amount of blood can reux into
the catheter using both the theo-
recal/mathemacal Experiment 1
(black) and actual in vitro venous
value Experiment 2 (red). On each
of the cross- seco nal views, the vol-
ume of blood reux is depicted in
red. Each NFC illustraon shows the
silicone elastomeric septum in yel-
low, outer NC housings in blue and
uid pathway in light blue.
Vrh=π
2
V = volume, r = radius, and Δh = change in height.
Figure 3 provides a visual representaon of the venous
simulaon experimental reux values as they appear inside a
20-gauge catheter connected to the respecve NFC, illustrat-
ing the implicaons of uid movement and reux within a
catheter.
Results of disconnecon displacement
Four-hundred and twenty NFC uid displacement mea-
surements were performed in vitro for negave, neutral and
an-reux NFC (30 actuaons for each of the rst 11 NFCs,
with a total of 14 NFCs for both experiments 2A and 2B). The
complete results of theorecal and actual venous simula-
on of uid reux are displayed in Table II. The results for
the three categories of negave, neutral and an-reux NFC
were reported per NFC, per category and in ranges of theo-
recal and actual. In the negave displacement group uid
displacement volumes were ranging from 9.73 to 50.37 μL for
all NFCs. This negave displacement group represented the
widest range of values in comparison to the four categories.
The theorecal calculaons were lower than the actual re-
sults by 10%-80%, with standard deviaon.
Table III represents the mean results of the top ve per-
forming NFCs as predicted by the quantave analysis versus
actual reux volumes based on the in vitro experiment.
Assessment of reux in needle-free connectors
6
© 2017 The Authors. Published by Wichg Publishing
TABLE III - Mean results of the top ve performing NFCs
Predicted volumes (μL) Actual volumes (μL)
Model Reux
volume
Model Reux
volume
Rymed InVision 2.93 Nexus TKO-6P 0.02
Nexus TKO-5 4.03 Nexus TKO-5 0.34
Nexus NIS 6P 5.21 ICU Neutron 1.73
ICU Neutron 5.21 Nexus NIS 6P 3.60
Nexus TKO-6P 5.26 Rymed InVision 6.54
Predictors of quantave analysis versus actual reux volumes of in vitro
experiments.
TABLE IV - Predicted versus actual reux volumes of the lowest ve
performers
Predicted volumes (μL) Actual volumes (μL)
Model Reux
volume
Model Reux
volume
BD Carefusion
MaxPlus
75.81 BD Carefusion
Smartsite
50.37
B. Braun Ultrasite 59.69 B.Braun Ultrasite 38.83
BD Carefusion
Smartsite
27. 92 BD Q-Syte 38.34
BD Q-Sy te 23.20 BD Carefusion
MaxPlus
23.7 3
Baxter One-Link 11.98 B. Braun Caresite 18.23
Predictors of quantave analysis versus actual reux volumes of in vitro
experiments with NFCs allowing the most reux.
Similarly, Table IV lists the predicted versus actual reux
volumes of the boom ve performers (NFCs allowing the
most amount of reux).
Figure 4 lists the reux volume range per NFC design cat-
egory from Table II.
Experiment 2B: Actual venous simulaon of posive
pressure NFC reux volume (Tab. II)
Purpose
Quanfy the amount of uid movement or reux associ-
ated with the connecon of a male luer lock to each of three
posive pressure NFC connectors.
To measure blood reux or uid movement associated
with each posive displacement NFC upon connecon of a
male luer, an in vitro venous model was created in the labora-
tory using the following:
Materials (Fig. 2)
• An industry standard 10-mL syringe (BD Luer-Lok™ tip
syringe)
• Needle-free connectors (NFCs)
• Clear PVC tubing and stopcock
• Glass capillary rod (6 mm OD × 1.2 mm ID × 12”)
• Metric ruler
• Collection bag with water (green food coloring added).
Method
The in vitro venous pressure model was designed to replicate
the condions that cause blood reux into an IV catheter during
connecon of a syringe from each NFC. This venous simulaon
apparatus/model was designed to replicate the peripheral ve-
nous pressure found in the human vasculature (2, 23). For the
purposes of this study, an average venous pressure of 8 mmHg
was used. A glass capillary rod was used to allow for visualiza-
on as well as accurate measurement of uid movement.
