Should we use closed or open infusion containers for prevention of bloodstream infections?

Abstract

Hospitalized patients in critical care settings are at risk for bloodstream infections (BSI). Most BSIs originate from a central line (CL), and they increase length of stay, cost, and mortality. Open infusion containers may increase the risk of contamination and administration-related (CLAB) because they allow the entry of air into the system, thereby also providing an opportunity for microbial entry. Closed infusion containers were designed to overcome this flaw. However, open infusion containers are still widely used throughout the world.The objective of the study was to determine the effect of switching from open (glass, burettes, and semi-rigid) infusion containers to closed, fully collapsible, plastic infusion containers (Viaflex) on the rate and time to onset of central line-associated bloodstream infections CLABs.
An open label, prospective cohort, active healthcare-associated infection surveillance, sequential study was conducted in four ICUs in Mexico. Centers for Disease Control National Nosocomial Infections Surveillance Systems definitions were used to define device-associated infections.
A total of 1,096 adult patients who had a central line in place for >24 hours were enrolled. The CLAB rate was significantly higher during the open versus the closed container period (16.1 versus 3.2 CLAB/1000 central line days; RR = 0.20, 95% CI = 0.11-0.36, P < 0.0001). The probability of developing CLAB remained relatively constant in the closed container period (1.4% Days 2-4 to 0.5% Days 8-10), but increased in the open container period (4.9% Days 2-4 to 5.4% Days 8-10). The chance of acquiring a CLAB was significantly decreased (81%) in the closed container period (Cox proportional hazard ratio 0.19, P < 0.0001). Mortality was statistically significantly lower during the closed versus the open container period (23.4% versus 16.1%; RR = 0.69, 95% CI = 0.54-0.88, P < 0.01).
Closed infusion containers significantly reduced CLAB rate, the probability of acquiring CLAB, and mortality.

Full text

Page 1

RESEARCHOpen Access
Should we use closed or open infusion
containers for prevention of bloodstream
infections?
Manuel S Rangel-Frausto1†, Francisco Higuera-Ramirez2†, Jose Martinez-Soto3†, Victor D Rosenthal4*
Abstract
Background: Hospitalized patients in critical care settings are at risk for bloodstream infections (BSI). Most BSIs
originate from a central line (CL), and they increase length of stay, cost, and mortality. Open infusion containers
may increase the risk of contamination and administration-related (CLAB) because they allow the entry of air into
the system, thereby also providing an opportunity for microbial entry. Closed infusion containers were designed to
overcome this flaw. However, open infusion containers are still widely used throughout the world.
The objective of the study was to determine the effect of switching from open (glass, burettes, and semi-rigid)
infusion containers to closed, fully collapsible, plastic infusion containers (Viaflex®) on the rate and time to onset of
central line-associated bloodstream infections CLABs.
Methods: An open label, prospective cohort, active healthcare-associated infection surveillance, sequential study
was conducted in four ICUs in Mexico. Centers for Disease Control National Nosocomial Infections Surveillance
Systems definitions were used to define device-associated infections.
Results: A total of 1,096 adult patients who had a central line in place for >24 hours were enrolled. The CLAB rate
was significantly higher during the open versus the closed container period (16.1 versus 3.2 CLAB/1000 central line
days; RR = 0.20, 95% CI = 0.11-0.36, P < 0.0001). The probability of developing CLAB remained relatively constant in
the closed container period (1.4% Days 2-4 to 0.5% Days 8-10), but increased in the open container period (4.9%
Days 2-4 to 5.4% Days 8-10). The chance of acquiring a CLAB was significantly decreased (81%) in the closed
container period (Cox proportional hazard ratio 0.19, P < 0.0001). Mortality was statistically significantly lower
during the closed versus the open container period (23.4% versus 16.1%; RR = 0.69, 95% CI = 0.54-0.88, P < 0.01).
Conclusions: Closed infusion containers significantly reduced CLAB rate, the probability of acquiring CLAB, and
mortality.
Background
Patients in hospitals are at risk for bloodstream infec-
tions (BSI), mainly in critical care settings. Most BSIs
originate from a central line (CL) [1], and they increase
length of stay, cost, and mortality [2-14].
During setup, admixture preparation, and administra-
tion [15,16], there is a high risk of contamination of
intravenous (IV) fluids. When the system is vented, as is
mandatory with open infusion containers, there are
extra risks of extrinsic contamination.
Both open and closed infusion containers are used
worldwide [5,17]. Open infusion containers consist of
rigid (glass, burette) containers or semi-rigid plastic con-
tainers that must admit air (air filter or needle) to empty
the contents (Figures 1, 2, 3). Closed infusion containers
consist of fully collapsible plastic containers that do not
require or use any external vent (air filter or needle) to
empty the solution, and have self-sealing injection ports
(Figure 4).
Standard practice incorporates the use of closed sys-
tems to prevent healthcare-associated infections (e.g.,
catheter-associated urinary tract infections [CAUTI])
[18], ventilator-associated pneumonia [VAP] [19], surgi-
cal site infections [20], and central line-associated-
* Correspondence: victor_rosenthal@inicc.org
† Contributed equally
4Medical College of Buenos Aires, Argentina
Rangel-Frausto et al. Annals of Clinical Microbiology and Antimicrobials 2010, 9:6
http://www.ann-clinmicrob.com/content/9/1/6
© 2010 Rangel-Frausto et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative
Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and
reproduction in any medium, provided the original work is properly cited.

