[Federal Register Volume 60, Number 187 (Wednesday, September 27, 1995)]
[Notices]
[Pages 49978-50006]
From the Federal Register Online via the Government Publishing Office [www.gpo.gov]
[FR Doc No: 95-23550]
[[Page 49977]]
_______________________________________________________________________
Part II
Department of Health and Human Services
_______________________________________________________________________
Centers for Disease Control and Prevention
_______________________________________________________________________
Intravascular Device-Related Infections Prevention; Guideline
Availability; Notice
Federal Register / Vol. 60, No. 187 / Wednesday, September 27, 1995 /
Notices
[[Page 49978]]
DEPARTMENT OF HEALTH AND HUMAN SERVICES
Centers for Disease Control and Prevention
Draft Guideline for Prevention of Intravascular Device-Related
Infections: Part 1. ``Intravascular Device-Related Infections: An
Overview'' and Part 2. Recommendations for Prevention of Intravascular
Device-Related Infections; Notice of Comment Period
AGENCY: Centers for Disease Control and Prevention (CDC), Public Health
Service (PHS), Department of Health and Human Services (DHHS).
ACTION: Notice.
-----------------------------------------------------------------------
SUMMARY: This notice is a request for review and comment of the draft
Guideline for Prevention of Intravascular Device-related Infections.
The Guideline consists of two parts: Part 1. ``Intravascular Device-
related Infections: An Overview'' and Part 2. ``Recommendations for
Prevention of Intravascular Device-related Infections,'' and was
prepared by the Hospital Infection Control Practices Advisory Committee
(HICPAC) and the National Center for Infectious Diseases (NCID), CDC.
DATES: Written comments on the draft document must be received on or
before October 30, 1995.
ADDRESSES: Comments on this document should be submitted in writing to
the CDC, Attention: IV Guideline Information Center, Mailstop E-69,
1600 Clifton Road, NE., Atlanta, Georgia 30333. To order copies of the
Federal Register containing the document, contact the U.S. Government
Printing Office, Order and Information Desk, Washington, DC 20402-9329,
telephone (202) 512-1800. Specify the date of the issue requested and
stock number 069-001-00089-1. See page II of the Federal Register for
additional ordering and cost information. In addition, the Federal
Register containing this draft document may be viewed and photocopied
at most libraries designated as U.S. Government Depository Libraries
and at many other public and academic libraries that receive the
Federal Register throughout the country. The order-desk operator can
tell you the location of the U.S. Government Depository Library nearest
you.
FOR FURTHER INFORMATION CONTACT: The IV Guideline Information Center,
telephone (404) 332-2569.
SUPPLEMENTARY INFORMATION: This 2-part document updates and replaces
the previously published CDC Guideline for Prevention of Intravascular
Infections (Am J Infect Control 1983;11:183-199). Part 1,
``Intravascular Device-related Infections: An Overview,'' reviews
issues relevant to intravascular device-related infections and serves
as the background for the consensus recommendations of the Hospital
Infection Control Practices Advisory Committee (HICPAC) that are
contained in Part 2, ``Recommendations for Prevention of Intravascular
Device-related Infection.''
HICPAC was established in 1991 to provide advice and guidance to
the Secretary and the Assistant Secretary for Health, DHHS; the
Director, CDC; and the Director, NCID regarding the practice of
hospital infection control and strategies for surveillance, prevention,
and control of nosocomial infections in U.S. hospitals. The committee
also advises CDC on periodic updating of guidelines and other policy
statements regarding prevention of nosocomial infections.
The Guideline for Prevention of Intravascular Device-related
Infections is the third in a series of CDC guidelines being revised by
HICPAC and NCID, CDC.
Dated: September 14, 1995.
Claire V. Broome,
Deputy Director, Centers for Disease Control and Prevention (CDC).
Guideline for Prevention of Intravascular Device-Related Infections
Executive Summary
The revised guideline is designed to reduce the incidence of
intravascular device-related infections and provides an overview of the
evidence for recommendations considered prudent by consensus of HICPAC
members. A working draft of the guideline was reviewed by experts in
hospital infection control, internal medicine, pediatrics, and
intravenous therapy; however, all recommendations contained in the
guideline may not reflect the opinion of all reviewers.
This document focuses largely on the epidemiology, pathogenesis and
diagnosis of, and preventive strategies for, infections associated with
the intravascular devices most commonly used in health care settings
and for which there is adequate scientific data on which to base
recommendations for device use and care. Such devices include
peripheral venous and arterial catheters, central venous and arterial
catheters, peripherally inserted central venous catheters, and pressure
monitoring systems. Newer devices (e.g., antimicrobial-impregnated
catheters, needleless infusion systems) are also discussed. However,
intraaortic balloon pumps, cardiac catheters, pacemakers, and
extracorporeal membrane oxygenators are not addressed in this document
because there is insufficient scientific data on which to base
recommendations for use and care.
The unique circumstances and special considerations related to
intravascular device-related infections in pediatric patients and
infections associated with parenteral nutrition and hemodialysis will
be addressed in separate sections.
Introduction
Intravascular devices are indispensable in modern-day medical
practice. However, the use of intravascular devices is frequently
complicated by a variety of local and/or systemic infectious
complications. Infections related to the use of intravascular devices,
particularly catheter-related bloodstream infections, are associated
with increased morbidity and mortality, prolonged hospitalization, and
increased medical costs.
Part 1, ``Intravascular Device-related Infections: An Overview''
addresses many of the issues and controversies in intravascular-device
use and maintenance. These issues include definitions and diagnosis of
catheter-related infection, barrier precautions during catheter
insertion, changes of catheters and administration sets, catheter-site
care, and the use of prophylactic antimicrobials, flush solutions and
anticoagulants. Part 2, ``Recommendations for Prevention of
Intravascular Device-related Infections'' provides consensus
recommendations of the HICPAC for the prevention and control of
infections related to the use of intravascular devices.
The Guideline for Prevention of Intravascular Device-related
Infections is intended for use by personnel who are responsible for
surveillance and control of infections in the acute-care, hospital-
based setting, but many of the recommendations may be adaptable for use
in the outpatient or home-care setting.
Part 1. Intravascular Device-Related Infections: An Overview
Contents
I. Background
II. Epidemiology
Devices Used for Short-term Vascular Access
Peripheral venous catheters
Peripheral arterial catheters
Midline catheters
Nontunneled central venous catheters (CVCs)
[[Page 49979]]
Central arterial catheters
Pressure monitoring systems
Peripherally Inserted CVCs
Devices Used for Long-term Vascular Access
Tunneled CVCs
Totally implantable intravascular devices
III. Microbiology
IV. Pathogenesis
V. Definitions and Diagnosis of Catheter-Related Infections
Infections Associated with Short-term Catheters
Infections Associated with Long-term Catheters
Catheter-related Bloodstream Infection
Infusate-related Bloodstream Infection
VI. Strategies for Prevention of Catheter-Related Infections
Site of Catheter Insertion
Type of Catheter Material
Barrier Precautions during Catheter Insertion
Changing Catheters and Administration Sets
Intravenous administration set changes
Intravenous catheter changes
Catheter-site Care
Cutaneous antiseptics and antimicrobial ointments
Catheter-site dressing regimens
In-line Filters
Silver-chelated Collagen Cuffs
Antimicrobial-Impregnated (Coated) Catheters
Intravenous Therapy Personnel
Prophylactic Antimicrobials
Flush Solutions, Anticoagulants, and Other Intravenous Additives
Needleless Intravascular Devices
Multidose Parenteral Medication Vials
VII. Intravascular Device-Related Infections Associated with Total
Parenteral Nutrition
Risk Factors
Surveillance and Diagnosis
Strategies for Prevention
Infusate preparation
Cutaneous antisepsis
Selection of catheter
Catheter-site dressings
Catheter changes
Specialized personnel
VIII. Intravascular Device-Related Infections Associated with
Hemodialysis Catheters
Epidemiology
Microbiology
Strategies for Prevention of Hemodialysis Catheter-related
Infections
Cutaneous antisepsis
Catheter changes
Prophylactic antimicrobials
IX. Intravascular Device-Related Infections in Pediatric Patients
Microbiology
Epidemiology
Peripheral venous catheters
Peripheral arterial catheters
Umbilical catheters
CVCs
Table 1. Definitions for Catheter-related Infection
Table 2. Factors Associated with Infusion-related Phlebitis among
Patients with Peripheral Venous Catheters
Figure 1. Potential Sources for Contamination of Intravascular
Devices
I. Background
Intravascular devices are indispensable in modern-day medical
practice. They are used to administer intravenous fluids, medications,
blood products, and parenteral nutrition fluids, and to monitor the
hemodynamic status of critically ill patients. However, the use of
intravascular-devices is frequently complicated by a variety of local
and/or systemic infectious complications (see definitions in Table 1),
including septic thrombophlebitis, endocarditis, bloodstream infection
(BSI), and metastatic infection (e.g., osteomyelitis, endophthalmitis,
arthritis) resulting from hematogenous seeding of another body site by
a colonized catheter. Catheter-related infections (CRIs), particularly
catheter-related BSIs (CR-BSIs), are associated with increased
morbidity; mortality of 10%-20%; prolonged hospitalization (mean of 7
days); and increased medical costs, in excess of $6,000 (1988 dollars)
per hospitalization.1-5
II. Epidemiology
An estimated 200,000 nosocomial BSIs occur each year.6 During
1980-1989, significant increases were detected in the rates of
nosocomial BSI reported from the National Nosocomial Infection
Surveillance (NNIS) System hospitals where hospital-wide surveillance
was conducted.7 Reported rates increased by 70%-279%, depending on
hospital size and affiliation.
Most nosocomial BSIs are related to the use of an intravascular
device, with BSI rates being substantially higher among patients with
intravascular devices than among those without such devices.8 As
with overall rates of nosocomial BSI, rates of device-related BSI vary
considerably by hospital size, hospital unit/service, and type of
device. During the years 1986-1990, NNIS hospitals conducting intensive
care unit (ICU) surveillance reported rates of central catheter-related
BSI ranging from 2.1 (respiratory ICU) to 30.2 (burn ICU) BSIs per
1,000 central catheter days. Rates of noncentral catheter-related BSI
were substantially lower, ranging from 0 (coronary, medical, and
medical/surgical ICU) to 2.0 (trauma ICU) BSIs per 1,000 noncentral
catheter-days.8
The incidence of and potential risk factors for intravascular-
device related infections may vary considerably with the type and
intended use of the device, and these factors should be considered when
selecting a device for use.
