Journal

of Hospital

Infection

(1994)

27, 263-273

Investigation
of contaminated
parenteral
nutrition
fluids associated
with an outbreak
of Serratia
odorifera
septicaemia
J. A. Frean,

L. Arntzen,

I. Rosekilly

Department
of Tropical Diseases,
Institute of Medical Research and
burg,
Isotec Nutrition
(Pty)
Accepted

and M. IsaQcson

School of Pathology
of the South African
University
of the Witwatersrand,
JohannesSouth Africa
Ltd, Johannesburg,
South Africa

for publication

2 February

1994

Summary:

An outbreak
of fatal septicaemia
caused
by Sewatia
odorifera
biotype
1 involved
infants
at several
hospitals;
the common
vehicle
of
infection
was contaminated
parenteral
nutrition
fluid.
The transfusate
had
been made up in a flexible
film isolator
system.
The implicated
organism
was
recovered
from surfaces
inside
the isolator,
despite
routine
decontamination
procedures
having
been carried
out shortly
before.
Our investigation
into the
origin
of contamination
revealed
several
shortcomings
in the infusate
compounding
process.
We noted
deficiencies
in cleaning
and decontamination
procedures,
and in storage
and sterility
testing
policies,
but the origin
and
mechanism
of the contamination
were
unclear.
Withdrawal
of parenteral
nutrition
products
and revision
of decontamination
procedures
terminated
the outbreak.
The
efficacy
of peracetic
acid
treatment
of flexible
film
isolators,
given the circumstances
of this outbreak,
may need further
investigation.
Regular
training
and assessment
of admixture
technicians
is important.
Keywords:
tia odorifeva;

Parenteral
nutrition
transfusate;
septicaemia;
infants.

bacterial

contamination;

Serra-

Introduction
Parenteral nutrition (PN) is frequently used to achieve an anabolic state in
patients who cannot otherwise maintain nitrogen balance. Infection ranks
high amongst the serious complications of PN.*v3 The potential sources of
bacteraemia associated with any intravascular line infection are related to
the cannula, and contamination
of the infusate.4 PN solutions may be
contaminated during manufacture (intrinsic contamination) or during their
compounding, or administration
in the hospital (extrinsic contamination).4
We report an outbreak of septicaemia in infants caused by contaminated PN
Correspondence
Johannesburg
0195-6701/94/080263+11

to: Dr J. A. Frean,
2000, South Africa.

Department

of Tropical

$08.00/O

263

Diseases,

SAIMR,

PO

Box

1038,

264

J. A. Frean

et al.

fluid; this is the second transfusate-transmitted

outbreak to occur in
Johannesburg in less than 3 years. 5 This article deals with our investigation
into the source of the contamination;
the clinical, microbiological
and
molecular epidemiological
analyses of the outbreak were done by the
Department of Medical Microbiology
of the School of Pathology, and have
been reported previously.j
The

outbreak

During a 2-week period in September 1992, eight infants died after

receiving intravenous parenteral nutrition (PN) fluids in four Johannesburg
hospitals. Serratia odorifera biotype 1 was isolated from blood cultures in
some cases, from liver and lung tissue removed at post-mortem in one case,
and from several sealed, unopened bags of transfusate belonging to the same
batches as had been administered to the patients. The isolation of the same
unusual species of bacterium, both from patients at several localities and
from parenteral nutrition
fluid of the corresponding
batches, clearly
indicated that an outbreak of transfusate-transmitted
infection had occurred
and an investigation was launched. Compounding of paediatric PN fluids
was stopped and all dispensed stock was recalled; no further cases occurred.
Subsequent ribotyping
showed that all the S. odorifera isolates were
identical.6
The implicated
vehicles of infection
were three batches of a
glucose-amino
acid solution and one of a glucose-amino
acid-lipid
emulsion mixture, supplied in 500 ml capacity bags for paediatric use. The
PN had been compounded in flexible film isolators (FFIs) by a private
company at its premises, and dispensed against prescriptions issued by
medical staff at the hospitals. Both private and government hospitals were
supplied by the company.
Methods