Page 2

bloodstream infections [CLAB]) [3,5,17]. Numerous
countries have reported outbreaks of infusion-related
CLAB traced to contamination of infusate in open infu-
sion systems [15,21-26]. Extrinsic or in-use contamina-
tion plays the most important role in bacterial
contamination of the infusion system [26-28].
In Argentina, switching from open, semi-rigid plastic
infusion containers to closed, fully collapsible plastic
infusion containers was cost effective and resulted in a
64% CLAB rate reduction [17]. However, the open con-
tainers studied were limited to semi-rigid plastic con-
tainers, and time to CLAB was not studied in Argentina.
In this study conducted in Mexico, 3 types of open
infusion containers (glass bottles, burettes, and semi-
rigid plastic) were utilized. A time-to-onset of CLAB
analysis was also included to evaluate the probability of
developing CLAB.
We report the results of a prospective, sequential
study undertaken to determine the impact of switching
from an open (glass bottles, burettes, and semi-rigid
plastic) to a closed, fully collapsible plastic infusion con-
tainer (Viaflex®) on the rate and time-to-onset of CLAB
in Mexico.
Methods
Setting
Three hospitals in Mexico City participated in the study:
General Hospital, Century XXI Specialties Instituto
Mexicano del Seguro Social (IMSS) Hospital, and
Mancera IMSS Hospital. In accordance with recommen-
dations [29], each has an active infection control pro-
gram, composed of a physician trained in internal
medicine and infectious diseases, and an infection con-
trol nurse. General Hospital is a public hospital, while
Century XXI Specialties IMSS Hospital and Mancera
IMSS Hospital are social security hospitals. The four
intensive care units (ICUs) in the study centers operate
at the highest level of complexity in Mexico (Level IV),
providing treatment for patients who have undergone
open-heart surgery, neurosurgery, gastrointestinal or
orthopedic surgery, as well as patients with trauma or
serious medical illnesses.
The study was conducted in accordance with the
Declaration of Helsinki. The study protocol was
approved by the ethics committee at General Hospital
and Century XXI Specialties IMSS Hospital. Approval
was not required by Mancera IMSS Hospital, as study
specific procedures did not exceed the scope of standard
medical care.
A subject informed consent letter, which detailed
information regarding adverse effects, was approved by
the ethics committee at General Hospital. However, sub-
ject informed consent was not required at either Cen-
tury XXI Specialties IMSS Hospital or Mancera IMSS
Hospital, as the study did not disclose the patient’s con-
fidentiality or privacy and did not involve any additional
risks beyond the usual medical interventions performed
in the participating ICUs.
Figure 1 Open Infusion Container - Glass container with air
filter.
Figure 2 Open Infusion Container - Semi-rigid container with
air filter.
Rangel-Frausto et al. Annals of Clinical Microbiology and Antimicrobials 2010, 9:6
http://www.ann-clinmicrob.com/content/9/1/6
Page 2 of 8