For the category of posive displacement NFCs:
To eecvely test reux, due to the design and funconal-
ity of the posive displacement NFCs, the tesng was reversed
from the test steps above to measure connecon reux (16).
Each of the posive displacement NFCs were connected to
the PVC tubing aached to a stopcock and the vercal glass
capillary rod on the model. The following procedure was per-
formed a total of 30 mes per NFC brand:
1. A 10-mL syringe was filled with water and attached di-
rectly to the NFC.
2. The stopcock on the model was turned “OFF” to the glass
capillary tube.
3. The 10-mL syringe plunger was depressed and all air was
purged from the NFC, tubing and stopcock on the model.
4. The stopcock was turned “ON” to the glass capillary rod.
5. The syringe plunger was slowly depressed allowing fluid
to fill the glass capillary rod until the fluid level reached
Fig. 4 - Reux volume range per needle-free connector (NFC) design
category from Table II.
Needleless connector design Range of reux
Negave displacement 9.73 μL to 50.37 μL
Neutral displacement 3.60 μL to 10.80 μL
Pressure acvated an-reux 0.02 μL to 1.73 μL
Posive displacement 18.23 μL to 38.83 μL
Hull et al 7
© 2017 The Authors. Published by Wichg Publishing
to provide comparison of potenal blood reux. It is not
known how long it takes for undisturbed blood in an NFC or
catheter to coagulate. It is not known the minimum blood
volume that will occlude an intravenous catheter. These is-
sues are the subject of future study. The funcon of nega-
ve, neutral and an-reux NFC is consistent with the uid
shi occurring primarily at disconnuaon. Posive displace-
ment NFCs have a dierent funcon with a mechanical valve
that required measurement of the uid movement at a dif-
ferent stage. In the posive displacement NFC, the uid shi
upon disconnecon is oset by the outward displacement
of uid, dierent from negave and neutral NFCs. This dif-
ference in the posive displacement NFC required the uid
displacement measurements to be performed at the me
of connecon, rather than disconnecon. Since intravenous
devices with posive displacement NFCs connue to have an
incidence of occlusion with the catheters, the hypothesis was
that signicant reux occurred at some point of connecon or
disconnecon, or even aer the posive pressure push of u-
id, jusfying the measurement and comparison with another
NFC. The researchers recognize that this created a variaon,
which should be considered as a signicant variable in com-
parisons between each of the NFC categories represented in
this study. Since all NFCs have some uid shi that may result
in blood reux upon connecon and/or disconnecon these
measurements do provide value in comparison, but require
an understanding of the disncon between the categories
and NFC funcons which may not facilitate an exact correla-
on.
The design of NFC has a signicant impact on the ability
to clear blood and control reux. Blood provides many of the
nutrients to support the growth of bacteria. Residual blood
inside the uid pathway of an NFC has the potenal to in-
crease the risk of occlusion of the device and may promote
bacterial growth (2, 24, 25). Each of the NFCs tested in this
study was designed to funcon in a specic way, leading it to
be classied as negave, posive, neutral or an-reux. For
the purposes of this study, the NFCs were categorized based
on design and funcon. A brief descripon follows:
Negave displacement NFCs allow uid displacement into
the catheter lumen during disconnecon from a male luer
syringe or IV tubing. This displacement occurs when uid
(blood) is mechanically pulled away from the paent and into
the catheter or NFC lumen based on pressure changes (9, 13).
Because blood is pulled toward the NFC through the catheter
upon disconnecon, protocol states the catheter be clamped
prior to luer-lock disconnecon (2, 6, 14). Apart from blunt
cannulas entering a split-septum, the general mechanism
consists of a plunger depressing a pre-slit septum facilitang
uid ow through the center of the device (6, 26).
Posive displacement NFCs allow uid displacement into
the catheter lumen during connecon of a male luer syringe
or IV tubing. Fluid movement or reux occurs upon connec-
on with sucon created as the syringe is pushed into the
NFC. Upon disconnecon of the syringe from the NFC a nal
uid push/displacement or uid movement occurs out as a
funcon of the posive displacement NFC (27). Several func-
onal characteriscs of posive NFCs dier from other NFCs.
For example, while an elasc or deformable plunger is sll
depressed during luer-lock connecon, the uid ow occurs
108 mm (equal to 8 mmHg of simulated venous pres-
sure) on the metric ruler.
