Eighteen studies (n=2,706) were included in the review. The included studies were case-control, prospective and retrospective: 14 studies used McMurray’s test, 14 studies used the JLT test, 7 studies used Apley’s Test and 5 studies used other tests.
The studies scored between 8 and 11 for validity using the QUADAS tool. Eight studies scored 10 or more, with six study arms evaluating the McMurray’s and JLT tests, four evaluating Apley’s test, and one evaluating other tests.
The diagnostic accuracy of all three tests was low, with Apley’s test showing the least accuracy.
McMurray’s test (14 studies): the pooled sensitivity was 70.5 (95% CI: 67.4, 73.4) and the pooled specificity 71.1 (95% CI: 69.3, 72.9). The DOR was 4.5 (95% CI: 3.7, 5.4) and the AUC 0.73 (95% CI: 0.71, 0.76).
JLT (14 studies): the pooled sensitivity was 63.3 (95% CI: 60.9, 65.7) and the pooled specificity 77.4 (95% CI: 75.6, 79.1). The DOR was reported as 4.5 (95% CI: 3.8, 5.4) and the AUC as 0.73 (95% CI: 0.71, 0.76).
Apley’s test (7 studies): the pooled sensitivity was 60.7 (95% CI: 55.7, 65.5) and the pooled specificity 70.2 (95% CI: 68.0, 72.4). The DOR was 3.4 (95% CI: 2.6, 4.4) and the AUC 0.69 (95% CI: 0.65, 0.73).
There was significant heterogeneity for the three main tests (p<0.01). Subgroup analyses did not significantly alter the results for the McMurray’s and JLT tests, but did suggest some homogeneity among the 3 studies using Apley’s test.