A Bidirectional C1q/TNF-Related Protein Signature Is Associated with Disease Activity in Behçet's Disease: A Cross-Sectional Study.

Background/

Objectives

Distinguishing active from quiescent Behçet's disease (BD) is difficult because conventional acute-phase reactants perform inconsistently. C1q/TNF-related proteins (CTRPs) are adiponectin paralogues; several members have been reported to exert opposing immunometabolic actions in experimental systems, although the evidence is uneven across the family and no member has been examined in BD.

Methods

In this cross-sectional study, eight CTRPs (CTRP1, CTRP4, CTRP6, CTRP9, CTRP12, CTRP13, CTRP14, CTRP15) were measured in serum using enzyme-linked immunosorbent assay in 30 patients with active BD, 30 with inactive BD and 30 healthy controls. Disease activity was defined by a Behçet's Disease Current Activity Form (BDCAF) score of ≥2. False discovery rate correction was applied separately to the eight omnibus and eight active-versus-inactive biomarker comparisons. Logistic models were adjusted for age, sex, body mass index and disease duration; a composite index combined the analytes changing in opposite directions.

Results

Six analytes differed between active and inactive disease after correction (q = 0.004-0.041): CTRP1, CTRP4, CTRP6 and CTRP14 were higher and CTRP12 and CTRP15 lower in active disease, with five correlating with the BDCAF score in the same directions. Four showed nominally significant associations in the adjusted models (CTRP12, CTRP14, CTRP6 and CTRP15), whereas CTRP1 and CTRP4 did not; adjusted odds ratios per doubling ranged from 0.27 (CTRP12, 95% CI 0.11-0.71) to 2.73 (CTRP14, 1.16-6.44). The composite index achieved an area under the curve of 0.866 (95% CI 0.765-0.944); C-reactive protein alone reached 0.810 (0.682-0.915) and the two combined 0.916 (0.839-0.974). Bootstrap internal validation, in which the selection of components was repeated in each resample, gave an out-of-bag area under the curve of 0.812 (0.649-0.946). The index remained associated with activity after adjustment for C-reactive protein (p = 0.002); adding the index to C-reactive protein did not significantly increase discrimination (ΔAUC +0.106, 95% CI -0.002 to +0.214; DeLong p = 0.055).

Conclusions

In this cross-sectional study, active BD was associated with a coordinated bidirectional CTRP pattern that paralleled clinical activity. The composite index remained associated with activity after adjustment for C-reactive protein, but its incremental discrimination over C-reactive protein did not reach statistical significance; longitudinal and external validation are required.