6. The syringe was disconnected from the NFC and the
amount of fluid in the glass capillary tube was recorded.
7. The syringe was again connected to the positive displace-
ment NFC and the amount of fluid reflux in the glass cap-
illary tube was recorded.
Three samples of each NFC were tested; steps 1-7 were re-
peated 10 mes for each sample for a total of 30 venous
simulaons per NFC.
All 30 tests were totaled and averaged to obtain a stas-
cally signicant total uid reux distance into the glass capil-
lary rod. In this study, all experiments were conducted by one
person, minimizing user variances, resulng in relavely low
standard deviaons.
The inside diameter of the glass capillary rod is 0.60 mm.
The average distance of uid reux volume of each of the 14
NFCs was used to calculate the total reux volume in micro-
liters (mm3 or µL)
Vrh=π
2
V = volume, r = radius, and Δh = change in height.
Figure 3 provides a visual representaon of the venous
simulaon experimental reux values as they appear inside a
20-gauge catheter connected to the respecve NFC, illustrat-
ing the implicaons of uid movement and reux within a
catheter.
Results of posive displacement
Four-hundred and twenty NFC uid displacement mea-
surements were performed in vitro (30 for each of the 3
posive pressure NFCs). The complete results of theorecal
and actual venous simulaon of uid reux are displayed in
Table II. The results for the last category of posive NFC were
reported in ranges of theorecal and actual. In the posive
displacement group were uid displacement volumes ranging
from 18.23-38.83 μL with displacement reux occurring upon
connecon. This posive displacement group represented
a narrow range of values in comparison to the other three
categories. The theorecal calculaons were generally higher
for this group than the actual results by approximately 35%,
apart from the B. Braun Caresite where the theorecal calcu-
laon was approximately 41% lower than the actual (Tab. II).
Table III represents the mean results of the top ve per-
forming NFCs as predicted by the quantave analysis versus
actual reux volumes based on the in vitro experiment.
Similarly, Table IV lists the predicted versus actual reux
volumes of the boom ve performers (NFC allowing the
most amount of reux).
Figure 4 lists the reux volume range per NFC design cat-
egory from Table II.
Discussion
All NFCs have some uid movement/reux either on con-
necon, disconnecon or both (16). In this study, we chose to
measure the uid movement for each category of connector
Assessment of reux in needle-free connectors
8
© 2017 The Authors. Published by Wichg Publishing
around the plunger (6, 26). This space between the plunger
and outer housing creates a reservoir where uid is gathered;
when the NFC is disconnected from a syringe or IV tubing,
uid movement occurs and is pushed outward, toward the
paent (13). This design is created to overcome potenal
blood reux that occurs upon disconnecon, but does not
prevent uid displacement associated with connecon (13).
Theorecally, reux in posive NFCs may also occur aer the
uid push occurs on disconnecon, rst shiing uid out and
then retracng uid back. The method to prevent the poten-
al uid movement aer the outward push of uid with dis-
connecon is through clamping (6, 14, 15). Measurement of
pressure variaons of posive displacement NFCs was only
performed in this study upon connecon, not aer posive
uid disconnecon.
Neutral displacement NFCs suggest the absence of uid
movement upon connecon or disconnecon. The mar-
keng term neutral indicates prevenon of blood reux
that is not substanated in other research (18). The name
‘neutral NFC’ inherently suggests that these devices elimi-
nate movement of uid typically observed in negave and
posive NFCs (13, 26). However, the internal mechanisms
that govern funcon of when and how uid movement and
ow is established appear to be like those of other negave
NFCs (13, 18).
An-reux NFCs suggest minimal uid movement upon
connecon or disconnecon. The funcon of an an-reux
NFC is through a 3-posion silicone diaphragm, which opens
and closes based upon uid or infusion pressure. The an-
reux diaphragm opens or closes based on uid pressure
changes from sources such as IV pump, IV bag, when ush-
ing with a syringe or from physiologic body pressure chang-
es (13, 17). Fluid movement and blood reux is minimized
with connecon or disconnecon of a male luer syringe or
IV tubing. The diaphragm within the an-reux NFC sup-
ports connuous bi-direconal uid pressure control when
aached to the hub of a catheter. When the uid pressure
drops, the an-reux diaphragm closes prevenng blood
reux into the catheters. NFCs designed with an an-reux
diaphragm provide connuous uid control while aached
to the catheter.