Review of compounding procedures

We obtained details of all aspects of the manufacturing
procedure by
interviewing
relevant staff members. Demonstrations
of the procedures
used in compounding
of PN products were given by the admixture
technicians who were routinely employed in this activity.
Microbiological investigations
Cotton wool swabs moistened with peptone-meat extract broth (nutrient
broth no. 2, Oxoid, Basingstoke) were used to sample numerous sites,
components and equipment surfaces inside the main FFI and in the rooms
comprising the compounding
facility. We also swabbed the throats and
hands of all members of staff (Table I). The swabs were broken off into
10ml quantities of nutrient broth and incubated overnight at 37C. The

transfusate

S. odorifera
Table

I. Microbiological

Specimen

investigations

type

PN

contamination

of parenteral
nutrition
ties and staff
Site

or origin

fluid (glucose-amino
acid)
Discard
bin swab
Peristaltic
pump
swab
Plastic
basket
swab
Throat
swabs

Issued
to hospital
but
unused
Interior
of FFI
Interior
of FFI
Interior
of FFI
Compounding
staff

Hand

Compounding

PN

swabs
fluid

components

Bags and tubes for PN compounding

Surface
swabs of doors,
grilles,
refrigerator,
disposal bins, exterior
of FFI;
drainage
water
samples

staff

Unused,
from
facilitys
stores
Unused,
from
facilitys
stores
Various
sites in compounding
facility

265

components,

compoundingfacili-

Culture

result

Serratia

odorifera

biotype

Serratia
odorzfera
biotype
1
Serratia
odorifera
bioty.pe
1
Staphylococcus
epidermzdis
Staphylococcus
aureus,
Staphylococcus
epidermidis,
Enterococcus
faecalis
Acinetobacter
lwofii,
Enterobatter
agglomerans
No growth
No

growth

Variety
teria;

of environmental
no Serratia
spp.

bac-

broth was subcultured

on MacConkey
and 5% horse blood agar and
incubated for 24 h, after which any growth was identified using standard
microbiological
techniques, and confirmed where appropriate by the API
20E identification
method (API System SA, France). The broth was
re-incubated
and subcultured
as above for 3 days, or until growth of
bacteria was detected. Unused samples of all the liquid components of PN
(amino acid solution, water for injection, glucose solution, electrolyte
concentrates, lipid emulsion), were obtained both from the facilitys stores,
and directly from the single supplier of components. The fluids were
inoculated in 5 ml volumes into aerobic and anaerobic Bactec (Becton
Dickinson) bottles and incubated for 7 days with daily readings. Fluid
samples were also filtered in 100 ml volumes through 0.45 pm filters which
were then placed in 100 ml of nutrient broth, incubated and subcultured
onto blood and MacConkey agar after 24 h if turbid; otherwise after 3 and
again after 7 days of incubation. Samples of unused tubing, 3 1 and 500 ml
capacity bags were filled with nutrient broth, and were incubated for up to a
week with daily inspections for broth turbidity.
Following removal of fluid
for culture (as described above), one unused bag of glucose-amino acid
solution from one of the implicated batches was stored at 4C for 47 days.
After this period of refrigeration,
the bacterial count was determined by
making lo-fold dilutions of the solution, and using a standard pour-plate
technique to count colonies.
Determination of pressure differential
A manometer, consisting of a U-shaped, graduated glass tube, partially
filled with water, was connected at one end to a rubber tube which entered

266

J. A. Frean

et al.

the main FFI through an air-tight gland (originally used for passage of
electrical wire into the interior). The other end of the glass tube was open to
room air. Differences in water level were measured:
(i) with the FFI blowers turned off
conditioning on;
(ii) with both blowers and air conditioning

but

the

facilitys

central

air

off.