In general, intravascular devices can be divided into two broad
categories, those used for short-term, or temporary, vascular access
and those used for long-term vascular access. Long-term (indwelling)
vascular devices usually require surgical insertion, while short-term
devices can be inserted percutaneously.
Devices Used for Short-Term Vascular Access
Peripheral venous catheters. Of all intravascular devices, the
peripheral venous catheter is the most commonly used. Phlebitis,
largely a physicochemical or mechanical rather than infectious
phenomenon, remains the most important complication associated with the
use of peripheral venous catheters. A number of factors, including type
of infusate and catheter material and size, influence a patient's risk
for developing phlebitis (Table 2); when phlebitis does occur, the risk
of local CRI may be increased.9-13 However, peripheral venous
catheters have rarely been associated with BSI; 9 14-17 this may
reflect the short duration of catheterization with these devices.
Peripheral arterial catheters. Peripheral arterial catheters are
commonly used in acute-care settings to monitor the hemodynamic status
of critically ill patients. Data suggest that peripheral arterial
catheters may be associated with a substantially lower risk of local
CRI and CR-BSI than are peripheral venous catheters left in place for a
comparable length of time.18 Although the reasons for the
differences in rates of CRI associated with these two types of
catheters are not clear, arterial catheters may be less prone to
colonization than are venous catheters because they are exposed to
higher vascular pressures.19 Factors shown to predispose patients
with peripheral arterial catheters to CRI are inflammation at the
catheter insertion site, catheterization >4 days, or catheter insertion
by cutdown.20 21 In contrast to peripheral venous catheters,
peripheral arterial catheters inserted in the lower extremities,
specifically the femoral area, do not clearly pose a greater risk of
infection than do peripheral arterial catheters inserted in upper
extremities or brachial areas.22
In addition to monitoring hemodynamic status, arterial catheters
may also be used to administer local intraarterial chemotherapy.
Although this is a well-established method for treating metastatic or
unresectable tumors, very little has been published on the infectious
complications associated with this form of therapy. Maki et al.
conducted an epidemiologic investigation of endarteritis associated
[[Page 49980]]
with intraarterial chemotherapy administration and identified several
risk factors for infection: leukopenia, hypoalbuminemia, prior
radiation therapy, difficult catheterization, and repeated manipulation
of the catheter.23
Midline catheters. Midline catheters are peripherally inserted
(into antecubital veins), six-inch elastomer catheters that do not
enter central veins, but have recently been used as an alternative to
central venous catheterization. Presently, there is little published
scientific data on which to assess the infectious risks posed by these
newer devices.
Nontunneled central venous catheters (CVCs). CVCs account for an
estimated 90% of all catheter-related bloodstream infections 6 and
nontunneled (percutaneously-inserted) CVCs are the most commonly used
central catheters. Among the factors that influence the risk of
infection associated with the use of CVCs are the number of catheter
lumens and the site at which the catheter is inserted.
Multilumen CVCs are often preferred by clinicians, because they
permit the concurrent administration of various fluids/medications and
hemodynamic monitoring among critically ill patients. In nonrandomized
trials, multilumen catheters have been associated with a higher risk of
infection than have their single-lumen counterparts.24-26 In two
of three randomized trials multilumen catheters were associated with an
increased risk of infection.27-29 Multilumen catheter insertion
sites may be particularly prone to infection because of increased
trauma at the insertion site and/or because multiple ports increase the
frequency of CVC manipulation.25 26 Although patients with
multilumen catheters tend to be more ill, the infection risk found with
the use of these catheters may be independent of the patient's
underlying disease severity.28
In addition to the number of lumens, the site at which a CVC is
inserted may play a major role in CVC-related infections. Five of six
studies have shown a significantly higher colonization or infection
rate with catheters inserted into the internal jugular vein compared
with those inserted into the subclavian vein, with a risk ratio as high
as 2.7.30-35 Other risk factors for CVC-related infections include
repeated catheterization, presence of a septic focus elsewhere in the
body, exposure of the catheter to bacteremia, absence of systemic
antimicrobial therapy,31 duration of catheterization, and type of
dressing.33
Central arterial catheters. Pulmonary artery catheters (PACs)
(i.e., Swan Ganz 1 catheters) differ from CVCs in that they are
inserted through a Teflon introducer and typically remain in place an
average of only 3 days. However, they carry many of the same risks and
have similar rates of BSI as do other central catheters. Risk factors
reported for CRI in patients with PACs include duration of
catheterization >3 days,36 >5 days,37 or >7 days;21
colonization of the skin insertion site;36 38 and catheter
insertion in the operating room using submaximal barrier precautions
(i.e., gloves, small-fenestrated drape).36 Site of insertion may
also influence the risk of infection associated with PACs. Two studies
suggest that PACs inserted into jugular veins have a higher rate of
infection compared with those inserted into subclavian veins;36,
39 three other studies found no difference in infection rates
associated with the two insertion sites.37 38 40
\1\ 1Use of trade names is for identification only and does not
imply endorsement by the U.S. Public Health Service or the U.S.
Department of Health and Human Services.
---------------------------------------------------------------------------
Pressure monitoring systems. Pressure monitoring systems used in
conjunction with arterial catheters have been associated with both
epidemic and endemic nosocomial BSIs.41 42 The first outbreak of
infections due to contamination of pressure monitoring systems was
reported in 1971;43 subsequently, 26 such outbreaks have been
reported.44-48 The final common pathway for microorganisms that
enter the bloodstream of patients and cause bacteremia is the fluid
column in the tubing between the patient's intravascular catheter and
the pressure monitoring apparatus. Microorganisms in a fluid filled
system may move from the pressure monitoring apparatus to the patient
or from the patient to the pressure monitoring system.42
The earliest outbreaks related to pressure monitoring were due to
contaminated infusate 43 or failure to sterilize the fluid pathway
in reusable transducers, particularly the chamber domes.49, 50
Because of the difficulties in sterilizing reusable transducers,
sterile disposable plastic chamber domes were developed. These domes
have a plastic membrane that makes contact with the sensor diaphragm on
the head of the transducer and isolates the sterile fluid pathway from
the transducer. However, systems containing these disposable domes have
also been associated with outbreaks.45 46 51 52 While
resterilization of disposable domes may damage the membrane and permit
ingress of microorganisms into the sterile fluid pathway,53 in
most outbreaks the membranes in the disposable domes remained
intact.46, 51 A study in 1979 showed that fluid used to fill the
space between the transducer head and the membrane of the disposable
dome frequently contaminated the hands of the operator and that the
system was inoculated by touch contamination during the subsequent
assembly of the pressure monitoring system.52 This mode of
contamination is most likely to occur when glucose solutions are used
between the transducer head and the chamber dome membrane and when
transducers are not effectively decontaminated between uses.54
Most outbreaks that have occurred since the introduction of the
disposable chamber dome have been due to this type of
contamination.54
Other mechanisms by which pressure monitoring systems have been
contaminated include contamination of infusate, 41 in-use
contamination of the system by nonsterile calibrating devices,55
contamination of the system by ice used to chill syringes,56
introduction of microorganisms into the system by contaminated
disinfectant 49 and in-use contamination of the system related to
blind, stagnant columns of fluid between the transducer and infusion
system.42 The importance of the latter mechanism in contamination
was shown by a substantial drop in contamination of the system after
introduction of a continuous flush device that eliminated the stagnant
column of fluid.57
To date, no outbreaks have been reported with the use of disposable
pressure transducers. A prospective study of disposable transducers has
shown a very low rate of associated infection (one case of bacteremia
in 157 courses of pressure monitoring).58 This study also showed
that disposable transducers can be safely used for 4 days.58
Disposable transducers were used as a control measure in one reported
outbreak caused by contaminated reusable transducers.45
Peripherally Inserted CVCs
Peripherally inserted CVCs (PICCs) are inserted into the right
atrium by way of the cephalic and basilar veins of the antecubital
space and provide an alternative to subclavian or jugular vein
catheterization and, because they do not require surgical insertion,
cost much less to insert than tunneled subclavian catheters or
subcutaneous ports. PICCs have been used for a variety of purposes,
including total parenteral nutrition (TPN) administration, and their
use appears to be associated with a rate of infection similar to that
reported with other percutaneously inserted CVCs.59 Further
studies are
[[Page 49981]]
needed to adequately determine how long PICCs can safely be left in
place 59 60 and to determine the epidemiology and microbiology of
associated infections.
Devices Used for Long-Term Vascular Access
Tunneled central venous catheters. Surgically implanted right
atrial catheters, including Hickmans, Broviacs, Groshongs, and
Quintons, are commonly used to provide vascular access to patients
requiring prolonged intravenous therapy (e.g., chemotherapy or home-
infusion therapy, hemodialysis). In contrast to percutaneously inserted
(nontunneled) CVCs, these catheters have a tunneled portion exiting the
skin and a Dacron cuff just inside the exit site. The cuff inhibits
migration of organisms into the catheter tract by stimulating growth of
the surrounding tissue, thus sealing the catheter tract and providing a
natural anchor for the catheter. In general, the rates of infections
reported with the use of tunneled catheters have been significantly
lower than those reported with the use of nontunneled CVCs;61-69
however, two recent studies, one randomized, found no significant
difference in the rates of infection among tunneled and nontunneled
catheters.59 70
Totally implantable intravascular devices (TIDs). TIDs are also
tunneled beneath the skin, but have a subcutaneous port or reservoir
with a self-sealing septum that is accessed by needle puncture through
intact skin. TIDs offer the advantage of improved patient image and
obviate the need for routine catheter-site care. Among devices used for
long-term vascular access, TIDs have the lowest reported rates of
catheter-related BSI,71-81 possibly because they are located
beneath the skin with no orifice for ingress of microorganisms.
