Limited data are available for the diagnosis of patients with tumors with infected endoprosthetic reconstructions. The purpose of this study was to evaluate whether sonication is effective for the diagnosis of infection and to compare it with tissue cultures. The files of 58 patients who underwent revision surgery for suspected infected endoprosthetic reconstructions were reviewed. Cultures were performed on 5 tissue samples obtained from each patient and on fluid obtained by sonication of the megaprosthesis. The sensitivity, specificity, and negative and positive predictive values of tissue and sonication fluid cultures were evaluated. Overall, tissue and sonication fluid cultures confirmed an infection in 42 of the 58 patients. In 36 of the 42 infected endoprosthetic reconstructions, tissue and sonication fluid cultures identified the same bacterial isolate. In 5 cases, a bacterial isolate was identified only in sonication fluid cultures, and in 1 case, a bacterial isolate was identified only in tissue cultures. The sensitivity and negative predictive value of sonication fluid cultures were statistically significantly better than those of tissue cultures, while the specificity and positive predictive value were not different between the 2 culture types. Compared with tissue cultures for the diagnosis of infected megaprostheses in patients with tumors, sonication fluid cultures are associated with a better sensitivity and negative predictive value and a similar specificity and positive predictive value. Therefore, sonication should be considered a useful adjunct for the optimal diagnosis and management of these patients.
ASJC Scopus subject areas