Diagnostic performance of swab PCR as an alternative to tissue culture methods for diagnosing infections associated with fracture fixation devices.

Abstract

Molecular procedures could potentially improve diagnoses of orthopaedic implant-related infections, but are not yet clinically implemented. Analysis of sonication fluid shows the highest sensitivity for diagnosing implant infections in cases of revision surgery with implant removal. However, there remains controversy regarding the best method for obtaining specimens in cases of revision surgery with implant retention. Tissue culture is the most common diagnostic method for pathogen identification in such cases. Here we aimed to assess the diagnostic performance of swab PCR analysis compared to tissue culture from patients undergoing revision surgery of fracture fixation devices. We prospectively investigated 62 consecutive subjects who underwent revision surgery of fracture fixation devices during a two-year period. Tissue samples were collected for cultures, and swabs from the implant surface were obtained for 16S rRNA PCR analysis. Subjects were classified as having an implant-related infection if (1) they presented with a sinus tract or open wound in communication with the implant; or (2) purulence was encountered intraoperatively; or (3) two out of three tissue cultures tested positive for the presence of the same pathogen. Tissue culture and swab PCR results from the subjects were used to calculate the sensitivity, specificity, accuracy, positive predictive value (PPV), negative predictive value (NPV), and area under the ROC curve (AUC) for identifying an orthopaedic implant-related infection. Orthopaedic implant-related infections were detected in 51 subjects. Tissue culture identified infections in 47 cases, and swab PCR in 35 cases. Among the 11 aseptic cases, tissue culture was positive in 2 cases and swab PCR in 4 cases. Tissue culture showed a significantly higher area under the ROC curve for diagnosing infection (AUC=0.89; 95% CI, 0.67-0.96) compared to swab PCR (AUC=0.66; 95% CI, 0.46-0.80) (p=0.033). Compared to swab PCR, tissue culture showed better performance for diagnosing orthopaedic implant-related infection. Although molecular methods are expected to yield higher diagnostic accuracy than cultures, it appears that the method of obtaining specimens plays an important role. Improved methods of specimen collection are required before swab PCR can become a reliable alternative to tissue-consumptive methods.