Recently, several investigators have attempted to compare the
infectious morbidity associated with TIDs and other tunneled catheters.
In one randomized study, TIDs and Hickman catheters had comparable
rates of infection.78 In another randomized study, TIDs had lower
rates of infection compared with other tunneled catheters.79
Groeger et al. conducted one of the largest comparisons of the
infectious complications associated with long-term vascular access
devices to date. In this prospective examination of 1431 devices in
patients with cancer, TIDs (0.21 infections per 1,000 device-days) had
a significantly lower rate of infectious complications compared with
other tunneled catheters (2.77 infections per 1,000 device days,
p0.001).80 However, the devices in Groeger's study were
not randomly assigned, thus the differences observed may be due to
factors other than those inherent to the devices. Existing data suggest
that either of the indwelling devices can be safely used with a low
risk of infection. The selection of a given device depends on the
intended use, patient population, and patient/practitioner preference.
III. Microbiology
Over the past two decades, there has been a marked change in the
distribution of pathogens reported to cause nosocomial BSIs.7 82
83 Since the mid-1980's, an increasing proportion of nosocomial BSIs
reported to NNIS have been due to gram-positive, rather than gram-
negative, species. Moreover, a major portion of the overall increase in
nosocomial BSIs reported to NNIS during the past decade was due to
significant increases in four pathogens: coagulase-negative
staphylococci (CoNS), Candida spp., enterococci, and Staphyloccocus
aureus. The distribution of these pathogens varied by hospital size and
affiliation (i.e., teaching, nonteaching).7
CoNS, particularly S. epidermidis, have become the most frequently
isolated pathogens in CRIs and accounted for an estimated 28% of all
nosocomial BSIs reported to NNIS during 1986-89.7 84 The emergence
of CoNS as the primary pathogen causing CRIs can be attributed to
several factors: (1) increased use of prosthetic/indwelling devices
(e.g., intravascular catheters);85 (2) improved survival of low
birthweight neonates and increased use of intralipids in these
patients;86 and (3) recognition of CoNS as true nosocomial
pathogens rather than harmless commensals.7 The prevalence of
these pathogens also shows that the hands of healthcare workers (HCWs)
and the flora of patients' skin are likely the predominant sources of
pathogens for most CRIs.
Prior to 1986, S. aureus was the most frequently reported pathogen
causing nosocomial BSIs.84 Now, S. aureus accounts for an
estimated 16% of reported nosocomial BSIs.87 S. aureus BSIs may be
complicated by metastatic foci of infection (e.g., vertebral
osteomyelitis) and endocarditis.88-90
Enterococci, another emerging nosocomial bloodstream pathogen,
accounted for 8% of nosocomial BSIs reported to NNIS during 1986-
1989.84 More alarming, has been the emergence of vancomycin-
resistant enterococci (VRE). During 1989-1993, 3.8% of the blood
isolates from BSIs reported to NNIS were vancomycin resistant. Although
data were not available to adequately assess the attributable mortality
of either the BSI or the antimicrobial resistance of the isolate,
mortality was significantly higher among patients whose isolates were
vancomycin resistant (36.6%) than among those whose isolates were
vancomycin susceptible (16.4%).91 Risk factors associated with VRE
BSIs include receipt of antimicrobials (including vancomycin),
gastrointestinal colonization with VRE, underlying disease severity
(e.g., in oncology or transplant patients), abdominal or cardiac
surgical procedures, use of indwelling devices, and prolonged hospital
stay.92-99 Although enterococcal BSIs may arise from the patients'
endogenous flora, nosocomial transmission of VRE via the hands of
HCWs,93 patient-care equipment,100 and contaminated
environmental surfaces 92 93 has also been suggested by the
findings of recent outbreak investigations. The emergence of
enterococci as significant nosocomial bloodstream pathogens is likely
due, in part, to the increased use of invasive devices and the
injudicious use of broad-spectrum antimicrobials for treatment and
prophylaxis of infections.101-105
Fungal pathogens represent an increasing proportion of nosocomial
BSIs. During 1980-1990, NNIS hospitals reported a nearly fivefold
increase in the rate of nosocomial fungal BSIs (1.0 to 4.9/10,000
discharges) and a nearly twofold increase in the proportion of BSIs due
to fungal pathogens (5.4 to 9.9%).106 Such increases were detected
for hospitals of all sizes and affiliations and on all major hospital
services. Candida spp., particularly C. albicans, accounted for >75% of
all nosocomial fungal infections reported to NNIS during this period.
Candidemia has traditionally been thought to arise from the endogenous
flora of colonized patients,107-109 but recent epidemiologic
studies, assisted by the use of molecular typing, show that exogenous
infection due to administration of contaminated fluids,110 111 use
of contaminated equipment,112 cross-infection,113-117 and the
colonized hands of HCWs 118-122 are also important contributors to
candidemia among hospitalized patients.
Although less commonly implicated than either gram-positive
bacterial or fungal species as a cause of BSI, gram-negative
microorganisms account for the majority of CRIs associated with the use
of arterial catheters. Moreover, it has been suggested that clusters of
infections caused by certain gram-negative species, such as
Enterobacter
[[Page 49982]]
spp., Acinetobacter spp., S. marcescens or non-aeruginosa pseudomonads,
should automatically raise suspicion of a common source, such as a
contaminated pressure monitoring device. The predominance of gram-
negative microorganisms in infections associated with pressure
monitoring devices may be due to concomitant receipt of broad-spectrum
antimicrobials by patients undergoing hemodynamic monitoring.
IV. Pathogenesis
The pathogenesis of CRIs is multifactorial and complex (Figure 1),
but available scientific data show most CRIs appear to result from
migration of skin organisms at the insertion site into the cutaneous
catheter tract with eventual colonization of the catheter tip.123-
126 However, there is a smaller, but growing, body of data to suggest
that hub contamination can be an important contributor to intraluminal
colonization of catheters, particularly long-term catheters.127-
130
The relative importance of these two mechanisms of catheter
contamination is the source of continuing debate. Recent findings
suggest that duration of catheterization influences which of the two
mechanisms predominates. Using electron microscopy, Raad demonstrated
that hub contamination was the more likely mechanism of infection for
long-term catheters (i.e., in place >30 days), while skin contamination
was the more likely mechanism for short-term catheters (i.e., <10>130 Although much less common than either of these two
mechanisms, hematogenous seeding of the catheter tip from a distant
focus of infection or administration of contaminated infusate may also
cause CRIs.128 131-134
Two other important pathogenic determinants of CRI are (1) the
material of which the device is made, and (2) the intrinsic properties
of the infecting organism. In vitro studies show that catheters made of
polyvinyl chloride or polyethylene appear to be less resistant to the
adherence of microorganisms than are newer catheters made of Teflon,
silicone elastomer, or polyurethane.135-137 Some catheter
materials also have surface irregularities that may further enhance the
microbial adherence of certain species (e.g., CoNS, Acinetobacter
calcoaceticus, and Pseudomonas aeruginosa).138 139 Thus, catheters
made of certain materials may be more prone to microbial colonization
and subsequent infection. Additionally, certain catheter materials are
more thrombogenic than others, a characteristic that also may
predispose to catheter colonization and catheter-related
infection.140
The adherence properties of a given microorganism are also
important in the pathogenesis of CRI. For example, S. aureus can adhere
to host proteins (e.g., fibronectin) commonly present on
catheters,141 142 and CoNS, the most frequent etiologic agents in
CRIs, adhere to polymer surfaces more readily than do other common
nosocomial pathogens such as E. coli or S. aureus.143
Additionally, certain strains of CoNS produce an extracellular
polysaccharide often referred to as ``slime.'' In the presence of
catheters, this slime potentiates the pathogenicity of CoNS by allowing
them to withstand host defense mechanisms 144 145 (e.g., acting as
a barrier to engulfment and killing by polymorphonuclear leukocytes) or
by making them less susceptible to antimicrobial agents 146 (e.g.,
forming a matrix that binds antimicrobials before their contact with
the organism cell wall). More recent studies suggest that certain
Candida spp., in the presence of glucose-containing fluids, may produce
``slime'' similar to that of their bacterial counterparts, potentially
explaining the increased proportion of BSIs due to fungal pathogens
among patients receiving parenteral nutrition fluids.147
V. Definitions and Diagnosis of Catheter-Related Infections
Establishing a clinical diagnosis of CRI, especially catheter-
related BSI, is often difficult. Diagnosis is typically based on
clinical and/or laboratory criteria, with each having significant
diagnostic limitations. The introduction of semiquantitative methods
for culturing catheters has greatly enhanced our ability to diagnose
CRIs. Both semiquantitative and quantitative methods have greater
specificity in identifying CRI than do traditional broth cultures,
where a clinically insignificant inocula of microorganisms can result
in a positive catheter culture.31 148
However, interpretation of the results of these culture methods may
vary depending on the type and location of the catheter and the culture
methodology used. The use of varying definitions in studies of CRI have
made it difficult to compare existing studies of these infections.
The predictive values of semiquantitative and quantitative methods
may vary, depending on the source of catheter colonization.\130\ For
example, if the skin is the primary source of catheter colonization,
methods that culture the external surface of the catheter may be
preferable. Conversely, if hub contamination is the primary mechanism
for catheter colonization, methods that culture both the external and
internal surfaces may have greater yield.\130\ As the use of
antimicrobial-coated catheters becomes more prevalent, existing
definitions of catheter colonization and CRI may need to be modified.
Infections Associated with Short-Term Catheters
The most widely used laboratory technique for diagnosis of CRI is
the roll-plate method described by Maki et al.\148\ This method
cultures a segment of the catheter after it has been removed from the
patient by rolling the catheter segment across the surface of an agar
plate and determining the number of bacterial colonies present after
overnight incubation. Growth of 15 colony forming units
(cfus) from a proximal or distal catheter segment by semiquantitative
culture in the absence of accompanying signs of inflammation at the
catheter site is considered indicative of catheter colonization. Growth
of 15 cfus from a catheter by semiquantitative culture with
accompanying signs of inflammation (e.g., erythema, warmth, swelling,
or tenderness) at the device site is indicative of local CRI. In the
absence of semiquantitative culture, CRI may be diagnosed when there is
purulent drainage from the skin-catheter junction. Limitations of the
roll plate method are that it requires removal of the catheter and
overnight incubation before results become available.
