Simulated joint infection assessment by rapid detection of live bacteria with real-time reverse transcription polymerase chain reaction.

BACKGROUND: Although microbiological bacterial culture is currently considered the gold standard for diagnosis of septic arthritis, many studies have documented substantial false-negative and false-positive rates. The objective of this study was to determine whether real-time quantitative reverse transcription polymerase chain reaction can be used to detect bacterial messenger RNA (mRNA) in synovial fluid as a way to distinguish live and dead bacteria as an indicator of active infection.
METHODS: Synovial fluid samples were obtained from twelve consecutive patients who presented with knee pain and effusion but no evidence of infection. Following assurance of sterility with plate cultures, each sample was inoculated with clinically relevant bacteria and incubated for twenty-four hours to simulate septic arthritis. Bacterial viability and load were assessed with cultures. Selected samples were also treated with a single dose of a combination of two antibiotics, vancomycin and gentamicin, and sampled at several time points. Total RNA isolated from each sample was analyzed in triplicate with one-step real-time quantitative reverse transcription polymerase chain reaction to detect mRNA encoding for the genes groEL or femC. Controls included sterile, uninoculated samples and inoculated samples analyzed with quantitative polymerase chain reaction without reverse transcription. mRNA content was estimated on the basis of detection limits as a function of serial dilutions and was expressed as a function of colony number in bacterial cultures and RNA content as determined spectrophotometrically.
RESULTS: All synovial fluid samples that had been inoculated with one of the four bacterial species, and analyzed in triplicate, were identified (distinguished from aseptic synovial fluid) with real-time quantitative reverse transcription polymerase chain reaction; there were no false-negative results. All inoculated samples produced bacterial colonies on culture plates, while cultures of the aseptic samples were negative for growth. The detection limit of the one-step bacterial mRNA-based real-time quantitative reverse transcription polymerase chain reaction varied depending on the bacterial species. A time-dependent decrease in the concentration of detectable bacterial mRNA was seen after incubation of bacteria with antibiotics.
CONCLUSIONS: The direct quantification of the concentration of viable bacterial mRNA with real-time quantitative reverse transcription polymerase chain reaction allows identification of both culture-positive bacterial infection and so-called unculturable bacterial infection in a simulated septic arthritis model. In contrast to conventional polymerase chain reaction, real-time quantitative reverse transcription polymerase chain reaction minimizes false-positive detection of nonviable bacteria and thus provides relevant information on the success or failure of antibiotic therapy.