Does intraoperative tissue sample enrichment help or hinder the identification of microorganisms in prosthetic joint infection?

Abstract

Objectives
Biofilm formation reduces the efficacy of standard microbiological techniques in prosthetic joint infection. This study aimed to investigate the sensitivity, specificity and predictive values of tissue sample enrichment as a means to increase diagnostic yield.

Methods
Patients undergoing revision arthroplasty surgery between May 2004 and January 2013 had intraoperative tissue samples cultured in standard media as well as enriched in brain heart infusion broth. Patients were separated into infected or non-infected groups according to modified criteria from the Musculoskeletal Infection Society.

Results
A total of 197 revision arthroplasties were included (non-infected, n = 165; proven infection, n = 32). The mean time until revision in non-infected and infected groups was 75.9 and 41.7 months, respectively. The commonest microorganisms cultured were coagulase-negative staphylococci (42.9 %) and Staphylococcus aureus (34.4 %). The sensitivity and specificity of standard tissue culture were 0.25 (CI 0.18–0.33) and 0.98 (CI 0.95–0.99), respectively. Including enrichment culture results increased the sensitivity to 0.45 (CI 0.37–0.54), but decreased specificity to 0.59 (CI 0.52–0.66).

Conclusion
Any potential increase in the sensitivity is far outweighed by the extremely high false-positive rate. Results of tissue samples cultured by enrichment should be used with caution and may lead to a worse outcome if incorrectly interpreted.

Full text

1
Authors pre-print manuscript. Published version available at Eur J Orthop Surg Traumatol. 1
2014 Nov 22. [Epub ahead of print] PMID: 25416208 2
3
Does intraoperative tissue sample enrichment help or hinder the 4
identification of microorganisms in prosthetic joint infection? 5
6
7
Running title: 8
Intra-operative tissue sample enrichment for diagnosis of prosthetic joint infection 9
10
11
12
Mr Robert W Jordan1, Specialist Registrar 13
14
Mr Adnan Saithna2*, Consultant Orthopaedic Surgeon 15
E-mail: Adnan.Saithna@nhs.net 16
Tel: +44(0)247 6351 351 17
18
Mr Nicholas Smith1, Specialist Registrar 19
20
Mr Rory Norris1, Specialist Registrar 21
22
Mr Andrew Sprowson3, Associate Professor 23
24
Mr Pedro Foguet1, Consultant Orthopaedic Surgeon 25
26
*Denotes corresponding author 27
28
29
30
1. University Hospitals Coventry and Warwickshire NHS Trust, Clifford Bridge Road, 31
Walsgrave, Coventry, CV2 2DX, UK 32
33
2. The George Eliot Hospital, College Street, Nuneaton, CV10 7DJ, UK 34
35
3. The University of Warwick, Coventry, CV4 7AL, UK 36
37
38
39
40
41
42
43
44

2
Abstract 45
Objectives: Biofilm formation reduces the efficacy of standard microbiological techniques in 46
prosthetic joint infection. This study aimed to investigate the sensitivity, specificity and 47
predictive values of tissue sample enrichment as a means to increase diagnostic yield. 48
49
Methods: Patients undergoing revision arthroplasty surgery between May 2004 and January 50
2013 had intra-operative tissue samples cultured in standard media as well as enriched in 51
Brain Heart Infusion Broth. Patients were separated into infected or non-infected groups 52
according to modified criteria from the Musculoskeletal Infection Society. 53
54
Results: 197 revision arthroplasties were included (non infected, n=165; proven infection, 55
n=32). The mean time until revision in non-infected and infected groups was 75.9 and 41.7 56
months respectively. The commonest microorganisms cultured were coagulase negative 57
staphylococci (42.9%) and staphylococcus aureus (34.4%). The sensitivity and specificity of 58
standard tissue culture was 0.25 (CI 0.18-0.33) and 0.98 (CI 0.95-0.99). Including enrichment 59
culture results increased the sensitivity to 0.45 (CI 0.37-0.54) but decreased specificity to 60
0.59 (CI 0.52-0.66). 61
62
Conclusion: Any potential increase in the sensitivity is far outweighed by the extremely high 63
false positive rate. Results of tissue samples cultured by enrichment should be used with 64
caution and may lead to a worse outcome if incorrectly interpreted. 65
66
Key words: 67
Arthroplasty; Prosthetic Joint Infection; Revision Surgery; Microbiology; Tissue Enrichment 68
Culture 69

