TRAUMA

Short duration of antibiotic prophylaxis in

open fractures does not enhance risk of
subsequent infection
N. Dunkel,
D. Pittet,
L. Tovmirzaeva,
D. Suv,
L. Bernard,
D. Lew,
P. Hoffmeyer,
I. Ukay
From Geneva
University Hospitals,
Geneva, Switzerland
N. Dunkel, PhD, Research
Assistant
D. Pittet, MD, MSc, Professor,
Head of Infection Control
Program
L. Tovmirzaeva, MD,
Research Assistant, Infectious
Diseases Fellow
D. Suv, MD, Attending
Septic Orthopaedic Surgery
P. Hoffmeyer, MD, Professor,
Head of Orthopaedic Surgery
I. Ukay, MD, , Attending
Septic Orthopaedic Surgery,
Attending Infectious Diseases,
Attending Infection Control
Program
Geneva University Hospitals, 4
rue Gabrielle Perret-Gentil, 1211
Geneva, Switzerland.
L. Bernard, MD, Professor,
Head of Infectious Diseases
CHU Tours, Hpital
Bretonneau, Boulevard
Tonnell 2 37044 Tours,
France.
D. Lew, MD, Professor, Head
of Infectious Diseases
Geneva University Hospitals,
Service of Infectious Diseases,
4 rue Gabrielle Perret-Gentil,
1211 Geneva, Switzerland.
Correspondence should be sent
to I. Uckay; e-mail:
ilker.Mruckay@hcuge.ch
2013 The British Editorial
Society of Bone & Joint
Surgery
doi:10.1302/0301-620X.95B6.
30114 $2.00
Bone Joint J
2013;95-B:8317.
Received 20 May 2012;
Accepted after revision 23
January 2013

VOL. 95-B, No. 6, JUNE 2013

We undertook a retrospective case-control study to assess the clinical variables associated

with infections in open fractures. A total of 1492 open fractures were retrieved; these were
Gustilo and Anderson grade I in 663 (44.4%), grade II in 370 (24.8%), grade III in 310 (20.8%)
and unclassifiable in 149 (10.0%). The median duration of prophylaxis was three days
(interquartile range (IQR) 1 to 3), and the median number of surgical interventions was two
(1 to 9). We identified 54 infections (3.6%) occurring at a median of ten days (IQR 5 to 20)
after trauma. Pathogens intrinsically resistant to the empirical antibiotic regimen used
(enterococci, Enterobacter spp, Pseudomonas spp) were documented in 35 of 49 cases
(71%). In multivariable regression analyses, grade III fractures and vascular injury or
compartment syndrome were significantly associated with infection. Overall, compared
with one day of antibiotic treatment, two to three days (odds ratio (OR) 0.6 (95% confidence
interval (CI) 0.2 to 2.0)), four to five days (OR 1.2 (95% CI 0.3 to 4.9)), or > five days (OR 1.4
(95% CI 0.4 to 4.4)) did not show any significant differences in the infection risk. These
results were similar when multivariable analysis was performed for grade III fractures only
(OR 0.3 (95% CI 0.1 to 3.4); OR 0.6 (95% CI 0.2 to 2.1); and OR 1.7 (95% CI 0.5 to 6.2),
respectively).
Infection in open fractures is related to the extent of tissue damage but not to the
duration of prophylactic antibiotic therapy. Even for grade III fractures, a one-day course of
prophylactic antibiotics might be as effective as prolonged prophylaxis.
Cite this article: Bone Joint J 2013;95-B:8317.

Open fractures are associated with a risk of

infection that may be as high as 30% for severe
injuries of the lower limb.1-3 In 1976, Gustilo
and Anderson1 established a system of classification that predicted the risk for infection in
open tibial fractures. This has been subsequently updated with a minor modification
related to type III fractures,2 but the grading
system remains ubiquitous and is often generalised to include all open fractures.3
When compared with the duration of preoperative antibiotic prophylaxis in surgery
for closed fractures where a single parenteral
dose is sufficient,4,5 open fractures remain one
of the few surgical fields where antibiotics are
often administered for several days, despite
the fact that almost no evaluation of the ideal
or minimal duration has been attempted.6,7 It
is generally accepted that antibiotic prophylaxis in grade I and II fractures should not be
administered for > 24 hours.5,6,8 However, the
minimum duration for grade III fractures varies between one and ten days,1,3,5,6,8-13 or even
several weeks.7 Guidelines based on expert
opinion and common sense advocate a

maximum of 48 hours6 to 72 hours3 for grade

III fractures.8 The British Association of Plastic, Reconstructive and Aesthetic Surgeons
and the British Orthopaedic Association
standards for the management of open
fractures of the tibia recommend parenteral
co-amoxiclav or cefuroxime for 72 hours
or definitive wound closure, whichever
is sooner.14
The choice of the antimicrobial agent is
also highly variable15 with usual recommendations for second-generation cephalosporins
alone1,6,8,9 or in combination with aminoglycosides,1,3,11 quinolones8,11,16 or regimens targeting anaerobic pathogens.5 Generally the
literature relies on historical controls6 or
underpowered studies involving < 200 infectious episodes.9,10,12,15
Very few studies have attempted to distinguish between infection due to inoculated
pathogens and potential hospital-acquired
organisms associated with the surgical intervention and prolonged hospital stay. In this
large 14-year study, we assess risk factors for
open fracture infection with an emphasis on
831

832

N. DUNKEL, D. PITTET, L. TOVMIRZAEVA, D. SUV, L. BERNARD, D. LEW, P. HOFFMEYER, I. UKAY

the duration of prophylactic antibiotic treatment and

excluding nosocomial infections as far as possible.