This study sought to quantavely evaluate uid re-
ux within NFCs. The results demonstrate a wide range of
displacement in dierent NFCs ranging from 10-50 µL for
‘negave’ NFCs, 3-10 µL for ‘neutral’ NFCs, and displace-
ment that approaches 0 with a range of 0.02-1.73 with pres-
sure acvated an-reux NFCs. Recent research validates
the results with similar ndings on some of the included
NFCs (20). Of the commercially available NFCs tested in Ex-
periment 2, pressure-acvated an-reux NFCs performed
best in terms of minimizing uid displacement. The results
suggest NFC designs and funconal variaons represent
potenal clong risk associated with blood reux.
Data shown in Table II and Figure 3 demonstrate the
following:
1. All needless connectors have a measurable volume of re-
flux on connection or disconnection, however small.
2. The amount of reflux within a catheter is dependent on
the individual design of the NFC.
3. The type of NFC device (negative, positive and neutral)
does not inherently guarantee against unintended or un-
controlled fluid movement or reflux of blood.
4. Anti-reflux connector had the lowest measurable volume
of fluid movement.
Clinical implicaons
Blood reux volumes as small as 4-30 µL may result in brin
formaon adequate to occlude the funcon of a catheter (28).
Body movements, muscle exing, respiraons, coughing, vom-
ing, crying, clamping, unclamping, syringe plunger rebound,
and connecon/disconnecon of syringes all cause mechanical
and physiological pressure changes within a catheter that typi-
cally pulls blood into the catheter p (11, 29). Short peripheral
catheters, PICCs and midlines are parcularly aected by blood
reux due to their small lumen diameter and high surface area.
Complicaons such as sluggish ow, inability to aspirate blood,
loss of patency, brin sheath formaon, catheter dysfuncon
and even catheter-related infecons are all complicaons
which may be related to blood reux (30, 31).
NFCs that permit a blood reux volume >10 μL allow
blood to move beyond the smooth-bore of the distal end of a
20-gauge catheter and into the wider lumen. Any amount of
blood moving into the lumen of a catheter may create the op-
portunity for paral or complete occlusion (13, 32); however,
as illustrated in Figure 3, reux volumes >10 μL provide great-
er risk due to the shape of the lumen. While the exact volume
of blood reux into each catheter resulng in occlusion is un-
known, greater blood reux volume and longer me in situ
will cause coagulaon within a catheter lumen. Smaller lumen
catheters, such as PICCs and midlines have higher incidence of
occlusion and may have greater impact from any amount of
blood reux. Catheter dysfuncon with loss of patency is the
most common complicaon of intravascular catheters result-
ing in signicant impact on connued catheter use with add-
ed cost associated with treatment or replacement (33-42).
According to Rupp, Jarvis and others, design features and
complexies of NFCs create higher or lower risk for vascular
access device infecon. Vascular access device infecon rates
have increased in some facilies with the advent of luer-
acvated mechanical valve NFCs (12, 23, 43, 44). The ability to
eecvely disinfect the surface area, gaps and hub designs of
each connector are listed by Jarvis as characteriscs aecng
risk of infecon (8). Ease of ushing and complete clearance of
all blood products and medicaons within the NFC is another
design feature contribung to infecon risk. CLABSIs impact
paent safety and nancial risk. To reduce CLABSIs and com-
plicaons of occlusion, it is crucial that all blood be adequately
ushed from NFCs and the reux of blood be minimized.
Methods to maintain patency and funcon of catheters
include consistent ushing, standardizaon of NFCs through-
out the facility and frequent educaon with competency vali-
daon on use of NFCs per manufacturer’s recommendaons.
The relaonship between blood reux and occlusion is not
claried in the research, although theorecally NFC pressure
control prevenng reux and minimizing blood within the
catheter would reduce occlusions while maintaining catheter
patency (18, 45, 46). In the Canadian Vascular Access Asso-
ciaon (CVAA hp://cvaa.info/PUBLICATIONS/OcclusionMan
Hull et al 9
© 2017 The Authors. Published by Wichg Publishing
agementGuideline(OMG)/tabid/229/Default.aspx ) Occlusion
Management Guidelines, prevenon strategies note the need
for educaon to: prevent reux of blood into the p of the
catheter by avoiding syringe rebound, keeping the infusion
rate at a level to avoid pressure changes or stasis, ushing
with at least twice the volume of the device and more aer
blood administraon or draws, and consideraon for using
technology designed to prevent catheter occlusions (39).
