Veterinary Surgery

33:636–643, 2004

A Prospective Comparison Between Stabilized Glutaraldehyde and

Chlorhexidine Gluconate for Preoperative Skin Antisepsis in Dogs

NICOLAAS E. LAMBRECHTS, MMedVet(Surg) Diplomate ECVS, KARIN HURTER, DrMedVet, JACKIE A. PICARD, BVSc,
JEREMY P. GOLDIN, BVSc, and PETER N. THOMPSON, MMedVet(Med)

Objective—To compare the efficacy of 0.3% stabilized glutaraldehyde and alcohol (SG þ A), 0.3%
SG and water (SG þ W), and 4% chlorhexidine gluconate tincture (CG þ A), as skin disinfectants in
dogs undergoing ovariohysterectomy.
Study Design—Prospective, blinded clinical study.
Animals—One hundred and twenty-one dogs.
Methods—Cutaneous bacterial colony forming units (CFU) from the perioperative site after skin
preparation, after antisepsis, and after surgery (incisional and paramedian), were quantified. The
influence of high initial bacterial counts (
150 CFU) and surgical time on antibacterial efficacy was
examined and the proportion of dogs from which Staphylococcus intermedius was cultured, deter-
mined. Perioperative skin reactions and wound infections were documented.
Results—All 3 antiseptic solutions significantly and equally reduced CFU to all post-antisepsis
sampling levels irrespective of surgical duration (mean surgical times 151.6, 136.2, and 149.6 min-
utes for CG þ A, SG þ A and SG þ W, respectively). Median percentage reductions in CFU ranged
between 99.3% and 100%. In dogs with initial high counts and disinfected with CG þ A and
SG þ W, the incisional samples had significantly higher counts than the post-antisepsis samples. In
the CG þ A and SG þ W groups, the proportion of post-surgery samples yielding S. intermedius was
significantly higher at the incisional than the paramedian sites. Eight mild cutaneous reactions were
recorded in equal proportions for the 3 solutions. There were no recorded infections.
Conclusions—All 3 preparations had an equal ability to reduce and maintain low CFU counts, with
minimal cutaneous reactions.
Clinical Relevance—SG solutions are safe and effective preoperative skin antiseptics for elective
clean-contaminated surgical procedures.
r Copyright 2004 by The American College of Veterinary Surgeons

Key words: antisepsis, bacterial colony-forming units, chlorhexidine gluconate, dog, glutaralde-
hyde, ovariohysterectomy, skin preparation, Staphylococcus intermedius.

INTRODUCTION of exogenous surgical wound contamination and infec-

T HE PRIMARY AIM of patient skin antisepsis is to
kill or incapacitate micro-organisms and reduce the
risk of post-operative infection.1 Most postoperative
wound infections are caused by endogenous micro-or-
ganisms of the host skin. Although use of caps, masks,
shoe covers, sterile gowns and gloves diminishes the risk
tion, the reduction of endogenous skin microflora is more
important.2,3
Desirable characteristics of an ideal skin antiseptic in-
clude a broad spectrum of antimicrobial activity (espe-
cially against vegetative and spore forms of bacteria),
rapid killing and a persistent lethal effect against micro-
organisms, detergent or cleansing capacity, lack of skin

From the Departments of Companion Animal Clinical Studies, Production Animal Studies and Veterinary Tropical Diseases, Faculty
of Veterinary Science, University of Pretoria, South Africa.
Address reprint requests to: Dr. Peter N. Thompson, MMedVet(Med), Department of Production Animal Studies, Faculty of
Veterinary Science, University of Pretoria, Private Bag X04, Onderstepoort, 0110, South Africa. E-mail: peter.thompson@up.ac.za
Submitted July 2004; Accepted August 2004
r Copyright 2004 by The American College of Veterinary Surgeons