Patients and Methods

The Orthopaedic and Traumatology Service at Geneva
University Hospitals, Geneva, Switzerland, is a level 1
trauma centre serving a population of approximately
800 000. For open fractures, priority is given to damage
control, vascular repair, skeletal stabilisation, copious irrigation with at least nine litres of normal saline, and empirical administration of antibiotic prophylaxis. Second-look
surgery is often performed after 48 hours and definitive
fracture fixation (with or without bone, skin, and muscular
grafting) is performed at a later stage.11 Quantitative cultures and antibiotic susceptibility testing is performed
according to the recommendations of the United States
Clinical and Laboratory Standards Institute.17 The microbiological procedures have remained unchanged during the
entire study period. The local ethics committee approved
the study as an internal quality assessment project and consequently no individual informed consent was required.
The databases of the hospital administrative coding centre and orthopaedic service were searched retrospectively to
identify all open bone fractures in adult trauma patients
admitted between 1 January 1996 and 31 December 2009.
For each episode 72 variables were assessed. Data collected
included details of the traumatic event, season of occurrence (winter vs summer the latter defined as March to
October), patient demographic characteristics and comorbidities, microbiology, surgical and antibiotic treatment
modalities, and outcome. The Gustilo and Anderson
classification1 was noted by the treating surgeon at the time
of first surgery.
A biologist with experience in clinical medical studies
(ND) and two physicians (LT, IU) reviewed all electronic
medical and nursing records. The infectious diseases consultant specialised in bone and joint infections (IU)
reviewed all infected cases a third time. Patients were followed until 31 December 2011, i.e., two years after the
inclusion of the last patient.
The definition of infection required the presence of pus
and surgical treatment, along with the prescription of antibiotics, targeting the infection. Infections occurring after two
months of the first surgical treatment or during a subsequent
hospitalisation were considered to be hospital-acquired and
excluded, as were patients with infections caused by methicillin-resistant Staphylococcus aureus (MRSA) or methicillin-resistant coagulase-negative staphylococci.18
Immunosuppression was defined as the presence of the
following co-morbidities: diabetes mellitus; human immunodeficiency virus infection with CD4 counts < 350/mm3;
transplantation; dialysis; Child-Pugh class B or C19 liver cirrhosis; and malignancy without remission. Advanced age
and polytrauma were not discounted. Compartment syndrome was defined clinically by the surgical need for fasciotomy to release tissue pressure.

Exclusion criteria include patients aged < 18 years, open

fracture outside the orthopaedic domain such as neurosurgical patients, incomplete medical documentation and
follow-up < 60 days.
Statistical analysis. Analyses were performed using STATA
v9 software (StataCorp, College Station, Texas). Group
comparisons were performed using the Pearson chi-squared
test or the Wilcoxon rank-sum test, as appropriate. Logistic
regression analyses determined associations with the primary outcome of infection. The analysis was performed for
all open fractures and for grade III fractures only (including
detailed surgical parameters). Variables with a p-value
0.05 in univariate analysis were introduced in a stepwise
multivariable model. We included between eight and ten
outcome events per predictor variable.20 As only a minority
of patients had more than one open fracture and all patients
were treated at our centre, cluster-controlled analysis was
waived. Key variables were checked for confounding, colinearity and interaction, the latter by Mantel-Haenszel
estimates, interaction terms and likelihood ratio tests. We
investigated a possible linear relationship between the
duration of antibiotic administration and infection by linear and logistic regression analyses with categorised variables and repeated with quadratic and logarithmic (ln)
transformation of these variables. We assessed also the
presence of a potential threshold for enhanced infection
risk by plotting on a graph all infections in relation to antibiotic duration. p-values 0.05 (all two-tailed) were considered significant.