3
Introduction 70
Infection after arthroplasty surgery is a devastating complication with an incidence of around 71
1%.1-4 Diagnosis can be challenging due to the formation of biofilms, the resulting low 72
diagnostic yield of standard microbiological techniques and the lack of consensus for 73
defining infection. In order to address this, the Musculoskeletal Infection Society has recently 74
published criteria for prosthetic joint infection (PJI) that uses a combination of clinical, 75
microbiological, biochemical and histological results.5 76
Correct identification of the causative microorganism is essential to optimize clinical 77
outcomes as this allows appropriately targeted antibiotic therapy based on culture 78
sensitivities. The culture of intraoperative tissue samples has traditionally been the gold 79
standard for identifying microorganisms in PJIs.6,7 Five or six tissue samples should be 80
harvested and a positive result is reflected by 3 or more positive cultures.8 However, the 81
sensitivity of these tests is often low in PJI due to biofilm formation.8-10 This has led to the 82
exploration of further techniques to increase the sensitivity of microorganism detection, such 83
as culture enrichment. This process involves placing the sample in a ‘broth’, which provides 84
ideal conditions for growth of microorganisms. Studies have reported that enrichment can 85
increase the sensitivity of both generic samples11 and within a PJI setting.12, 13 However some 86
authors argue that the increase in positive results is in fact a result of sample contamination.14 87
Therefore the exact role of enrichment in helping identify microorganisms in PJI remains 88
unclear and this study aims to assess the role of enrichment in terms of sensitivity, specificity, 89
and predictive values. 90
91
Methods 92
We retrospectively reviewed all revision arthroplasty procedures performed by the senior 93
surgeon (PF) for any indication between May 2004 and January 2013. During this period 94

4
both one and two stage revision surgeries were undertaken. Those procedures performed as a 95
one-stage revision or the first stage of a two-stage revision were included for analysis in this 96
study. The Musculoskeletal Infection Society published criteria for diagnosis of PJI in 2011, 97
however a modification of this criteria was used to classify patients as either infected or non-98
infected post-operatively (see Table 1).5 This modification was necessary as during the study 99
period as neither histology nor synovial white blood cell counts were performed on intra-100
operative tissue samples at our centre. It is our understanding that this reflects practice in 101
many centres and therefore represents a pragmatic study design. 102
103
Table 1 - Criteria used for diagnosis of peri-prosthetic joint infection 104
The$presence$of$a$major$factor:
1)# Sinus#tract#communicating#with#the#prosthesis.
2)## Pathogen#isolated#by#culture#from#2#or#more#####
###########separate#tissue#or#fluid#samples
The$presence$of$3$out$of$the$5$
minor$factors:
1)# Elevated#erythrocyte#sedimentation#rate#and#
serum#C?reactive#protein#concentration.
2)# ‘Profuse’#or#‘high’#synovial#white#blood#cell##
###########count.
3)# ‘Majority’#of#polymorphonuclear#celles#in####
###########synovial#sample.
4)# Presence#of#purulence#in#the#affect#joint.
5)# Isolation#of#a#microorganism#in#one#culture#of##
###########peri?prosthetic#tissue#or#fluid.
105

