[Molecular diagnostics for the detection of prosthetic joint infection].
PURPOSE OF THE STUDY: Ten years after inauguration of the molecular methods into the orthopaedic practice for diagnosing Prosthetic Joint Infection (PJI), this approach is still in the limelight of research and discussion. The aim of the current study was to determine the diagnostic power of our Polymerase Chain Reaction (PCR) protocol for preoperative detection of bacterial nucleic acid in the synovial joint fluid. MATERIAL AND METHODS: Synovial fluids obtained from thirty-five septic hip or knee arthroplasties and sixty-six aseptic controls were investigated by the conventional PCR technique. Two subgroups were established with regard to antibiotic administration before sample collection; with (n=13) and without (n=22) previous antibiotic exposition, respectively. All the surgeries were performed under the identical conditions with strictly established and fulfilled inclusion criteria. The study design applied was a prospective cohort trial. Primers targeting phylogenetically conserved regions of the bacterial gene were used to detect the bacterial 16SrRNA gene in the retrieved samples. If this was positive, a restriction endonuclease treatment of the amplified DNA was performed to reveal the PJI pathogen. Current guidelines were used to evaluate the test performance, including the confidence intervals. The concordance between the culture and PCR identification of PJI pathogens was estimated providing both of the relevant data were available. RESULTS: A qualitative analysis showed the following figures in the subgroup without previous antibiotic exposition: sensitivity (0.64), specificity (0.97), accuracy (0.89), positive predictive value (0.88), negative predictive value (0.89), likelihood ratio for positive result (21.0), and likelihood ratio for negative result (0.38). In the second subgroup the corresponding figures were as follows: 0.85, 0.97, 0.95, 0.85, 0.97, 27.9, and 0.16. The rate of concordance between the microbial and PCR findings was almost identical in both of the subgroups. DISCUSSION: There were large discrepancies in sensitivity and positive predictive values found in the published results. The earlier studies have had various methodological weaknesses, including a lack of strictly formulated inclusion criteria and control groups. In addition, there are differences among research centers in PCR laboratory procedures and specimen retrieval tactics which may potentially have an impact on PCR results. Low sensitivity and high specificity of our PCR technique may be explained by both the intrinsic (DNA extraction protocol, configuration of inner controls, choice of detection threshold, etc.) and extrinsic factors (in particular intra-operative retrieval of specimens). A hypothesis on the inadequacy of PCR techniques for PJI detection still remains to be excluded. CONCLUSION: Based on the current study, the positive results of our PCR technique may be perceived as a mild criterion from the point of power for PJI diagnosis support. However, its clinical utility should be significantly increased in cases with higher pretest probability of PJI, but negative cultures.