0161-3499/04
doi:10.1111/j.1532-950X.2004.04086.x
636
 LAMBRECHTS ET AL
irritation or toxicity, and an ability to retain antimicro-
bial efficacy in the presence of organic material.4–6 Such
an antiseptic would also have no teratogenic, mutagenic
or carcinogenic effects, and would have no systemic ef-
fects in the patient or for the user. An ideal skin antiseptic
possessing all of these desirable characteristics does not
exist.4
Chlorhexidine gluconate is considered the skin anti-
septic of choice, because it possesses some of these char-
acteristics,4,6,7 and is effective against most Gram-positive
and Gram-negative bacteria, and some fungi.3,8
Glutaraldehyde (1,5-pentanedial) is a potent biocidal
agent that has gained widespread popularity for use on
inanimate objects as a disinfectant and a ‘‘cold sterilant’’
in an alkaline fluid formulation.9–12 Glutaraldehyde’s
spectrum of activity is very wide and it is rapidly bac-
tericidal, fungicidal, sporicidal and virucidal.9,12 It is con-
sidered stable but inactive at acid pH, and requires
addition of an alkaline ‘‘activator’’ (usually sodium bi-
carbonate) for its cidal activity.9,12,13 At alkaline pH
however, it becomes unstable, loses its cidal activity over
2 weeks, and thus has a short ‘‘shelf-life’’.9 Glutaralde-
hyde also retains its activity in the presence of organic
material.9,12,13 Glutaraldehyde’s mechanism of action is
associated with the reactive aldehyde groups that occur at
either end of the molecule. They are potent seekers of
nitrogen and bind irreversibly to proteins, which become
cross-linked.12 This has the effect of sealing the microbial
cell envelope, cutting the cell off from nutrient sources
and thus killing the cell.9,12 Glutaraldehyde is also be-
lieved to irreversibly damage microbial cell enzymes.9
A recently developed, glutaraldehyde product stabi-
lized glutaraldehyde (SG) (G-Cide, G-Cide [Pty] Ltd,
Gauteng, South Africa) is a combination of glutaralde-
hyde and a non-toxic, stabilizing surfactant molecule. It is
non-corrosive, non-volatile, non-toxic, biodegradable,
stable, and highly microbicidal at neutral pH. The man-
ufacturer claims that the formulation process results in
glutaraldehyde molecules having predominately linear
chain structures (the most active form), making the prod-
uct more effective than other glutaraldehyde products
which have less effective structures (G-Cide Users Manual,
G-Cide [Pty] Ltd). It requires no activation before use.
In in vitro testing, a skin detergent containing 0.3%
SG was found to effect a 99.9% reduction in numbers of
viable Staphylococcus aureus (SATCC Sta 53), Es-
cherichia coli (SATCC Esc 25), and Pseudomonas aerugi-
nosa (SATCC Pse 16) at 221C after 60 seconds contact
time (South African Bureau of Standards: test 261).
Glutaraldehyde’s wide range of cidal action, together
with its rapid activity, could make it an effective and
competitive preoperative topical antiseptic in veterinary
and medical use. Whether this product is a safe and
effective preoperative topical antiseptic, and has any
637
persistent antimicrobial activity on animate objects, re-
mains to be investigated.
Our objective was to determine the efficacy of 0.3% SG
and alcohol (SG þ A) or water (SG þ W), for preoperative
skin antisepsis and to compare that with 4% chlorhexi-
dine tincture (CG þ A), currently in use in our hospital.
There were 2 reasons why the 2 formulations of SG were
tested. SG þ A was tested to determine whether the in-
clusion of alcohol would enhance or diminish the efficacy
of SG compared with SG þ W. Comparing the 2 tinctures
was believed to constitute a balanced and fair trial.
MATERIALS AND METHODS
Dogs
Intact female dogs (n ¼ 121), admitted for ovariohysterec-
tomy (OVH), were included in the study after determination
of clinical health and absence of visible skin conditions. Dogs
o 5 kg were excluded because their limited abdominal surface
area precluded sample taking without overlapping sample
sites, and thus the associated risk of influencing subsequent
results. Using computer-generated random numbers, dogs
were randomly assigned to 1 of 3 skin antiseptic groups. For
all dogs, mean ( SD) age and weight were 23.0  21.0
months and 18.5  11.4 kg, respectively. Treatment group
mean age and weights were: CG þ A, 27.0  27.2 months,
22.8  17.9 kg; SG þ A, 19.2  15.6 months, 21.2  14.5 kg;
SG þ W, 16.8  9.3 months, 17.5  9.3 kg.
Anesthesia
After premedication with acepromazine (0.05 mg/kg sub-
cutaneously [SC]) and morphine sulfate (0.4 mg/kg SC), an-
esthesia was induced with thiopentone (8–10 mg/kg
intravenously [IV]) and maintained with halothane and oxy-
gen. Ringer’s lactate (10 mL/kg/h IV) was administered until
the dog recovered and morphine sulfate (0.2 mg/kg SC) was
administered again immediately after extubation and again
2–4 hours later if needed.
Skin Preparation
The entire ventral abdominal area was clipped, vacuumed,
and detergent solution without proven antibacterial properties
(10% alcohol ethoxylate [DOW Chemicals, Houston, TX])
and distilled water) to remove surface dirt. Residual detergent
was removed using moistened gauze swabs wiping from the
midline, outwards. Dogs were transported to the operating
room and positioned in dorsal recumbency.
Before draping, the ventral abdomen was sprayed until the
skin was saturated with 1 of 3 antiseptics: 4% chlorhexidine
gluconate and 70% ethanol (CG þ A), 0.3% SG and 70%
ethanol (SG þ A) or 0.3% SG and distilled water (SG þ W).
The 2 SG solutions were pigmented with the same colorant
(carmoisine; PCM Chemicals, Altrinchem, UK) used in com-
mercially available CG to yield a similar appearance and
638 PREOPERATIVE SKIN ANTISEPSIS IN DOGS
decanted into unmarked, opaque spray bottles, which were
alphabetically identified on their bases, by a staff member not