Page 3

Data Collection
Patients who had a CL in place for >24 hours were
enrolled from each of the study ICUs. A trained nurse
prospectively recorded on case report forms the patient’s
gender, average severity-of-illness score (ASIS) on ICU
entry [30], device utilization, antibiotic exposure, and all
active infections identified while in the ICU. The deci-
sion to obtain blood cultures was made independently
by the patient’s physicians. Standard laboratory methods
were used to identify microorganisms recovered from
positive blood cultures [31].
Definitions
United States Centers for Disease Control National
Nosocomial Infections Surveillance Systems (CDC-
NNIS) program definitions were used to define device-
associated infections: CLAB (both laboratory-confirmed
infection and clinical primary nosocomial sepsis), cathe-
ter-associated urinary tract infection CAUTI, and venti-
lator-associated pneumonia VAP [32].
An open infusion container was defined as a rigid
(glass, burette) or semi-rigid plastic container that must
admit air to empty (air filter or needle). A closed infu-
sion container was defined as a fully collapsible, plastic
container that does not require or use any external vent
(air filter or needle) to empty the solution, and has
injection ports that are self-sealing.
Study Design
Active surveillance for CLAB and compliance with
infection control practices continued throughout the
study using CDC-NNIS methodologies, definitions, and
criteria [30]. The open container period lasted 6 months
(December 2002 to May 2003) at General Hospital, 5
months (February 2003 to June 2003) at Century XXI
Specialties IMSS Hospital, and 4 months (April 2003 to
July 2003) at Mancera IMSS Hospital. The closed con-
tainer period lasted 6 months (June 2003 to November
2003), 5 months (July 2003 to November 2003), and 4
months (August 2003 to November 2003), respectively.
Baxter Viaflex®, a fully collapsible plastic closed infu-
sion container, was used during the closed container
period. Commercially available open infusion containers
(glass container, semi-rigid plastic container, and burette
products) were used during the open container period.
Protocol specified target hand hygiene (HH) and CL
care compliance was set at ≥70% and ≥ 95%, respec-
tively. We assessed HH (which included the use of alco-
hol based hand rub at one hospital) compliance [33],
placement of gauze on CL insertion sites, condition of
gauze dressing (absence of blood, moisture, and gross
soilage; occlusive coverage of insertion site) [34,35], and
documentation for date of CL insertion. A research
nurse observed healthcare workers (physicians, nurses,
and paramedical staff) three times per week across all
work shifts and recorded information on a standard
form. In addition, we conducted active surveillance for
other common healthcare-acquired infections such as
VAP and CAUTI. Mortality data also were collected.
Data Analysis
Outcomes measured during the open and closed con-
tainer periods included the incidence density rate of
CLAB (number of cases/1000 CL days) and time to
CLAB. Chi-square analyses for dichotomous variables
and t-test for continuous variables were used to analyze
baseline differences between periods. Relative risk (RR)
Figure 3 Open Infusion Container - Burette with air filter.
Figure 4 Closed Infusion Container - Fully collapsible plastic
container without air filter.
Rangel-Frausto et al. Annals of Clinical Microbiology and Antimicrobials 2010, 9:6
http://www.ann-clinmicrob.com/content/9/1/6
Page 3 of 8

