Eleven studies (nine cohort and two case-control) with a total of 1,768 participants were included in the review.
The quality score and level of evidence associated with each study included in the review was reported in the article; all included studies had quality scores of 50 per cent or more.
The pooled estimates of sensitivity, specificity and DOR (three studies, 731 participants) for Apley's test were: 0.22 (95% CI: 0.17, 0.28) for sensitivity, 0.88 (95% CI: 0.72, 0.96) for specificity and 2.20 (95% CI: 0.27, 17.66) for DOR.
The pooled estimates of sensitivity, specificity and DOR ( eight studies, 1,374 participants) for the joint line tenderness test were 0.76 (95% CI: 0.73, 0.80) for sensitivity, 0.77 (95% CI: 0.64, 0.87) for specificity and 10.98 (95% CI: 3.02, 39.95) for DOR.
The pooled estimates of sensitivity, specificity and DOR (eight studies, 1,475 participants) for McMurray's test were 0.55 (95% CI: 0.50, 0.60) for sensitivity, 0.77 (95% CI: 0.62, 0.87) for specificity and 3.99 (95% CI: 1.04, 15.31) for DOR.
Accuracy data for all tests assessed in individual included studies were reported in full in the article.
Summary ROC curves and pooled estimates of accuracy showed joint line tenderness to be the best test, followed by McMurray's test then Apley's test. Thessaly's test appeared to have the best overall accuracy, but data for this test were derived from a single case-control study.