For those NFC manufacturers who indicate clamping upon
disconnecon within the instrucons for use, the NFCs are
not intended to be used without a clamping sequence. The
signicance of microliter reux volumes on each catheter re-
mains speculaon; results are currently unknown and are the
subject of research to follow this invesgaon.
Limitaons
While the theorecal/mathemacal calculaons were
useful in idenfying best and worst uid reux in this in vitro
invesgaon, several limitaons prevented the exact predic-
on of expected reux with a greater degree of accuracy. The
theorecal calculaons were based purely on the 3-D model
changes in volume and uid displacement created when the
so silicone septum was compressed by the male luer-lock
connector. Addional variables inuencing the actual amount
of uid-movement/reux into the catheter included:
• Amount of pressure or distortion of the soft silicone ma-
terial within each NFC
• Amount of deflection and speed of contraction of the
soft silicone material when engaging luer-lock or cannula
to the needle-free IV connector
• The opening and closing of the split septum seals in the
soft silicone septum
• The potential for mechanically created fluid movement
in and out of the system at different rates due to internal
mechanisms, pressures and freely moving bi-directional
flow
• Manner in which some irregularly shaped compressible
parts fold, and the amounts of residual fluids these parts
trap as they fold.
Research performed with theorec calculaons and in vitro
tesng is a limitaon where clinical implicaons are dicult
to dene. Clinical relevance of reux volume in intravascular
catheters has yet to be determined. The venous simulaon
experiments yield insight into the amount of reux likely to
occur with a standard venous pressure of 8 mmHg; however,
actual reux in the clinical seng is inuenced by a variety
of other factors. Paents are likely to have condions result-
ing in abnormal blood pressures (both high and low) (47), ab-
normal blood viscosity (usually higher) (48) and are likely to
experience temporary, oen acute, changes in bloodstream
pressure due to factors such as bodily movement, coughing,
sneezing, etc.
The experimental values of reux also have their own
limitaons. First, these values are for unclamped operaon.
Surprisingly, the clamping sequence in some manufacturer
informaon for use is not specied or not clearly described,
whereas others are clear in the instrucons for clamping
aer disconnecon (13, 22, 49, 50). In addion, we have
compared the reux upon disconnecon for negave, neu-
tral and an-reux connectors to the reux upon connecon
for posive connectors. In doing so, we chose to compare the
maximum inward movement of uid into the catheter at any
point during the usage of the NFC, which (as highlighted in
Tab. I) funconally occurs during connecon of a syringe for
posive displacement NFCs, and during disconnecon for all
other NFC types. This dierence in funconal uid movement
establishes a variable and limitaon in direct correlaon of
results from negave, neutral, an-reux and posive dis-
placement NFCs. Surfaces that contact blood (especially sur-
faces in irregularly shaped regions from which blood may not
be completely expelled upon ushing) serve as zones where
adverse events such as occlusion formaon and bacterial
colonizaon can occur. Thus, our work idenes the way in
which NFCs may be expected to perform best, in relaon to
uid movement, through both theorecal and actual quan-
tave measurement methods.
Conclusion
In conclusion, this study serves as a necessary stepping
stone to quantavely inspect and evaluate commercially
available NFCs, while also establishing evidence for educa-
on of healthcare providers regarding risk associated with
NFCs. These results indicate incorporaon of NFC designs
with pressure acvated an-reux diaphragm, which may
minimize blood reux and potenally contribute to the re-
ducon of lumen occlusion. Overall, the results demonstrat-
ed signicant dierences in the volume of uid reux based
on NFC design. More comparave research on the impact of
blood reux and associated outcomes in intravascular cath-
eters is needed.
Disclosures
Financial support: This study was funded by a grant to the Univer-
sity of Missouri from Nexus Medical, LLC. Shramik Sengupta PhD,
Department of Bioengineering, served as principal invesgator on
this grant. The funding source played no role in the research results
or reporng of the data. The results and conclusions of the research
are the work product of the authors. Graphic contribuons were
received from Nexus Medical.
