Aims: The diagnosis of a prosthetic joint infection is difþcult, but crucial for appropriate treatment. Scintigraphy with speciþc markers for infection (labeled white cells or immunoglobulin-G) has been reported as a more reliable diagnostic tool than clinical assessment (fever, þstula), laboratory studies [polynuclear neutrophils blood count (PNC), erythrocyte rate sedimentation (ESR), and C-reactive protein (CRP)], and preoperative aspiration. Methods: In the þrst part of this study, we retrospectively reviewed 230 patients admitted with a suspected prosthetic joint infection and compared the validity of these different diagnostic tools. 209 patients had an infection. Results: Pain, fever, ESR, and PNC are unreliable for identifying occult infection. The presence of a þstula is inconstant, but when present is very reliable to detect infection. Our study revealed sensitivity, speciþcity, positive and negative predictive value as follows: CRP: 97%, 81%, 98%, 71% respectively; aspiration: 82%, 94%, 99%, 43% respectively, and labelled scintigraphy 74%, 76%, 91%, 44% respectively. In the second part, we reviewed 23 articles which included 1,722 prosthetic joints with preoperative evaluation of infection. Conclusions: Both our study and the literature review indicate that CRP and joint aspiration are the best tools to diagnose prosthetic joint infection.

Purpose: The diagnosis of chronic bone and joint infections, particularly in patients with implants, can be a difficult task. Among the clinical and laboratory tests proposed for the diagnosis of infection, 99mTc HMPOA labelled leukocyte scintigraphy is one of the least invasive examinations available. We evaluated its efficacy in terms of reliability.

Material and methods: Ninety patients with suspected bone and joint infections were included in this study: 53% men and 47% women. Mean age was 56.6 years and 80% had osteosynthesis implants. Mean duration of clinical signs before scintigraphy was 6.5 months. The suspected site was the hip in 49%, the knee in 28% and another in 23%. Physical examination (local aspect, temperature) and laboratory tests (differential count, platelets, CRP, ESR) as well as standard radiographs were performed in addition to labelled scintigraphy. These patients were operated and bone samples were taken for bacteriology studies to confirm or infirm the presence of infection. In this series, 73% of the patients were found to have a real infection (73% staphylococcal, 17% multiple germs, 20% other).

Results: The following variables were included in the multivariate analysis: fever, standard radiographs, polynuclear neutrophil count, CRP, ESR, leukocyte-labelled scintigraphy. Sensitivity (Se), specificity (Sp), and odds ratio (OR) were determined. The multivariate analysis showed: fever (Se=0.48; Sp=0.59; OR=1.3); abnormal radiograph (Se=0.71; Sp=0.62; OR=4; p=0.02); polynuclear neutrophil count (OR=1; p=0.19); CRP (OR=1.02; p=0.06); ESR (OR=1.03; p=0.04); leukocyte-labelled scintigraphy (Se=0.71; Sp=0.82; OR=11.6; p< 0.001).

Discussion and conclusion: These findings demonstrate the efficacy of 99mTc HMPOA-labelled leukocyte scintigraphy in terms of reliability for the diagnosis of chronic bone infection compared with other clinical (fever), laboratory (ESR, CRP), and radiographic indicators.