Medical Trade International · Product assistant10096 2018 Article 3406
Readable documentation excerpt. Open the original PDF for all tables, figures and complete instructions.
Text extraction may differ from the original layout. Follow the manufacturer’s current instructions.Page 1
European Journal of Clinical Microbiology & Infectious Diseases (2019) 38:149–155
https://doi.org/10.1007/s10096-018-3406-4
ORIGINAL ARTICLE
Comparison of different human tissue processing
methods for maximization of bacterial recovery
Mohamed Askar 1 1
& Waheed Ashraf & Brigitte Scammell & Roger Bayston
1 1
Received: 26 July 2018 / Accepted: 1 October 2018 / Published online: 24 October 2018
# The Author(s) 2018
Abstract
Tissues are valuable microbiological samples that have proved superiority over swabs. Culture of tissue samples is
used in the diagnosis of a variety of infections. However, as well as factors such as the site of obtaining the sample,
the number of samples, and previous antibiotic use, the method of tissue processing may have an important effect on
sensitivity. Data from the literature comparing different tissue processing methods is very limited. This study aimed
to compare different mechanical and chemical methods of tissue processing in terms of efficacy and retaining the
viability of the bacteria in the tissues. Standard suspensions of Staphylococcus aureus and Escherichia coli were
prepared and treated differently to test the effect of that treatment on bacterial viability. Artificially inoculated pork
tissue and known infected human tissue samples were then processed by different methods prior to culture, and
results were compared. Percentages of reduction in the number of viable bacteria compared to the control by
homogenization was similar to 5-min dithiothreitol treatment but significantly lower than bead beating. Bacterial
recovery from homogenized human tissues was significantly higher than from any other method of treatment.
Although bead beating could be the most efficient method in obtaining a homogeneous tissue product, it signifi-
cantly reduces the number of viable bacteria within tissues. Homogenization offers the most effective easily con-
trollable retrieval of bacteria from tissue and retains their viability. Guidelines for diagnosing infections using tissue
samples should include a standardized processing method.
Keywords Homogenization . Tissue processing
Introduction diagnosing infections, e.g., method of sample collection
and number of samples [3]. Previous antibiotic use and
Tissue samples are among the most valuable samples the method of tissue processing could also have a det-
for microbiological testing, particularly in musculoskel- rimental effect on sensitivity [4]. The quality of every
etal infections. Recovery of bacteria from tissues has subsequent step in diagnosing infections is dependent on
been included in many guidelines for diagnosis of dif- the effectiveness of the initial step of tissue processing.
f e r e n t i n f e c t i o u s d i s e a se s . A n e x a m p l e i s t h e For highly sensitive microbiological analyses, it is cru-
Musculoskele
Page 1
on
been included in many guidelines for diagnosis of dif- the effectiveness of the initial step of tissue processing.
f e r e n t i n f e c t i o u s d i s e a se s . A n e x a m p l e i s t h e For highly sensitive microbiological analyses, it is cru-
Musculoskeletal Infection Society (MSIS) definition of cial to ensure maximal bacterial release from tissues.
prosthetic joint infection (PJI) [1]. Furthermore, tissue There are currently several tissue processing methods
samples have a proven superiority over swabs as a di- ranging from simple direct streaking on agar plates or
agnostic material [2]. Several factors can affect the qual- manual partitioning with scalpels to more automated ho-
ity and reproducibility when using tissues for mogenizers and chemical lysis, yet there is no firm con-
sensus regarding the efficacy of such processing
methods in terms of tissue disruption, releasing of bac-
* Roger Bayston teria from tissues, and the compromise of bacterial sur-
Roger.Bayston@nottingham.ac.uk vival with each method.
1 In this study, different mechanical and chemical processing
Department of Academic Orthopaedics, Queen’s Medical Centre,
University of Nottingham, C Floor, West Block, Derby Road, methods were compared in vitro and ex vivo to determine
Nottingham NG7 2UH, UK their efficacy and reproducibility.