Results
Of 1522 open fractures identified, 30 were excluded owing
to infection already present on admission (n = 11) or MRSA
infection (n = 7) or infections occurring more than two
months after surgery (n = 6) or fractures in an area of prior
chronic osteomyelitis (n = 6). The remaining 1492 fractures
(in 1290 patients) were retained for final analysis. The
median age of the patients was 41 years (18 to 101);
1010 (78.3%) were male. The median duration of active
follow-up was 35 months (2 to 120) for the 1492 fractures.
The characteristics and distribution of the fractures are
given in Table I. The severity and incidence of open fractures did not differ substantially between seasons (56%
occurred in summer) or across the study period.
All patients were treated surgically with a median number of two interventions (1 to 9) and a median delay
between trauma and first surgery of 0 days (interquartile
range (IQR) 0 to 1). Grades I and II fractures were almost
always closed primarily. If definitive fracture fixation was
performed in a second step, mostly for grade III fractures, it
occurred after a median interval of 2.5 days (1 to 8).
All but two patients received systemic prophylactic parenteral antibiotics for a median duration of three days (0 to 90;
IQR 1 to 3). Median duration was one day (IQR 1 to 3) for
grade I, three days (IQR 2 to 3) for grade II, and three days
(IQR 3 to 6) for grade III fractures. No devices delivering
THE BONE & JOINT JOURNAL

SHORT DURATION OF ANTIBIOTIC PROPHYLAXIS IN OPEN FRACTURES DOES NOT ENHANCE RISK OF SUBSEQUENT INFECTION

Table I. Characteristics of the 1492 included fractures

Characteristic
Fractures / patients (n)
Patients (n)
One fracture
Two fractures
Three fractures
Fracture site (n, %)
Tibia
Pilon tibia
Mid-femur
Trochanter
Humerus
Patella
Ulna
Radius
Toes
Tarsus
Hallux
Fibula
Calcaneum
Clavicle
Femoral neck
Scapula
Cuboid
Gustilo & Anderson grade (n, %)
I
II
III
IIIa
IIIb
IIIc
Not classifiable*

1492 / 1290
1109
160
21

452 (30.3)
165 (11.1)
158 (10.6)
119 (8.0)
101 (6.8)
97 (6.5)
86 (5.8)
81 (5.4)
55 (3.7)
41 (2.7)
41 (2.7)
33 (2.2)
30 (2.0)
19 (1.3)
11 (0.7)
2 (0.01)
1 (0.01)
663 (44.4)
370 (24.8)
310 (20.8)
63
53
63
149 (10.0)

* no clear grading noted in the surgical records

local antibiotic therapy were used. In 119 episodes (8%),

patients were treated for more than five days. For 33 individuals, there was an associated remote infection outside the
trauma site (mostly pneumonia and urinary tract infection).
Cefuroxime was the most frequently prescribed antibiotic regimen (n = 1067; 72%), followed by amoxicillinclavulanate (n = 63; 4%). Overall, we noted 40 different
regimens. In all but two cases, once prescribed the regimens
remained unchanged throughout the course. Immunosuppressed patients did not receive longer therapy than immunocompetent individuals (median 3 vs 3 days; p = 0.12,
Wilcoxon rank-sum test).
Infections of open fracture sites. Infection
occurred at
54 fracture sites (3.6%) in 54 patients with a median delay
of ten days between trauma and the onset of clinical infection (IQR 5 to 20). The rate of infection in grade I fractures
was 0.9% (six of 663), compared with 1.9% (seven of 370)
and 12% (37 of 310) in grade II and III fractures, respectively (Tables II and III). The incidence of infection varied
according to the grade III sub-type: 3.2% (2 of 63) in grade
IIIa, 5.7% (3 of 53) in grade IIIb and 33.3% (21 of 63) in
grade IIIc. Although the rate of infection was low for grades
VOL. 95-B, No. 6, JUNE 2013

833

IIIa and IIIb compared with IIIc, both grade I and II fractures still revealed higher infection risks than grade IIIa and
grade IIIb fractures (13/1020 vs 5/111; p = 0.012, chisquared test).
Risk factors for infection are shown in Tables III and IV.
Men and young adults (considered < 30 years of age) were
more frequently infected than women. This difference is
probably due to a higher incidence of grade III fractures in
men compared with women: of the 310 grade III fractures,
227 (73.2%) were in males (p = 0.01, chi-squared test).
Infected patients had a higher median hospital stay than
non-infected patients (46 days (3 to 349) vs 15 days (2 to
345); p < 0.0001), but were not exposed to a higher median
number of previous surgical interventions before onset of
infection (3 vs 2; p = 0.245, Wilcoxon rank-sum test).
Due to differences in group comparisons (Tables II and
III) and significant associations by univariate analysis
(Table IV), we adjusted for case mix (Tables IV and V). The
choice of empirical antibiotic therapy and its duration were
not related to infection. In particular, combinations of
cephalosporins with aminoglycosides, fluoroquinolones,
metronidazole, vancomycin or carbapenems were frequently associated with grade IIIc fracture, and failed to
achieve a protective effect. Of note, when analysed as a
continuous variable, the odds ratio (OR) and the 95% confidence interval (CI) for the duration of antibiotic administration were around one (Tables IV and V).
There was no threshold in the duration of total antibiotic
treatment beneath which the infection risk was enhanced.
Likewise, there was no linear, quadratic or logarithmic relationship between antibiotic duration and infection risk. For
these multivariable models, areas under the receiver-operating curve (ROC) were > 0.80, indicating that the modelling
was acceptable.
Pathogens were identified for 49 of the 54 infections generally in four of five (2 to 12) intra-operative samples. Enterobacter spp was the most frequently identified microorganism (15 of 49; 31%), followed by Pseudomonas spp (n
= 15) and Enterococcus spp (n = 13). In 30 episodes (60%),
Gram-negative non-fermenting rods were identified. No epidemiological link could be established between cases. Staph.
aureus was involved only in eight cases and other Gram-positives in 13 infections, with polymicrobial infection in 28 episodes (57%). There was no infection caused by fungi,
mycobacteria, vancomycin-resistant enterococci (VRE), Acinetobacter or multi-resistant Pseudomonas spp, or extended
spectrum beta-lactamase (ESBL)-producing rods. In one
case, the non-dominant pathogen was Clostridium spp
without further specification.
In two-thirds of cases (35 of 49; 71%), the selected
empirical antibiotic regimen did not cover the infecting
pathogen. For example, among all episodes involving grade
III fractures, 90 intra-operative samples (90 of 310; 29%)
were collected at first surgical treatment (pre-infection). Of
these, 29 (32%) grew bacteria but none was responsible for
later diagnosed infection.

