Remnant cholesterol inflammation index (RCII) as a potential biomarker for ACPA-negative rheumatoid arthritis: evidence from dual-cohort analysis and serum lipidomics.

introductionCompared with anti-citrullinated protein autoantibody (ACPA)-positive RA, the etiology of ACPA-negative RA remains unclear, posing significant challenges to diagnosis. Recently, evidence has emerged that ACPA-negative RA is characterized by metabolic reprogramming, as well as dysregulated lipid metabolism. The remnant cholesterol inflammation index (RCII) is a comprehensive index of lipid burden and inflammation; this study evaluated the predictive value of RCII as a novel biomarker for ACPA-negative RA.

methodsIn the Beijing Hospital (BJH) cohort, we first compared RCII across various autoimmune diseases and undifferentiated inflammatory arthritis versus healthy controls (HCs). Subsequently, we assessed the independent association between RCII and ACPA-negative RA in the BJH cohort and UK Biobank (UKB) cohort using multivariable logistic regression, adjusting for confounders including age, sex, BMI, and cardiovascular comorbidities. Diagnostic performance was evaluated via receiver operating characteristic (ROC) analysis, and optimal diagnostic RCII intervals were defined using a stratified 5-fold cross-validation framework. Furthermore, the possible underlying mechanisms were investigated using untargeted serum lipidomics in a matched subset.

resultsIn the BJH cohort, RCII levels were significantly elevated in ACPA-negative RA, ACPA-positive RA, undifferentiated inflammatory arthritis, dermatomyositis, and systemic lupus erythematosus compared with HCs, with the highest levels observed in ACPA-negative RA. Dual-cohort analysis confirmed elevated RCII in ACPA-negative RA, with independent associations (BJH: OR = 26.11; UKB: OR = 5.59) even after adjusting for age, sex, BMI, and cardiovascular disease (CVD). In the ROC analysis, RCII demonstrated relatively high diagnostic accuracy, achieving an area under the curve (AUC) of 0.929 in the BJH cohort and 0.785 in the UKB cohort. Cross-validation established stable diagnostic intervals, effectively minimizing the indeterminate decision zone, or "gray zone" to less than 10% of cases. Lipidomic profiling revealed a possible metabolic signature characterized by the broad suppression of phosphatidylcholines (PCs) and triacylglycerols (TAGs) alongside the specific upregulation of pro-inflammatory ceramides (Cers), preliminarily providing biological plausibility for the lipid-inflammation interplay reflected by RCII.

conclusionRCII served as an integrative biomarker, demonstrating improved diagnostic performance compared with isolated lipid or inflammatory measures in identifying ACPA-negative RA. By reflecting the metabolic-inflammatory burden, the interval-based RCII thresholds established in this study provided a practical tool with the potential to help address the current diagnostic gap in seronegative patients. Key Points • This dual-cohort study shows elevated RCII in ACPA-negative RA (median 3.13 vs 0.32 in BJH; 1.91 vs 0.64 in UKB, both P< 0.001) and independent associations after adjustment for age, sex, BMI, and CVD (OR 26.11 in BJH; OR 5.59 in UKB). • Stratified 5-fold cross-validation with reciprocal validation produced RCII diagnostic intervals (0.211-1.744 from BJH; 0.120-1.873 from UKB) that placed all cases in the positive interval while restricting the grey zone to 8.3-9.8% across cohorts. • Serum lipidomics in matched subsets (n = 25 per group) identified 64 differential lipids (61 downregulated, 3 upregulated), featuring reduced TAGs and PCs along with increased pro-inflammatory ceramides (e.g., Cer d18:1/18:0, FC = 1.91), preliminarily providing biological plausibility for the observed RCII elevations in ACPA-negative RA.