The microbiological diagnosis of prosthetic joint infection is essential for optimal therapeutic decision. The reliability of the bacteriological diagnosis of all joint infections depends on the quality of samples, terms of transportation to the laboratory and the techniques used in the laboratory. For each step, protocols must be drafted.
The microbiological diagnosis of prosthetic joint infections is tricky because of the isolation of a bacterium in a peri-prosthetic levy does not mean that isolated germ is responsible for the known or suspected infection. Only the positivity several deep samples, performed in the site of infection possible to ensure that isolated bacteria have a formal value. Superficial swab samples as a fistula or a broken scar should not be made as they are misleading and have no reasonable intelligence value. Moreover, they remain in the files and are a source of conflict.
I - Samples
They must be performed after judgment (at least 15 days and sometimes benefit) of any antibiotics to avoid false negative results.
· Preoperative samples :
o joint puncture with joint lavage addressed separately in the laboratory . It is best to leave these samples in their sterile syringe sealed by a sealing stopper. Decanting in a pot is source of contamination and exposure to air can kill anaerobic bacteria. It is possible to inoculate the joint aspiration of fluid in two ways aerobic blood culture and anérobie ; but it is recommended to keep a few drops of joint fluid in the syringe (or a heparin tube) for cytology and colorations.
o abscess puncture .
The results of these samples will define the modalities of surgical management and initial antibiotic therapy
· In case of fever , it is important to take three blood cultures in aerobic and anaerobic and systematically take antibiotics before: urine, skin lesions, vaginal swab, nasal swab and throat ...) to detect the gateway or the head of infection spread of the bacterium
· Deep intraoperative samples confirm the preoperative diagnosis or help to establish it. They should be done without any antibiotics or antibiotic prophylaxis. It is recommended to take at least five samples from the liquid macroscopically pathological areas (pus, joint fluid) or solid (tissue interposition, granuloma, pathological bone ...) .
All samples should be taken under strict aseptic conditions and arrive quickly to the laboratory. The transport of these samples should meet the requirements of best practice transport of organic products with risk of infection.
II - Conventional microbiological techniques
All samples should be technical in safety cabinet Type 2, with cassock disposable, sterile gloves, disposable materials and instruments to avoid maximum contaminant that can cause confusion.
A / joint fluid
The study should include joint fluid cytology (numeration and formula nucleated elements). For knee prostheses, more than 1700 cells / mm3 (sensitivity 94% and specificity 88%) and more than 65% neutrophils are very suggestive of prosthesis infection . For the hip and shoulder, counting and formula nucleated elements has not yet been evaluated. The examination of a pellet cytocentrifugation Gram stain can afford to see bacteria.
B / tissue samples
Tissue fragments must be coarsely crushed to release cells of bacteria and biofilm. Grinding can be done by hand in a mortar or using shredders that dilacèrent but do not reduce solid samples into mush. Examination of tissue fragments smear must look for the presence of neutrophils (better identified by staining May Grunwald Giemsa stain) and the presence of bacteria (Gram stain). The sensitivity of direct examination is low (6%) but the specificity close to 100% .
C / Cultures
Samples must be seeded on enriched agar (blood agar in aerobic, anaerobic columbia, chocolate and multivitamins in 5% CO2) in liquid media aerobic and anaerobic enrichment that allow the detection of bacteria growth slow due to the nutritional richness of the medium used and the greater amount of inoculated sample [3,5].
The media of blood cultures containing resins adsorbing antibiotioques and lytic agents for releasing bacteria cells can be interesting, but some bacteria (eg Propionibacterium acnes and Peptostreptococcus spp ) can not grow there.
The examining cultures must be daily for agar media under aerobic and C02 as well as liquid media; agar in anaerobic are examined every 48 hours. Once a disorder appears in a broth, it must be transplanted.
Incubation should be continued at least ten days to allow growth of some colonies of different species or micro-colonies that appear on agar offset in time. Recently it has been recommended to keep cultures for 14 days. 
Some bacteria can not disturb the liquid media, it is strongly recommended at the end of the incubation, transplanting systematically enrichment broth on agar media described above, and incubated for 48 hours before being discarded.
Search for mycobacteria is performed only on specific indication.
III - The bacteriological results
The bacteriological results can be fast (48h) in acute infection caused by virulent bacteria, often visible from the Gram stain of pus smear.
Most often, the result is long to get because:
· Bacterial colonies appear slowly and often in a manner offset in time, giving a polymorphic nature to crops,
· Presence of different aspects of settlements that require isolates every aspect; interest loupe [7,8]
· Late onset (D + 5 to D + 7) Propionibacterium spp , of Peptostreptococcus and nutritional variants or " small colony variants " have lost some characters making their identification problem [9,10].