Page 2
150 Eur J Clin Microbiol Infect Dis (2019) 38:149–155
Methods tubes contained 0.1-mm glass beads (Cambio: UC-13118-
50; Glass Beads; 0.1 mm). Four cycles of homogenization
Exposure of bacterial suspensions to processing and bead beating were run, 45 s each at 4500 rpm with 45 s
in between. Vortexing was used to simulate the routine tissue
In order to determine the impact of three different methods on processing in many microbiology laboratories. This was the
viability of planktonic bacteria, suspensions were subjected to same as the tissue homogenization tube but used for vortexing
homogenization or bead beating for various times. for 20 s instead of homogenization.
Staphylococcus aureus and Escherichia coli previously One hundred microliters of each product was plated out and
isolated in the Biomaterials-Related Infection laboratory, incubated overnight. Colonies were counted and recorded. For
University of Nottingham, UK, were used for the preparation the results to be considered valid, the blank piece plates had to
of bacterial suspensions in tryptic soy broth (TSB, Oxoid, grow neither S.aureus nor E.coli, though they may have other
Basingstoke, UK) at a concentration of 10 5 cfu/mL. irrelevant bacteria despite the prior extensive rinse.
Concentration was controlled by light absorbance calibration
and plating of serial dilutions. Known infected human tissues
One hundred microliters of each suspension was added to
1 mL of phosphate-buffered saline (PBS) and treated by either To compare the various processing methods when applied to
homogenization for two or four cycles (each cycle lasting for infected human clinical material, 26 tissue samples were col-
45 s at a speed of 4500 rpm with 45 s between cycles) or bead lected from six patients diagnosed with PJI according to the
beating for two or four cycles in the same way as homogeni- MSIS criteria. 0.5 g of each sample was added to 1 mL of PBS
zation using 0.1-mm glass beads. The MagNA Lyser homog- and processed by either homogenization, bead beating,
enizer (Roche, Basel, Switzerland) was used. vortexing, sonication, DTT, or proteinase K treatment.
One hundred microliters of bacterial suspensions were also Homogenization, bead beating, vortexing, and DTT were
added to 0.1% w/v dithiothreitol (DTT, Thermo Fisher run as described above. For proteinase K (Roche,
Scientific, Vilnius, Lithuania) in sterile water for 5 min before Mannheim, Germany), 20 μL were added 10 min before plat-
plating to be compared to homogenized samples for four cy- ing. One tube was sonicated for 5 min at 50 Hz. After it
cles as described above. became clear that homogenization and DTT processing were
One hundred microliters of each of processed and contr
Page 2
to be compared to homogenized samples for four cy- ing. One tube was sonicated for 5 min at 50 Hz. After it
cles as described above. became clear that homogenization and DTT processing were
One hundred microliters of each of processed and control superior, 14 known infected human tissue samples were col-
(unprocessed) bacterial suspension was spread on blood agar lected from either diabetic feet [4] or prosthetic joint infections
plates (Oxoid), and colonies were counted after overnight with pus-draining sinuses [5] and were processed by either
incubation. homogenization or DTT only.
The liquid product of each processing was cultured on
Mechanical processing of artificially inoculated pork blood agar plates and incubated aerobically and anaerobically
meat for 72 h and 14 days respectively.
To determine the influence of tissue on bacterial viability dur- Statistical analysis
ing processing, pork tissue samples were Bspiked^ with bac-
terial suspensions and processed by three methods. Pork bone- Graphpad Prism 7 was used to analyze data and produce
less meat was obtained from a local grocery store. Equal charts. Appropriate statistical tests were used according to
pieces weighing 7.5 g were prepared. Each piece was thor- the distribution of data. P values < 0.05 were considered
oughly and repeatedly washed in sterile water and further significant.
divided into equal thirds (2.5 g each), and stored in a −
20 °C freezer for future use. Ethics
One hundred microliters of 105 cfu/mL bacterial suspen-
sion (S. aureus or E.coli) was used for inoculation at the center Human tissue samples were collected under the ethical ap-
of the meat piece using a 1-mL plastic syringe (Becton proval of the Nottingham Health Science Biobank.
