ABSTRACT
BACKGROUND: The urinary leukocyte esterase (LE) test strip has been suggested as a good screening test for periprosthetic joint infection (PJI). The purpose of this study is to compare the diagnostic profile of LE assays from different manufacturers and determine whether the LE test strip is a good rule-out test. METHODS: Synovial fluid samples (N = 344), sent to 1 laboratory for PJI testing, were used in this prospective study. Four different tests for synovial fluid LE were simultaneously evaluated for their performance in detecting white blood cell (WBC) positive samples (>3000 cells/µL). RESULTS: Both neutrophil elastase immunoassays demonstrated greater sensitivity than urinary LE test strips (92.0% and 90.8% vs 72.4% and 80.3%; all P < 0.011). Fifty-three percent of false-negative urinary LE test strip results clearly missed the presence of elevated levels of synovial fluid LE. Invalid urinary LE test strip results were 4-fold more likely among WBC (+) compared with WBC (-) samples (27.0% vs 6.8%; P < 0.0001). The combined failure to detect an elevated WBC count, because of either false-negative or invalid results, was 47.1% and 41.4% for the Roche and Siemens test strips, respectively. CONCLUSIONS: This study agrees with the existing literature demonstrating that the LE test strips are among the lowest sensitivity tests for PJI. The urinary LE tests strips should not be used to rule-out PJI, as they often fail to detect abundant levels of LE in synovial fluid. Instead, it is more appropriate to use the (++) LE test strip result as a secondary confirmatory rule-in test for PJI because of its high specificity.
Subject(s)
Arthritis, Infectious/diagnosis , Carboxylic Ester Hydrolases/urine , Prosthesis-Related Infections/diagnosis , Synovial Fluid/chemistry , Arthritis, Infectious/urine , Biomarkers/urine , Humans , Prospective Studies , Prosthesis-Related Infections/urine , Sensitivity and SpecificityABSTRACT
BACKGROUND: The aim of the present study was to evaluate the influence of organism type on the performance of the synovial fluid C-reactive protein (CRP) test. METHODS: We retrospectively reviewed the results of 21,422 synovial fluid samples sent to one common laboratory for the purpose of diagnostic testing for periprosthetic joint infection. Both a synovial fluid CRP result and a positive culture were present for 1789 submitted samples. The cultured organisms were grouped by species, virulence, and gram type; and the median CRP level was determined for each group. RESULTS: The median synovial fluid CRP level was significantly lower for less-virulent organisms, when compared to those organisms classified as virulent (15.10 mg/L vs 32.70 mg/L; P < .0001). Some less-virulent species such as yeast and Staphylococcus epidermidis were associated with a 4-10 times lower CRP response than those of virulent organisms such as Streptococcus agalactiae and Staphylococcus aureus (P < .0001). Bacterial gram type had no influence on the median CRP result. The rate of false-negative CRP values was 50.9% for yeast, 29.4% for S. epidermidis, 28.5% for all less-virulent organisms, and 11.6% for all virulent organisms. CONCLUSION: The CRP response appears to be highly dependent on the infecting organism and is more likely to provide false-negative results in the setting of less-virulent organisms. Although the use of a CRP level is an important part of the workup for periprosthetic joint infection, surgeons must be aware that this protein may yield a false-negative result in the setting of less-virulent organisms.