Page 4

ratios, 95% confidence intervals (CIs), and P-values were
determined for all outcomes. Time to first CLAB was
analyzed using a log-rank test and is presented graphi-
cally using Kaplan-Meier curves. In addition, simple life
table conditional probabilities are presented graphically
to help explain the changing risk of infection over time
(Figure 5).
Results
During the study, 1096 patients were enrolled, 548 dur-
ing the open container period and 548 during the closed
container period. Patients in both periods were statisti-
cally similar regarding patient demographics, underlying
illness (except for endocrine diseases, cardiac failure,
angina pectoris, chronic obstructive pulmonary disease,
hepatic failure, and presence of previous infection),
length of stay, ASIS, device utilization, and antibiotic
usage (Table 1).
No new infection control interventions, training pro-
grams, products or technologies were introduced during
the study periods and all of the investigators, key study
personnel, classifications, and diagnostics techniques
remained constant throughout the entire study. A lead-
in period of at least 3 months duration was conducted
to standardize HH and CL care compliance practice.
Compliance with HH during both periods was above
70% (81.0% and 74.7% during the open and closed con-
tainer periods, respectively; RR = 0.92; 95% CI = 0.90 -
0.95).
Presence of gauze at CL site was 99.2% and 99.0%
during the open and closed container periods, respec-
tively (RR = 1.00; 95% CI = 0.99 - 1.00). Correct condi-
tion of gauze was 98.3% and 97.9% during the open and
closed container periods, respectively (RR = 1.00; 95%
CI = 0.99 - 1.00). Presence of the date at the CL inser-
tion site/administration set was 100% during both the
open and closed container periods.
The incidence density rate and percentage of patients
with CLAB were both significantly lower in the closed
container period compared to the open container period
(Table 2). The majority of laboratory confirmed CLAB
isolates in both the open (53.3%, 16/30) and closed
(66.6%, 4/6) infusion container period were Gram-posi-
tive; Gram-negative isolates represented 43.3% (13/30)
in the open container period and 33.3% (2/6) in the
closed container period. The distribution of microorgan-
isms for both container periods is shown on Table 3.
In this study, we examined the timing of when the
first CLAB was acquired comparing the open and closed
container periods (Figure 5). The majority (62%) of
patients had a CL in place for ≤4 days. When examining
three-day intervals, the conditional probability of acquir-
ing a CLAB in the closed container period was observed
to be relatively constant (1.4% at Days 2-4 to 0.5% at
Days 8-10). In the open container period, the condi-
tional probability of acquiring a CLAB was higher in
each three-day interval compared to the corresponding
three-day interval in the closed container period. The
conditional probability of acquiring a CLAB in the open
container period ranged from 4.9% at Days 2-4 to 5.4%
at Days 8-10.
Overall, the chance of a patient acquiring a CLAB
decreased significantly-by 81%- in the closed container
period (Cox proportional hazard ratio 0.19, P < 0.0001).
There was no statistically significant difference
between the two periods with respect to incidence of
CAUTI or VAP rate (Table 2).
Mortality during the closed container period (16.1%)
was statistically significantly lower than during the open
container period (23.4%) (RR = 0.69; 95% CI = 0.54 -
0.88; P = 0.002) (Table 2).
Discussion
Critically ill patients commonly require CL access for
administration of large volumes of IV fluid, medications,
blood products, or for hemodynamic monitoring. All of
these carry a greater risk of CLAB [4,5,10,11].
Studies have shown that CLABs increase length of
stay, cost, and attributable mortality [2,3,5,36]. In 2005,
Stone conducted a meta-analysis of all the cost studies
published over a five-year period and found that the
average cost of one CLAB was $36,441 (US) [36]. Pre-
viously, Rosenthal et al in Argentina published that
CLAB resulted in an extra 12 days of hospitalization
and $4,888 (US) [3]. Similarly, Higuera et al reported
that CLAB in Mexico resulted in an extra six days of
hospitalization and $11,560 (US) [2].
Moreover, CLABs are apparently related to increased
attributable mortality: In an Australian study, CLABs
resulted in excess mortality of 12% [37], and in a study
from the United States, Pittet et al reported an
Figure 5 Cumulative probability of 1stCLAB displayed by CL
days.
Rangel-Frausto et al. Annals of Clinical Microbiology and Antimicrobials 2010, 9:6
http://www.ann-clinmicrob.com/content/9/1/6
Page 4 of 8