involved in the trial. The investigators were unaware of the
identity of the specific antiseptic used until completion of the
trial.
Staff, wearing surgical caps and masks, performed skin
antisepsis. After 3 minutes, the skin was resprayed using the
same antiseptic. After an additional 3 minutes, dry sterile
gauze swabs were used to soak up pooled, excess antiseptic
and the abdomen was draped using a standard cloth quarter
drape technique. Veterinary students performed routine,
standardized, ventral median OVH after appropriate surgical
hand scrubbing, gowning, and gloving.
Agar Plate Preparation and Quality Control
Quantification of skin bacterial numbers was performed
using sterile Replicating Organism Detection and Counting
(RODAC).14,15 RODAC plates were prepared in-house and
contained Columbia agar (Oxoid, Basingstoke, UK), 5%
equine blood, 2% lecithin (Sigma-Aldrich, St Louis, MO), and
5% Tween 80 (PDH Chemicals Ltd, Poole, UK) to neutralize
residual antiseptic on the skin. Inclusion of blood allowed for
detection of Streptococcus canis, common b-hemolytic skin
commensals and potential pathogens. Agar was added at 3%
to inhibit swarming by Proteus spp. and thus overgrowth of
other bacteria. Each batch of agar was checked for typical
bacterial growth using S. aureus (ATCC 25213) and E. coli
(ATCC 25922). Five randomly selected plates from each batch
were also checked for sterility by incubating at 371C in 5%
CO2 for 72 hours.
Skin Sampling
For most dogs, RODAC plates were pressed onto the skin
at 3 different sampling times: after initial scrubbing (post-
scrub), after skin antisepsis before draping (post-antisepsis) and
at the end of surgery (post-surgery). All preoperative sites were
adjacent but not on the intended incision site. Two post-surgery
samples were taken: across the incision line (incisional) and at a
random adjacent site (paramedian). For the first 16 dogs in the
trial, an additional sample was taken before initial scrubbing
and only the incisional sample was taken post-surgery.
Bacterial numbers from the prescrub sample were high and
repeatedly produced a confluent growth on the plate, making
accurate colony forming units (CFUs) counting impossible,
therefore this sample was not taken in subsequent dogs. The
post-surgery paramedian sample was introduced to compare
bacterial growth between the immediate surgical site and sur-
gically prepared area under the drapes. The location of the
extraincisional sampling sites was randomized to reduce the
impact of inadvertent overlapping of subsequent samplings.
The concern was that overlapping of sampling sites would
result in the inhibition of the antiseptics by Tween 80 and
lecithin residues and that mechanical stripping of bacteria
could occur, thereby artificially influencing subsequent CFU
counts. This effect would have been influenced in part by the
abdominal surface area, overlapping of samples being more
likely in smaller individuals (weight range of the dogs was
5.3–65 kg).
Incubation and CFU Counting
Standard plate incubation and counting of CFU followed
methods described elsewhere.14–16 Counts were performed at
48–72 hours by staff who, like the investigators, were not
aware of which antiseptic solution had been used. Bacteria
isolated on RODAC plates were identified to at least the genus
level using published methods.17 Bacterial species and their
prevalence were noted and described.
Adverse Effects
Other documented variables were the presence of any
visible skin reactions immediately after skin antisepsis, imme-
diately after surgery, and 24 hours after surgery, assessed in-
hospital. Wound infection, characterized by wound exudation
and signs of inflammation was assessed and documented be-
fore discharge (24 hours after surgery), by telephone interview
at 3–5 days and at suture removal (10–14 days after surgery).
Investigators removed skin sutures, however if dogs were not
returned, documentation was completed after telephone in-
terview. Any intraoperative incidents that may have led to
skin or wound contamination and the use of perioperative
antibiotics, not routinely used, were also recorded.
Data Analysis
Median CFU counts were compared between sampling
times (post-scrub, post-antisepsis, post-surgery—incisional
and post-surgery—paramedian) using the Friedman rank-
sum test and the Wilcoxon signed-rank test. This was done
within each of the 3 groups, as well as within low (o150 CFU)
and high (
150 CFU) strata of post-scrub (preantisepsis)
CFU count within each group, to evaluate the influence of
initially high bacterial levels on antibacterial efficacy. At each
sampling time, differences between antiseptic groups were
evaluated using Kruskal–Wallis 1-way ANOVA on ranks,
both overall and within low and high post-scrub CFU strata.
Use of non-parametric procedures was necessary because the
assumption of normality was not satisfied.
Spearman’s rank correlation was used to evaluate the as-
sociation between surgical time and post-surgery CFU counts
and between surgical time and changes in CFU counts from
pre- to post-surgery. The proportions of samples with
45 CFU (considered excessive bacterial growth14), with no
bacterial growth and with S. intermedius present, were com-
pared between sampling times within antiseptic groups, and
between antiseptic groups within sampling times using Fish-
er’s exact test. The proportions of dogs with cutaneous reac-
tions and wound infections were also compared between
antiseptic groups using Fisher’s exact test. The data were
analyzed using NCSS 2001 statistical software (NCSS, Kays-
ville, UT) and a public-domain statistical calculator EpiCalc
2000 (http://www.brixtonhealth.com/epicalc.html). The signif-
icance level (a) was set at .05 for all analyses.
 LAMBRECHTS ET AL 639