5
Revision surgery was performed by the senior author (PF) in a theatre with laminar airflow. 106
A minimum of 3 different tissue samples were taken in all infected cases, taken with a fresh 107
sterile scalpel and forceps and put into separate containers. Tissue samples were taken in an 108
identical fashion for non-infected revisions but cases where no tissue samples were taken 109
were excluded from the study. On arrival in the laboratory tissue samples were divided into 110
two; the first part was cultured on standard agar plates for a total of 5 days and the second 111
part underwent enrichment. Results from the two culturing processes were reported 112
separately; one for standard culture and the other for enrichment. The enrichment process 113
involved placement of the sample in a Brain Heart Infusion Broth (Oxoid Microbiology 114
Products, Hampshire, UK).15 After 24 hours in the enrichment broth the sample was then 115
placed on standard agar plates for a further five days. 116
117
For those in the infected group empirical intravenous antibiotics were commenced post-118
operatively until microbiological results allowed tailoring of the treatment strategy. Once the 119
patients C-reactive protein had normalised they were converted to an oral antibiotic regimen 120
whilst an inpatient. If after 48 hours of oral antibiotics the patient remained well and the C-121
reactive protein normal they could be discharged with outpatient appointments at 1, 3 and 6 122
weeks after which antibiotics were stopped if progress was satisfactory. 123
124
Hospital electronic records were used to retrieve data including demographics, primary 125
diagnosis, date of surgery, reason for failure of arthroplasty, pathology and microbiology 126
results, clinic letters, theatre records, the use of antibiotics prior to revision surgery and the 127
requirement for further surgery. Microbiological results from all tissue cultures were 128
recorded, assessing both the number of specimens taken and the number of positive results by 129
culture and enrichment. Isolation of a microorganism in the infected group was taken as a 130

6
true positive result and in the non-infected as a false positive. A negative culture was 131
regarded as a false negative in the infected group and a true negative in the non-infected 132
group. Patients were followed up after their revision surgery until July 2013 to analyse 133
whether further surgery was required and to determine whether there was any evidence of 134
infection. When further surgery had been undertaken, both the indication for surgery and the 135
microbiological results from tissue samples taken were collected. Sensitivity, specificity and 136
positive and negative predictive values, with their respective binomial 95% confidence 137
intervals (CIs), were calculated by using standard formulae. 138
139
Results 140
217 revisions were performed during the study period. Using the modified criteria 32 were 141
classified as infected (16 hips and 16 knees), and 185 cases as non-infected. All 32 infected 142
cases had at least 3 tissue samples taken but 20 of those in the non-infected group had no 143
tissue samples taken and were excluded from the study. After exclusion, a total of 197 cases 144
were included in the study, of which 165 were non-infected (53 knee and 112 hip revisions) 145
and 32 were infected. The demographics of the two groups are shown in Table 2. 146
147
148
149
150
151
152
153
154
155

7
Table 2 - Patient demographics 156
Non-infected group
Infected group
Age (years)
Mean
63.9
69.5
Range
20-93
46-83
Gender (%)
Male
30.3
43.8
Female
63.6
56.3
Primary diagnosis
(%)
Osteoarthritis
86.7
75
Rheumatoid
4.8
15.6
Post fracture
6.1
6.3
Dysplastic
1.8
3.1
Perthes
0.6
0
157
In the infected group, the mean age was 69.5 years and 56.3% were female. 24 were 158
performed as a one-stage revision and 8 were the first stage of a two-stage revision. The 159
mean time from original surgery to revision was 41.7 months (range 7 to 148 months) and 160
this is further illustrated in Table 3. A total of 141 tissue samples were collected with 35 161
positive on standard cultures and 29 by enrichment. The commonest microorganisms grown 162
in standard cultures were coagulase negative staphylococci (42.9%) and staphylococcus 163
aureus (34.4%). The microorganisms isolated in all samples are shown in Table 4. Twelve 164
patients had received oral antibiotics within six months of their revision surgery and in these 165
cases 49 tissue samples were harvested of which 6 (10%) were positive on standard cultures 166
and 11 (27%) by enrichment, further details are shown in Table 5. 167
168
169

8
Table 3 – Stratification of patients by time elapsed from original surgery until revision 170
Time
Non-infected group (%)
Infected group (%)
Under 1 year
9.1
18.8
1 to 2 years
13.3
21.9
2 to 3 years
13.3
12.5
3 to 4 years
12.7
9.4
4 to 5 years
8.5
15.6
5 to 10 years
23.6
15.6
10 to 15 years
12.7
6.3
15 to 20 years
3.0
0
Over 20 years
5.5
0
171
172
Table 4 - Microorganisms isolated from tissue culture in infected cases 173
Positive standard culture
Positive enrichment culture
Coagulase negative
staphylococcus
15
28
Staphylococcus aureus
11
0
Enterobacter cloacae
3
0
Pseudomonas
3
0
Enterococcus
2
0
Alpha haemolytic
streptococcus
0
1
174
175