Cooper et al. proposed direct gram-staining of catheters on removal
as a rapid way to diagnose catheter infection and as a complement to
semiquantitative culture.\126\ However, this method appears to be
considerably more time-consuming than semiquantitative culture and,
thus, may be impractical for routine diagnostic use.
Acridine-orange staining of catheters has been proposed as a
modification of the gram-staining technique.\149\ Although similar to
gram-staining, acridine-orange staining is a single-step procedure that
uses a fluorescent dye to enhance detection of microorganisms in
clinical specimens. This procedure avoids many of the technical
shortcomings encountered with the direct gram-staining technique, but
confirmatory studies documenting its quantitative test performance are
needed before it can be recommended.
The most sensitive technique for diagnosis of CRI is quantitative
culture. To culture a catheter quantitatively, the catheter segment is
either flushed with and then immersed in broth \150\ or placed in broth
and sonicated; 151 152 the
[[Page 49983]]
broth recovered from these procedures is cultured quantitatively.
Sonication releases microorganisms from both the luminal and external
surfaces of the catheter and thus may have greater sensitivity for
diagnosing CRIs, especially those associated with central venous and
arterial catheters, than do methods that only culture the external
surface of the catheter.\152\
All semiquantitative and quantitative catheter culture methods
require removal of the implicated catheter, but the venous access site
can be preserved by removing the catheter over a guidewire and
inserting a new catheter over the guidewire. The proximal and distal
segments of the catheter removed over the guidewire are cultured using
the semiquantitative technique.\153\ If a catheter is removed over a
guidewire and has a negative culture, the catheter inserted over the
guidewire may be left in place. If the catheter removed over a
guidewire has a culture result suggesting colonization/infection, the
second catheter should be removed, and a new catheter inserted at a new
site.59 131 153
Quantitative blood culturing techniques have been developed for
diagnosis of CR-BSI in patients where catheter removal is undesirable
because of limited vascular access. These techniques rely on
quantitative culture of paired blood samples, one obtained through the
central catheter and the other from a peripheral venipuncture site. In
most studies, a colony count from the blood obtained from the catheter
that is five to tenfold greater than the colony count from the blood
obtained from a peripheral vein has been predictive of CR-BSI.154-
156
Infections Associated With Long-Term Catheters
The use of these indwelling catheters may be complicated by a
variety of local infectious complications: exit-site, tunnel, or pocket
infections, as defined in Table 1.\69\ However, clinical diagnosis of
CRI involving the intravascular portion of indwelling catheters is
particularly difficult; thus, laboratory diagnosis is important. The
utility of the roll-plate method for diagnosis of infection associated
with long-term vascular access devices has not been evaluated, but
recovery of 15 cfus on semiquantitative culture of a
catheter segment may be diagnostic of colonization of the intravascular
segment. BSI resulting from a colonized intravascular segment may also
be suspected if 10-fold higher concentration of
microorganisms on quantitative culture of blood obtained from the
catheter compared with the concentration of microorganisms in blood
obtained from a peripheral venous site.157-159
Catheter-Related Bloodstream Infection
CR-BSI is most stringently defined as isolation of the same
organism (i.e., identical species, antibiogram) from semiquantitative
or quantitative cultures of both a catheter segment and the blood
(preferably drawn from a peripheral vein) of a patient with
accompanying clinical symptoms of BSI and no other apparent source of
infection. In the absence of laboratory confirmation, defervescence
after removal of an implicated catheter from a patient with BSI is also
considered indirect evidence of CR-BSI.
Infusate-Related Bloodstream Infection
Since BSI may result from the administration of contaminated
intravenous fluids, culturing intravenous fluids should be part of an
investigation of potential sources of infection. Infusate-related BSI
is usually defined as the isolation of the same organism from both
infusate and separate percutaneous blood cultures, with no other
identifiable source of infection.
VI. Strategies for Prevention of Catheter-Related Infections
Strict adherence to handwashing and aseptic technique remains the
cornerstone of prevention of CRIs; however, other measures may confer
additional protection and must be considered when formulating
preventive strategies. These measures include the selection of an
appropriate site of catheter insertion, selection of appropriate
catheter material(s), use of barrier precautions during catheter
insertion, change of catheters and administration sets at appropriate
intervals, catheter-site care, and the use of filters, flush solutions,
prophylactic antimicrobials, and newer intravascular devices (e.g.,
impregnated catheters, needleless infusion systems).
Site of Catheter Insertion
The site at which a catheter is placed may influence the subsequent
risk of CRI. For peripheral venous catheters, lower extremity
insertions pose a greater risk of phlebitis than do those inserted in
the upper extremity, and upper extremity sites differ in their risk for
phlebitis.160-164 Peripheral venous catheters inserted into hand
veins have a lower risk of phlebitis than do those inserted in upper
arm or wrist veins.\6\
Among CVCs, catheters inserted into subclavian veins have a lower
risk for infection than do those inserted in either jugular or femoral
veins. 31-36 39 Internal jugular insertion sites may pose a
greater risk for infection because of their proximity to oropharyngeal
secretions, and because catheters at internal jugular sites are
difficult to immobilize. However, mechanical complications associated
with insertion are less common with internal jugular vein insertion
than with subclavian venous catheterization.
Type of Catheter Material
The relationship between catheter material and infectious morbidity
has been largely examined by the study of peripheral venous catheters.
The majority of peripheral venous catheters in the U.S. are made of
Teflon or polyurethane, and these catheters appear to be associated
with fewer infectious complications than are catheters made of
polyvinyl chloride or polyethylene.17 135 165 In one large,
randomized prospective study of Teflon and polyurethane catheters, the
two types of catheters had comparable rates of local infection, 5.4%
and 6.9%, respectively,\17\ but polyurethane catheters were associated
with a nearly 30% lower risk of phlebitis when compared with Teflon
catheters. In this trial, neither the Teflon nor polyurethane catheter
was associated with BSI.\17\ By contrast, polyvinyl chloride or
polyethylene catheters have been associated with BSI rates ranging from
0%-5%.166 167
Steel needles, used as an alternative to synthetic catheters for
peripheral venous access, have the same rate of infectious
complications as do Teflon catheters. 168 169
However, the use of steel needles is frequently complicated by
infiltration of intravenous fluids into the subcutaneous tissues, a
potentially serious complication if the infused fluid is a
vesicant.\169\ In view of the low rates of BSI seen with newer Teflon
and polyurethane catheters, the relative risks and benefits of using
steel needles must be evaluated on an individual patient basis.
Catheter material seems to also be an important determinant in the
risk of infection associated with CVCs. Most CVCs used in the U.S. are
made of polyurethane, polyvinyl chloride, polyethylene, or silicone. In
one small, prospective trial comparing silicone with polyvinyl TPN
catheters, silicone catheters had a significantly lower rate of CR-BSI
than did polyvinyl chloride catheters, 0.83 and 19 per 1,000 catheter
days, respectively; however, the silicone catheters were tunneled, and
the polyvinyl chloride catheters were largely nontunneled. The
polyvinyl
[[Page 49984]]
chloride catheters also were associated with a higher risk of
mechanical complications (i.e., breakage, blockage, displacement, and
thrombosis).\170\ Because of the potential confounding caused by the
different types of catheters in this comparison (i.e., tunneled vs.
nontunneled), appropriate conclusions about the contribution of
catheter material to CVC-related infections can not be drawn.
Barrier Precautions During Catheter Insertion
It is generally accepted that good handwashing before and attention
to aseptic technique during insertion of peripheral venous catheters
provide adequate protection against infection. Central venous
catheterization, however, carries a significantly greater risk of
infection, and the level of barrier precautions needed to prevent
infection during insertion of CVCs has been a source of debate.
Until recently, it was assumed that catheters inserted in the
operating room posed a lower risk of infection than did those inserted
on inpatient wards or other patient-care areas. However, data from two
recent prospective studies suggest that the difference in risk of
infection depends largely on the magnitude of barrier protection used
during catheter insertion, rather than the sterility of the surrounding
environment (i.e., ward vs. operating room) 36 171; CVCs or PACs
inserted in the operating room using submaximal barrier precautions
(i.e., gloves, small fenestrated drape) were more likely to become
colonized and to be associated with subsequent BSI than were those
inserted on the ward or in the ICU using maximal barrier precautions
(i.e., gloves, gown, large drape, masks). These data suggest that if
maximal barrier precautions are used during CVC insertion, catheter
contamination and subsequent CVC-related infections can be minimized,
irrespective of whether the catheter is inserted in the operating room
or at the patient's bedside.171 172
Changing Catheters and Administration Sets
Intravenous administration set changes. The optimal interval for
routinely changing intravenous administration sets used for patient
care has been examined in three well- controlled studies. Data from
each of these studies show that changing administration sets
72-hours after initiation of use is not only safe, but cost-
beneficial.173-175 However, because certain fluids (i.e., blood,
blood products, TPN, and lipid emulsions) are more likely than other
parenteral fluids to support microbial growth if contaminated, 132
176-179 more frequent tubing changes may be required when such fluids
are administered.
A common component of intravenous administration sets is the
stopcock. Stopcocks are used for injection of medications,
administration of intravenous infusions, or collection of blood samples
and, thus, represent a potential portal of entry for microorganisms
into vascular catheters or intravenous fluids. Although stopcock
contamination is common, ranging between 45% and 50% in most series,
the relative contribution of stopcock contamination to intravascular
catheter or intravenous fluid contamination is unclear. Few studies
have been able to demonstrate that the organism(s) colonizing stopcocks
is the same one responsible for CRI.180 181 Data suggest that the
use of a closed-needle sampling system can significantly reduce
sampling-port and intravenous fluid contamination.182 183
``Piggyback'' systems may be used as an alternative to stopcocks.
