Evaluation of imaging in loosening of hip arthroplasty in 326 consecutive cases.

Between 1989 and 1995, 326 hip arthroplasties were surgically revised in our clinic because of loosening. For evaluation of loosening, we used history, clinical examination, radiographs, bone scans and arthrography with digital subtraction technique. In 130 cases (40%), surgery was planned by plain radiographs. In 116 cases (89%) a correct diagnosis of loosened components was made. Of the 72 patients (22%) with an additional bone scan, 87% (63 patients) had a correct diagnosis. Fifty-four patients (17%) with different findings were examined by hip arthrography; in 83% (45 patients) the diagnosis was confirmed intraoperatively. In 93% (65 patients) of 70 patients (21%) with bone scans and arthrography, surgery confirmed the diagnosis. Sensitivity for loosening of the socket was 61% by bone scan and 85% by digital subtraction arthrography; specificity was 75% in each case. Sensitivity for loosening of the stem was 71% by bone scan and 65% by digital subtraction arthrography; specificity was 80% and 88%, respectively. Our results showed that in most cases of suspected loosening of total hip arthroplasties, analysis of plain radiograph series can lead to an accurate diagnosis in 85%-90%. In cases of uncertainty, we suggest the additional use of digital subtraction arthrography, which has a similarly high diagnostic quality. Because of the low sensitivity of nuclide bone scanning (about 60%) and its dependence on prosthetic material and fixation, we do not recommend primary scintigraphy. It can be helpful in problematic cases with additional questions like activity of ectopic bone formation.