In both cases the exhaust vent of the isolator was open to the exterior
normal operational fashion.

in the

Results

Transfusate manufacture
PN solutions were compounded in a totally enclosed flexible film isolator
(FFI) made of polyvinyl
chloride sheeting supported by steel tubing. The
operators worked inside half-suits,
entered from underneath,
which
protruded
into the FFI, and which allowed items in the latter to be
manipulated, whilst isolating the operator from the interior. The interior of
the FFI and all items used in the compounding process were treated with
peracetic acid vapour. The PN components were treated in a separate,
smaller isolator fitted with glove ports, and then transferred via a lock to the
main FFI. This was being done several times per day at the time of the
outbreak. Another small emergency container was used similarly for more
rapid transfer of limited amounts of material into and out of the main FFI.
Both isolators were pressurized by integral blowers with high efficiency
particulate air (HEPA) filters to a pressure of about 4 mm of water. PN
solutions were compounded from components supplied in flexible plastic
containers (bags), vials and glass bottles. The tubing and 3 1 motherbags
were usually supplied in sealed ethylene oxide-sterilized
pouches, although
the 500 ml bags for paediatric
solutions were supplied unsealed, in
cardboard
boxes. The compounding
of the PN solutions involved
connecting the bags and bottles containing the components (glucose, amino
acid and lipid emulsion) via plastic connectors to a 3 1motherbag, which was
allowed to fill by gravity. When required, supplementary
vitamins or
electrolytes were injected with needle and syringe into the motherbag via an
injection port, after spraying this with an alcohol solution. The paediatric
bags were filled by a peristaltic pump from a single 3 1 capacity motherbag.
This was done by connecting them one by one to a port in the tubing
attached to the pump, filling, then disconnecting the bag and heat sealing its
filling tube. The final volume varied between 150 and 450 ml, depending on
its formulation.
Adult PN solutions were supplied in 3 1 capacity bags.
Discarded material (used bags, bottles, tubing, ampoules and syringes) were
placed in a bin attached via a flexible collar to the FFI and fitted with an air
lock which allowed it to be disconnected, cleaned, decontaminated with

S. odorifera

peracetic
FFI.

transfusate

acid, and reconnected,

contamination

whilst allowing

work to continue

267

in the

Storage
Compounded
PN solutions were refrigerated
in a commercial upright
refrigerator before despatch. The shelf life in the compounding facility was
30 days but most bags were sent out within a few days of production. On
being dispensed, bags were labelled with an expiry date of 7 days
(refrigerated) f rom issue, unless vitamins had been added, in which case a
4-day expiry was set. Delivery to hospital was by insulated truck, the bags
having been transferred from the refrigerator in an insulated box.
Decontamination procedures
The main FFI was treated twice weekly with peracetic acid vapour for 5.5 h
at a time, in accordance with the FFI manufacturers recommendations (La
Calhene, France). A commercially manufactured 15% solution of peracetic
acid (Perasan; Chemrite,
Johannesburg),
diluted to produce a final
concentration
of hydrogen peroxide of 3.5%, was vaporized at a rate of
about 60 ml h- and carried by air flow (40 1min-) into the isolator. The
smaller FFI, and the emergency container, together with components
required for PN compounding, as well as the waste bin, were treated with
peracetic acid vapour as often as production demanded. The inside walls,
floor, and shelving in the FFIs were wiped down with 10% peracetic acid in
a detergent (Teepol)
solution.
Sterility testing
The isolators had been tested for effectiveness of decontamination
by the
use of Bacillus subtilis spore killing. This was done on a 6-monthly schedule;
since the duration of operation had been less than a year, this testing had
been done once, with satisfactory results. Settle plate testing had also been
satisfactory. Sterility testing of final products was performed only on the
paediatric bags. One bag from each batch of 12 filled from a 3 1 motherbag
was sent for culture; these samples would usually be accumulated over a
working week and be sent at its end; however this policy was not strictly
adhered to. The samples from two of the incriminated
batches had been
sent to the laboratory on the day of compounding, but another (containing
lipid emulsion) had been stored under refrigeration for 2 weeks before being
submitted for culture.
Compounding demonstration
We requested one of the technicians involved in routine compounding to
demonstrate the production process, and noted several shortcomings. The
wearing of three sets of gloves (one pair of latex surgical gloves put on before
entering the suit; the suits neoprene gloves; a final pair of sterile latex
surgical gloves worn over the suits gloves inside the isolator) made for

268

J. A. Frean

et al.

clumsiness in handling items such as connectors, syringes and needles.

