Synovial fluid leukocytes as diagnostic marker in periprosthetic shoulder infection

Abstract

Background

Limited data exist regarding the diagnostic accuracy of synovial fluid leukocyte count (SFLC) in diagnosing periprosthetic shoulder infection (PSI). The main objective of this study was to determine the diagnostic value of leukocyte count at a common threshold of 3 cells/nL and the optimal cut-off value.

Methods

Patients who underwent shoulder arthroplasty revision surgery and aspiration with SFLC between 2012 and 2023 were retrospectively included. The International Consensus Meeting 2018 definition was used to characterize infection status for SFLC threshold and synovial fluid neutrophil percentage (SFNP). Sensitivity and specificity were presented using cross tabulation. The area under the curve was calculated, and the optimal cut-off was determined using maximized Youden Index.

Results

35 cases with an average age of 71 years (43% male) were included in our study. At a threshold of 3 cells/nL, SFLC showed a sensitivity of 70% and specificity of 83%. The corresponding positive and negative predictive values (PPV and NPV) were 89% and 59%, respectively. We found the optimal cut-off for our cohort at 4.7 cells/nL, increasing specificity to 92% while maintaining sensitivity at 70% (PPV = 94%, NPV = 61%). SFNP at a cut-off of 80% demonstrated 50% sensitivity and 91% specificity, with corresponding PPV and NPV of 92% and 48%, respectively. The optimum threshold for SFNP was 54%, which had a sensitivity of 77% and a specificity of 64%, as well as a PPV of 81%, and NPV of 58%. The area under the curve was 0.72 for SFLC and 0.74 for SFNP. Sonication detected pathogens in 63% of cases, while 57% of all cases showed positive tissue cultures and 43% positive aspirate cultures. Especially, the most frequently found microorganism, Cutibacterium acnes, was detected less often in aspirate culture.

Conclusions

SFLC shows good specificity but moderate sensitivity for diagnosing PSI when using a threshold of 4.7 cells/nL. Therefore, it can serve as a confirmatory test for diagnosing PSI but not for ruling out infection.