Page 5

Table 1 Patient demographics, underlying illness, length of stay, device utilization and antibiotic usage during the
two study periods
Open infusion container
N = 548
Closed infusion container
N = 548
RR95% CI P-value
% (n/N)% (n/N)
Sex (Male)
Sex (Female)
Endocrine Disease
Cardiac Failure
Angina Pectoris
Cardiac Surgery
COPD
Cancer
Renal Impairment
Hepatic Failure
Abdominal Surgery
Thoracic Surgery
Trauma
Previous Infection
Stroke
Immunodeficiency
Urinary catheter
Mechanical ventilator
43.8% (240/548)
56.2% (308/548)
21.0% (45/214)
27.1% (58/214)
2.8% (6/214)
0.0% (0/214)
0.0% (0/214)
4.2% (23/548)
4.7% (10/214)
0.9% (2/214)
23.4% (128/548)
0.0% (0/214)
0.0% (0/214)
0.0% (0/214)
0.0% (0/214)
0.9% (2/214)
91.2% (500/548)
66.4% (364/548)
Mean ± SD
48.0% (263/548)
52.0% (285/548)
30.3% (154/508)
42.5% (216/508)
10.0% (51/508)
0.4% (2/508)
15.0% (76/508)
4.2% (23/548)
7.5% (38/508)
3.9% (20/508)
22.1% (121/548)
0.8% (4/508)
1.2% (6/508)
21.5% (109/508)
1.0% (5/508)
1.6% (8/507)
94.0% (515/548)
68.2% (374/548)
Mean ± SD
1.10
0.93
1.44
1.57
3.58
-
-
1.00
1.60
4.21
0.95
-
-
-
-
1.69
1.03
1.03
0.96 - 1.25
0.83 - 1.03
1.08 - 1.93
1.23 - 2.00
1.56 - 8.22
-
-
0.57 - 1.76
0.81 - 3.15
0.99 - 17.86
0.76 - 1.18
-
-
-
-
0.36 - 7.89
1.00 - 1.07
0.95 - 1.12
0.16
-
0.01
<0.01
<0.01
0.36
<0.01
1.0
0.17
0.03
0.61
0.19
0.11
<0.01
0.15
0.50
0.08
0.52
ICU stay (days)
Age (yrs)
Severity-of-illness score
CL utilization per patient (days)
Mechanical ventilator utilization per patient (days)
Urinary catheter utilization per patient (days)
6.5 ± 6.56
54.1 ± 18.29
3.7 ± 0.90
6.7 + 7.35
4.3 ± 6.10
6.0 ± 6.64
Defined daily dose
7.1 ± 6.69
54.4 ± 18.58
3.8 ± 0.82
7.4 + 7.54
4.7 ± 6.61
6.6 ± 6.65
Defined daily dose
-
-
-
-
-
-
-
-
-
-
-
-
0.13
0.78
0.45
0.11
0.24
0.18
Antibiotic use per 1000 days1485 1540--0.056
n, number of subjects with the characteristic present at baseline; N, total number of subjects with the with available (non-missing) results for the characteristic at
baseline.
Table 2 Incidence of CLAB, CAUTI, VAP, and mortality during the two study periods
Open infusion container
N = 548
Closed infusion container
N = 548
RR95% CIP-value
CL days no.
CLAB no.
CLAB per 1000 CL days
Percentage of patients with CLAB
Urinary catheter days no.
CAUTI no.
CAUTI per 1000 catheter days
Mechanical ventilator days no.
VAP no.
VAP per 1000 mechanical ventilator days
Deaths no.
Percentage of patients who died
3661
59
16.1
10.8
3302
30
9.1
2344
66
28.2
128
23.4
4055
13
3.2
2.4
3590
36
10.0
2592
71
27.4
88
16.1
-
-
-
-
-
-
0.20
0.22
-
-
1.10
-
-
0.97
-
0.69
0.11 - 0.36
0.12 - 0.40
-
-
0.68 - 1.79
-
-
0.70 - 1.35
-
0.54 - 0.88
<0.01
<0.01
-
-
0.69
-
-
0.87
-
<0.01
no., number; CL, central line; CLAB, central line-associated bloodstream infection; CAUTI, catheter-associated urinary tract infection; VAP, ventilator-associated
pneumonia
Rangel-Frausto et al. Annals of Clinical Microbiology and Antimicrobials 2010, 9:6
http://www.ann-clinmicrob.com/content/9/1/6
Page 5 of 8