There were several small holes in the half suits. The connection between the
3 1motherbag and the transfer set tubing (run through the peristaltic pump)
leaked fluid out and allowed air into the system. Spillage of liquid
components, mainly from partially emptied vials and bottles, was evident,
especially around the mouth and collar of the discard bin. We were also
shown how the discard bin was cleaned. This involved first emptying its
contents into a transfer container (which was subsequently unloaded into an
outdoor skip), and then washing the bin components (bin, collar and
connector) in a mixture of peracetic acid and detergent (Teepol).
The
disinfectant
and detergent volumes in the cleaning solution were not
accurately measured and the container holding the cleaning solution was too
small to allow immersion of the components.
Microbiological investigations
The results of the microbiological investigations are summarized in Table I.
One sample of PN fluid from one of the already implicated batches yielded
S. odorifera at a low concentration (estimated as < 1 bacterium ml-). After
storage of this bag at 4C for nearly 7 weeks, the count of S. odorifera had
increased to 1.6 x 10 ml-. Swabs taken inside the FFI from the rim of the
discard bin and the surface of the peristaltic pump grew S. odorifera
bioptype 1, with the same antibiotic susceptibility
pattern as the isolate
from PN fluid.
Determination of pressure dijfeerential
Results were as follows:
Experiment (i). The interior pressure dropped with respect to the outside
and became negative, reaching - 2 mm H,O. This took about 5 min.
Experiment (ii). The interior pressure dropped and equilibrated with the
exterior, i.e. did not become negative with respect to the exterior, within the
limits of detection of the rather crude instrument used.
Discussion

Most infusate-related
infections are due to colonization
of intravenous
cannulae.247 Infections associated with PN administration have frequently
been described,4v8 but a search of the recent literature did not reveal any
outbreaks of infection associated with intrinsically-contaminated
products
used for PN. Relatively few cases of extrinsic contamination
occurring
during compounding
or storage of PN transfusate (before the product
reached the patient) have been well documented. The previous outbreak in
Johannesburg involved apparent extrinsic contamination of PN fluid during
its compounding
by a different
manufacturer.
The origin of the

S. odorifera

transfusate

contamination

269

contamination
with a variety of Gram-negative
organisms was never
established. Extrinsic
contamination
by Enterobacter
cloacae during
compounding has occurred while manipulating connections; two outbreaks
of Candida parapsilosis infection resulted from the use of vacuum pumps in
compounding of PN solutions.
Dugleux et al.* implicated a refrigerator
as a source of contamination of PN. The scarcity of reports probably reflects
the fact that many episodes are not recognized as such. In many cases the
origin of extrinsic contamination
(whether hospital pharmacy, patient or
ward staff) is not known.3,
Although we found the implicated organisms inside the FFI, we can only
speculate on their source. Ihe normal habitat of Serratia spp. (soil and
water) suggests a natural environmental
source, but we w:ere not able to
isolate the organism from the immediate surroundings of the production
facility. We noted small breaches in the integrity of the FFI, which, under
certain circumstances,
could have allowed unsterile air to enter. The
bacteria could have been transferred from contaminated surfaces in the FFI
into the PN fluid during compounding
via the operators gloved hands
touching
connections,
or via aerosols drawn in through
the leaking
connection. The clumsiness inherent in the wearing of three pairs of gloves
made the touching of connections inevitable, which our own observations
confirmed. Another possibility is that an intrinsically
contaminated fluid
component was brought into the FFI and was used to make the implicated
batches of PN, and that the organisms found in the isolator resulted from
spillage during compounding
and discarding of used components. No
evidence of intrinsic contamination was found, but relatively small samples
were tested. We considered the possibility that electrical power failures
might compromise the integrity of the FFI, given the fact that small holes in
it had been noted at several sites. We saw that with the blowers (used
normally to create a 335 mm H,O positive pressure) switched off, the
pressure inside the canopy appeared to be negative with respect to the
exterior. This impression was created by the apparent sucking in of the
flexible film forming the roof and walls of the isolator. Had the interior and
exterior
pressures equilibrated,
the film would have merely drooped
because of its own weight. We attempted to quantitate any pressure
differential,
in order to assessthe potential effects of power failures. This
was considered in view of the fact that no emergency electrical power
backup was available, nor was any power failure monitoring
or warning
device in use. The findings indicated that isolated blower power supply
interruption
could lead to the entry of unsterile room air through any
breaches in the isolator structure, due to the air conditioner
creating a
pressure gradient. However a general power failure (which would affect
both blower and air conditioner)
would not cause ingress of unsterile air
across a pressure gradient. Nevertheless,
in this event the protective
positive pressure inside the canopy would be lost, and the exterior would
communicate with the exterior through any holes or defects in the structure.

