Fourteen studies (3,142 participants) were included in the review. Sample size ranged from 37 to 544. The quality of study reporting was poor with only four studies that reported over 50% of the STARD items and only three studies that reported blinding of examiners.
The pooled sensitivity for all studies was 62% (95% CI, 60 to 64; Ι²= 98.2%) and the pooled specificity was 73% (95% CI, 72% to 75%; Ι²= 98.8%). The analysis restricted to studies using an instrument gold standard showed a pooled sensitivity of 74% (95% CI 71 to 76; Ι² = 96.6%) and specificity of 70% (95% CI, 68 to 72; Ι²= 98.7%). The pooled DOR was 17.34 (95% CI, 8.75 to 34.39; Ι²= 88.4%) for all studies and 19.97 (95% CI, 8.11 to 49.18; Ι² not reported) for the ten studies using an instrument gold standard. The area under the curve was 0.88 for the full set of studies and 0.90 for the restricted set.
In the meta-regression, only patient age and the technique for monofilament assessment impacted on results.