9
Table 5 - Effect of pre-operative antibiotics on standard tissue culture results 176
Antibiotics
administered
Number of
patients
Number of
samples
Positive
standard
samples (%)
Positive
enrichment
samples
Until surgery
3
14
0
50
1 month prior to
surgery
2
10
20
0
1-3 months
prior to surgery
4
16
18.8
0
3-6 months
prior to surgery
2
9
0
66.7
Any antibiotics
within 6 months
11
49
10.2
26.5
No antibiotics
20
92
32.6
17.4
177
The non-infected group contained 165 patients with a mean age of 63.9 years. The mean time 178
until revision was 75.9 months (range 6 months to 434 months) as detailed in Table 3. The 179
indications for revision surgery are illustrated in Table 6 with the commonest being aseptic 180
loosening (44.8%), problematic resurfacing (21.1%) and recurrent dislocations (9.7%). In this 181
group a total of 215 tissue samples were taken with 4 positive in standard cultures and 84 by 182
enrichment. The microorganisms grown are shown in Table 7 with coagulase negative 183
staphylococcus (72.6%) being isolated in the majority of positive enrichment samples 184
185
186
187

10
Table 6 - Reason for prosthesis failure in the non-infected group 188
Reason
Percentage
Aseptic loosening
44.8
Painful hip resurfacing
21.1
Dislocation
9.7
Pain with no obvious cause
6.1
Malpositioning
4.8
Progression of arthritis to other knee
compartments
4.8
Failed cannulated screw
2.4
Pain post hemiarthroplasty
2.4
Wear
1.8
Impending fracture
1.8
189
The sensitivity and specificity of standard tissue culture was 0.25 (CI 0.18-0.33) and 0.98 (CI 190
0.95-0.99). Including data from enrichment studies increased the sensitivity to 0.45 (CI 0.37-191
0.54) but decreased specificity to 0.59 (CI 0.52-0.66). Similarly, the positive predictive value 192
decreased from 0.90 (CI 0.75-0.97) to 0.42 (CI 0.34-0.50). The negative predictive value was 193
virtually unchanged, from 0.67 (CI 0.61-0.77) to 0.62 (CI 0.55-0.69). These results are shown 194
in Table 8 and similar findings were demonstrated in both hip and knee cases. 195
196
Patients were followed up after revision surgery for a mean of 45.7 months in the infected 197
group (range 6 to 105) and 40.8 months in the non-infected group (range 6 to 108.7). During 198
this follow up period, 7 of the 32 infected cases required further surgery with the details 199
shown in Table 9. Of these, four required further surgery for infection. In the non-infected 200

11
group 19 (11.5%) required further surgery during follow up; 10 for recurrent dislocations, 5 201
for aseptic loosening and 4 for infection. The mean time to further surgery in the four 202
infected cases was 2.7 months, in three of these cases more than 2 tissue cultures were 203
positive and in the other all samples were negative. None of these four patients had grown 204
any organism by either enrichment or standard culture at their initial revision surgery. 205
206
Table 7 - Microorganisms isolated from tissue culture in non-infected cases 207
Positive standard culture
Positive enrichment
culture
Coagulase negative
staphylococcus
1
61
Staphylococcus aureus
3
0
Micrococcus
0
5
Skin flora
0
3
Alpha haemolytic
streptococcus
0
2
Enterococcus
0
2
Diptheroids
0
2
Pseudomonas
0
2
Streptococcus mitis
0
2
Coliforms
0
2
Escherichia coli
0
1
Klebsiella
0
1
Bacillus oxidase
0
1
208