Conict of interest: All authors submied ICMJE Form for Disclosure
of Potenal Conicts of Interest. G Hull reported no conicts of inter-
est, reported no employment and was a full me student. Dr. Sengup-
ta reported serving as a consultant to Fresenius. N Moureau reported
employment with PICC Excellence, Inc, Greenville Memorial Uni-
versity and Medical Center, Greenville, South Carolina; educaonal
speaker and consultant to 3M, Angiodynamics, Access Scienc, B
Braun, BD Carefusion, Chiesi, Entrotech, Excelsior, Fresenius, Linear
Health Sciences, Nexus, Parker Labs, Teleex, and research grant re-
cipient from 3M, Cook and Entrotech.
This research represents potenal limitaon and bias due to com-
mercial funding. Experiments and research process was performed
independently and exclusively by the University of Missouri research
department; the manufacturer had no role in the research results
or reporng of the data. The role of the manufacturer was in pro-
viding the funding, the product, reviewing the inial objecves and
protocol developed by the research department prior to iniaon
of the study.
Clave, Microclave and Neutron are registered trademarks of ICU
Medical, Interlink and One Link are registered trademarks of Baxter
Assessment of reux in needle-free connectors
10
© 2017 The Authors. Published by Wichg Publishing
Healthcare, Smartsite and MaxPlus are registered trademarks of BD/
Carefusion, Q-Syte is a trademark of Becton Dickinson, Ultrasite and
Caresite are registered trademarks of B. Braun, Invision Plus is a reg-
istered trademark of Rymed, and Nexus TKO-5 and Nexus TKO-6P
are registered trademarks of Nexus Medical LLC.
References
1. Hadaway L. Needlesck injuries, short peripheral cathe-
ters, and health care worker risks. J Infus Nurs. 2012;35(3):
164-178.
2. Jarvis W. Choosing the best design for intravenous needlefree
connectors to prevent healthcare-associated bloodstream in-
fecons. Infecon Control Today. 2010;14(8):30-31. Available
from: hp://www.infeconcontroltoday.com/arcles/2010/07/
choosing-the-best-design-for-intravenous-needleless-connec-
tors-to-prevent-bloodstream-infecons.aspx?pg=2#. Accessed
July 12, 2017.
3. US Congress. Needlesck Safety and Prevenon Act. 2000(HR
5178): p. 4. Available from: hps://www.govtrack.us/con-
gress/bills/106/hr5178. Accessed July 12, 2017.
4. O’Grady NP, Alexander M, Burns LA, et al. Guidelines for the
prevenon of intravascular catheter-related infecons 2011.
Centers for Disease Control. 2011:1-83. Available from: hps://
www.cdc.gov/hai/pdfs/bsi-guidelines-2011.pdf. Accessed July
12, 2017.
5. Yébenes JC, Serra-Prat M. Clinical use of disinfectable needle-
free connectors. Am J Infect Control. 2008.36(10): 36(10):
S175.e1-4.
6. Hadaway L. Needleless connectors for IV catheters. Am J Nurs.
2012;112(11):32-44, quiz 45.
7. Btaiche IF, Kovacevich DS, Khalidi N, Papke LF. The eects of
needleless connectors on catheter-related bloodstream infec-
ons. Am J Infect Control. 2011;39(4):277-283.
8. Jarvis W. Needlefree connectors and the improvement of paent
and healthcare professional safety. Infecon Control Today. 2013.
17(12). Available from: hp://www.infeconcontroltoday.com/
arcles/2013/12/needleless-connectors-and-the-improvement-
of-paent-and-healthcare-professional-safety.aspx. Accessed
July 12, 2017.
9. Hadaway L. Needlefree Connectors: Improving Pracce, Re-
ducing Risks. J Assoc Vasc Access. 2011;16(1):20-33.
10. Cookson ST, Ihrig M, OMara EM, et al. Increased bloodstream
infecon rates in surgical paents associated with variaon
from recommended use and care following implementa-
on of a needleless device. Infect Control Hosp Epidemiol.
1998;19(1):23-27.
11. Schilling S, Doellman D, Hutchinson N, Jacobs BR. The impact
of needleless connector device design on central venous cath-
eter occlusion in children: a prospecve, controlled trial. JPEN
J Parenter Enteral Nutr. 2006;30(2):85-90.
