Abstract
The accuracy of preoperative synovial fluid culture for microbe detection in shoulder periprosthetic joint infection (PJI) is poorly described. To evaluate the utility of preoperative culture data for early pathogen identification for shoulder PJI, we determined the concordance between preoperative synovial fluid culture results and intraoperative tissue culture results.
Fifty patients who met the 2014 Musculoskeletal Infection Society criteria for shoulder PJI between January 2016 and December 2019 were retrospectively reviewed for clinical and demographic data. This cohort of patients was divided into 2 groups based on the concordance between preoperative and intraoperative culture results. The pathogens identified on preoperative and intraoperative cultures were classified as high-virulence or low-virulence. Student’s t tests and Mann-Whitney U tests were used as appropriate for continuous variables, and χ2 and Fisher’s exact tests were used as appropriate for categorical variables.
Concordance between preoperative aspiration and intraoperative tissue culture was identified in 28 of 50 patients (56%). Preoperative cultures positive for Gram-positive species were more likely to be concordant than discordant (P = .015). Preoperative cultures positive for Cutibacterium acnes were more likely to be concordant with intraoperative cultures (P = .022). There were more patients with polymicrobial infection in the discordant group compared with the concordant group (P < .001). No statistically significant correlation between the preoperative serum C-reactive protein level and the intraoperative category of bacteria was reported. Staphylococcus aureus and coagulase-negative Staphylococci were associated with high specificity and negative predictive value. Preoperative cultures positive for C. acnes demonstrated sensitivity, specificity, positive predictive value, and negative predictive value lower than 0.8. Gram-negative pathogens demonstrated the highest sensitivity (1) and specificity (1), whereas polymicrobial infections exhibited the lowest sensitivity and positive predictive value.
Preoperative synovial fluid aspiration for shoulder PJI poorly predicts intraoperative culture results, with a discordance of 44%. More favorable concordance was observed for monomicrobial preoperative cultures, particularly for Gram-negative organisms and methicillin-sensitive S. aureus. The overall high rate of discordance between preoperative and intraoperative culture may prompt surgeons to base medical and surgical management on patient history and other factors and avoid relying solely on preoperative synovial fluid culture data.