12
Table 8 - Sensitivity, specificity, positive and negative predictive values 209
Sensitivity
Specificity
Positive
predictive
value
Negative
predictive
value
All Tissue
samples (CI)
0.25
(0.18-0.33)
0.98
(0.95-0.99)
0.90
(0.75-0.97)
0.67
(0.61-0.77)
All Tissue
samples +
enrichment(CI)
0.45
(0.37-0.54)
0.59
(0.52-0.66)
0.42
(0.34-0.50)
0.62
(0.55-0.69)
TKR tissue
samples (CI)
0.30
(0.20-0.43)
0.98
(0.91-1.0)
0.95
(0.73-1.0)
0.61
(0.51-0.7)
TKR tissue +
enrichment
samples (CI)
0.52
(0.40-0.65)
0.40
(0.29-0.52)
0.44
(0.33-0.56)
0.48
(0.35-0.62)
THR tissue
samples (CI)
0.21
(0.13-0.31)
0.99
(0.94-1.0)
0.84
(0.60-0.96)
0.70
(0.63-0.76)
THR tissue +
enrichment
samples (CI)
0.46
(0.34-0.58)
0.68
(0.60-0.76)
0.40
(0.29-0.52)
0.73
(0.64-0.80)
210
211
212
213
214

13
Table 9 - Further surgery during follow up of infected cases 215
Original revision
surgery
Reason for
further surgery
Details
One stage hip
(n=13)
Instability
Required revision surgery – intraoperative
samples negative
Instability
Required revision surgery – intraoperative
samples negative
Infection
3 months post procedure - same organism grown
as original procedure. Treated with suppression
antibiotics.
Two stage hip
(n=3)
None
One stage knee
(n=11)
Patella tendon
rupture
Required reconstructive surgery – intraoperative
samples negative
Two stage knee
(n=5)
Infection
Nine months post surgery - grew streptococcus,
previous samples negative.
Treated with debridement, change of spacer and
antibiotics successfully.
Infection
44 months post surgery - grew pseudomonas,
same organisms as initial surgery
Infection
28 months post surgery - grew pseudomonas,
previous samples negative. Treated by above
knee amputation after failed debridements.
216
217

14
Discussion 218
Biofilm formation in PJI reduces the diagnostic yield of standard microbiological techniques. 219
Enrichment has been used to try and increase the detection rate as identifying the causative 220
organism allows targeted antibiotic therapy and improved outcomes.11-13,16 However in our 221
study 39% of samples taken in non-infected cases were positive by enrichment and this 222
supports concerns that sample contamination is responsible for the increased number of 223
microorganisms isolated and an unacceptable increase in the rate of false positive results.14 224
225
The primary goal of identifying the responsible microorganism intra-operatively is to allow 226
targeted antibiotic treatment.16 If a microorganism is not identified, the patient is usually 227
treated with broad spectrum antibiotics. If a microorganism is detected, targeted antibiotics 228
can be used. However if the incorrect microorganism is identified, targeted antibiotics may 229
not cover the ‘true’ microorganism. Therefore in cases of already known infection, the 230
specificity can be considered more important. Inclusion of results from enrichment reduced 231
the specificity in our study from 0.98 to 0.59 with a similar reduction reported in both THR 232
and TKR cases. This suggests using enrichment results would increase the number of patients 233
receiving a false positive diagnosis (and unnecessary antibiotic treatment) and also increase 234
the risk of prescribing incorrectly targeted antibiotics in cases of true infection. 235
236
The commonest microorganisms cultured in our study were coagulase negative 237
staphylococcus and staphylococcus aureus and this reflects their known incidence in PJI.17-20 238
In those patients receiving antibiotics within 6 months of revision surgery the proportion of 239
positive standard tissue cultures was 10.2% compared to 32.6% in the those patients who did 240
not receive antibiotics. This is consistent with previous work which demonstrated that over 241
50% of culture negative PJIs follow treatment with antibiotics.21 As, there is no single test 242