12. Jarvis WR, Murphy C, Hall KK, et al. Health care-associated
bloodstream infecons associated with negave- or posive-
pressure or displacement mechanical valve needleless connec-
tors. Clin Infect Dis. 2009;49(12):1821-1827.
13. Hadaway L, Richardson D. Needleless connectors: a primer on
terminology. J Infus Nurs. 2010;33(1):22-31.
14. Chernecky C, Macklin D, Casella L, Jarvis E. Caring for paents
with cancer through nursing knowledge of IV connectors. Clin J
Oncol Nurs. 2009;13(6):630-633.
15. Logan R. Neutral displacement intravenous connectors: Evalu-
ang new technology. JAVA. 2013;18(1):31-36.
16. Macklin D. The impact of IV connectors on clinical pracce
and paent outcomes. JAVA. 2014;15(3):139.
17. Jasinsky LM, Wurster J. Occlusion reducon and heparin elimi-
naon trial using an anreux device on peripheral and central
venous catheters. J Infus Nurs. 2009;32(1):33-39.
18. Elli S, Abbruzzese C, Cannizzo L, Lucchini A. In vitro evaluaon
of uid reux aer ushing dierent types of needlefree con-
nectors. J Vasc Access. 2016;17(5):429-434.
19. Hadaway LC. Major thromboc and nonthromboc compli-
caons. Loss of patency. J Intraven Nurs. 1998;21(5)(Suppl):
S143-S160
20. McGee WT. Central venous catheterizaon: beer and worse.
J Intensive Care Med. 2006;21(1):51-53.
21. Jarvis WR, Murphy C, Hall KK, et al. Health care-associated
bloodstream infecons associated with negave- or posive-
pressure or displacement mechanical valve needleless connec-
tors. Clin Infect Dis. 2009;49(12):1821-1827.
22. Carefusion BD. MaxPlus clear needlefree connector - Instruc-
ons for use. Available from: hp://www.carefusion.com/
our-products/infusion/iv-therapy/needlefree-connectors/
maxplus-clear-needlefree-connector. Accessed July 17, 2017.
23. Field K, McFarlane C, Cheng AC, et al. Incidence of catheter-re-
lated bloodstream infecon among paents with a needleless,
mechanical valve-based intravenous connector in an Austra-
lian hematology-oncology unit. Infect Control Hosp Epidemiol.
2007;28(5):610-613.
24. Hanche M. Visualizing the IV uid path as an emerging con-
cept in infecon control. Infecon Control Today. 2004. 1-2.
Available from: hp://www.infeconcontroltoday.com/ar-
cles/2015/05/expanding-intravascular-catheter-surveillance-
issues-and-obstacles.aspx. Accessed July 12, 2017.
25. Maki DG, Kluger DM, Crnich CJ. The risk of bloodstream infec-
on in adults with dierent intravascular devices: a systemac
review of 200 published prospecve studies. Mayo Clin Proc.
2006;81(9):1159-1171.
26. Richardson D. Vascular Access Nursing Pracce: Standards of
Care and Strategies to Prevent Infecon: A Review of Flush-
ing Soluons and Injecon Caps (Part 3 of a 3-Part Series).
Journal of the Associaon for Vascular Access. 2007;12(2):
74-84.
27. Chernecky CC, Macklin D, Jarvis WR, Joshua TV. Comparison
of central line-associated bloodstream infecon rates when
changing to a zero uid displacement intravenous needle-
less connector in acute care sengs. Am J Infect Control.
2014;42(2):200-202.
28. Faintuch S, Salazar GM. Malfuncon of dialysis catheters: man-
agement of brin sheath and related problems. Tech Vasc In-
terv Radiol. 2008;11(3):195-200.
29. Jacobs BR, Schilling S, Doellman D, Hutchinson N, Rickey M,
Nelson S. Central venous catheter occlusion: a prospecve,
controlled trial examining the impact of a posive-pres-
sure valve device. JPEN J Parenter Enteral Nutr. 2004;28(2):
113-118.
30. Mayo D. Catheter-related thrombosis. J Infus Nurs.
2001;24(3S):S13-S22. Available from: hp://journals.lww.
com/journalonfusionnursing/Abstract/2001/05001/Cath-
eter_Related_Thrombosis.5.aspx. Accessed July 12, 2017.