Dickinson, Drogheda, Ireland) and 16-mm, 5/8″ needles
(Becton Dickinson).
A third meat piece was left Bblank^ without inoculation to Results
report on possibility of contamination.
After inoculation, samples were left on the bench for 2 h Bacterial suspensions exposed to processing
then processed by either homogenization, bead beating, or
vortexing. Aliquots of 0.5 g of each piece were transferred Quantitative bacterial recovery from any processed bacterial
to 2-mL tubes along with 1 mL of PBS. For bead beating, suspension was significantly lower than from the unprocessed
Page 3
Eur J Clin Microbiol Infect Dis (2019) 38:149–155 151
control group. The percent reductions compared to unpro- S.aureus
400
cessed suspensions are shown in Table 1. Bacterial recovery
from bead-beaten samples was lower than from homogenized
samples. The difference in reduction between two and four 300
cycles of homogenization was statistically significant for
P=0.004
S.aureus but not for E.coli (Fig. 1). There was no statistically
P=0.08
CFU
significant difference between bacterial suspension samples 200
treated by either homogenization or DTT as shown in Fig. 2.
100
Mechanical processing of artificially inoculated pork
meat
0
tr 2 4 2 4
Quantitative bacterial recovery from artificially inoculated on
ol H H B
B
B
B
pork samples differed among samples depending on whether C
Processing
they were homogenized, bead beaten, or vortexed. For
S.aureus, means of recovered bacteria by homogenization,
bead beating, and vortexing were 394, 36, and 136 cfu/mL 150 E.coli
respectively. For E.coli, means were 448, 70, and 166 cfu/mL
respectively. Results are shown in Fig. 3.
P=0.4
100
Known infected human tissues
CFU
P<0.001
Nine out of the 26 tissue samples were positive by at least one
processing method. Only one patient with PJI had negative 50
culture results with all processing methods. Quantitative bac-
terial recovery from infected human tissues were signi
Page 3
P<0.001
Nine out of the 26 tissue samples were positive by at least one
processing method. Only one patient with PJI had negative 50
culture results with all processing methods. Quantitative bac-
terial recovery from infected human tissues were significantly
higher (p = 0.0239) after homogenization than all other sam- 0
ples (Table 2 and Fig. 4). Bacterial recovery from human- ol
tr H2 H 4 B 2 B 4
on B B
infected tissues from cases of diabetic feet and PJI was signif- C
Processing
icantly higher with homogenization than DTT (Fig. 5).
Table 3 summarizes the results of homogenization versus Fig. 1 Bacterial recovery from S.aureus and E.coli suspension after
homogenization and bead beating. H2 and H4 are two and four cycles
DTT. of homogenization respectively; BB2 and BB4 are two and four cycles of
bead beating respectively. CFU colony forming unit
Discussion bacteria is crucial. As the majority of bacteria are not expected
to be on the outside surfaces of collected tissue samples, it is
The first steps of tissue sample processing have a great impact not sufficient just to streak tissues on agar plates or to immerse
on the downstream steps in the microbiological workflow. For them in culture broth. Microbial distribution and loads within
diagnosis of infections in which recovery of bacteria from the potentially infected tissues can vary widely. Depending on
infected tissues is a gold standard, the ability to release the whether the dissection and harvest planes will cross a focus of
bacterial collection, some microbes could be located on the
surface of the tissues and might be retrieved by direct streak-
Table 1 Percentages of reduction in bacterial count after different
processing
ing or swabbing of the tissues, but it is just as likely that
surface contaminants are isolated.