15
that can diagnose PJI, it is possible that some of the non-infected group were suffering from 243
low grade infections that were not identified at the time of revision surgery. We followed up 244
these patients for a mean of 75.9 months to assess whether they had required subsequent 245
surgery and if there was any evidence of on-going infection. During this follow up period 246
four patients (2.4%) required further surgical intervention for infection with three having 247
positive microbiological results at the time of this surgery. Interestingly, none had previously 248
grown organisms by either standard culture or enrichment at their original revision surgery 249
suggesting that these infections were either de novo from the time of revision surgery or were 250
unidentified infections. Previous studies have reported infection rates in aseptic revisions of 251
between 1.3% and 5% and the fact our results are similar would suggest that these may also 252
be de novo infections.22-24 253
254
As well as the difficulty in definitively diagnosing or excluding PJI, there are some 255
limitations to this study. The retrospective design resulted in some slight variability in 256
practice. There were a total of 20 non-infected cases where tissue samples were not harvested 257
and in the remainder of this group the numbers of samples varied, which may bias the results. 258
A positive culture in the infected group for both standard and enrichment cultures was 259
assumed to be a true positive and this is likely to have over-estimated the sensitivity of these 260
techniques. We know from the non-infected group that enrichment produced a high false 261
positive rate and therefore a proportion of positive cultures in the infected group may also 262
have been secondary to contamination. In addition, we only analysed tissue sample cultures. 263
Therefore the results of this study are only valid when considering tissue samples intra-264
operatively. 265
266
267

16
Conclusion 268
This study has produced some important results that can help clinicians determine whether 269
the use of BHIB enrichment cultures are indicated in the investigation of a particular patient 270
and how to interpret the results. We conclude that there is an unacceptably high false 271
positive rate in presumed non-infected cases and great caution should be taken in diagnosing 272
infection in these patients on the basis of BHIB enrichment cultures alone. In those patients 273
with a clinical PJI, standard microbiological techniques have low yield and using BHIB 274
enrichment cultures can increase sensitivity. However false positive results are still a concern 275
and therefore results should be interpreted with caution. In this scenario it may be advisable 276
to avoid the use of an antibiotic regimen based solely upon the BHIB enrichment culture if it 277
excludes other possible causative organisms. Further studies are needed to either support or 278
refute these findings. 279
280
281
282
283
284
285
286
287
288
289
290
291
292

17
References 293
1) Kurtz SM, Lau E, Schmier J, Ong KL, Zhao K, Parvizi J. Infection burden for hip and 294
knee arthroplasty in the United States. J Arthroplasty 2008; 23(7): 984-991. 295
2) Jamsen E, Varonen M, Huhtala H, Lehto MU, Lumio J, Konttinen YT, Moilanen T. 296
Incidence of prosthetic joint infections after primary knee arthroplasty. J Arthroplasty 2010; 297
25(1): 87-92. 298
3) Pulido L, Ghanem E, Joshi A, Purtill JJ, Parvizi J. Periprosthetic joint infection: the 299
incidence, timing, and predisposing factors. Clin Orthop 2008; 466(7): 1710-1715. 300
4) Phillips JE, Crane TP, Noy M, Elliott TS, Grimer RJ. The incidence of deep prosthetic 301
infections in a specialist orthopaedic hospital: a 15 year prospective survey. J Bone Joint Surg 302
Br 2006; 88(7): 943-948. 303
5) Parvizi J, Zmistowksi B, Berbar EF, Bauer TW, Springer BD, Della Valle CJ, Garvin 304
KL, Mont MA, Wongworawat MD, Zalavras CG. New definition for periprosthetic joint 305
infection: from the workgroup of the Musculoskeletal Infection Society. Clin Orthop Relat 306
Res 2011; 469: 2992-2994. 307
6) Bauer TW, Parvizi J, Kobayashi N, Krebs V. Current Concepts Review – Diagnosis of 308
periprosthetic infection. J Bone Joint Surg Am 2006; 88(4): 869-881. 309
7) Parvizi J, Adeli B, Zmistowski B, Restrepo C, Greenwald A. Management of 310
Periprosthetic Joint Infection: The Current Knowledge. AAOS Exhibit Selection. J Bone 311
Joint Surg Am 2012; 18;94(14): 104. 312
8) Atkins BL, Athanasou N, Deeks JJ, Crook DWM, Simpson H, Peto TEA, Mclardy-313
Smith P, Berendt AR. Prospective evaluation of criteria for microbiological diagnosis of 314
prosthetic-joint infection at revision arthroplasty. J Clin Microbiol 1998; 36: 2932-2939. 315
9) Parvizi J, Ghanem E, Sharkey P, Aggarwal A, Burnett SJ, Barrack RL. Diagnosis of 316
infected total knee: findings of a multicenter database. Clin Orthop Relat Res 2008; 466(11): 317