Table 1. Prevalence of the Most Common Bacteria Isolated at Various Sampling Times After Abdominal Skin Preparation with Antiseptic Solution in
Dogs That Had Ovariohysterectomy
Sampling Time
Post-Scrub Post-Antisepsis Post-Surgery Post-Surgery
Bacterium identification (n ¼ 119) (n ¼ 119) (Incisional) (n ¼ 118) (Paramedian) (n ¼ 101)

Staphylococcus intermedius 85 (71.4%) 11 (9.2%) 46 (39.0%) 13 (12.9%)

Coagulase-negative Staphylococcus spp. 53 (44.5%) 10 (8.4%) 22 (18.6%) 11 (10.9%)
(excluding S. intermedius and S. aureus)
Bacillus spp. 56 (47.1%) 29 (24.4%) 8 (6.8%) 16 (15.8%)
Acinetobacter spp. 15 (12.6%) 1 (0.8%) 2 (1.7%) 2 (2.0%)
S. aureus 7 (5.9%) 0 0 0

RESULTS antibiotics (amoxicillin IV) was administered because of

Data were incomplete for a number of dogs. The first
16 dogs did not have the post-surgery paramedian sample
taken (see Materials and Methods). In 1 additional dog,
the post-surgery incisional sample and in another, the
paramedian sample was not taken. The post-surgery
paramedian sample was accidentally contaminated in 1
other dog and these data were recorded as missing. In 2
other dogs, all the RODAC data were lost and recorded
as missing.
In 2 plates taken post-scrub, there was bacterial over-
growth; making counting impossible although bacterial
identification was still possible. For statistical purposes, a
count of 1 greater than the highest recorded count (2985)
was allocated. In 2 dogs, a single dose of post-operative
possible intraoperative contamination.
Table 1 summarizes the prevalence of the most com-
mon bacteria that were isolated at the various sampling
stages during the trial. Twelve additional species of bac-
teria were isolated on o 3 occasions. Table 2 summarizes
CFU counts for the sampling times. There were no sig-
nificant differences in CFU count between the 3 groups,
either overall or within the low (n ¼ 85) or high (n ¼ 34)
post-scrub CFU strata (P ¼ .27–.87).
For all 3 groups, median CFU counts were signifi-
cantly reduced post-antisepsis and at both post-surgery
sampling sites, compared with the post-scrub (preanti-
sepsis) sample. This was true overall, as well as within
both the low and high post-scrub CFU strata. In most
cases, there were no significant differences between CFU

Table 2. Median Colony Forming Unit (CFU) Counts (Range) at Different Sampling Times for 3 Antiseptic Groups, Overall, and Stratified by Post-
Scrub CFU from the Abdominal Skin of Dogs That Had Ovariohysterectomy
Sampling Time
Post-Scrub Post-Antisepsis Post-Surgery Post-Surgery
Stratum Antiseptic (n ¼ 119) (n ¼ 119) (Incisional) (n ¼ 118) (Paramedian) (n ¼ 101)

Total CG þ A 44a 0b 1b 0b
(0–1856) (0–15) (0–84) (0–19)
SG þ A 26a 0b 1b 0b
(0–2985) (0–13) (0–21) (0–80)
SG þ W 25a 0b 1c 0b
(0–1312) (0–33) (0–146) (0–33)
o150 CG þ A 23a 0b 0b 0b
(0–122) (0–15) (0–19) (0–16)
SG þ A 16a 0b 0b 0b
(0–123) (0–13) (0–21) (0–13)
SG þ W 13a 0b 0b 0b
(0–147) (0–33) (0–146) (0–33)

150 CG þ A 253a 0b 3c 2b,c
(150–1856) (0–2) (0–84) (0–19)
SG þ A 327.5a 0b 1b 1b
(176–2985) (0–12) (0–10) (0–80)
SG þ W 471.5a 0b 3c 0b,c
(153–1312) (0–4) (0–109) (0–5)