270

J. A. Frean

et al.

The FFI, with the pump installed inside, had undergone two full cycles
of peracetic acid treatment just prior to our isolation of S. odorifera from its
interior. This raises questions as to the efficacy of the decontamination
procedure.
Peracetic acid, which dissociates to produce acetic acid,
hydrogen peroxide, oxygen and water,i6 has broad-spectrum
bactericidal,
fungicidal and sporicidal activity at low concentrations and temperatures,
maintaining
its properties in the presence of organic material.6-8 The
Centers for Disease Control and Prevention (CDC), Atlanta, have listed it
as a chemical sterilant and high-level disinfectant.i6 One of peracetic acids
disadvantages is thermodynamic
instability,
but commercial preparations
contain stabilizers to reduce this. In this outbreak, the level of hydrogen
peroxide in the peracetic acid mixture was monitored in accordance with the
manufacturers
recommendations
(La Calhene, France) at about 1-week
intervals and the dilution was adjusted to give a final concentration of 3.5%
in the evaporator used to produce the vapour. The final concentration of
peracetic acid in the vapour, however, was not known with any accuracy.
A possible reason for the failure of the peracetic acid treatment to
eliminate the S. odorifera contamination might be a protective effect of high
concentrations of glucose, amino acids and other PN components, drying
around the bacteria where the fluid was spilt which, as we observed, was a
frequent occurrence at sites at which we isolated the organism. This
hypothesis needs to be tested, in view of the alleged activity of peracetic acid
in the presence of organic material.j
However,
the relatively
poor
penetrating ability of peracetic acid vapour has been noted previously.
Trexler and Reynolds2 ascribed inability to sterilize animal cages with
peracetic acid fog, to the presence of dirt. It appears that adequate cleaning
of surfaces is a prerequisite for effective peracetic acid vapour treatment.
The hypertonicity
and low pH of crystalline amino acid-glucose solutions
inhibit growth and survival of most bacteria 21,22although Candida albicans is
less inhibited.1,23 Several species of bacteria and fungi are able to grow at
room temperature or higher in these mixtures; however, storage at 4C
generally inhibits microbial growth.24 Lipid emulsions are excellent culture
media for Gram-positive
and Gram-negative
bacteria and fungi.* Total
nutrient admixtures (lipid emulsion combined with dextrose and amino
acids) support the growth of most microorganisms better than PN without
lipid. 22 Lipid emulsion was ruled out as the vehicle of infection in our
outbreak, as it was used in only one of three batches of infected PN solution.
CDC guidelines allow refrigerated storage of PN solutions;22 the company
involved in the current outbreak refrigerated admixtures for up to 30 days,
but gave up to a weeks expiry period once bags were issued to hospitals.
Psychrophilic
organisms such as the S. odorifera involved in this outbreak
obviously pose potentially
greater problems than conventional
bacteria
under these circumstances. Dugleux et ~1.~ implicated a refrigerator as the
source for E. cloacae which contaminated PN solutions. In this outbreak the
refrigerator
was also used to store cartons of material used in the