Organism Processing variable Homogenization Bead beating Prior steps of releasing bacteria should be carried out first
to enhance the sensitivity of the microbiological tests. There
2 cycles 4 cycles 2 cycles 4 cycles
Page 3
Bead beating Prior steps of releasing bacteria should be carried out first
to enhance the sensitivity of the microbiological tests. There
2 cycles 4 cycles 2 cycles 4 cycles
are two prerequisites for an ideal processing method: to re-
S.aureus Median 14 35 48 80 lease as many bacteria as possible and not to adversely affect
interquartile range 2–24 24–43 33–60 70–84 bacterial viability. Here, we demonstrate a Btissue processing
E.coli Median 10 13 67 92 paradox^ in which the higher the force delivered to tissues for
interquartile range 0.5–20 5–20 54–76 83–96 processing, the more homogeneous the end product, and ap-
parently, the more bacterial release, but the lower bacterial
Page 4
152 Eur J Clin Microbiol Infect Dis (2019) 38:149–155
S.aureus S.aureus
80
P=0.006
P=0.7
60 P<0.001
500
CFU/mL (X10)
40 400
300 P=0.03
CFU
20
200
0
tio T 100
n D
ni D
za
om 0
og
e za
n
be
g ex
H tio at
rt
Vo
E.coli en i
in
og ea
d
om B
60 P=0.9 H
E.coli
P=0.017
CFU/mL (X10)
40
800 P=0.03
20 600
Page 4
P=0.017
CFU/mL (X10)
40
800 P=0.03
20 600
CFU 400 P=0.12
0
n
io T
at D
D 200
iz
en
og
om
H
0
Fig. 2 Bacterial recovery from S.aureus and E.coli suspension after tio
n g ex
in rt
homogenization versus DTT ni
za
be
at Vo
og B
e ea
om d
survival. A balance should be achieved to ensure both maxi- H
mal release and maximal viability of bacteria. As quantitative Fig. 3 Quantitative recovery of S.aureus and E.coli from inoculated pork
rather than qualitative estimation of bacteria in tissues is cru- samples by homogenization, bead beating, and vortexing. CFU colony
forming unit
cial for guiding the treatment, the method of initial tissue
processing is therefore important.
Diagnosis of prosthetic joint infection (PJI), for instance, is made of tissue processing methods. The tissue processing
has been and is continuing to be challenging for both micro- method could have a detrimental effect on the accuracy of
biologists and surgeons with tireless efforts to improve and microbiological diagnosis of PJI, but data in the literature on
standardize it. Periprosthetic tissue samples collected at the the optimal tissue processing method are limited.
time of surgery are among the most valuable materials for The optimal number of samples for diagnosis of PJI has
PJI diagnosis [6, 7]. One of the two major criteria of the been addressed by a number of studies. Atkins et al. 1998 [3],
MSIS definit
Page 4
essing method are limited.
time of surgery are among the most valuable materials for The optimal number of samples for diagnosis of PJI has
PJI diagnosis [6, 7]. One of the two major criteria of the been addressed by a number of studies. Atkins et al. 1998 [3],
MSIS definition of PJI depends on the isolation of the patho- Bemer et al. 2016 [8], and Peel et al. 2017 [9] recommended
gen from tissue or fluid samples [1]. In that consensus, differ- five to six, four, and three to four samples respectively. They
ent tissue-related aspects were discussed such as how many processed tissue samples by manual bead milling, automated
samples should be collected, how they should be collected, bead milling, and homogenization respectively. Different tis-
and for how long cultures should be incubated, but no mention sue processing could be one possible explanation for the
Page 5
Eur J Clin Microbiol Infect Dis (2019) 38:149–155 153
Table 2 Bacterial recovery in cfu/100 μL from infected human tissue Infected tissues
150
samples processed by different methods (H, homogenization; BB, bead
beating; S, sonication; V, vortexing; PK, proteinase K; DTT, dithiothreitol)
P<0.001
Sample Organism H BB S V PK DTT
CFU/mL (X10)
100
1 Staphylococcus epidermidis 5 0 0 0 1 0
2 Enterococcus faecalis 180 68 15 92 40 91
3 Enterococcus faecalis 139 68 19 21 44 22
50
4 Enterococcus faecalis 174 60 48 32 32 36
5 Enterococcus faecalis 300 280 16 59 71 26
6 Staphylococcus epidermidis 4 0 0 1 1 0
7 Enterococcus faecalis 5 4 0 0 3 0 0
8 Enterococcus faecalis 66 10 35 41 38 31 H DT
om D
9 Staphylococcus epidermidis 142 8 0 13 14 2 og
en
iz
at
io
n
variability of their results. Atkins et al. [3] developed a math- Fig. 5 Quantitative recovery from infected human tissue samples (from
Page 5
n
variability of their results. Atkins et al. [3] developed a math- Fig. 5 Quantitative recovery from infected human tissue samples (from
four diabetic feet and ten PJI cases) processed by homogenization versus
ematical model to reach their conclusion and recommendation
DTT
of five to six tissue samples to be collected with a diagnostic
cutoff of three or more positive cultures for satisfactory sensi-
tivity and specificity. Bemer et al. [8] conducted a multicenter disruption has not received the same attention in spite of the
study and recommended four samples to be collected and potentially equal impact.
