Twenty-five studies met the inclusion criteria (2,502 patients with SLE of which 1,317 had lupus nephritis; range 12 to 245). Twenty-two studies compared patients with SLE with and without lupus nephritis and nine compared active and inactive lupus nephritis.
All 25 studies avoided incorporation bias and clinical review bias, 22 studies reported appropriate patient selection, 18 had a representative patient spectrum, 23 studies each used an appropriate reference standard, avoided partial and differential verification bias, reported blinding of interpreters of the reference standard and reported on withdrawals, 10 studies avoided progression bias, five studies reported blinding of interpreters of index test and 19 reported uninterpretable and/or intermediate test results.
Overall, when comparing SLE patients with and without lupus nephritis anti-C1q antibodies had a pooled sensitivity of 0.58 (95% CI 0.56 to 0.61), specificity of 0.75 (95% CI 0.72 to 0.77), LR+ of 2.60 (95% CI 2.06 to 3.28), LR- of 0.51 (95% CI 0.41 to 0.63) and DOR of 6.08 (95% CI 3.91 to 9.47).
When comparing patients with active and inactive lupus nephritis, anti-C1q antibodies had a pooled sensitivity of 0.74 (95% CI 0.68 to 0.79), specificity of 0.77 (95% CI 0.71 to 0.82), LR+ of 2.91 (95% CI 1.83 to 4.65), LR- of 0.33 (95% CI 0.19 to 0.56) and DOR of 10.56 (95% CI 4.56 to 24.46).
Significant heterogeneity was observed for all analyses. Results for the multiple regression analysis were reported. There was no indication of threshold effect. Significant publication bias was detected.