transfusate

S. odorifera

contamination

271

compounding
of PN, and the investigators isolated E. cZoacae from the
exterior of the packaging. These workers were not able to explain the
mechanism of contamination
of the PN fluid. The temperature in their
refrigerator was as high as 15C in places, which emphasizes the importance
of measuring and recording the storage temperature of PN solutions. In the
we noted an absence of refrigeration
course of our investigation,
temperature monitoring. Although culturing small samples of compounded
PN fluid may not be sensitive enough to detect low-level contamination,22
had the sterility-monitoring
policy of the company been more rigorously
applied in our case, the problem might have been curtailed earlier.
In summary, we noted several deficiencies related to the use of FFI
technology,
namely:
spillage of PN components,
without
adequate
subsequent cleaning; small defects compromising
the integrity
of the
isolator envelope; clumsiness of operators because of the glove system;
inadequate cleaning facilities for the discard bin; demonstrable inadequacy
of peracetic acid treatment in the presence of spilled fluids; lack of
refrigerator temperature monitoring; and delayed sterility testing. Multiple
factors, culminating in the contamination of the solutions, were probably
involved, but it was not possible to determine the exact sequence of events.
The immediate withdrawal of PN products ended the outbreak. Facilities
and procedures were revised over a period of 2 months before production
was restarted. Revisions included installation of a larger sink for washing
FFI bin elements, improvements
in the loading pattern of the FFI to
prevent occlusion of areas during decontamination,
and more frequent and
regular sterility
testing of larger samples of compounded
PN fluid.
Investigation
of the potential
for terminal gamma-irradiation
of PN
products was initiated. No recurrence of contamination has been apparent.
Polyvinyl
chloride FFIs have been used for production
of germ-free
animals, safe autopsy of infectious bodies,26 as well as for production
of
pharmaceuticals and PN compounding. 27In all these examples of the use of
isolator
technology,
peracetic
acid is an appropriate
method
of
decontamination of the enclosure. This outbreak demonstrates that even the
use of proven technology is not a substitute for adherence to the basic
principles of aseptic technique. The difficulties involved in manipulating
components inside the isolator makes regular training and education in the
use and upkeep of the equipment essential. Complacency and blind faith
in any technology, especially in such a sensitive area as parenteral nutrition
production, must be avoided.

References
1. Goldman
DA, Maki
223: 136&l
364.
2. Freund
HR, Rimon
1990; 14: 39-41.

DG.

Infection

B. Sepsis

during

control
total

in total
parenteral

parenteral
nutrition.

nutrition.

J Parenter

JAMA

Enteral

1973;

Nutr

J. A. Frean

272

et al.

3. Henderson
DK.
Of snakes and bugs: nosocomial
infection
associated
with nutritional
support.
Nutr Clin Pratt 1991; 6: 3941.
4. Maki
DG. Infections
associated
with intravascular
lines. In: Remington
JS, Swartz
MN,
Eds. Current Clinical
Topics in Infectious Diseases. New York:
McGraw-Hill
1982; 3:
309-363.
5. Rothbere
AD.
Lessons
from
the investigation
into intravenous
fluid-related
neonatal
deaths.
5 Afr MedJ
1992; 81: 397-398. 6. Jentsch
U, Duse A, Smith
A, Koornhof
HJ, Klugman
KP. Analysis
of a transfusate
related
outbreak
of Serratia
odorifera bioptype
1 using
RFLPs
and ribotyping.
In:
Proceedings of the 4th Joint Congress of the Sexually Transmitted Diseases and Infectious
Diseases Societies of Southern Africa. Sun City, August
1993; 33.
7. Maki
DG,
Ringer
M. Evaluation
of dressing
regimens
for prevention
of infection
with
peripheral
intravenous
catheters.
JAMA
1987; 258: 2396-2403.
8. Maki
DG,
Goldmann
DA, Rhame
FS. Infection
control
in intravenous
therapy.
Ann

Intern Med 1973; 79: 867-887.

9. Verschraegen
identify

G,

Claeys

G,

Delanghe

M,

Enterobacter
cloacae contaminating
Microbial Infect Dis 1988; 7: 306-307.

10.

Plouffe

JF,

Brown

Candida parapsilosis
137: 1686-1689.
11.

12.

