Epidemiological and microbiological study of hip and knee infected arthroplasties using the sonication technique

In the absence of a gold standard criterion for diagnosing prosthetic joint infections (PJI), some authors have shown that sonication of the removed implant may provide superior microbiological identification to synovial fluid and peri-implant tissue cultures, thereby improving PJI diagnosis. In this scenario, we propose to evaluate the performance of sonication cultures compared to tissue cultures in the PJI diagnosis and the epidemiological and microbiological profile of these infections. Methods: A retrospective study of 146 patients undergoing removal of hip or knee prosthetic joints due to any reason and submitted to implant sonication between 2010 and 2018 was conducted at a tertiary hospital in São Paulo, Brazil. In the absence of a gold standard criterion for diagnosing PJI, the following diagnostic criteria were used to compare the performance of sonication fluid culture versus periprosthetic tissue cultures: ICM-2018, IDSA, EBJIS criteria, and a modified clinical criterion. Sonication fluid and tissue cultures were performed according to published standardized protocols. Statistical analysis was carried out using McNemar´s test and proportion comparison was employed to calculate p-value. Results: Of total patients assessed, 56% (82/146) were diagnosed with PJI using the clinical criteria. Out of these cases, 57% (47/82) tested positive on tissue culture versus 93% (76/82) on sonication culture. Applying the clinical criteria, sonication fluid cultures showed 92.7% sensitivity (95% CI: 87.1%- 98.3%), tissue cultures 57.3% (95% CI: 46.6%- 68.0%) (p<0.001) and the two methods combined showed 96,3% sensitivity (95% CI: 91.5%- 100%). Sonication culture and the combination of the sonication cultures with tissue cultures showed a higher sensitivity than tissue cultures alone for all diagnostic criteria used. Sonication cultures were also more sensitive than tissue cultures for some subgroups of patients with clinically PJI: patients with prothesis placed more than 24 months, visible purulent secretion, positive histology and previous antimicrobial use. Use of sonication fluid culture in association with tissue culture improved diagnostic accuracy. Conclusions: The sonication technique improved microbiological identification in patients undergoing hip or knee prosthetic joint revision for any reason. In a context where diagnostic criteria available have shortcomings and tissue cultures remain the gold standard, tissue sonication can aid PJI diagnosis. Sonication is especially useful when diagnosis employing the criteria is inconclusive due to missing data (joint puncture, histology, or preoperative laboratory exams), the periprosthetic tissue sample available for collection is small, and the sample processing method is not ideal.