[Usefulness of 16S rDNA polymerase-chain-reaction (PCR) in the intraoperative detection of infection in revision of failed arthroplasties].

AIM: In this study the accuracy of the 16S DNA polymerase chain reaction (PCR) in revision arthroplasties was compared to that of conventional bacterial culture when correlated to intraoperative histological findings. Furthermore, the influence of antibiotic treatment and different ways of collecting samples was evaluated.

METHOD: In a prospective study we collected samples of tissues, aspiration fluids and swabs during revision arthroplasty surgery and examined them with PCR as well as conventional bacterial culturing methods. Also, we correlated these two methods with the histological findings of intraoperative tissue samples. Two independent examiners evaluated the samples according to the criteria of Mirra et al. Sensitivity, specificity, positive and negative prediction value and the accuracy were calculated for the molecular biological and culture methods. Three groups were defined to evaluate the influence of accompanying antibiotic treatment and the way of collecting the microbiological samples.

RESULTS: Nine periprosthetic infections could be detected by PCR as well as by conventional bacterial culturing. Correlated with the 25 positive histological findings this resulted in a sensitivity of 0.36, a specificity of 1.0, a negative prediction value of 0.61, a positive prediction value of 1.0 and an accuracy of 0.68 for both methods. Swabs compared to aspiration fluids and tissues samples showed the highest sensitivity with both methods. No higher sensitivity of PCR compared to conventional bacterial culturing could be observed in patients with accompanying antibiotic treatment.

CONCLUSION: Although PCR is more rapidly available for the diagnosis of periprosthetic infection, a definite advantage of this more expensive method could not be demonstrated in view of the same low sensitivity of PCR and conventional bacterial culturing.