CG þ A, chlorhexidine gluconate and alcohol; SG þ A, stabilized glutaraldehyde and alcohol; SG þ W, stabilized glutaraldehyde and water.
a,b,c
Interpret row-wise: medians with no common superscripts differ significantly (Po.05).
640 PREOPERATIVE SKIN ANTISEPSIS IN DOGS

counts post-antisepsis and post-surgery. However, in the 90

high post-scrub CFU stratum there was a tendency for 80 b

the post-surgery incisional sample to have slightly higher b
70 CG+A b
CFU counts than the post-antisepsis sample, and in the SG+A
bc
b
CG þ A and SG þ W groups this achieved statistical sig- 60 SG+W b

Percent of samples
nificance (P ¼ .009 and .03, respectively). b
50 b
Median percentage reductions in CFU count from c
post-scrub to post-antisepsis for the groups were all 40

100%. From post-scrub to post-surgery (incisional) the 30

median percentage reductions in CFU count were 99.7%
for the CG þ A group, 99.9% for the SG þ A group, and
99.3% for the SG þ W group, and from post-scrub to
post-surgery (paramedian) they were all 100%. There
were no significant differences in percentage CFU count
reduction between the groups.
For all groups there was a large and highly significant
(Po.0001) reduction in the proportion of samples yield-
ing excessive bacterial growth (45 CFU) between the
post-scrub sample and all 3 subsequent samples (Fig 1).
Although there was a tendency for a higher proportion of
post-surgery incisional samples to yield excessive growth,
this was not statistically significant. At no stage were
there significant differences in this regard between any of
the groups.
The proportion of negative cultures (samples with no
bacterial growth) increased significantly (Po.001) be-
tween the post-scrub sample and all 3 subsequent samples
(Fig 2). Again, there were no significant differences in the
proportion of negative cultures between any of the
groups at any sampling time.
20
a
10
a a
0
Post-scrub Post- Post-surgery Post-surgery
antisepsis (incisional) (paramedian)
Sampling time
Fig 2. Proportion of samples with no bacterial growth for the
3 antiseptic groups at each sampling time. Within antiseptic
groups, different letters denote significant differences (Fisher’s
exact test, Po.05). CG þ A, chlorhexidine gluconate and al-
cohol; SG þ A, stabilized glutaraldehyde and alcohol; SG þ W,
stabilized glutaraldehyde and water.
There was a significant decrease in the proportion of
samples from which S. intermedius was isolated, between
the post-scrub sample and all 3 subsequent samples (P-
values from .02 to o.0001; Fig 3). In the CG þ A and
SG þ W groups, but not in the SG þ A group, the pro-
portion of post-surgery samples yielding S. intermedius
was significantly higher at the incisional than at the

90 90

80 a 80
a a
a a
70 70 a CG+A
SG+A
60 60 SG+W
Percent of samples

CG+A
Percent of samples

SG+A
50 SG+W 50 b
b
40 40
b
30 30
b b
bc
20 20 c
b c
b c
b b b
10 b 10 c c
b

0 0
Post-scrub Post- Post-surgery Post-surgery Post-scrub Post- Post-surgery Post-surgery
antisepsis (incisional) (paramedian) antisepsis (incisional) (paramedian)
Sampling time Sample time

