External validation of the alternative Ankylosing Spondylitis Disease Activity Score in three randomized clinical trials of ixekizumab.

Abstract

The alternative ASDAS (altASDAS) is an index that can be used when patient global assessment is unavailable. Our aim was to test the truth and discrimination aspects according to OMERACT filter 2.0 of the altASDAS in an external cohort. Cohorts from the COAST trials of ixekizumab (COAST-V, -W, -X; 16-week primary endpoint) enrolling radiographic/non-radiographic axial SpA patients were pooled. The ASDAS [original formula with patient global assessment (PGA)] and altASDAS were calculated. Truth was assessed by agreement with the continuous ASDAS [intraclass correlation coefficients (ICCs)] and ASDAS disease activity (DA) states (weighted κ), Bland-Altman plots [mean difference (MD) and 95% limits of agreement (LoA)] and Pearson's correlations between altASDAS/ASDAS and other constructs. Discrimination was tested by the ability of altASDAS to distinguish high/low DA according to nocturnal pain >6/10 as an external anchor and agreement (κ) with ASDAS in major improvement (MI) and clinically important improvement (CII). A total of 958 patients were included. For truth, agreement with ASDAS was very good (ICC = 0.99, κ = 0.91), MD with ASDAS was 0.03 (95% LoA -0.31-0.24) and correlation coefficients of altASDAS with related constructs were within a prespecified 0.3-wide band around those between ASDAS and the same construct. For discrimination, the altASDAS discriminated between DA states and agreed with ASDAS response (κ MI = 0.91, CII = 0.93). The altASDAS was truthful and discriminative in an external cohort and as such has been fully validated to be used in cases when PGA is unavailable.