Abstract
Background: The aim of this study was to investigate which markers in serum and in the synovial fluid have the highest diagnostic potential for predicting septic arthritis and periprosthetic joint infections (PJIs). The likelihood ratio assesses the diagnostic utility of a diagnostic test and the likelihood that a patient has a disease. Methods: The levels of inflammatory markers in serum [white blood cells, C-reactive protein (CRPS)] and synovial fluid [synovial fluid white blood cell count (SFWBC), percentage of polymorphonuclear cells (%PMN), lactic acid, lactate dehydrogenase (LDH), glucose, total protein] of patients suffering from septic arthritis (n = 114), PJI (n = 67), non-infectious joint diseases (n = 495) and arthralgia after total joint arthroplasty (n = 43) were determined. The arithmetical means, cut-off values, sensitivities, specificities, positive and negative likelihood ratios (+ LR, –LR), interval likelihood ratios and receiver operating characteristic curves with corresponding area under the curve (AUC) of inflammatory markers were calculated. Results: The parameters with the highest diagnostic potential for differing between septic arthritis and non-infectious arthritis were the SFWBC (AUC = 0.850, cut-off value = 6.7 × 103/μl, sensitivity = 81.8%, specificity = 76.5%,+ LR = 3.41, –LR = 0.24), CRPS (AUC = 0.797), %PMN (AUC = 0.766) and synovial lactate (AUC = 0.760). The highest diagnostic potential for predicting a PJI was shown by LDH (AUC = 0.833) and the SFWBC (AUC = 0.828). Conclusions: The SFWBC, CRPS, %PMN and synovial lactate were the best inflammatory markers in predicting septic arthritis. Synovial lactate levels > 10 mmol/l or an SFWBC > 50 × 103/μl substantially increased disease probability, while SFWBC < 1.0 × 103/μl or CRPS < 0.5 mg/dl diminished the post-test probability of septic arthritis considerably. An SFWBC < 1.1 × 103/μl or a %PMN < 70% made a PJI unlikely, while SFWBC > 20 × 103/μl or %PMN > 86% increased the post-test probability of a PJI. The use of the corresponding interval likelihood ratios could help physicians to estimate the probability of septic arthritis and PJI more accurately.
Declaration of interest: This retrospective study was performed without financial support or funding. The authors have no professional and financial affiliations that may have biased this study.