Performance of implant sonication culture for the diagnosis of periprosthetic shoulder infection

Abstract

Diagnosing infection after shoulder arthroplasty can be a challenge because of the high prevalence of low-virulence organisms, such as Propionibacterium acnes. The purpose of this study was to evaluate the utility of implant sonication fluid cultures in the diagnosis of periprosthetic joint infection compared with standard culture techniques in patients undergoing revision shoulder arthroplasty. Routine perioperative testing was performed in 53 patients who underwent revision shoulder arthroplasty. In addition to routine tissue and fluid culture, the retrieved shoulder implants underwent sonication with culture of the sonicate fluid. Diagnostic performance of implant sonication culture was determined on the basis of previously defined infection criteria and compared with standard intraoperative cultures. Of the 53 revision cases that underwent implant sonication fluid culture, 25 (47%) were classified as infected. Intraoperative culture (tissue and fluid) sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy were 96%, 75%, 77%, 95%, and 85%, respectively. Using a cutoff of >20 colony-forming units per milliliter to exclude contaminants, the sensitivity, specificity, PPV, NPV, and accuracy of implant sonicate culture were 56% (P < .001, compared with standard intraoperative cultures), 93% (P = .07), 88% (P = .4), 70% (P = .02), and 75% (P = .22), respectively. Without use of a sonication fluid culture cutoff value, the sensitivity, specificity, PPV, NPV, and accuracy of implant sonicate culture were 96% (P = 1.0, compared with standard intraoperative cultures), 64% (P = .38), 71% (P = .53), 95% (P = .9), and 79% (P = .45). Implant sonication fluid culture in revision shoulder arthroplasty showed no significant benefits over standard intraoperative cultures in diagnostic utility for periprosthetic joint infection.