However, they also pose a risk for contamination of the intravascular
fluid if the needle entering the rubber membrane of an injection port
is partially exposed to air, or comes into direct contact with the tape
used to fix the needle to the port. A recently described ``piggyback''
system appears to prevent contamination at these sites and reduces the
incidence of CR-BSI sixfold compared with conventional stopcock and
``piggyback'' systems.182
Intravenous catheter changes. Routine or scheduled change of
intravascular catheters has been advocated as a method to reduce CRIs.
Studies of peripheral venous catheters show that the incidences of
thrombophlebitis and bacterial colonization of catheters seem to
increase dramatically when catheters are left in place >72
hours.12 168 Both phlebitis and catheter colonization have been
associated with an increased risk of CRI. Because of the increased risk
of infection, as well as patient discomfort associated with phlebitis,
peripheral catheter sites are commonly rotated at 48-72 hour intervals
to reduce the risk of phlebitis.
In the maintenance of CVCs, decisions regarding the frequency of
catheter change are substantially more complicated. Some investigators
have shown duration of catheterization to be a risk factor for
infection, 33 35 184 185 and routine change of CVCs at specified
intervals has been advocated as a measure to reduce infection. However,
more recent data suggest that the daily risk of infection remains
constant and show that routine changes of CVCs, without a clinical
indication, do not reduce the rate of catheter colonization or the rate
of catheter-related BSI. 186, 187
The method of replacing CVCs has also been a topic of controversy
and intensive study. CVCs can be changed by placing a new catheter over
a guidewire at the existing site or by inserting the new catheter at
another site. Catheter replacement over a guidewire has become an
accepted technique for changing a malfunctioning catheter or exchanging
a PAC for a CVC when invasive monitoring is no longer needed. Catheters
inserted over a guidewire are associated with less discomfort and a
significantly lower rate of mechanical complications than are those
percutaneously inserted at a new site.131 186 188 189 Guidewire-
assisted exchange may, however, be accompanied by complications, most
notably bleeding at the site, hydrothorax, and subsequent infection of
the newly placed catheter.131 189
Studies examining the infectious risks associated with guidewire
insertions have yielded conflicting results. Three prospective studies
(two randomized) have shown no significant difference in infection
rates between catheters inserted percutaneously and those inserted over
a guidewire.153 187 190 One prospective randomized study has shown
a significantly higher rate of BSIs associated with catheters changed
over a guidewire compared with catheters inserted
percutaneously.186 Most investigators agree that if guidewire-
assisted catheter change occurs in the setting of an CRI, the newly
placed catheter should be removed (131,153,187,188).
Catheter-Site Care
Cutaneous antiseptics and antimicrobial ointments. Skin cleansing/
antisepsis of the insertion site is regarded as one of the most
important measures for preventing CRI, but comparative studies of
cutaneous antisepsis have largely examined its efficacy in eradicating
bacterial flora from the hands of hospital personnel.191 192
However, in one trial, the effectiveness of 2% chlorhexidine, 10%
povidone-iodine, and 70% alcohol 193 as cutaneous antiseptics were
compared in preventing central venous and arterial CRIs. The rate of
catheter-related BSI when chlorhexidine was used for catheter site
preparation was 84% lower than the rates when the other two antiseptic
regimens were used; however, the 2% chlorhexidine preparation used in
this trial is not currently available in the U.S. More recently, a
sustained-release chlorhexidine gluconate patch (250 mu/
[[Page 49985]]
mg dressing) has been introduced as a dressing for catheter insertion
sites. In one randomized trial of epidural catheters, the use of these
patches significantly reduced the incidence of catheter
colonization.194 However, the efficacy of the chlorhexidine patch
in reducing intravascular device-related infection still needs to be
determined.
Tincture of iodine also has been widely used in hospitals for skin
antisepsis before catheter insertion, but its efficacy in reducing
catheter colonization and infection have not been thoroughly evaluated.
Data derived from examining its use as an antiseptic prior to blood
culturing suggest that it, like 70% alcohol and 10% povidone iodine,
may be an effective cutaneous antiseptic for preparation of the skin
prior to insertion of intravascular catheters.195 However,
tincture of iodine may cause skin irritation.195
The application of antimicrobial ointments to the catheter site at
the time of catheter insertion and/or during routine dressing changes
has also been used to reduce microbial contamination of catheter-
insertion sites. Studies of the efficacy of this practice in preventing
CRIs have yielded contradictory findings.30 196-200 Moreover, the
use of polyantibiotic ointments that are not fungicidal may
significantly increase the rate of colonization of the catheter by
Candida spp.198 200 201
Recently, topical mupirocin, a nonsystemic anti-staphylococcal
antimicrobial with documented efficacy in reducing nasal staphylococcal
spp. carriage,202 has been used for cutaneous antisepsis in
conjunction with 2.5% tincture of iodine prior to catheter insertion.
Used in this way, mupirocin was reported to reduce the incidence of
internal jugular catheter colonization among cardiac surgery patients.
However, the utility of mupirocin in reducing the rate of colonization
of peripheral or arterial catheters has not been demonstrated 203
and its use on catheter sites has not been approved. Moreover,
mupirocin resistance has been reported (204-206). Controlled studies
are needed to fully evaluate the effectiveness and potential adverse
effects of mupirocin use for catheter-site maintenance.
Catheter-site dressing regimens. Transparent, semipermeable,
polyurethane dressings have become a popular means of dressing
catheter-insertion sites. These transparent dressings reliably secure
the device, permit continuous visual inspection of the catheter site,
permit patients to bathe and shower without saturating the dressing,
and require less frequent changes than do standard gauze and tape
dressings, thus saving personnel time. Nevertheless, the use of
transparent dressings remains one of the most actively researched, and
controversial, areas of catheter site care. Some studies suggest that
their use increases both microbial colonization of the catheter site
and the risk of subsequent CRI,15 207-210 while other studies have
shown no difference in catheter colonization and infection rates
between the use of transparent dressings and gauze and tape
dressings.10 165 211 The potential risk of infection posed by
transparent dressings appears to vary with the type of catheter
(peripheral or central venous catheter) they are used to dress and,
perhaps, with the season of the year.10 15 209
In the largest controlled trial of dressing regimens to date, Maki
et al. examined the infectious morbidity associated with the use of
transparent dressings on >2,000 peripheral catheters.165 Their
findings suggest that the rate of catheter colonization among catheters
dressed with transparent dressings (5.7%) is comparable to that of
those dressed with gauze (4.6%) and that there are no clinically
important differences in either the incidences of catheter-site
colonization or phlebitis between the two groups. Further, these data
suggest that transparent dressings can be safely left on peripheral
venous catheters for the duration of catheter insertion without
increasing the risk of thrombophlebitis.165
Studies of the use of transparent dressings on CVCs have also
yielded contradictory findings. Some investigators have found an
increased risk of CRI among CVCs with a transparent dressing compared
with those gauze; 209 210 others have found the risk of infection
posed by these two types of dressings to be comparable.211 212
Most of the data on the use of transparent dressings on CVCs are
derived from studies of short-term nontunneled devices and little data
have been published regarding the use of transparent dressings on long-
term, tunneled CVCs.213 In a metaanalysis of catheter dressing
regimens, CVCs on which a transparent dressing was used had a
significantly higher incidence of catheter tip colonization, but a
nonsignificant increase in the incidence of CR-BSI.214 Preliminary
data suggest that newer transparent dressings that permit the escape of
moisture from beneath the dressing may be associated with lower rates
of skin colonization and CRI,213 215 but the length of time that a
transparent dressing can be safely left on a CVC catheter site is
unknown.
Collodion has also been evaluated for use as a potential dressing
for catheter sites. One small (n=34), retrospective study of its use on
CVCs reported a low incidence of CRIs, despite catheters remaining in
place an average of 16.5 days.216 However, before collodion can be
recommended for routine use as a catheter site dressing, randomized
trials comparing collodion to existing dressings should be done.
In-Line Filters
In-line filters may reduce the incidence of infusion-related
phlebitis (217-220), but there are no data to support their efficacy in
preventing infections associated with intravascular devices and
infusion systems. Proponents of the use of filters cite a number of
potential benefits: (1) reducing the risk of infection from
contaminated infusate or proximal contamination (i.e., introduced
proximal to the filter); (2) reducing the risk of phlebitis in patients
who require high doses of medication (e.g., antimicrobials) or in those
in whom infusion-related phlebitis has already occurred; (3) removing
particulate matter that may contaminate intravenous fluids; 221
and (4) filtering endotoxin produced by gram-negative organisms in
contaminated infusates.222 These theoretical advantages must be
tempered by the knowledge that infusate-related BSI rarely occurs and
that pre-use filtration in the pharmacy is a more practical, and less
costly, way to remove particulates from infusates. Furthermore, in-line
filters may become blocked, especially with certain solutions (dextran,
lipids, mannitol), and consequently increase line manipulations and/or
decrease the availability of administered drugs.223 Because of
these potential untoward effects, the routine use of in-line filters
may increase cost, personnel time, and possible infections.224
Silver-Chelated Collagen Cuffs
Since 1987, a silver-chelated, collagen cuff that is attachable to
percutaneously inserted CVCs has been commercially available. Similar
to the cuff used on Hickman and Broviac catheters, this cuff is
designed to form a mechanical barrier to skin microorganisms migrating
into the cutaneous catheter tract; 201 225 the silver provides an
additional antimicrobial barrier.201 225 Two randomized controlled
trials examining the efficacy of silver-chelated collagen cuffs have
been published. In the first trial, cuffed CVCs were associated with a
threefold lower risk of catheter colonization and a nearly fourfold
lower risk of CR-BSI compared with traditional noncuffed CVCs.225
In the second trial, a 78% reduction in
[[Page 49986]]
catheter colonization and a 100% reduction in CR-BSI were observed with
these devices.201 The relative contribution of the cuff versus the
antimicrobial properties of the silver preventing CRI is uncertain. No
controlled trials examining the efficacy of cuffs without antiseptic or
antimicrobial coating have been published.
The protective effect of these cuffed CVCs appears to be immediate
and exceeds that seen with the use of antimicrobial ointment
alone.201 However, cuffs appear to be most beneficial with
catheters left in place for >4 days.225 Studies on the efficacy of
these cuffs in preventing infection with longer-term CVCs (i.e., >20
days) have not been published.