cultured on three different media for an accurate cost- Furthermore, culture-negative PJI can occur in approxi-
effective diagnosis of PJI. Peel et al. [10] found that culturing mately 7% of all PJI cases, and it remains a challenge and a
of tissue samples in blood culture bottles could enhance the source of frustration for both microbiologists and orthopedic
sensitivity of PJI diagnosis without compromising the speci- surgeons [5]. In our study, one patient had negative culture
ficity. Later, they concluded that three tissue samples cultured results despite the processing method. A well-studied reason
in blood culture bottles can give the same level of accuracy as for failure of culture to retrieve the causative organism is prior
four samples cultured conventionally [9]. It is clear that much antibiotic use [5, 11]. However, suboptimal microbiological
care was given to the culture media which are unarguably methods could also be implicated especially with low tissue
crucial for enhanced bacterial recovery. Yet, initial tissue microbial load. In our case, the patient had not been on anti-
biotics prior to surgery, and different processing methods did
not improved the detection of causative organism which might
Infected tissues have been in a viable but non-culturable state [12].
250 The ideal tissue processing method should have the highest
possible bacterial release and survival. In other words, it
200
Page 5
The ideal tissue processing method should have the highest
possible bacterial release and survival. In other words, it
200 should not be too harsh for the bacteria to survive. In addition,
it should be less liable to contamination and not over-
CFU/mL (X10)
150 dependent on worker’s skills.
In the present study, the ability of the processing method to
100 keep the bacteria alive and cultivable was firstly addressed by
processing bacterial suspensions and comparing them to the
control non-processed group. Homogenization showed the
50
highest bacterial survival.
An example of gram-positive and gram-negative common
0
pathogens was chosen to be tested. The effect of processing
R G N G K TT
IZ E TI N
ATI
O
TE
N
A
SE D can vary among different bacterial species and strains, most
EN EA XI IN
G B N
IC R TE probably because of variability in their cell envelope structure
O D SO VO O
M
O EA PR and hence resistance to mechanical stress. Two and four cycles
H B
Processing of homogenization and bead beating were tested. As tough-
Fig. 4 Quantitative recovery from human
Page 5
and hence resistance to mechanical stress. Two and four cycles
H B
Processing of homogenization and bead beating were tested. As tough-
Fig. 4 Quantitative recovery from human infected tissue samples (from
ness of different tissues is variable, the amount and duration of
PJI cases) processed by different methods. Bacterial recovery was the force that needs to be delivered to obtain a homogeneous
significantly higher with homogenization (p = 0.024) mixture can vary. Tissue processing should always be kept to
Page 6
154 Eur J Clin Microbiol Infect Dis (2019) 38:149–155
Table 3 Bacterial recovery from
infected tissue samples processed Sample Infection Organism Homogenization DTT
by homogenization versus DTT
(cfu/100 μL) 1 Diabetic feet S.epidermidis 53 4
Corynebacterium pseudotuberculosis
2 Diabetic feet S.aureus 45 2
3 Diabetic feet Streptococcus dysgalactiae 128 12
4 Diabetic feet Serratia marcescens 140 13
Staphylococcus capitis
5 PJI Staphylococcus epidermidis 3 1
6 PJI Enterococcus faecalis 125 74
7 PJI Enterococcus faecalis 56 16
8 PJI Enterococcus faecalis 78 7
9 PJI Enterococcus faecalis 35 32
10 PJI Enterococcus faecalis 2 0
11 PJI Enterococcus faecalis 12 0
S.aureus
12 PJI Enterococcus faecalis 58 17
S.aureus
13 PJI Enterococcus faecalis 73 13
S.aureus
14 PJI Enterococcus faecalis 68 16
S.aureus
the minimum required as the longer the processing, the lower reputation in releasing bacteria from biofilms from the sur-
the viability of the released bacteria. faces of the implants [17–20] but this method does not give
In order to determine whether this conclusion held true in a homogeneous product in the case of tis
Page 6
ion in releasing bacteria from biofilms from the sur-
the viability of the released bacteria. faces of the implants [17–20] but this method does not give
In order to determine whether this conclusion held true in a homogeneous product in the case of tissues.