834

N. DUNKEL, D. PITTET, L. TOVMIRZAEVA, D. SUV, L. BERNARD, D. LEW, P. HOFFMEYER, I. UKAY

Table II. Incidence of bone infections in open fractures, stratified by grade and localisation for 1343 classified fractures

Infection (n, %)
Upper extremity
Humerus
Radius
Ulna
Above knee
Femur (incl. trochanter)
Patella
Below knee
Tibia, fibula, ankle
Calcaneum
Tarsal, metatarsal
Toes

Gustilo grade I (n = 663)

Gustilo grade II (n = 370)

Gustilo grade III (n = 310)

Incidence summary

6 (0.9)

7 (1.9)

37 (12)

0% (0 of 55)
0% (0 of 27)
0% (0 of 32)

6% (1 of 18)
0% (0 of 8)
0% (0 of 7)

14% (4 of 28)
50% (1 of 2)
33% (1 of 3)

5% (5 of 101)
2.7% (1 of 37)
2.4% (1 of 42)

0.7% (1 of 136)
0% (0 of 33)

1.9% (1 of 54)
2.4% (1 of 41)

9.4% (5 of 53)
15.4% (2 of 13)

2.9% (7 of 243)
3.4% (3 of 87)

0% (1 of 234)
14% (1 of 7)
7.7% (1 of 13)
11.8% (2 of 17)

0.8% (1 of 132)
0% (0 of 9)
14.3% (2 of 14)
8.3% (1 of 12)

17.4% (23 of 132)

0% (0 of 10)
0% (0 of 14)
7.1% (1 of 14)

5.0% (25 of 498)

0% (1 of 26)
7.3% (3 of 41)
9.5% (4 of 43)

Table III. Characteristics of 1492 open fractures stratified by infection (p-values by logistic regression)
Characteristic

Infection

No infection

p-value

Fractures (n)
Female gender (n, %)
Median age (yrs)
Gustilo and Anderson grade (n, %)
Grade I
Grade II
Grade III
Non-identifiable
Open fracture of long bones*
Fracture site
Upper extremity
Below knee
Median delay traumasurgery (days) (IQR)
Median surgical interventions (n)
Median duration of pre-emptive antibiotic therapy (days) (IQR)

54
8 (15)
38

1438
474 (33)
42

0.005
0.075

6 (11)
7 (14)
37 (70)
4 (7)
42 (78)

657 (46)
363 (25)
272 (19)
145 (10)
1109 (77)

0.043
0.663
0.206
0.008
0.910

5 (9)
31 (57)
1 (0 to 6)
3
3 (0 to 21)

222 (15)
653 (45)
0 (0 to 1)
2
3 (0 to 90)

0.215
0.016
0.120
0.111
0.247

* including humerus, radius, ulna, femur, tibia and fibula

Discussion
We report an overall infection risk of 3.6% secondary to
1492 open fractures, which is similar to that reported in the
current literature.1,9,13,16,21 The 12% incidence among
grade III fractures is lower than previously reported,12,22,23
The areas under ROC curves were good, which tends to
confirm we have recorded acceptable predictive values. We
consider that we have made a reasonable attempt to separate off a substantial part of nosocomial infections by
excluding methicillin-resistant staphylococcal infections
and those occurring 60 days or more after surgery.
Carsenti-Etesse et al24 suggested that up to 92% of infected
open fractures could be caused by hospital-acquired bacteria. This proportion seems largely overestimated according
to our data and our hospital does not have such a predominance of multi-susceptible Enterobacter and Pseudomonas, regardless of the site of infection. Using
definitions based on time from admission, all infections
might be termed nosocomial. However, contamination