18
2628–2633. 318
10) Trampuz A, Piper KE, Jacobson MJ, Hanssen AD, Unni KK, Osmon DR, Mandrekar 319
JN, Cockerill FR, Steckelberg JM, Greenleaf JF, Patel R. Sonication of removed hip and knee 320
prostheses for diagnosis of infection. N Engl J Med 2007; 357(7): 654-663. 321
11) Daur AV, Klimak F Jr, Cogo LL, Botão GD, Monteiro CL, Dalla Costa LM. 322
Enrichment methodology to increase the positivity of cultures from body fluids. Braz J Infect 323
Dis 2006; 10(6): 372-373. 324
12) Hughes HC, Newnham R, Athanasou N, Atkins BL, Bejon P, Bowler IC. 325
Microbiological diagnosis of prosthetic joint infections: a prospective evaluation of four 326
bacterial culture media in the routine laboratory. Clin Microbiol Infect 2011; 17(10): 1528-327
1530. 328
13) Velay A, Schramm F, Gaudias J, Jaulhac B, Riegel P. Culture with BACTEC Peds Plus 329
bottle compared with conventional media for the detection of bacteria in tissue samples from 330
orthopedic surgery. Diagn Microbiol Infect Dis 2010; 68(1): 83-85. 331
14) Trampuz A, Steckelberg JM, Osmon DR, Cockerill FR, Hanssen AD, Patel R. 332
Advances in the laboratory diagnosis of prosthetic joint infection. Rev Clin Microbiol 2003; 333
14:1-14. 334
15) http://www.oxoid.com/UK/blue/prod_detail/prod_detail.asp?pr=CM1135&c Accessed 335
on 1st July 2014 336
16) Larsen LH, Lange J, Xu Y, Schonheyder HC. Optimizing culture methods for diagnosis 337
of prosthetic joint infections: a summary of modifications and improvements reported since 338
1995. J Med Microbiol 2012; 61(3): 309-316. 339
17) Sharma D, Douglas J, Coulter C, Weinrauch P, Crawford R. Microbiology of infected 340
arthroplasty: implications for empiric peri-operative antibiotics. Journal of Orthopaedic 341
Surgery 2008; 16(3): 339-342 342

19
18) Nickinson RSJ, Board TN, Gambhir AK, Porter ML, Kay PR.The microbiology of the 343
infected knee arthroplasty. Int Orthop 2010; 34(4): 505–510. 344
19) Stefánsdóttir A, Johansson D, Knutson K, Lidgren L, Robertsson O. Microbiology of 345
the infected knee arthroplasty: report from the Swedish Knee Arthroplasty Register on 426 346
surgically revised cases. Scand J Infect Dis 2009; 41(11-12): 831-840. 347
20) Tsukayama DT, Estrada R, Gustilo RB. Infection after total hip arthroplasty: a study of 348
the treatment of one hundred and six infections. J Bone Joint Surg Am 1996; 78: 512–523. 349
21) Berbari EF, Marculescu C, Sia I, Lahr BD, Hanssen AD, Steckelberg JM, Gullerud R, 350
Osmon DR. Culture-negative prosthetic joint infection. Clin. Infect. Dis 2007; 45: 1113-351
1119. 352
22) Katz RP, Callaghan JJ, Sullivan PKM, Johnston RC. Long-term results of revision total 353
hip arthroplasty with improved cementing technique. J Bone Joint Surg Br 1997; 79(2): 322-354
326. 355
23) Kosashvili Y, Backstein D, Safir O, Lakstein D, Gross AE. Dislocation and infection 356
after revision total hip arthroplasty: comparison between the first and multiply revised total 357
hip arthroplasty. J Arthroplasty 2011; 26(8): 1170-1175. 358
24) Mortazavi SM, Schwartzenberger J, Austin MS, Purtill JJ, Parvizi J. Revision total knee 359
arthroplasty infection: incidence and predictors. Clin Orthop Relat Res 2010; 468(8): 2052-360
2059. 361
362
Conflicts of Interest 363
All authors confirm that there are no conflicts of interest to declare. 364
365
Funding and Sponsorship 366
No funding or sponsorship was received for this study. 367