Indium-111 chloride imaging in the detection of infected prostheses.

Thirty-three patients with painful joint prostheses and a suspicion of infection were imaged with [111In]chloride. A final diagnosis was established by culture in 19. Of these, 12 were categorized as true positives and three as true negatives. There were two false-positive studies, occurring in patients with knee prostheses. In both, the culture was obtained by aspiration. Two false negatives were in patients with hip prostheses, one of whom had been on long-term antibiotic suppressive therapy. The sensitivity was 86%, specificity 60%, and accuracy 79%. Seventeen of the proven cases had bone imaging prior to [111In]chloride imaging. All 17 static images were positive and were not helpful in differentiating loosening from infection. Using increased uptake on the blood-pool image as a criteria for infection, the sensitivity was 89%, but the specificity was 0. Adding flow studies made little difference in interpreting the blood-pool images. This study shows that [111In]chloride accurate in evaluating infection in prosthesis than bone imaging.