Background: The noninvasive diagnosis of chronic musculoskeletal infections remains a challenge. Recent studies have indicated that fluorine-18 fluorodeoxyglucose-positron emission tomography is a highly accurate imaging technique and is significantly more accurate than the combination of a bone scan and a white blood-cell scan for the diagnosis of chronic infection in the central skeleton (p < 0.05). However, patients who had had surgery within the previous two years were excluded from study. It was our aim to evaluate the technique in an unselected, clinically representative population.

Methods: Sixty patients with a suspected chronic musculoskeletal infection involving the central skeleton (thirty-three patients) or the peripheral skeleton (twenty-seven patients) were studied with fluorine-18 fluorodeoxyglucose-positron emission tomography. Thirty-five patients had had surgery within the previous two years. The fluorine-18 fluorodeoxyglucose-positron emission tomography studies were read in a blinded, independent manner by two experienced readers. The final diagnosis was based on histopathological studies or microbiological culture (eighteen patients) or on clinical findings after at least six months of follow-up (forty-two patients).

Results: On the final composite assessment, twenty-five patients had infection and thirty-five did not. All twenty-five infections were correctly identified by both readers. There were four false-positive findings; in two of these cases, surgery had been performed less than six months prior to the study. The sensitivity, specificity, and accuracy were 100%, 88%, and 93% for the whole group; 100%, 90%, and 94% for the subgroup of patients with a suspected infection of the central skeleton; and 100%, 86%, and 93% for the subgroup of patients with a suspected infection of the peripheral skeleton. Interobserver agreement was excellent (kappa = 0.97).

Conclusions: Fluorine-18 fluorodeoxyglucose-positron emission tomography is highly accurate as a single technique for the evaluation of chronic musculoskeletal infections. It is especially valuable in the evaluation of the central skeleton, where white blood-cell scans are less useful. Because of its simplicity and high degree of accuracy, it has the potential to become a standard technique for the diagnosis of chronic musculoskeletal infections. Further studies are needed to assess its ability to identify infections at the sites of total joint replacements and to distinguish infection from aseptic loosening of these prostheses.