Imaging of low-grade bone infection with a technetium-99m labelled monoclonal anti-NCA-90 Fab' fragment in patients with previous joint surgery.

Low-grade bone infection represents a serious clinical problem. Diagnostic options are often insufficient, yet the therapeutic implications of proven disease are important, especially in patients with prosthetic joint replacement. Technetium-99m labelled monoclonal anti-NCA-90 granulocyte antibody Fab' fragment (MN3 Fab') has been shown to be useful in bone and joint infection, but there are no data specifically referring to low-grade bone infection. We therefore analysed 38 scans in 30 consecutive patients (age range, 30-85 years; median age, 62 years) referred for suspected low-grade bone infection. There were 17 patients (21 scans) with total hip arthroplasty (THA), six with total knee arthroplasty (TKA), three who had undergone hip or knee surgery for trauma and five (seven scans) with resected hips and no endoprostheses (Girdlestone situations); one of these five patients had been investigated before with THA in situ and another prior to surgery for low-grade coxitis. There were no patients with rheumatoid arthritis as the underlying disease. Results were verified by means of bacteriological cultures, histopathological findings and/or follow-up and compared with the respective Zimmerli scores, which were used for clinical assessment of inflammatory activity. In one patient, the final diagnosis could not be established. One, 5 and 24 h after intravenous injection of up to 1.1 GBq of MN3 Fab', whole-body and planar scans were performed using a dual-head gamma camera. Scans were analysed visually and semiquantitatively adopting an arbitrary score ranging from 0 to 3. There were 13 true positive, 14 true negative and 10 false positive outcomes, yielding an overall sensitivity of 100%, an overall specificity of 58%, an accuracy of 73% and positive and negative predictive values of 57% and 100%, respectively. In patients with THA or TKA, accuracy was 81% and 80%, respectively, while it dropped to 43% in patients with Girdlestone situations owing to a high proportion of false positive findings (4/7) in this subgroup. Scintigraphic score was 1 in all of the false positive and in 11/13 true positive findings. The two remaining true positive findings displayed scintigraphic scores of 2 and 3, respectively. Scintigraphic and Zimmerli scores were loosely correlated (Spearman rho=0.38, P<0.05). Infection was excluded in 22/24 investigations with Zimmerli scores of <6. In this group, there were 13 scintigraphically true negative, nine false positive outcomes, and just two true positive outcomes. In 11/12 investigations with Zimmerli scores of 6 or 7, infection was verified and scintigraphic outcome was accordingly true positive, while the remaining patient was true negative. In conclusion, MN3 Fab' scintigraphy proved to be highly sensitive but not specific in diagnosing low-grade infections of the hip and knee regions in patients with previous joint surgery. The method seems reliable in excluding but not in proving the presence of infection. MN3 Fab' scintigraphy should not be applied in patients with Girdlestone situations. Assessment of infection using the Zimmerli score was more reliable than MN3 Fab' scintigraphy in this group of patients without rheumatoid arthritis as the underlying disease. Considering results from the literature concerning leucocyte scintigraphy, MN3 Fab' scintigraphy may be clinically useful in evaluating low-grade bone infection in THA and TKA patients with Zimmerli scores above 5 and concomitant rheumatoid arthritis or other inflammatory diseases.