DG,
Silva
fungaemia

Pattyn
P. Serotyping
total
parenteral

J, Eck T, Stricof
RL,
related
to intravenous

Fekety
FR.
infusions.

and phage
typing
nutrition.
Eur J
Nosocomial

to

Clin

outbreak

of

Arch Intern Med 1977;

Solomon
SL, Khabbaz
RF,
Parker
RH,
et al. An outbreak
of Candida parapsilosis
bloodstream
infections
in patients
receiving
parenteral
nutrition.
J Infect Dis 1984; 149:
98-102.
Dugleux
G, Le Coutour
X, Hecquard
C, Oblin
I. Septicemia
caused by contaminated
parenteral
nutrition
pouches:
the refrigerator
as an unusual
cause. J Parenter Enteral

Nutr 1991; 15: 474-475.

13.

14.

15.

16.
17.
18.
19.
20.
21.
22.

Lifiares
J, Sitges-Serra
A, Garau
J, Perez J, Martin
R. Pathogenesis
of catheter
sepsis:
prospective
study
with quantitative
and semiquantitative
cultures
of catheter
hub and
segments.
J Clin Microbial 1985; 21: 357-360.
Ng PC, Herrington
RA,
Beane
CA,
Ghoneim
ATM,
Dear
PRF.
An outbreak
of
ancinetobacter
septicaemia
in a neonatal
intensive
care unit. J Hasp Infect 1989;
14:
363-368.
Holmes
B,
Gross
RJ.
Coliform
bacteria;
various
other
members
of
the
Enterobacteriaceae.
In: Parker
MT,
Collier
LH, Eds. Topley and Wilsons Principles of
Bacteriology,
Virology and Immunology,
8th edn.
London:
Edward
Arnold
1990;
4288431.
Block
SS. Peroxygen
compounds.
In: Block
SS, Ed. Disinfection, Sterilization,
and
Preservation, 4th edn. Philadelphia:
Lea and Febiger
1991; 167-181.
Crow
S. Peracetic
acid sterilization:
a timely
development
for a busy healthcare
industry.
Infect Control Hasp Epidemiol 1992; 13: 11 l-l 13.
Baldry
MGC.
The
bactericidal,
fungicidal
and sporicidal
properties
of hydrogen
peroxide
and peracetic
acid. J Appl Microbial
1983; 54: 417423.
Portner
DM,
Hoffman
RK.
Sporicidal
effect of peracetic
acid vapor.
Appl Microbial
1968; 16: 1782-1785.
Trexler
PC, Reynolds
LI. Flexible
film apparatus
for the rearing
and use of germfree
animals.
Appl Microbial 1957; 5: 4066412.
Gelbart
SM,
Reinhart
GF,
Greenlee
HB. Multiplication
of nosocomial
pathogens
in
intravenous
feeding
solutions.
Appl Microbial 1973; 26: 874-879.
Thompson
B, Robinson
LA. Infection
control
of parenteral
nutrition
solutions.
Nutr

Clin Practice
23.

24.

Sherertz
infections
care unit.
Weil DC,

Microbial
25.

1991; 6: 49-54.

KS, Hampton
KD
et al. Outbreak
with retrograde
medication
administration
J Pediatr 1992; 120: 455461.
Arnow
PM. Safety
of refrigerated
storage
of admixed

RJ, Gledhill
associated

McKee
KT,
Melly
MA,
Greene
HL,
associated
with the use of lipid emulsion
Ohnishi

in a neonatal
parenteral

bloodstream
intensive

fluids.

J Clin

1988; 26: 1787-1790.

Schaffner
W.
in parenteral

Gram-negative
nutrition.

133: 649-650.
26.

Candida

of

N,

Sasaki

S.

Germ-free

technology.

In:

Block

bacillary

sepsis

Am J Dis Child 1979;

SS,

Ed.

Disinfection,

S. odorifera
27.

transfusate

contamination

273

Sterilization,
and Preservation,
4th edn. Philadelphia:
Lea and Febiger
1991; 696-712.
Florent
M.
Packaging
and
sterilising
of parenteral
mixtures.
In:
Sitges-Serra,
Sitges-Creus,
Schwartz-Riera,
Eds. Clinical
Progress
in Nutrition
Research.
Proceedings
of the 9th Congress,
ESPEN,
Barcelona
1987. Base]: Karger
1988; 198-203.