Abstract
Aims
Microbiological diagnosis of bone and joint infections (BJIs) is pivotal. However, no consensus exists about the best choice for techniques to be used and the best indications for molecular methods. Our objectives were: (i) to compare the performance of various microbiological diagnostic methods (cultural and molecular) on synovial fluid specimens and (ii) to select an algorithm for optimizing the diagnosis of BJIs in adults.
Methods
This prospective multicentric study (in Lyon and Saint-Etienne, France) included 423 joint fluid samples, collected from 333 adult patients (median age 69 years) suspected for BJI on the basis of medical history and clinical symptoms. For each inclusion, joint fluid and blood culture were collected concomitantly. The synovial fluid was also inoculated into blood culture bottles. Cytology, culture (using 5 solid media and an enrichment broth, incubated for 15 days), universal 16S rRNA PCR and PCR targeting Staphylococcus spp, S.aureus, Streptococcus spp, S.pneumoniae, Kingella kingae, Borrelia burgdorferi and Propionibacterium acnes were systematically performed on synovial fluid.
Results
Prosthetic materials were present in 65.0% of the cases and 31.7% of the patients had received antibiotics in the 15 days before puncture. Out of 423 joint fluids, 265 (62.6%) were positive by at least one diagnostic technique (cultural or molecular): 219 mono- and 46 poly-microbial, for a total of 322 bacteria. Identified bacteria were staphylococci in 54.0%, streptococci-enterococci in 15.2%, Gram-negative bacilli in 14.0%, anaerobic species in 10.9% and other bacteria in 5.9% of cases. Comparing the individual performance of each cultural technique, blood culture bottles showed the highest rate of positivity (detecting 61.4 and 58.4% of the bacteria, for the paediatric and anaerobic bottles, respectively) but cannot be performed alone and require to be combined with solid media. The 16S rDNA PCR was positive in only 49.2% of the cases whereas higher detection was obtained with specific PCR. Blood cultures performed concomitantly with joint puncture were positive in only 9.7% of the cases.
Conclusions
In order to simplify the culture procedures and to precise the place of PCR for synovial fluid, we propose the following algorithm: joint fluids should be inoculated onto 3 solid media (blood and chocolate agars for 2 days, anaerobic blood agar for 10 days), associated with inoculation into blood culture bottles for 10 days. If culture remains negative, 16S rDNA PCR and/or Staphylococcus PCR should be added. Applying this algorithm on our cohort, 93.6% of the bacteria would have been detected.