most likely occurs before hospital admission at the time of

trauma and bone fracture.
Our study confirms many aspects of previous reports,
such as a predominance of male gender, long bones, tibial
fractures6,10,25-27 and a short median delay between injury
and infection, ranging between seven and 15 days.1,12 As
observed by other authors,10,25 time between injury and
surgery was not associated with increased risk, even if
others disagree.28 Intramedullary nailing or reaming
was not associated with more infections than plates or
external fixation.29
Grade III fractures were predominantly associated with
infection, particularly fractures with vascular lesions (grade
IIIc) of which one-third became infected. Conversely, no
association was observed with immunosuppressive status,
number of surgical interventions, delay between trauma
and surgery,7,30-32 seasonal variation, and duration of antibiotic therapy. Whether assessed as a continuous variable
or when categorised, duration of therapy was equivalent
THE BONE & JOINT JOURNAL

SHORT DURATION OF ANTIBIOTIC PROPHYLAXIS IN OPEN FRACTURES DOES NOT ENHANCE RISK OF SUBSEQUENT INFECTION

835

Table IV. Risk factors for infection in the 1492 open fractures. Variables in bold are statistically significant (two-tailed
p value < 0.05, unconditional logistic regression) (CI, confidence interval)
Univariate analysis

Multivariable analysis

Variable

Odds ratio (95% CI)

p-value

Odds ratio (95% CI)

p-value

Median age (yrs)

Age group
< 30 years
30 to 41 years
42 to 59 years
60 years

0.99 (0.97 to 0.99)

0.032

1.0 (1.0 to 1.0)

0.456

Reference
1.4 (0.7 to 2.9)
1.2 (0.6 to 2.6)
0.3 (0.1 to 0.9)

0.312
0.581
0.032

Reference
1.6 (0.8 to 3.5)
1.1 (0.5 to 2.6)
0.5 (0.2 to 1.7)

0.189
0.756
0.273

Immunosuppression*

1.6 (0.5 to 5.2)

0.409

1.2 (0.6 to 2.4)

0.562

Gustilo and Anderson grade

Grade I
Grade II
Grade III

Reference
2.1 (0.7 to 6.3)
14.9 (6.2 to 36.7)

0.182
< 0.001

Reference
1.8 (0.6 to 5.3)
12.5 (5.2 to 30.4)

0.348
< 0.001

Fracture of long bones

1.0 (0.5 to 2.0)

0.910

n.a

Upper extremity

0.6 (0.2 to 1.4)

0.221

n.a

Tibia

1.6 (0.9 to 2.8)

0.085

1.1 (0.6 to 2.0)

0.437

Per additional surgical intervention

1.1 (0.7 to 1.8)

0.626

n.a.

Median duration of antibiotics (days)

Antibiotic duration
1 day
2 to 3 days
4 to 5 days
> 5 days

1.0 (1.0 to 1.0)

0.150

1.0 (1.0 to 1.1)

0.664

Reference
2.4 (0.9 to 2.6)
8.9 (2.9 to 27.1)
9.8 (3.4 to 28.4)

0.086
< 0.001
< 0.001

Reference
0.6 (0.2 to 2.0)
1.2 (0.3 to 4.9)
1.4 (0.4 to 4.4)

0.651
0.207
0.261

Coverage for MRSA

1.2 (0.2 to 9.2)

0.851

n.a.

Median delay to surgery (days)

Delay to surgery
1 day
2 days
3 days
4 days
5 days

1.0 (0.99 to 1.02)

0.307

1.0 (0.99 to 1.02)

0.308

Reference
0.9 (0.2 to 3.7)
2.4 (0.5 to 10.4)
2.2 (0.7 to 6.3)
4.2 (2.2 to 8.3)

0.849
0.246
0.156
< 0.001

Reference
n.a
n.a
n.a
n.a

* including diabetes mellitus, human immunodeficiency virus infection, dialysis, cancer without remission
including humerus, radius, ulna, femur, tibia, and fibula
MRSA, methicillin-resistant Staphylococcus aureus

compared with one day in terms of incidence of infection.

This confirms the study by Dellinger et al9 who showed the
equivalence of five days of cefonicid or cefamandole treatment compared with one day of cefonicid in open extremity
fractures.
The choice of antibiotic agents or targeted empirical preventive therapy towards pathogens found on initial
debridement might theoretically play a role. In our retrospective study, 71% of infections complicating grade III
fractures were due to pathogens presumably selected by the
antibiotic agents used. However, a combination of cephalosporins with aminoglycosides, fluoroquinolones, metronidazole, vancomycin or carbapenems equally failed to
demonstrate a protective effect, underlining the fact that
VOL. 95-B, No. 6, JUNE 2013

the use of broader antibiotic coverage cannot guarantee the

absence of subsequent infection. Open fractures are contaminated by a large variety of antibiotic-susceptible pathogens,1,7,27,28,33 including P. aeruginosa and Staph. aureus.
Predicting which pathogen will emerge as the result of
ongoing selective antibiotic therapy is impossible, independent of the initial choice of the agent.
We investigated the congruence of pre-infection microbiological samples collected during the first surgical exploration with the infectious pathogen and found an absence
of any relationship. The proportion of positive pre-infectious cultures ranges between 60%21,34 and 83%33 in the
literature. This finding is shared by other groups,7,11,16,33,34
and very few reports support a possible correlation between