31. Steiger E. Dysfuncon and thromboc complicaons of vas-
cular access devices. JPEN J Parenter Enteral Nutr. 2006;30(1)
(Suppl):S70-S72.
32. Hadaway LC. Reopen the pipeline for I. V. therapy. Nursing.
2005;35(8):54-61, quiz, 61-63.
33. Hadaway L. Loss of catheter patency. Thromboc and non-
thromboc occlusions. 1999: p. 1-8.
34. Moureau N, Poole S, Murdock MA, Gray SM, Semba CP.
Central venous catheters in home infusion care: outcomes
analysis in 50,470 paents. J Vasc Interv Radiol. 2002;13(10):
1009-1016.
Hull et al 11
© 2017 The Authors. Published by Wichg Publishing
35. Tripathi S, Kaushik V, Singh V. Peripheral IVs: factors aecng
complicaons and patencya randomized controlled trial. J In-
fus Nurs. 2008;31(3):182-188.
36. Keogh S, Flynn J. Maintenance of intravascular device patency:
a survey of nursing and midwifery ushing pracce. Qld Nurse.
2014;33(2):30-31.
37. van Miert C, Hill R, Jones L. Intervenons for restoring patency
of occluded central venous catheter lumens (Review). [Re-
view]. Evid Based Child Health. 2013;8(2):695-749.
38. Ernst FR, Chen E, Lipkin C, Tayama D, Amin AN. Comparison of
hospital length of stay, costs, and readmissions of alteplase
versus catheter replacement among paents with occluded
central venous catheters. J Hosp Med. 2014;9(8):490-496.
39. Hill J, Broadhurst D, Miller K, et al. Occlusion management
guidelines for CVAD. Vascular Access. 2013;7(Suppl. 1):1-34.
Available from: hp://www.cvaa.info/Portals/0/documents/
OMG%202013%20Final%20Revised.pdf. Accessed July 12, 2017.
40. Wallis MC, McGrail M, Webster J, et al. Risk factors for periph-
eral intravenous catheter failure: a mulvariate analysis of data
from a randomized controlled trial. Infect Control Hosp Epide-
miol. 2014;35(1):63-68.
41. Baskin JL, Reiss U, Wilimas JA, et al. Thrombolyc therapy
for central venous catheter occlusion. Haematologica. 2012;
97(5):641-650.
42. Hadaway LC. Managing vascular access device occlusions, part
2. Nursing. 2009;39(3):13-14.
43. Rupp ME, Sholtz LA, Jourdan DR, et al. Outbreak of bloodstream
infecon temporally associated with the use of an intravascular
needleless valve. Clin Infect Dis. 2007;44(11):1408-1414.
44. Salgado CD, Chinnes L, Paczesny TH, Cantey JR. Increased rate of
catheter-related bloodstream infecon associated with use of
a needleless mechanical valve device at a long-term acute care
hospital. Infect Control Hosp Epidemiol. 2007;28(6):684-688.
45. Hadaway L. Technology of ushing vascular access devices.
J Infus Nurs. 2006;29(3):137-145.
46. Khalidi N, Kovacevich DS, Papke-ODonnell LF, Btaiche I. Impact
of the Posive Pressure Valve on Vascular Access Device Occlu-
sions and Bloodstream Infecons. Journal of the Associaon
for Vascular Access. 2009;14(2):84-91.
47. Mayo Clinic. High and low blood pressure (hyper/hypoten-
sion). 2015. Available from: hp://www.mayoclinic.org/
diseases-condions/high-blood-pressure/basics/denion/
con-20019580?reDate=16032016. Accessed July 12, 2017.
48. Kwaan HC, Bongu A. The hyperviscosity syndromes. Semin
Thromb Hemost. 1999;25(2):199-208.
49. ICU Medical Inc. Neutron and MaxPlus comparave matrix.
2012, ICU Medical Inc. Available from: hp://www.icumed.
com/media/155735/M1-1372%20Neutron%20vs%20MaxPlus
%20Combat%20Rev.01_email.pdf. Accessed July 12, 2017.
50. Nexus Medical LLC. Nexus TKO®-6P Luer Acvated An-
Reux Device. Available from: hp://www.nexusmedical.
com/tko-6p.htm. Accessed July 12, 2017.
A preview of this full-text is provided by SAGE Publications Inc.
Content available from The Journal of Vascular Access
This content is subject to copyright.