case of infected tissues, artificially inoculated pork tissues Our study confirms the results of Redanz et al. 2015 [4] but
were processed, and the same result was found. Pork was gives a clearer view of the superiority of homogenization
chosen because of presumptive similarity with human tissues firstly by giving an estimate of loss of bacterial viability using
[13]. different processing methods. Secondly, this is a comparative
DTT has been recommended for tissue processing because study between different published mechanical and chemical
of its high sensitivity, specificity, practicability, and relatively methods. The study by Redanz et al. 2015 [4] compared ho-
low cost [14, 15]. In this study, DTT was found to be as safe as mogenization to direct plating which is not a suitable alterna-
homogenization in terms of bacterial viability. However, when tive because of lack of actual processing.
they were compared in terms of bacterial release from known The current study has number of limitations. First, the
infected tissue samples, homogenization was significantly bet- small number of samples might make it difficult to generalize
ter. Lack of this comparative study was admitted as a limita- the conclusion. Second, the study did not include fungi or
tion by the authors recommending use of DTT [15]. anaerobic bacteria which are uncommon, yet reported, causa-
Nevertheless, DTT could be considered as an alternative in tive organisms of PJI, and other infections. Lack of isolation
case of unavailability of homogenization because of its pres- of gram-negative bacteria from infected human tissues also
ervation of bacterial viability. might make it hard to generalize these findings. However,
As a final step, tissue samples collected from PJI patients these results give good evidence of superiority of homogeni-
diagnosed according to the MSIS criteria were used to com- zation in releasing bacteria while retaining their viability.
pare homogenization, bead beating, sonication, vortexing, In conclusion, homogenization offers the most effective
proteinase K, and DTT treatment. Homogenization showed retrieval of bacteria from tissue and retains their viability.
significantly higher bacterial release. The enhanced quantita- Tissue processing should be standardized and included in
tive and qualitative microbial yield using homogenization has guidelines of diagnosis and definition of PJI and other infec-
been reported previously [4]. Manual and automated bead tions because of the detrimental effect of some methods.
beating has also been described in PJI diagnosis [3, 8, 16].
In our s
Page 6
d using homogenization has guidelines of diagnosis and definition of PJI and other infec-
been reported previously [4]. Manual and automated bead tions because of the detrimental effect of some methods.
beating has also been described in PJI diagnosis [3, 8, 16].
In our study, bead beating, in spite of a very homogeneous
Funding This work was supported by Newton-Mosharafa Fund through
product, seems to be destructive to the released bacteria with the mission sector of the Egyptian ministry of higher education. MAwas a
low bacterial count end result. Sonication has a good PhD student in our institute supported by the funder.