836

N. DUNKEL, D. PITTET, L. TOVMIRZAEVA, D. SUV, L. BERNARD, D. LEW, P. HOFFMEYER, I. UKAY

Table V. Risk factors for infection across all Gustilo grade III open fractures and grade IIIc fractures (extensive damage, vascular injury)
(CI, confidence interval). Variables in bold are statistically significant (two-tailed p value < 0.05, all unconditional logistic regression)
Grade III fractures (n = 310)
Univariate analysis

Grade IIIc fractures (n = 63)

Multivariable analysis

Multivariable analysis

Variable

Odds ratio (95% CI)

p-value

Odds ratio (95% CI)

p-value

OR (95% CI)

Female gender

0.5 (0.2 to 1.2)

0.101

n.a

Median age (yrs)

Age group
< 30 years
30 to 41 years
42 to 59 years
60 years

1.0 (1.0 to 1.0)

0.927

1.0 (1.0 to 1.0)

0.654

1.0 (1.0 to 1.0)

Reference
1.7 (0.8 to 4.0)
1.0 (0.4 to 2.6)
0.3 (0.1 to 1.6)

0.202
0.949
0.176

Reference
1.7 (0.7 to 4.3)
0.9 (0.3 to 2.6)
0.2 (0.1 to 1.4)

0.255
0.893
0.104

Reference
n.a
n.a
n.a

Immune suppression*

0.5 (0.1 to 4.2)

0.554

0.8 (0.1 to 6.3)

0.800

n.a

Psychiatric comorbidity

2.6 (1.1 to 6.0)

0.027

n.a.

Gustilo and Anderson grade

Grade IIIa
Grade IIIb
Grade IIIc

Reference
1.8 (0.3 to 11.4)
15.2 (3.4 to 68.5)

0.517
< 0.001

Reference
1.7 (0.3 to 10.6)
12.5 (2.7 to 57.8)

Long bone fracture

Upper extremity fracture
Tibial fracture
Vascular injury
Compartment syndrome
No. of surgical interventions
Delay trauma to surgery
Delay first to second surgery
Wound closure during first surgery
Intramedullar reaming
Plate osteosynthesis

1.8 (0.7 to 4.8)

1.1 (0.3 to 4.0)
1.6 (0.8 to 3.2)
5.1 (2.4 to 10.7)
7.8 (3.2 to 18.9)
1.2 (0.6 to 2.5)
1.0 (1.0 to 1.0)
0.9 (0.8 to 1.0)
0.4 (0.2 to 0.98)
1.6 (0.5 to 5.0)
0.8 (0.3 to 2.4)

0.247
0.838
0.203
< 0.001
< 0.001
0.871
0.571
0.768
0.047
0.420
0.695

n.a.
n.a.
1.0 (0.4 to 2.7)
n.a
n.a
1.2 (0.5 to 2.6)
1.0 (0.99 to 1.02)
0.9 (0.8 to 1.0)
0.4 (0.1 to 1.1)
1.1 (0.3 to 3.9)
0.6 (0.2 to 2.1)

0.691
0.716
0.824
0.624
0.940
0.742

n.a
n.a
0.5 (0.2 to 2.0)
n.a
1.0 (0.2 to 5.2)
n.a
1.0 (0.94 to 1.09)
n.a
n.a
n.a
n.a

Median duration of antibiotic therapy

(days)
Antibiotic therapy
1 day
2 to 3 days
4 to 5 days
> 5 days

1.0 (0.97 to 1.08)

0.923

1.0 (0.96 to 1.09)

0.910

n.a

Reference
0.3 (0.1 to 1.7)
0.6 (0.1 to 3.3)
1.0 (0.2 to 5.3)

0.463
0.935
0.352

Reference
0.3 (0.1 to 3.3)
0.6 (0.2 to 2.1)
1.6 (0.5 to 6.1)

0.949
0.243
0.432

Reference
1.5 (0.1 to 21.1)
2.2 (0.4 to 12.4)
6.9 (0.9 to 52.0)

p-value

n.a
0.324

n.a

0.590
0.001

0.973

Reference
-

0.232
0.675
0.321

0.876
0.546
0.134

* including diabetes mellitus, human immunodeficiency virus infection, dialysis, cancer without remission
including humerus, radius, ulna, femur, tibia and fibula

initial wound samples and further infection.21 There is the

possibility that secondary or repeat cultures may predict
infection. In theory, secondary cultures taken one or two
days after debridement might indicate more reliably the
potential pathogen/s. One study advocated a repeat culture
the day after debridement and found that a concordance
between these consecutive cultures could help eventually to
predict future infection.33 However, there is no widelyaccepted agreement on this at present among experts. As
we did not perform repeat cultures, we cannot comment on
this issue.
In our study, results of intra-operative samples were not
consistent with isolates from subsequent infection, possibly
because they were fully susceptible to the prophylaxis used.

