CAN LEUCOCYTE/BONE MARROW SPECT CT DIAGNOSE DEEP INFECTION OF SHOULDER ARTHROPLASTIES?

Background Periprostetic joint infections (PJI) are often difficult to diagnose, to treat and often leave the patient with severe impaired function. The presence of low virulent bacteria is frequently discovered in apparent aseptic revisions of shoulder arthroplasties and pose a challenge to diagnose preoperatively. Dual Isotope In111 Leucocyte/ Tc99 Bone Marrow SPECT CT scan (L/BMS) is considered the radionuclide gold standard in preoperative diagnosing PJI with reported high specificity and sensitivity in hip and knee arthroplasties. Unfortunately, it is labour-intensive and expensive to perform and documentation using L/BMS on shoulder arthroplasties lack. Aim To investigate if L/BMS succeeds in detecting shoulder PJI compared to tissue cultures obtained perioperatively. Method All patients referred to a highly-specialised shoulder department with a painful or stiff shoulder-arthroplasty were included in the cohort. To diagnose infection as a possible cause of arthroplasty failure a L/BMS was planned for all patients. If the arthroplasty was revised, 5 tissue biopsies were obtained from the most infection-suspicious site during revision. Biopsies were cultured in broth and on plates for 14 days due to the high frequency of low virulent infection in shoulder revisions. Infection was defined as growth of the same bacteria in 3 or more of 5 the biopsies. Results During the observation period 71 patients were referred. Revision surgery was performed in 62% of the patients (44/71) of which 29 also had been examined by L/BMS. A microbiological diagnose was available for all. The most predominant organism isolated was P. Acnes. Two patients both had a positive L/BMS and positive cultures. Negative L/BMS and negative cultures were found in 20 patients. The remaining 7 patients had negative L/BMS, but positive cultures. The two patients with a positive L/BMS both showed overt clinical signs of infection. L/BMS show a sensitivity 0.22 95%CI(0–0.49) and specificity 1.00 95%CI(1.00–1.00) in detecting shoulder PJI. The Positive Predictive Value is 1.00 95%CI(1.00–1.00) and Negative Predictive Value 0.74 95%CI(0.57–0.90). No patients infected with P. Acnes resulted in a positive scintigraphy nor had they preoperative or perioperative signs of infection. Conclusion Only patients with severe infectious symptoms of shoulder PJI resulted in positive L/BMS. Hence, the scan added nothing to the preoperative clinical diagnose. In111 Leucocyte/ Tc99 Bone Marrow SPECT CT scan cannot be recommended as a standard screening procedure when evaluating failed shoulder arthroplasties for possible infection.