A systematic review 1 including 14 studies with a total of 6242 subjects was abstracted in DARE. Signs and symptoms (7 studies). The only symptoms with sensitivity greater than 50% were joint pain (sensitivity 0.85, 95% CI: 0.78, 0.90; 2 studies), a history of joint swelling (sensitivity 0.78, 95% CI: 0.71, 0.85; 2 studies) and fever (sensitivity 0.57, 95% CI: 0.52, 0.62; 7 studies). None of these studies assessed specificity. Serum laboratory values (2 studies). Abnormal peripheral WBC count, erythrocyte sedimentation rate and C-reactive protein had limited accuracy (the positive LRs ranged from 1.3 to 1.6). Synovial WBC counts (5 studies). WBC counts above 100,000/microL showed good accuracy for ruling in septic arthritis but poor accuracy for ruling out the condition: the sensitivity ranged from 13 to 40% and specificity from 93 to 100%. As the threshold to define a positive WBC count decreased, sensitivity increased but at the expense of specificity. Polymorphonuclear cells (4 studies). The presence of 90% or more polymorphonuclear cells showed poor accuracy, both for ruling in and ruling out septic arthritis: the sensitivity ranged from 57 to 92% and specificity from 68 to 83%. Other synovial fluid laboratory test results (3 studies). Low glucose had poor sensitivity (38 to 64%) and moderate specificity (85%). The presence of lactate dehydrogenase in synovial fluid showed excellent sensitivity (100%) but poor specificity (51%).
Comment: The quality of evidence is downgraded by limitations in review quality and limitations in study quality (poorly defined gold standard). The ability of clinical findings, laboratory tests, and their combinations in ruling out septic arthritis was poorly reported.
The following decision support rules contain links to this evidence summary:
Primary/Secondary Keywords