· Culture solely enrichment broth ( Propionibacterium sp and Peptostreptococcus will not be detected in blood culture bottles for PLC
· Susceptibility difficult to obtain and tedious dentification of demanding bacteria.
Interpretation of results
To involve a bacterium responsible for the infection of a prosthesis, it is necessary that:
· Several samples are positive :
· At least 3 samples (pre- or intraoperative) with positive bacterial skin flora: SCN, Propionibacterium , corynébatérie or
· At least 2 positive samples with bacteria not belonging to the skin flora and which possible contamination can not be discussed (eg Staphylococcus aureus , Enterobacteriaceae: E. coli, Klebsiella, Proteus spp ... ) or
· One positive sample (pre-operative (joint aspiration or blood culture or intraoperative) with a bacteria that can not be a contaminant (eg, pneumococcal, Salmonella, Listeria, Campylobacter, Pasteurella ...).
In general, the isolation of a bacterium belonging to the skin flora in a single sample is not significant especially when using enrichment media. It is then necessary to compare the bacteriological information, clinical, radiological and histopathological to decide if the isolated bacterium must be taken into account or not.
The result obtaining bacteriological period may not be less than 48 hours and is often greater than 10 days. This shows the importance of preoperative diagnosis for not wasting time postoperatively.
Bacterial strains responsible for bone and joint infections, particularly those responsible for prosthesis infections should be stored at -80 ° C
IV - Epidemiology
The most frequently isolated bacteria from the genus Staphylococcus (about 60% of prosthesis infection): S.aureus and coagulase negative staphylococci in particular S. epidermidis are in equal proportion . streptococci (beta-hemolytic streptococci groups B, C, G), enterococci, gram-negative rods (Enterobacteriaceae andPseudomonas aeruginosa ), anaerobic (usually Gram positive) is also isolated . About 10% of hip prostheses infections are plurimicrobiennes.
Of skin flora bacteria (coagulase negative staphylococci, Propionibacterium acnes ) are more often found in chronic forms. 
Numerous other species were isolated: Haemophilus sp. Campylobacter sp. Clostridium difficile, Brucella, Francisella, Pasteurella, ...
The presence of Mycoplasma. Tropheryma whipplei can be detected by molecular biology techniques.
Rare fungal infections are highlighted: Candida albicans, Aspergillus sp. .
Mycobacteria are perhaps not sufficiently researched. 
Published series reported between 5 to 15% of sterile cultures in prosthetic joint infections. This can be explained by an antibiotic that was not long enough Arrested, poor quality samples (swabs or sampling microscopic size), too long transportation in inadequate conditions (mailing ...), media unsuitable for cultivation of fastidious microorganisms, inadequate incubation period, a mycobacterium infection that grows only in special environments [11,12]. Finally, in this rare case, sometimes fistulized puriform reaction may be associated with a granuloma or a foreign body reaction.
V - The activity of antibiotics
It is determined by susceptibility testing, but can be difficult:
· Same patient, several samples can be positive with several aspects of colonies for the same bacterial species (eg Staphylococcus epidermidis ) susceptibility testing should be performed on different aspects because it is common to observe changes in activity antibiotics ; some strains are very slow to grow, the standardizing susceptibility testing is not always easy to follow.
· Many strains of staphylococci are resistant to oxacillin: detecting the gene mec A will often performed for coagulase-negative staphylococci.
· MIC glycopeptides are frequently required to select the glycopeptide, adjust the dosage and identify GISA (glycopeptide intermediate Staphylococcus aureus ) or GISE (glycopeptide intermediate Staphylococcus epidermidis ).
· On viridans streptococci, the activity of beta-lactam antibiotics (amoxicillin, cefotaxime) should be checked (CMI) even if they appear active on classical susceptibility.
Bacteriological diagnosis of an articular prosthesis infection based on the presence of the same bacterium (or bacteria of the same) and the same resistance phenotypes.
The microbiological diagnosis of prosthetic joint infections on is difficult. In the literature, the sensitivity and specificity are very variable from 65 to 94% and from 81 to 100%, respectively. 
The quality of bacteriological diagnosis depends on the quality of harvest, transport, technical treatment in the laboratory of biological validation and multidisciplinary interpretation. The microbiological diagnosis can only be reliable if the laboratory complies with the methodology for the detection of bacteria with growing requirements involving: the conservation of culture media for at least 10 days, the careful study of the various colonies and the realization of multiple susceptibility which are expensive and are not counted in this bacteriological examination codified in the standard level nomenclature, that the sample is positive or sterile with a bacteria that has several susceptibility.
We must no longer accept bacteriological results " sterile in 48 hours " since the majority of chronic prosthetic joint infections are due to bacteria that slowly grow by 5 to 10 days or more and are in half of the cases, isolated only enrichment broths after prolonged incubation longer than a week.
Full conference title:
- RICAI 29th (2009)