Background: Adult septic arthritis can be challenging to differentiate from other causes of acute joint pain. The diagnostic accuracy of synovial lactate and polymerase chain reaction (PCR) remains uncertain.
Objective: Our aim was to quantify the diagnostic accuracy of synovial lactate, PCR, and clinical evaluation for adults with possible septic arthritis in the emergency department (ED).
Methods: We report a prospective sampling of ED patients aged ≥ 18 years with knee symptoms concerning for septic arthritis. Clinicians and research assistants independently performed history and physical examination. Serum and synovial laboratory testing was ordered at the discretion of the clinician. We analyzed frozen synovial fluid specimens for l- and d-lactate and PCR. The criterion standard for septic arthritis was bacterial growth on synovial culture and treated by consultants with operative drainage, prolonged antibiotics, or both. Diagnostic accuracy measures included sensitivity, specificity, likelihood ratios, interval likelihood ratios, and receiver operating characteristic area under the curve.
Results: Seventy-one patients were included with septic arthritis prevalence of 7%. No finding on history or physical examination accurately ruled in or ruled out septic arthritis. Synovial l- and d-lactate and PCR were inaccurate for the diagnosis of septic arthritis. Synovial white blood cell count and synovial Gram stain most accurately rule in and rule out septic arthritis.
Conclusions: Septic arthritis prevalence in ED adults is lower than reported previously. History and physical examination, synovial lactate, and PCR are inadequate for the diagnosis of septic arthritis. Synovial white blood cell count and Gram stain are the most accurate tests available for septic arthritis.
Keywords: diagnosis; emergency medicine; infectious disease; laboratory medicine; orthopedic; septic arthritis.
Copyright © 2020 Elsevier Inc. All rights reserved.