Page 7
Eur J Clin Microbiol Infect Dis (2019) 38:149–155 155
Compliance with ethical standards 7. Osmon DR et al (2013) Diagnosis and management of prosthetic
joint infection: clinical practice guidelines by the Infectious
Diseases Society of America. Clin Infect Dis 56(1):e1–e25
Conflict of interest The authors declare that they have no conflicts of
8. Bemer P et al (2016) How many samples and how many culture
interest.
media to diagnose a prosthetic joint infection: a clinical and micro-
biological prospective multicenter study. J Clin Microbiol 54(2):
Disclaimer The funders had no role in study design, data collection and 385–391
interpretation, or the decision to submit the work for publication. 9. Peel TN et al (2017) Optimal Periprosthetic tissue specimen num-
ber for diagnosis of prosthetic joint infection. J Clin Microbiol
Open Access This article is distributed under the terms of the Creative
55(1):234–243
Commons Attribution 4.0 International License (http://
10. Peel TN et al (2016) Improved diagnosis of prosthetic joint infec-
creativecommons.org/licenses/by/4.0/), which permits unrestricted use,
tion by culturing periprosthetic tissue specimens in blood culture
distribution, and reproduction in any medium, provided you give appro-
bottles. MBio 7(1):e01776–e01715
priate credit to the original author(s) and the source, provide a link to the
Creative Commons license, and indicate if changes were made. 11. Malekzadeh D et al (2010) Prior use of antimicrobial therapy is a
risk factor for culture-negative prosthetic joint infection. Clin
Orthop Relat Res 468(8):2039–2045
12. Li L et al (2014) The importance of the viable but non-culturable
state in human bacterial pathogens. Front Microbiol 5:258
References 13. Swindle MM et al (2012) Swine as models in biomedical research
Page 7
state in human bacterial pathogens. Front Microbiol 5:258
References 13. Swindle MM et al (2012) Swine as models in biomedical research
and toxicology testing. Vet Pathol 49(2):344–356
1. Parvizi J, Gehrke T, Chen AF (2013) Proceedings of the interna- 14. Drago L et al (2013) Use of dithiothreitol to improve the diagnosis
tional consensus on Periprosthetic joint infection. Bone Joint J 95- of prosthetic joint infections. J Orthop Res 31(11):1694–1699
B(11):1450–1452 15. De Vecchi E et al (2016) Treatment with dithiothreitol improves
2. Aggarwal VK et al (2013) Swab cultures are not as effective as bacterial recovery from tissue samples in osteoarticular and joint
tissue cultures for diagnosis of periprosthetic joint infection. Clin infections. J Arthroplast 31(12):2867–2870
Orthop Relat Res 471(10):3196–3203 16. Roux AL et al (2011) Diagnosis of prosthetic joint infection by
3. Atkins BL et al (1998) Prospective evaluation of criteria for micro- beadmill processing of a periprosthetic specimen. Clin Microbiol
biological diagnosis of prosthetic-joint infection at revision Infect 17(3):447–450
arthroplasty. The OSIRIS collaborative study group. J Clin 17. Janz V et al (2015) Improved identification of polymicrobial infec-
Microbiol 36(10):2932–2939 tion in total knee arthroplasty through sonicate fluid cultures. Arch
4. Redanz S, Podbielski A, Warnke P (2015) Improved microbiolog- Orthop Trauma Surg 135(10):1453–1457
ical diagnostic due to utilization of a high-throughput homogenizer 18. Trampuz A et al (2007) Sonication of removed hip and knee pros-
for routine tissue processing. Diagn Microbiol Infect Dis 82(3): theses for diagnosis of infection. N Engl J Med 357(7):654–663
189–193 19. Renz N et al (2015) Sonication in the diagnosis of periprosthetic
5. Berbari EF et al (2007) Culture-negative prosthetic joint infection. infections : Significance and practical implementation. Orthopade
Clin Infect Dis 45(9):1113–1119 44(12):942–945
6. Tande AJ et al (2014) Clinical characteristics and outcomes of pros- 20. Portillo ME et al (2013) Sonication versus vortexing of implants for
thetic joint infection caused by small colony variant staphylococci. diagnosis of prosthetic joint infection. J Clin Microbiol 51(2):591–
MBio 5(5):e01910–e01914 594
Page 7
MBio 5(5):e01910–e01914 594