Fig 1. Proportion of samples with excessive bacterial growth
(45 colony forming units ) for the 3 antiseptic groups at each
sampling time. Within antiseptic groups, different letters de-
note significant differences (Fisher’s exact test, Po.05).
CG þ A, chlorhexidine gluconate and alcohol; SG þ A, stabi-
lized glutaraldehyde and alcohol; SG þ W, stabilized glutaral-
dehyde and water.
Fig 3. Proportion of samples from which Staphylococcus
intermedius was isolated for the 3 antiseptic groups at each
sampling time. Within antiseptic groups, different letters de-
note significant differences (Fisher’s exact test, Po.05).
CG þ A, chlorhexidine gluconate and alcohol; SG þ A, stabi-
lized glutaraldehyde and alcohol; SG þ W, stabilized glutaral-
dehyde and water.
 LAMBRECHTS ET AL
paramedian site (P ¼ .0004, .01 and .28, respectively). At
no sampling time were there significant differences in the
frequency of S. intermedius isolation between any of the
groups.
Surgical times ranged from 54 to 288 min (mean,
145.8 min). Mean (  SD) surgical times were: CG þ A,
151.6  44.3 min; SG þ A, 136.2  48.2 min; and SG þ W,
149.6  43.9 min; these results were not significantly dif-
ferent. There were no significant correlations between
surgical time and post-surgery CFU counts, either over-
all, within groups, or within pre-scrub CFU strata.
Eight cutaneous reactions, all noted at the post-anti-
sepsis stage, were recorded: mild erythema (6), mild skin
swelling (1), and rash (1). Three reactions occurred in
each of the CG þ A and SG þ A groups, and 2 in the
SG þ W group; there were no significant differences be-
tween groups. All reactions resolved completely within 24
hours, without requiring treatment. Wound infection did
not occur and all wounds appeared or were reported to
be in an advanced state of healing at suture removal.
DISCUSSION
Our study differs somewhat from similar studies.14,15
The initial skin scrub was performed using a detergent
solution without specific antibacterial properties. We felt
that this separated the mechanical, bacteria-reducing ef-
fect of skin scrubbing from the specific effects of the an-
tiseptic solutions. In a clinical setting, one would use a
surgical detergent with antibacterial properties. Secondly,
in all of the first 16 dogs, bacterial overgrowth of the pre-
scrub sample prevented accurate CFU counting and this
sampling stage was omitted in the remaining dogs. Fi-
nally, we only examined antiseptic performance during 1
type of clean–contaminated procedure, namely OVH.
This eliminated some of the variation described in similar
studies,14,15 where orthopedic and neurologic procedures
were included. This could have involved procedures of
considerably differing durations and this may have in-
fluenced reported wound infection rates. It also implied
that skin incisions were made into different body regions
with different bacterial population densities, which in
turn could have affected CFU counts and infection rates.
Bacterial species cultured and identified at the various
sampling stages paralleled those found in similar studies,
where Staphylococcus spp., especially S. intermedius, pre-
dominated.14,15 S. intermedius, an important cause of ca-
nine pyoderma and other infections in dogs, is considered
a permanent resident of the mucosa but transient on the
hair and skin surface.18 Allaker et al 19 found low num-
bers of staphylococci on the skin of normal dogs. In that
study, the highest levels of S. intermedius isolated were
found on the abdominal skin, where they were thought to
641
be contaminants from grooming or the environment.19
S. aureus and Bacillus spp are also considered transient
on the skin surface.18,20 The status of the Gram-negative
bacteria including Acinetobacter spp. is uncertain al-
though Harvey and Lloyd infer that they could be part of
the normal canine skin microflora.20 After antisepsis, the
numbers of transient surface flora isolated in our study,
remained low for the duration of the surgical period, ex-
cept for S. intermedius. Prevalence at the post-surgery
incisional site increased to 39.0% from the 9.2% recov-
ered post-antisepsis (Table 1). A possible explanation is
that another population of S. intermedius resides in hair
follicles18 and migrates to the skin surface during surgery
because of skin manipulation and removal of the anti-
septics by the surgeon. This effect would be less pro-
nounced at the paramedian site, protected by the drapes.
Rapid bacterial multiplication, nosocomial repopulation,
and contamination from the genital tract may also have
played a role. As expected, the resident coagulase-nega-
tive staphylococci also behaved in this fashion.20
All 3 antiseptic solutions significantly and equally re-
duced CFU counts from post-scrub levels to post-anti-
sepsis and both post-surgery levels. Stratifying the data,
to determine whether high bacterial counts (arbitrarily
defined as
150 CFU) would influence the antiseptic
ability of the 3 solutions, did not alter this outcome. In
the high CFU stratum, both CG þ A and SG þ W were
unable to maintain post-surgery incisional counts as low
as the post-antisepsis levels. The clinical significance of
this is unknown, although probably small, as the actual
counts were very low. A previous study14 found that in-
itial high CFU counts did not affect the performance of
CG þ A.
There were no statistical differences between the per-
formances of SG þ A and SG þ W, suggesting that the
