|Abstract / Summary|
Infection after any joint replacement can cause detrimental effects to function. Diagnose is often straightforward if the infection is acute, but a clear-cut diagnosis of chronic periprosthetic joint infection (PJI) can be a challenge. Further, up to 30% of aseptic revised shoulder replacements exhibit positive cultures. The interpretation and impact of such Unexpected Positive Cultures (UPC) are still debated.
To extend the knowledge of diagnosis of PJI and the impact of UPC on outcome, this thesis evaluated the diagnostic accuracy of 111Indium-labelled White blood cells/99Technesium labelled bone marrow SPECT CT (Study I) and 18F-FDG PET/CT (Study II) to diagnose chronic PJI. Furthermore, patient reported outcome after revisions with emergence of UPC was assessed (Study III). All three studies were based on a prospective cohort of patients referred with a failed shoulder replacement from April 1st, 2014 to September 30th, 2017. All patients were followed at least two years after revision.
During a one-year period 28 patients were included in Study I of which 11 were infected. The scans were true positive in two cases and false positive in zero cases, false negative in 9 cases and true negative in 17 cases. Sensitivity was 0.18 (95% CI: 0.00-0.41), specificity 1.00 (95% CI: 1.00-1.00), PPV 1.00 (95% CI: 1.00-1.00) and NPV 0.67 (95% CI: 0.49-0.84). Despite the study was prematurely ceased due to reduced scanning capacity, results showed a clear trend of low diagnostic performance.
In Study II 86 patients with FDG-PET prior to revision surgery were included. Three scans were true positive, 6 false positive, 51 true negative and 19 false negative; corresponding to a sensitivity of 0.14 (95% CI: 0.03–0.35), specificity 0.91 (95% CI: 0.81–0.97), PPV 0.38 (95% CI: 0.15–0.70) and NPV 0.71 (95% CI: 0.67–0.75). To increase the homogeneity of the FDG-PET diagnoses, a criterion of three distinct patterns of tracer uptake each defining infection were used. Despite these well-defined criteria, the interobserver agreement between the three reviewers was only moderate.
Study III compared outcome described by Oxford Shoulder Scores (OSS) of 126 patients who underwent a standard revision. UPC was found in 28 patients and all cases were due to low-virulent bacteria. At baseline an identical OSS score of 20 (95% CI: 18-22) in the culture-negative group and UPC group 20 (95% CI: 17-23) was found. The 2-year follow-up did also reveal nearly identical scores; 33 (95% CI: 31-36) in the culture-negative and 34 (95% CI: 29-38) in the UPC group. Regardless of culture result a clinically relevant increase in OSS score and decrease in pain were seen. Similarly, most patients experienced a gain in forward elevation. In contrast, the external rotation was unchanged in all groups.
In conclusion, neither WBC/BM SPECT CT nor FDG-PET performs adequately to detect PJI to justify a routine use in the preoperative workup. If no preoperative suspicion of infection exists most patients gain from the revision. If UPC emerges with bacteria of low virulence outcome of the procedure does not seem to be affected.