Antimicrobial-Impregnated (Coated) Catheters
In animal models, antimicrobial or antiseptic impregnation of
catheters appears to reduce bacterial adherence and biofilm
formation,226 227 but the utility of these impregnated catheters
in clinical settings has only recently been evaluated. Kamal et al.
conducted a large, randomized, prospective trial among SICU patients to
evaluate a CVC bonded with cefazolin for the entire length of its
external and luminal surfaces.228 The authors found a sevenfold
reduction in the incidence of catheter colonization (2% vs 14%), but no
difference in catheter-site inflammation (i.e., culture-negative
inflammation of the insertion site). No bacteremias occurred in either
group. The authors suggest that antimicrobial coating of the luminal
surfaces of catheters may be particularly beneficial in reducing the
risk of infection resulting from hub contamination.
Data supporting the utility of antimicrobial coating for peripheral
catheters are much less conclusive. Kamal et al. also studied a small
number of peripheral arterial catheters as part of their evaluation of
the cefazolin-impregnated catheter.228 Although impregnated
peripheral arterial catheters had a fivefold lower incidence of CRI
compared with noncoated catheters (3% vs 15%), this difference was not
statistically significant. The lack of demonstrable efficacy of
antimicrobial coating of peripheral arterial catheters in reducing CRI
may be due, in part, to the inherently low incidence of CRI associated
with the use of peripheral arterial catheters.
Of the studies reported to date, antimicrobial-coated catheters do
not appear to pose any greater risk of adverse effects than do
noncoated catheters, but additional controlled trials need to be done
to fully evaluate their efficacy, determine the appropriate situations
for their use, and assess the risk of emergence of resistant
bloodstream pathogens.
Intravenous Therapy Personnel
Because insertion and maintenance of intravascular catheters by
inexperienced staff may increase the risk of catheter colonization
153 and CR-BSI, many institutions have established infusion
therapy teams. Available data suggest that trained personnel designated
with the responsibility for insertion and maintenance of intravascular
devices provide a service that effectively reduces CRIs and overall
costs.229-231
Prophylactic Antimicrobials
Prophylactic administration of antimicrobials has been used to
reduce the incidence of CR-BSIs, but scientific studies on the efficacy
of this practice are inconclusive. Two published studies, one
randomized 232 and one nonrandomized,233 suggest that
antimicrobials administered systemically at the time of (or immediately
after) insertion of a CVC may reduce the incidence of CR-BSI. Two
randomized trials of systemically administered antibiotics demonstrated
no benefit of such prophylaxis.234 235 One randomized controlled
trial showed a significant protective effect of a heparin-vancomycin
flush solution used daily in immunocompromised patients with tunneled
CVCs.236 Two other randomized controlled trials have examined the
effect of continuous low dose (25g) vancomycin, added to TPN
fluids, in reducing the incidence of CoNS BSI in low birthweight
infants.237 238 In one of these trials, the incidence of CoNS BSI
decreased from 34% to 1.4% (P<0.001) among="" neonates="" weighing=""><1500>237 However, 4/71 (5.6%) treated neonates developed a BSI due
to gram-positive cocci after vancomycin prophylaxis was completed. The
other trial studied neonates weighing <1000 gm="" and="" found="" that="" the="" use="" of="" vancomycin="" was="" associated="" with="" a="" significantly="" lower="" incidence="" (0%="" vs="" 15%)="" of="" cons="">238 Although prophylactic administration of
vancomycin decreased the incidence of CoNS BSI, it did not decrease
overall mortality among low birth weight infants in either study.
Further studies are needed to assess the additional benefit afforded by
prophylactic antimicrobials in reducing CRIs when standard infection
control measures are adhered to and to assess the concern that such
prophylaxis may select for resistant microorganisms, particularly those
resistant to vancomycin.
Flush Solutions, Anticoagulants, and Other Intravenous Additives
Flush solutions are designed to prevent thrombosis, rather than
infection, but thrombi and fibrin deposits on catheters may serve as a
nidus for microbial colonization of the intravascular devices.
Furthermore, catheter thrombosis appears to be one of the most
important factors associated with infection of long-term
catheters.69 239 Thus, the use of anticoagulants (e.g., heparin)
or thrombolytic agents may have a role in the prevention of CR-BSI.
However, several recent studies suggest that 0.9% saline is as
effective as heparin in maintaining catheter patency and reducing
phlebitis among peripheral catheters.137 240 241 Furthermore,
recent in vitro studies suggest that the growth of CoNS on catheters
may be enhanced in the presence of heparin. In contrast, the growth of
CoNS on catheters can be inhibited by edetic acid (EDTA),242
suggesting that EDTA, rather than heparin, may decrease the incidence
of CoNS CR-BSIs. Also, the routine use of heparin to maintain catheter
patency, even at doses as low as 250-500 units/day, has been associated
with thrombocytopenia and thromboembolic and hemorrhagic
complications.243-246 Clinical trials are needed to further assess
the relative efficacy, risks, and benefits of the routine use of
various anticoagulants (e.g., EDTA) in preventing CRI.
The risk of phlebitis associated with the infusion of certain
fluids (e.g., potassium chloride,247 lidocaine,247 248
antimicrobials,247 also may be reduced by the use of certain
intravenous additives, such as hydrocortisone.247 Bassan et al. in
a prospective, controlled trial of patients being evaluated for
possible myocardial infarction found that heparin and/or hydrocortisone
significantly reduced the incidence of phlebitis in veins infused with
lidocaine.248 In other trials, topical application of
venodialators such as glycerol trinitate,249 250 or anti-
inflammatory agents such as cortisone near the catheter site,251
has effectively reduced the incidence of infusion-related
thrombophlebitis and increased the life span of the catheters.251
252 Larger, controlled trials are needed to assess the advisability of
the routine use of these agents to reduce phlebitis.
Needleless Intravascular Devices
Attempts to reduce the incidence of sharps injuries and the
resultant risk of transmission of bloodborne infections to
[[Page 49987]]
HCWs have led to the design and introduction of needleless intravenous
systems. However, there are limited data by which to assess the
potential risk of contamination of the catheter and infusate and
subsequent CRI that may be associated with the use of these devices. In
one trial where conventional and needleless heparin-lock systems were
compared, the rates of infection were comparable.\253\ However, in
another investigation, the combined use of a needleless infusion system
and TPN was associated with an increased rate of BSIs among patients
receiving home infusion therapy.\254\ As the use of these systems
becomes more widespread, the potential infectious risks associated with
their use can be more fully evaluated.
Multidose Parenteral Medication Vials (MDVs)
Parenteral medications are commonly dispensed in MDVs that may be
used for prolonged periods for one or more patients. Although the
overall risk of extrinsic contamination of MDVs appears to be small, an
estimated 0.5 per 1,000 vials,\255\ the consequences of contamination
may be serious. Contamination of MDVs due to breaks in aseptic
technique have resulted in several nosocomial outbreaks. The implicated
vehicles in these outbreaks have been lipids infused intravenously from
multidose containers\177\ and medications used for intra-articular
injections.256 257 However, when bacteria or yeasts were
inoculated into some commonly used medications, such as heparin,
potassium chloride, procainamide, methohexital, succinylcholine
chloride, and sodium thiopental, and left at room temperature, no
microorganisms could be cultured from these medications after 96 hours,
with rare exceptions, irrespective of whether they contained a
preservative.\258\ Microorganisms could proliferate in lidocaine and
insulin only if the inocula were prepared in peptone water (with one
exception), which allowed for transfer of nutrients to the vials. Even
under these conditions, when vials were kept at 4 deg.C (the
recommended storage temperature), microorganisms did not proliferate in
the insulin. There is one report of hepatitis B virus transmission
related to the use of a contaminated vial of bupivacaine in a
hemodialysis unit.\259\
VII. Intravascular Device-Related Infections Associated With Total
Parenteral Nutrition
Catheter-related BSI remains one of the most important
complications of TPN therapy and reported rates of infection during TPN
vary widely depending on the population studied and the definitions
used. Because TPN solutions commonly contain dextrose, amino acids,
and/or lipid emulsions, they are more likely than conventional
intravenous fluids to support microbial growth if contaminated.177
179 260-263 Lipid emulsions are particularly suited for the growth of
specific bacteria and yeasts,176 177 with microbial growth
occurring as early as 6 hours after inoculation of a lipid emulsion and
reaching clinically significant levels (>10\6\ CFU/ml) within 24
hours.178 Newer combined TPN solutions (e.g., 3-in-1 system) which
use glucose, amino acids, lipid emulsion, and additives in one
multiliter administration bag, may increase the risk of infection
associated with TPN, but data on which to assess this risk are not
available.
Although TPN solutions are particularly suited for microbial
growth, most infections that occur during the administration of TPN
result from contamination of the catheter. TPN- related CRI result much
less commonly from infusion of contaminated fluids or from hematogenous
seeding of the catheter.
The microbiology of TPN-related CR-BSIs is similar to that of other
CR-BSIs, with gram-positive species, particularly CoNS or S. aureus,
being the predominant pathogens. However, the proportion of BSIs due to
fungal pathogens, particularly Candida spp., are significantly greater
in patients receiving TPN.\106\
Risk Factors
A number of factors have been associated with the development of
CRI during TPN therapy, including catheter-site colonization,123
125 155 method and site of catheter insertion, the experience of the
personnel inserting the catheter,\153\ the use of the TPN line for
purposes other than administration of parenteral nutrition fluids,\264\
breaks in the protocol for aseptic maintenance of the infusion
systems,167 223 264 265 and the use of triple-lumen
catheters.24 25 27 28
Surveillance and Diagnosis
Surveillance for CRI during TPN administration should be the same
as during the administration of other types of infusion therapy.
Although culturing the skin adjacent to the catheter insertion site may
help predict BSI in patients who are receiving TPN,123 125 155
routine microbiologic surveillance can not be advocated. As with other
suspected CRIs, semiquantitative and quantitative catheter cultures may
also be useful for the diagnosis of TPN-related CRIs. Vanhuynegem et
al. evaluated the efficacy of semiquantitative cultures of blood drawn
through in place TPN catheters in febrile patients for diagnosing CR-
BSI.\266\ Comparing their methodology to the semiquantitative culture
technique of Maki, they found that such cultures had a positive
predictive value of 60%, and a negative predictive value of 100%.