However, the chosen prophylactic agent did not cover the

later infecting pathogen in 71% of cases. This could be the
result of low inoculum present at the site of injury and their
different virulence, or the pathogen(s) isolated might have
been selected by the prophylactic agent. If we infer that
infectious pathogens were indeed selected by the antibiotic
regimen or its duration, this would further support the use
of a short course therapy to avoid selection.
Our study has several limitations. First, it involves a
heterogeneous adult population treated in a single centre in a
high-income country, and this might limit the generalisability
of the findings to other settings.6 Our patients were mostly
victims of road accidents with no open fractures related to
gunshots or explosions.12,35 Fracture grading was performed
THE BONE & JOINT JOURNAL

SHORT DURATION OF ANTIBIOTIC PROPHYLAXIS IN OPEN FRACTURES DOES NOT ENHANCE RISK OF SUBSEQUENT INFECTION

by different surgeons. This might be important as the agreement on the Gustilo classification of open tibial fractures
among orthopaedic surgeons was only 60% in one study.36
To overcome this and other issues, a new classification
scheme was proposed by the Orthopaedic Trauma Association in 2010.37 There is little reported on its usefulness.
Although many key variables and risk factors have been
included, some were not, including body mass index,7 smoking history,7 trauma energy7,28 and injury causes.10 Infections
may have been missed in patients who were subsequently
treated elsewhere but as the Geneva University Hospitals is
the only public hospital in the area, we consider this to be
unlikely. Finally, surgeons differ in terms of surgical technique and usage of antibiotics depending on their reaction to
the injury they are treating. These are weaknesses inherent to
all observational studies and reinforces the need for randomised, controlled multicentre trials.9
In conclusion, most infections in open grade fractures
occur because of extensive tissue damage. Prophylaxis
beyond one day does not seem to counterbalance the negative effects of injury when decontamination fails in the early
initial management. If confirmed in prospective trials, what
is already known for grade I or II fractures could be
extended to grade III fractures.
The authors would like to thank the teams of the Orthopaedic Service, the Laboratory of Microbiology, and the Infection Control Programme of the Geneva
University Hospitals for their support. They would also like to thank R. Sudan
for expert editorial assistance.
No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article.
This article was primary edited by D. Rowley and first-proof edited by G. Scott.

References
1. Gustilo RB, Anderson JT. Prevention of infection in the treatment of one thousand
and twenty-five open fractures of long bones: retrospective and prospective analyses.
J Bone Joint Surg [Am] 1976;58-A:453458.
2. Gustilo RB, Gruninger RP, Davis T. Classification of type III (severe) open fractures
relative to treatment and results. Orthopedics 1987;10:17811788.
3. Hoff WS, Bonadies JA, Cachecho R, Dorlac WC. East practice management
guidelines work group: Update to practice management guidelines for prophylactic
antibiotic use in open fractures. J Trauma 2011;70:751754.
4. Ukay I, Harbarth S, Peter R, et al. Preventing surgical site infections. Expert Rev
Anti Infect Ther 2010;8:657670.
5. Trampuz A, Zimmerli W. Antimicrobial agents in orthopaedic surgery: prophylaxis
and treatment. Drugs 2006;66:10891105.
6. Hauser CJ, Adams CA, Jr, Eachempati SR. Surgical Infection Society guideline:
prophylactic antibiotic use in open fractures: an evidence-based guideline. Surg Infect
(Larchmt) 2006;7:379405.
7. Lenarz CJ, Watson JT, Moed BR, et al. Timing of wound closure in open fractures
based on cultures obtained after debridement. J Bone Joint Surg [Am] 2010;92A:19211926.
8. Barie PS. Breaking with tradition: evidence-based antibiotic prophylaxis of open
fractures. Surg Infect (Larchmt) 2006;7:327329.
9. Dellinger EP, Caplan ES, Weaver LD, et al. Duration of preventive antibiotic
administration for open extremity fractures. Arch Surg 1988;123:333339.
10. Khatod M, Botte MJ, Hoyt DB, et al. Outcomes in open tibia fractures: relationship
between delay in treatment and infection. J Trauma 2003;55:949954.
11. Zalavras CG, Patzakis MJ. Open fractures: evaluation and management. J Am
Acad Orthop Surg 2003;11:212219.

