# Value of Fluorodeoxyglucose Positron Emission Tomography/Computed Tomography in Identifying Osteoarticular Septic Grafts in Suspected Infective Endocarditis: Results from a Large Monocentric Cohort

## TL;DR

This study shows that FDG-PET/CT scans can detect hidden joint or bone infections in nearly 28% of patients suspected of having heart infections, often without symptoms.

## Contribution

The study demonstrates the high prevalence and clinical relevance of osteoarticular septic grafts detected by FDG-PET/CT in infective endocarditis patients.

## Key findings

- FDG-PET/CT identified osteoarticular septic grafts in 27.6% of patients with suspected infective endocarditis.
- 43.8% of detected osteoarticular septic graft locations were asymptomatic.
- Osteoarticular septic grafts were associated with musculoskeletal pain and tricuspid valve involvement.

## Abstract

Background: 18F-fluorodeoxyglucose positron emission tomography–CT (FDG-PET/CT) is useful for identifying infective endocarditis (IE) but also the detection of other concomitant septic foci. Previously, we found that FDG-PET/CT identified an osteoarthritic septic graft (OASG) in 19.1% of IE patients, frequently asymptomatic. These preliminary results encouraged us to extend our analyses to a larger population, including all patients initially explored for suspected IE, to assess the prevalence, characteristics, and OASG locations brought out by FDG-PET/CT and to identify predictive factors. Methods: From a single-center cohort of patients referred for a clinical and/or biological suspicion of IE, we included all patients who underwent FDG-PET/CT, mainly performed to confirm a prosthesis heart valve or a foreign cardiac device infection. We excluded those who did not meet the 2015 modified Duke Criteria and those for whom another infectious diagnosis was finally retained or for whom all bacterial samples were negative. Demographic, clinical, bacteriological, imaging, and therapeutic data were collected. FDG-PET/CT images were retrospectively analyzed by three blinded nuclear medicine specialists to identify OASGs. Results: We identified 72 distinct OASG locations by FDG-PET/CT in 48 of 174 patients (27.6%), mainly located in the spine (21 OASGs in 20 patients); 14 patients (8.0%) had several OASG locations. In total, 43.8% of OASG locations were asymptomatic. In multivariate analysis, the presence of OASGs was associated with musculoskeletal pain (p < 0.001) and tricuspid valve involvement (p = 0.002). Conclusions: FDG-PET/CT is useful for identifying OASGs in patients with suspected IE, especially those with tricuspid IE or musculoskeletal pain. The identification of OASGs could impact antibiotic therapy and would allow adapted orthopedic management to be proposed.