inclusion of alcohol did not enhance or diminish the ef-
fectiveness of SG. This would support the claim of a
rapid and broad spectrum of antibacterial action. Alco-
hol has a reported persistent action of some 3 hours8 and
it will effectively kill S. intermedius. The only indication
of a possible advantage of the tincture over the aqueous
preparation of SG is that in the SG þ A treatment group,
there was no significant difference in the frequency of S.
intermedius isolated between the 2 post-surgery sites. In
the SG þ W group, S. intermedius was isolated more fre-
quently at the incisional site (Fig 3).
A similar ability to maintain the low post-antisepsis
CFU levels until the post-surgery samples were taken
occurred with all 3 solutions. This would support the
presence of persistent antibacterial activity. CG is report-
ed to bind protein in the stratum corneum of the skin,
leaving a residue that is not removed by alcohol. This
results in persistent activity by way of a free chlorine
group on 1 side of the chlorhexidine molecule.3–5,7 The
642 PREOPERATIVE SKIN ANTISEPSIS IN DOGS
exact duration of activity is unknown but is affected by
the concentration used and the repeated application of
the product on the patient’s skin. It maintained low
numbers of bacteria under surgical gloves for up to 6
hours,3 an effect that may however be diminished by the
inclusion of alcohol.15 We did not investigate this pos-
sibility but it could mean that CG and water would have
performed better than the tincture in this regard and that
it should have been included in our comparison. Whether
SG retains any reactive unused aldehyde groups and has
any persistent antimicrobial activity on the skin remains
uninvestigated. Although the manufacturer of SG makes
no claim of such activity, our findings do suggest the
possibility of such a characteristic and warrant further
investigation.
There were very few skin reactions that could have
been attributed to the various antiseptics we used. All 3
solutions were associated with an equally low incidence of
skin reactions. CG has rarely been reported to provoke
skin reactions.7,14,16 Systemic absorption is low and other
toxic reactions are infrequently reported.4
Side effects, reported in studies describing the topical
application of formulations of glutaraldehyde in humans
and animals, have either been completely absent or min-
imal, and when present, include mild skin irritation on
broken skin and slight yellow or brown skin discoloration
at high concentrations.12,21–23 At low concentrations
(o2%), glutaraldehyde has been used as a disinfectant
both on inanimate and animate objects with few reported
side effects, and has been widely used as a preservative in
skin cosmetics both in the USA and the EU countries.9
Topical application at high concentrations has been suc-
cessful in humans in the treatment of plantar viral
warts,9,22 superficial onychomycosis,22 and plantar and
palmar hyperhydrosis.23 At lower concentrations, it has
been used in human anti-caries dental formulations and
in anti-caries chewing gum and mouth wash.9,12 It has
also been used (at low concentrations) as an antiseptic
teat dip in dairy cattle to prevent mastitis after milk-
ing,9,24 and has been used as an ingredient in antiseptic
hand soaps.12
Glutaraldehyde has been used historically as a volatile,
high-concentration (
2%) sterilizing solution for inani-
mate objects, and most reports indicate toxic effects only
at these concentrations. These include allergic contact
sensitization and dermatitis,6,25,26 ocular, nasal, and
throat irritation, and nausea and headaches.27 Tests con-
ducted to determine glutaraldehyde’s carcinogenicity
have been inconclusive.28 Even at high doses, glutaralde-
hyde has been judged non-teratogenic in mice, nor mu-
tagenic when using the Ames test.9
Independent, primary dermal irritancy studies using
SG indicate that the product is non-irritant on both intact
and abraded skin in rabbits (Roodeplaat Research Lab-
oratories [Pty] Ltd, South Africa). Glutaraldehyde binds
strongly to keratin in the stratum corneum of the skin,
limiting systemic absorption.25
No cases of wound infection were observed or report-
ed, despite frequent lapses in the application of Hall-
stead’s principles by inexperienced surgeons in our study.
Infection rates after clean–contaminated procedures have
been reported as 4.5%29 and 5.0%.30
The inclusion of CFU and bacterial identity data of
2 dogs that were administered antibiotics may be con-
tentious, although any influence would have been limited
to the growth from the post-surgery samples. As we were
measuring only skin surface bacteria and very shortly
after administration of any antibiotics, we considered
that its influence would have been negligible. However,
wound infections related to antisepsis failure may have
been prevented. The reasons for randomizing the extra-
incisional sample sites were given earlier; nevertheless, it
is possible that this may have introduced additional
variation into the data. However, at no sampling
stage were there significant differences in CFU counts
between randomly selected sampling sites, nor were there
differences in the distribution of sampling sites between
antiseptic groups. Two similar studies14,15 only describe
their sampling sites as being adjacent the incisional site,
but do not mention how overlapping of subsequent
samples was avoided. CFU counts of S. intermedius were
not performed; this might have uncovered additional in-