Moreover, using this technique, they were able to prevent unnecessary
removal of 87% of the catheters in which infection was suspected.
Strategies for Prevention
The strategies previously outlined for the prevention of CRIs are
also effective in reducing the risk of infections associated with TPN,
and rigorous aseptic nursing care has been shown to greatly reduce the
incidence for TPN-related infection.265 267 268 Nevertheless, a
number of supplemental preventive measures that have been proposed to
reduce the risk for TPN-related CRIs bear discussion, including special
precautions for infusate preparation, cutaneous antisepsis, and
catheter selection and care.
Infusate preparation. Since TPN solutions are prone to microbial
growth if contaminated, strict attention must be given to asepsis
during the compounding of TPN solutions. Although controlled trials
have not been done, centralized preparation of TPN solutions in
hospital pharmacies, using a laminar flow hood, has generally been
regarded as the safest method of preparation.
Cutaneous antisepsis. Findings on the efficacy of various
antiseptic skin preparations on decreasing the incidence of CRI during
TPN suggest that tincture of iodine and chlorhexidine in ethyl alcohol
are superior to povidone-iodine as a skin antiseptic during TPN
catheter care.\269\ Furthermore, in one prospective randomized study,
the application of povidone-iodine ointment to the insertion sites of
subclavian catheters used for TPN was not associated with a decrease in
CRIs when compared with catheters on which povidone-iodine was not
used.\268\
The application of organic solvents, such as acetone or ether, to
``defat'' (remove skin lipids) the skin prior to catheter insertion and
during routine dressing changes has been a standard component of many
hyperalimentation protocols. However, these agents appear neither to
confer additional protection against skin colonization nor
significantly decrease the incidence of CRI. Moreover, their use can
greatly increase local inflammation and patient discomfort.\270\
[[Page 49988]]
Selection of catheter. Tunnelling of TPN catheters has been
proposed for three reasons: (1) to prevent dislodgement of the
catheter; (2) to reduce the incidence of CR-BSI by increasing the
distance between the sites where the catheter exits the skin and where
it enters the subclavian vein; and (3) to protect the catheter from
potentially contaminated sites such as tracheostomies. However, few
prospective randomized studies have been done to evaluate the efficacy
of this practice. When Koehane et al. assessed the risk of BSI among
patients with short-term, noncuffed, tunneled and nontunneled TPN
catheters, they demonstrated a reduction in the incidence of CR-BSI
among tunneled catheters as compared with nontunneled catheters.\267\
However, this reduction was greatest when a designated nutrition nurse
was used to maintain the catheter; after improved adherence to the
infection control protocol, short-term, noncuffed, tunneled and
nontunneled catheters were associated with a similar rate of BSI. The
only other controlled trial of short-term, noncuffed, tunneled and
nontunneled catheters similarly failed to demonstrate a beneficial
effect of tunnelling after rigorous attention to infection
control,\127\ suggesting that if strict infection control practices are
adhered to, short-term, noncuffed, tunneled and nontunneled TPN
catheters have a similar risk of infection.
Catheter-site dressings. The use of occlusive dressings on
catheters used for TPN has been a continuing source of debate. Two
controlled studies suggest that, with adherence to strict infection
control protocols, semipermeable, transparent dressings are a safe,
cost- effective alternative to gauze and tape for dressing TPN
catheter-insertion sites.212 268 Moreover, data suggest that
transparent dressings used on TPN catheter sites can be safely changed
at 7-day intervals.212 268 271
Catheter changes. Prospective, randomized trials examining the
frequency of TPN catheter changes have not been published. However,
data from a study in 1974 suggest that the rate of infection (6.2%) for
TPN catheters in place for >30 days is similar to the rate of infection
(7%) for all catheters.\265\
Specialized personnel. Many institutions have protocols and a
nutritional support team for insertion and maintenance of catheters
used for TPN. As with vascular devices used for other purposes, the use
of specially trained personnel to insert and maintain the catheters
appears to reduce the rate of infection in patients receiving
TPN.230, 231, 267
VIII. Intravascular Device-Related Infections Associated With
Hemodialysis Catheters
Epidemiology
Each year approximately 150,000 patients undergo maintenance
hemodialysis for chronic renal failure. Since 1979, when the Uldall
subclavian catheter was introduced, CVCs have gained popularity as a
convenient, rapid way of establishing temporary vascular hemodialysis
access until placement or maturation of a permanent arteriovenous
fistula or permanent access for patients without alternative vascular
access.\272\ In 1990, an estimated 73% of centers participating in the
National Surveillance System for Hemodialysis Associated Diseases had
1 patients in whom CVCs were used for permanent vascular
access.\273\ However, only a limited number of controlled trials
examining the infectious risk associated with the use of CVCs for
hemodialysis have been published; most data are derived from small
studies at individual institutions.
Subclavian hemodialysis catheters have been associated with a rate
of BSI that exceeds that reported for virtually all other subclavian
catheters274-283 or for alternative forms of hemodialysis vascular
access275 284 and their use may be complicated by bacterial
endocarditis, septic pulmonary emboli,274 275 282 284 and/or
thrombosis (e.g., venous thrombosis, catheter occlusion). The factors
contributing to the increased rate of infection experienced with CVCs
used for hemodialysis have not been fully elucidated,277 278 but
manipulations and dressing changes of dialysis catheters by
inadequately trained personnel,\285\ duration of catheterization and
mean number of hemodialysis runs,\277\ and cutdown insertion of the
catheter\286\ may increase the risk of CRI among hemodialysis patients.
More recently, jugular vein catheters have been used for
hemodialysis access because descriptive studies indicate that they are
associated with fewer mechanical complications than subclavian
catheters, including subclavian thrombosis, stenosis, and
perforation.287-294 These double-lumen, Dacron-cuffed, silicone
CVCs have been used for exclusive, or prolonged, vascular access in
chronic hemodialysis patients286 295 and appear to have a longer
median use-life and fewer insertion complications than do either of
their single-lumened Teflon or polyurethane counterparts.280 295
296 Moss et al. recently reviewed the 4-year experience with double-
lumen, cuffed, silicone catheters at their institution. All catheters
(n=168) had been placed for long-term use (1 month) and were
the sole vascular access for hemodialysis.\286\ The median life span
for these catheters was 18.5 months, with 12- and 24-month catheter
survival being 65% and 30%, respectively. As with subclavian
hemodialysis catheters, thrombosis (catheter and vein) and infection
were the most frequent catheter complications. BSI occurred in 16/131
(12%) patients and exit-site infections in 28/131 (21%); diabetics
(33%) were significantly more likely to develop exit-site infections
than were nondiabetics (11%). Based on the duration of catheterization,
the authors determined the following rates of CRIs associated with the
use of double-lumen CVCs: 0.25 BSIs per patient-year, 0.36 exit-site
infections per patient-year (nondiabetics), and 0.87 exit-site
infections per patient-year (diabetics). The BSI rates reported in this
review were comparable to those reported for more conventional forms of
hemodialysis vascular access (0.09-0.20 BSIs per patient-year).284
297-299
Two studies have examined the potential impact of tunneled
hemodialysis catheters on the risk of subsequent CRI. In a
nonrandomized study, Hickman catheters used for prolonged hemodialysis
access was associated with a significantly lower rate of BSI (0.08 BSIs
per 100 catheter-days) than were nontunneled hemodialysis
catheter.\300\ Schwab et al. prospectively examined the use of cuffed,
tunneled, double-lumen jugular venous catheters for prolonged
hemodialysis access. Compared with percutaneously inserted, noncuffed
subclavian dialysis catheters, double-lumen jugular venous catheters
had a longer live span, a lower (1.3% vs 3.6%) incidence of associated
BSIs, but a significantly higher incidence of exit-site infection (29%
vs 9%).\295\
Hemodialysis catheters may become contaminated by a variety of
proposed mechanisms: (1) penetration of organisms from the skin due to
the pulsatile action of the dialysis pump; (2) manipulation of catheter
connections by medical personnel with contaminated hands; (3) leakage
of contaminated hemodialysis fluid into the blood compartment; or (4)
administration of contaminated blood or other solutions through the
catheter during the dialysis session.
Microbiology
CR-BSIs in hemodialysis patients, as in other patient populations,
are most frequently caused by S. epidermidis.274-276 281-283 285
However, because of their high rates of
[[Page 49989]]
colonization with S. aureus,\301\ hemodialysis patients have a greater
proportion of CR-BSIs due to S. aureus \284\ than among other patient
populations.
Strategies for Prevention of Hemodialysis Catheter-Related Infections
Strategies for the prevention of infections associated with the use
of hemodialysis catheters have not been as rigorously examined as those
proposed for the prevention of infections associated with CVCs used for
other purposes. Although there are limited data on infectious
complications in hemodialysis settings associated with various types of
catheters, frequency of catheter change, cutaneous antisepsis, and
prophylactic administration of antimicrobials, no studies examining
catheter-site dressing regimens, or the utility of newer devices, such
as antimicrobial-impregnated hemodialysis catheters have been
published.
Cutaneous antisepsis. In some series, as many as 50 to 62% of
hemodialysis patients have been found to be carriers of S.
aureus.301-304 Therefore, skin antisepsis is a crucial component
for the prevention of hemodialysis catheter-associated infections. In
one randomized, controlled study of 129 subclavian dialysis catheters,
the routine application of povidone-iodine ointment to catheter-
insertion sites was more effective than plain gauze in reducing the
incidence of exit-site infections (5% vs 18%), catheter-tip
colonization (17% vs 36%), and BSIs (2% vs 17%);304 duration of
catheterization was comparable for treated (mean, 38.6 days) and
nontreated (mean, 36.2 days) catheters, each ranging from 2-210 days.
The beneficial effect of povidone-iodine ointment was most evident
among patients with S. aureus nasal carriage where its use reduced the
incidences of BSI and exit-site infection by 100% and catheter-tip
colonization by 71%. No adverse effects were detected with the routine
application of povidone-iodine ointment to subclavian dialysis
catheter- insertion sites.
