Shoulder periprosthetic joint infection (PJI) is difficult to diagnose with traditional methods. Leukocyte esterase (LE) has recently proven to be reliable in knee arthroplasty; however, its value in the shoulder has not been explored. We hypothesized that LE would display high sensitivity and specificity in shoulder PJI.Two groups were prospectively evaluated: 45 primary and 40 revision shoulder arthroplasties. Synovial fluid and soft tissue cultures were obtained at surgery. Synovial fluid was evaluated with LE test strips. Any aspiration that contained erythrocytes was centrifuged and retested. Shoulder PJI was defined by modified Musculoskeletal Infection Society (MSIS) criteria.Of 5 primaries with positive tissue cultures (11%), only 1 was positive for LE. Of 16 revisions with positive cultures (40%), 4 had positive LE results. Among all patients with bacterial isolates, 6 aspirates were not interpretable (29%), despite centrifugation. LE had sensitivity of 25% and specificity of 75% to predict positive cultures in revisions. Ten revision patients met modified MSIS criteria for PJI. The sensitivity of LE in these patients was 30%, and the specificity was 67% (positive predictive value, 43%; negative predictive value, 83%). If bloody aspirates were considered positive, LE sensitivity in MSIS PJI increased to 60%, but the positive predictive value fell to 37.5%.LE is an unreliable diagnostic measure in shoulder PJI. The presence of erythrocytes within aspirates further decreased its accuracy. We conclude that LE should not be used for the routine identification of shoulder PJI.