formation not identified by our method of simply re-
cording their presence or absence. Finally, the accuracy
of details about wound infection and skin reaction, ob-
tained by telephone interview is problematic and it is
possible that such information was not accurately com-
municated.
We found that all 3 antiseptics solutions had a similar
ability to significantly reduce the numbers of cutaneous
surface bacteria and maintain these levels, when used
preoperatively in an elective, clean–contaminated proce-
dure. The 2 SG formulations can therefore be recom-
mended for procedures like OVH where infection
would likely not have catastrophic consequences. Fur-
ther study is required to determine the suitability of SG
for neurosurgical and orthopedic procedures, where in-
fection would pose severe risk to the patient, and to
investigate whether SG has any persistent or residual
activity.
ACKNOWLEDGMENT
The authors thank Johan Carstens (posthumously) and
Elize Pretorius for the bacteriological work and Treacey
Cable for patient care.
 LAMBRECHTS ET AL
REFERENCES
1. Bruch MK: Methods of testing antiseptics: antimicrobials
used topically in humans and procedures for hand scrubs,
in Block SS (ed): Disinfection, Sterilization, and Preserva-
tion (ed 4). Philadelphia, PA, Lea & Febiger, 1991, pp
1028–1046
2. Polk HC, Simpson CJ, Simmons BP, et al: Guidelines for
prevention of surgical wound infection. Arch Surg
118:1213–1217, 1983
3. Denton GW: Chlorhexidine, in Block SS (ed): Disinfection,
Sterilization, and Preservation (ed 5). Philadelphia, PA,
Lea & Febiger, 2001, pp 321–335
4. Lemarié RJ, Hosgood G: Antiseptics and disinfectants in
small animal practice. Compend Contin Educ Pract Vet
17:1339–1351, 1995
5. Paul JW, Gordon MA: Efficacy of a chlorhexidine surgical
scrub compared to that of hexachlorophene and
povidone-iodine. Vet Med Small Anim Clin May:573–
579, 1978
6. Hess JA, Molinari JA, Gleason MJ, et al: Epidermal toxicity
of disinfectants. Am J Dent 4:51–56, 1991
7. Rosenberg A, Alatary SD, Peterson AF: Safety and efficacy
of the antiseptic chlorhexidine gluconate. Surg Gynecol
Obstet 143:789–792, 1976
8. Shmon C: Assessment and preparation of the surgical patient,
in Slatter D (ed): Textbook of Small Animal Surgery
(ed 3). Philadelphia, PA, W.B. Saunders Company, 2003,
pp 162–178
9. Scott EM, Gorman SP: Glutaraldehyde, in Block SS (ed):
Disinfection, Sterilization, and Preservation (ed 4). Phila-
delphia, PA, Lea & Febiger, 1991, pp 596–614
10. Russell AD: Glutaraldehyde: current status and uses. Infect
Control Hosp Epidemiol 15:724–733, 1994
11. Power EGM, Russell AD: Glutaraldehyde: its uptake by
sporing and non-sporing bacteria, rubber, plastic and an
endoscope. J Appl Bacteriol 67:329–342, 1989
12. Gorman SP, Scott EM: Antimicrobial activity, uses and
mechanism of action of glutaraldehyde. J Appl Bacteriol
48:161–190, 1980
13. Borick PM, Dondershine FH, Chandler VL: Alkalinized
glutaraldehyde, a new antimicrobial agent. J Pharm Sci
53:1273–1275, 1964
14. Stubbs WP, Bellah JR, Vermaas-Hekman D, et al: Chlor-
hexidine gluconate versus chloroxylenol for preoperative
skin preparation in dogs. Vet Surg 25:487–494, 1996
643
15. Osuna DJ, DeYoung DJ, Walker RL: Comparison of three
skin preparation techniques. Part 2: clinical trial in 100
dogs. Vet Surg 19:20–23, 1990
16. Osuna DJ, DeYoung DJ, Walker RL: Comparison of three
skin preparation techniques in the dog. Part 1: Experimen-
tal trial. Vet Surg 19:14–19, 1990
17. Quinn PJ, Carter ME, Markey B, et al: Clinical Veterinary
Microbiology. London, Wolfe Publishing, 1994, pp 6–628
18. Harvey RG, Lloyd DH: The distribution of Staphylococcus
intermedius and coagulase-negative staphylococci on the
hair, skin surface, within the hair follicles and on the mu-
cous membranes of dogs. Vet Dermatol 5:75–81, 1994
19. Allaker RP, Lloyd DH, Simpson AI: Occurrence of Staphylo-
coccus intermedius on the hair and skin of normal dogs. Res
Vet Sci 52:174–176, 1992
20. Harvey RG, Lloyd DH: The distribution of bacteria (other
than staphylococci and Propionibacterium acnes) on the
hair, at the skin surface and within the hair follicles of dogs.
Vet Dermatol 6:79–84, 1995
21. Scott KW: Glutaraldehyde gel for warts. Practitioner 226:
1342–1343, 1982
22. Suringa DWR: Treatment of superficial onychomycosis with
topically applied glutaraldehyde. Arch Dermatol 102:
163–167, 1970
23. Juhlin L, Hansson H: Topical glutaraldehyde for plantar
hyperhidrosis. Arch Dermatol 97:327–330, 1968
24. Murdough PA, Pankey JW: Evaluation of 57 teat sanitizers
using excised cow teats. J Dairy Sci 76:2033–2038, 1993
25. Maibach HI, Prystowsky SD: Glutaraldehyde (pentanedial)
allergic contact dermatitis. Arch Dermatol 113:170–171,
1977
26. Bardazzi F, Melino M, Alagna G, et al: Glutaraldehyde der-
matitis in nurses. Contact Dermatitis 14:319–320, 1986
27. Jachuck SJ, Bound CL, Steel J, et al: Occupational hazard in
hospital staff exposed to 2 percent glutaraldehyde in an
endoscopy unit. J Soc Occup Med 39:69–71, 1989
28. National Occupational Health and Safety Commission
Approved criteria for classifying hazardous substances
[NOHSC:0006(1993)]. Canberra, Australia, AGPS, 1993
(http://www.worksafe.gov.au/)
29. Vasseur PB, Levy J, Dowd E, et al: Surgical wound infection
rates in dogs and cats. Data from a teaching hospital. Vet
Surg 17:60–64, 1988
30. Brown DC, Conzemius MG, Shofer F, et al: Epidemiological
evaluation of postoperative wound infections in dogs and
cats. J Am Vet Med Assoc 210:1302–1306, 1997