Catheter changes. Since attainment and preservation of vascular
access in patients with chronic renal failure are often difficult, the
frequency of catheter change and the role of guidewire catheter
exchange are of utmost importance. However, to date, there are limited
data on which to base recommendations for either of these issues in
hemodialysis patients. One prospective, randomized trial of subclavian
dialysis catheters using guidewire exchange suggested that the rate of
BSIs was comparable when catheters were changed weekly or when
clinically indicated.305 One recent study examined the role of
guidewire exchange in the treatment of infected jugular vein
hemodialysis catheters. In this study, a 92% one-year catheter survival
was observed with the combined use of guidewire exchange and
administration of antimicrobials 48 hours before and 2 weeks after
guidewire exchange, when frank pus was not present at the exit
site.306 These findings, however, are contrary to a large body of
data suggesting that guidewire exchange should not be done in the
setting of documented CRI.59 131 153 307 308
Prophylactic antimicrobials. Hemodialysis patients receiving
antistaphylococcal antimicrobials at the time of catheter placement
have been shown to have a lower incidence of CRI.274 276 277 309
However, the role of prophylactic antimicrobials has not been directly
studied.
Whether hemodialysis catheters can be treated in the same way as
CVCs used for other purposes is unclear. Prospective, controlled trials
of hemodialysis catheters are needed to determine the epidemiology of
CRIs associated with their use and to evaluate the role of preventive
role of different types of catheter materials, appropriate insertion
sites, intervals for catheter change, guidewire exchange, catheter-site
dressing regimens, and the use of newer modalities (e.g., such as
antimicrobial-impregnated hemodialysis catheters).
IX. Intravascular Device-Related Infections in Pediatric Patients
This section addresses some of the specific issues relevant to
intravascular access and intravascular device-related infections among
the pediatric population. However, the epidemiology of intravascular-
device related infections in pediatric patients is less well-described
than that in adults, and there are limitations to the existing data.
First, few controlled trials of intravascular devices in children have
been reported; most published data are derived from uncontrolled
retrospective or prospective studies. Second, pediatric data that are
available were derived, largely, from studies in neonatal (NICU) or
pediatric intensive care units (PICU) where rates of infection are
usually higher than on general pediatric wards. Finally,
semiquantitative culture methods have, in large part, not been used in
the studies of CRIs in children because such cultures require catheter
removal.
Microbiology
As in adults, most CR-BSIs in children are caused by staphylococcal
spp., with S. epidermidis being the predominant species.310 311
Other species of gram-positive cocci and fungi are the next most
frequently isolated pathogens, with Malassezia furfur being an
especially common pathogen in neonates receiving intravenous
intralipids.311-319
Bertone et al. performed quantitative skin cultures on 50 neonates
to determine the microbial flora present at commonly used catheter-
insertion sites.320 Only 33 neonates had an intravascular device
in place at the time of culturing; 25 had peripheral venous catheters
and eight had CVCs. The highest mean colony counts were found at
jugular sites (2.7 x 104 cfus/10cm2) and the lowest at
subclavian sites (5.2 x 103 cfus/10-cm2). However, femoral
and jugular sites had similar mean colony counts as did subclavian and
umbilical sites. Although CoNS was the pathogen most frequently
cultured from all body sites, other microbial species (e.g., aerobic
gram-negative bacilli, yeast, and Enterococcus spp.) were more commonly
cultured from umbilical and femoral sites.320
Epidemiology
The majority of nosocomial BSIs in children are also related to the
use of an intravascular device. During 1985-1990, children's hospitals
participating in NNIS and conducting ICU surveillance reported
significantly higher rates of BSI among PICU patients with CVCs (11.4
BSIs per 1,000 central-catheter days) compared with those without CVCs
(0.4 BSIs per 1,000 noncentral-catheter days).8 Participating
Level III NICUs reported a median of 5.1 BSIs per 1,000 umbilical or
central-catheter days for the 1,500 gram birthweight group
and 14.6 BSIs per 1,000 umbilical or central-catheter days for the
<1,500 gram="" birthweight="" group="" over="" the="" same="">321 Birthweight
and device utilization were important determinants of a NICU infant's
risk for acquiring BSI.321 Others have shown receipt of
intravenous lipids to also be an important risk factor for the
acquisition of CR-BSI, particularly CoNS BSIs, among neonates.86
Cronin studied 376 catheters, of varying types, to determine the
incidence of catheter colonization and CR-BSI among NICU
patients.322 The incidence of catheter colonization varied by type
of catheter, site of insertion, and duration of catheterization.
Consistent with the findings of other investigators, the rate
[[Page 49990]]
of catheter colonization was significantly lower among patients
receiving systemic antimicrobials, having birthweight 1500
gm, and not receiving parenteral nutrition. In general, the
colonization rates detected in this study were higher than those
previously reported for catheters in adults and children.17 148
311 312 However, the authors could not conclusively determine the
relationship of catheter colonization to BSI.
Peripheral venous catheters. As in adults, the use of peripheral
venous catheters in pediatric patients may be complicated by phlebitis,
extravasation, and catheter colonization. Garland et al. prospectively
studied 654 peripheral Teflon catheters in PICU patients to determine
the incidence of and risk factors for each of these
complications.311 Of the 654 catheters studied, 83 (13%) were
associated with phlebitis. Catheter location, infusion of
hyperalimentation fluids with continuous intravenous lipid emulsions,
and length of ICU stay before catheter insertion were all factors that
increased a patient's risk for phlebitis. However, contrary to the
studies among adults, the risk of phlebitis did not increase with the
duration of cannulation. The overall incidence of phlebitis in this ICU
population (13%) was comparable to that reported in general pediatric
patients (10%); for children >10 years of age the incidence of
phlebitis (21%) was comparable to that reported for adults 169 and
older children.323
Of 459 peripheral venous catheters cultured by Garland, 54 (11.8%)
were colonized. However, only one (1.9%) of these colonized catheters
was associated with CR-BSI. In an earlier study, comparable rates of
catheter colonization (10.4%) were found for Teflon peripheral
catheters (n=115) used in patients on general pediatric wards.312
Time in place was the single most important predictor of subsequent
catheter colonization, with the incidence of colonization increasing
threefold after catheters remained in place >144 hours.311 Between
48 and 144 hours, the catheter colonization rate was stable at 11%.
Other factors significantly but less strongly associated with catheter
colonization were patient age and receipt of lipid emulsions. Catheters
inserted emergently were no more prone to colonization than were those
inserted electively.311
Extravasation, the most frequent complication, occurred with 28% of
catheters. Several risk factors for extravasation were identified,
including patient age (1 year), receipt of anticonvulsant,
and duration of catheterization (72 hours); the risk of
extravasation decreased significantly after the catheter was in place
for 72 hours.311
There are limited data examining the relationship of catheter
material to the risk of infection among pediatric patients. In one
study of premature infants, Teflon catheters and steel needles used in
scalp veins had a comparable risk of infection. However, Teflon
catheters had a significantly longer survival than did steel
needles.313
Peripheral arterial catheters. In a prospective study using
semiquantitative culture of 340 peripheral arterial catheters, Furfaro
identified two risk factors for CRI: (1) use of an arterial system of a
certain design, and (2) duration of catheterization.314 The
implicated arterial system (system A) contained a stopcock and a 120-cm
pressure tubing through which blood was drawn back to clear the line of
heparin before taking a sample. The alternate system (system B), with a
significantly lower risk of infection, contained a one-way valve that
did not permit blood backflow into the tubing. The authors noted that
the implicated arterial system (A) was the design most widely used in
U.S. hospitals.314
Although there was a correlation between duration of
catheterization and risk of catheter colonization, the risk remained
constant for 2-20 days at 6.2%. Catheters in place 48 hours
had a zero risk of colonization.314
Umbilical catheters (UCs). Although the umbilical stump becomes
heavily colonized soon after birth, umbilical vessel catheterization is
often used for vascular access in newborn infants because umbilical
vessels are easily cannulated, allow for delivery of intravenous
fluids/medications, permit easy collection of blood samples, and permit
measurement of hemodynamic status. Studies of the infectious
complications associated with UCs indicate that the incidences of
catheter colonization and BSI appear to be similar for umbilical vein
catheters (UVC) and umbilical artery catheters (UAC). The incidences of
colonization reported among UACs have ranged from 40 to 55%;324
325 those among UVCs have varied between 22% and 59%.324-326 The
incidences of BSI detected for the two types of catheters are also
similar, 5% for UACs and 3%-8% for UVCs.324 326 However, the risk
factors for infection appear to differ for the two types of catheters.
Landers et al. found that neonates with very low birthweight and
prolonged receipt of antimicrobials were at increased risk for UAC-
related BSIs. In contrast, those with higher birthweight and receipt of
parenteral nutrition fluids were at increased risk for UVC-related BSI;
duration of catheterization was not an independent risk factor for
infection either type of umbilical catheter.324
In addition to the risk of endemic infection, umbilical vessel
catheterization has been associated with epidemics among critically ill
NICU infants. Solomon et al. reported an outbreak of C. parapsilosis
fungemia among NICU infants 41 in which duration of umbilical
artery catheterization, prolonged receipt of parenteral nutrition, and
low gestational age were risk factors for fungemia.41
Several investigators have reported lower rates of UC colonization
among infants or neonates receiving systemic antimicrobials during
umbilical catheterization.315 325 326 However, the one prospective
study of prophylactic antimicrobials in patients with chronic UACs
found no clear benefit to this therapy.327
Central venous catheters. The use of indwelling catheters (e.g.,
Hickmans and Broviacs, TIDs) in children has become increasingly
important over the past decade for the treatment of children with
chronic medical conditions, especially malignancies. The Broviac,
rather than the Hickman, catheter is preferentially used in children
because of its smaller diameter; TIDs may be particularly advantageous
in younger pediatric patients (73 328 329
Although data from the Children's Cancer Study Group suggest that
as many as 18% of all chronic venous access devices in children are
removed due to infection,330 the use of these devices in children
have generally been associated with low rates of infections.