VOL. 95-B, No. 6, JUNE 2013

837

12. Mody RM, Zapor M, Hartzell JD, et al. Infectious complications of damage control
orthopedics in war trauma. J Trauma 2009;67:758761.
13. Braun R, Enzler MA, Rittmann WW. A double-blind clinical trial of prophylactic
cloxacillin in open fractures. J Orthop Trauma 1987;1:1217.
14. Nanchahal J, Nayagam S, Khan U, et al. BAPRAS/BOA Working Group. Standards
for the management of open fractures of the lower limb, 2009. http://
www.bapras.org.uk/downloaddoc.asp?id=141 (date last accessed 24 January 2013).
15. Gosselin RA, Roberts I, Gillespie WJ. Antibiotics for preventing infection in open
limb fractures. Cochrane Database Syst Rev 2004;1:CD003764.
16. Patzakis MJ, Bains RS, Lee J, et al. Prospective, randomized, double-blind study
comparing single-agent antibiotic therapy, ciprofloxacin, to combination antibiotic
therapy in open fracture wounds. J Orthop Trauma 2000;14:529533.
17. No authors listed. Clinical and Laboratory Standards Institute. Performance standards for antimicrobial susceptibility testing, seventeenth informational supplement,
2007. http://www.microbiolab-bg.com/CLSI.pdf (date last accessed 24 January
2013).
18. Horan TC, Gaynes RP, Martone WJ, Jarvis WR, Emori TG. CDC definitions of
nosocomial surgical site infection, 1992: a modification of CDC definitions of surgical
wound infections. Am J Infect Control 1992;20:271274.
19. Pugh RN, Murray-Lyon IM, Dawson JL, Pietroni MC, Williams R. Transection
of the oesophagus for bleeding oesophageal varices. Br J Surg 1973;60:646649.
20. Vittinghoff E, McCulloch CE. Relaxing the rule of ten events per variable in logistic
and Cox regression. Am J Epidemiol 2007;165:710718.
21. Ojo OD, Oluwadiya KS, Ikem IC, et al. Superficial swab cultures in open fracture
management: insights from a resource-poor setting. J Wound Care 2010;19:432438.
22. Ikem IC, Oginni LM, Bamgboye EA, et al. The bacteriology of open fractures in IleIfe, Nigeria. Niger J Med 2004;13:359365.
23. Vasenius J, Tulikoura I, Vainionp S, Rokkanen P. Clindamycin versus cloxacillin in the treatment of 240 open fractures: a randomized prospective study. Ann Chir
Gynaecol 1998;87:224228.
24. Carsenti-Etesse H, Doyon F, Desplaces N, et al. Epidemiology of bacterial infection during management of open leg fractures. Eur J Clin Microbiol Infect Dis
1999;18:315323.
25. Sungaran J, Harris I, Mourad M. The effect of time to theatre on infection rate for
open tibia fractures. ANZ J Surg 2007;77:886888.
26. Ostermann PA, Seligson D, Henry SL. Local antibiotic therapy for severe open
fractures: a review of 1085 consecutive cases. J Bone Joint Surg [Br] 1995;77-B:93
97.
27. Ako-Nai AK, Ikem IC, Daniel FV, Ojo DO, Oginni LM. A comparison of superficial
and deep bacterial presence in open fractures of the lower extremities. Int J Low
Extrem Wounds 2009;8:197202.
28. Pollak AN, Jones AL, Castillo RC, et al. The relationship between time to surgical
debridement and incidence of infection after open high-energy lower extremity
trauma. J Bone Joint Surg [Am] 2010;92-A:715.
29. Henley MB, Chapman JR, Agel J, et al. Treatment of type II, IIIA, and IIIB open
fractures of the tibial shaft: a prospective comparison of unreamed interlocking
intramedullary nails and half-pin external fixators. J Orthop Trauma 1998;12:17.
30. Al-Hilli AB, Salih DS. Early or delayed surgical treatment in compound limb fractures due to high velocity missile injuries: a 5-year retrospective study from Medical
City in Baghdad. Iowa Orthop J 2010;30:9498.
31. Charalambous CP, Siddique I, Zenios M, et al. Early versus delayed surgical
treatment of open tibial fractures: effect on the rates of infection and need of secondary surgical procedures to promote bone union. Injury 2005;36:656661.
32. Spencer J, Smith A, Woods D. The effect of time delay on infection in open longbone fractures: a 5-year prospective audit from a district general hospital. Ann R Coll
Surg Engl 2004;86:108112.
33. Robinson D, On E, Hadas N, et al. Microbiologic flora contaminating open fractures: its significance in the choice of primary antibiotic agents and the likelihood of
deep wound infection. J Orthop Trauma 1989;3:283286.
34. Faisham WI, Nordin S, Aidura M. Bacteriological study and its role in the management of open tibial fracture. Med J Malaysia 2001;56:201206.
35. Peleg K, Aharonson-Daniel L, Stein M et al. Gunshot and explosion injuries: characteristics, outcomes, and implications for care of terror-related injuries in Israel. Ann
Surg 2004;239:311318.
36. Brumback RJ, Jones AL. Interobserver agreement in the classification of open
fractures of the tibia: the results of a survey of two hundred and forty-five orthopaedic
surgeons. J Bone Joint Surg [Am] 1994;76-A:11621166.
37. Orthopaedic Trauma Association: Open Fracture Study Group. A new classification scheme for open fractures. J Orthop Trauma 2010;24:457464.