Anti-Carbamylated Protein Antibody Identifies Seronegative Early Rheumatoid Arthritis Missed by Anti-CCP, Independently Predicts Joint Deformity, and Extends ACR/EULAR Classification: A Dose-Response and Bootstrap-Validated Cross-Sectional Study from North India

Background Anti-cyclic citrullinated peptide (anti-CCP) antibodies miss 30–50% of early rheumatoid arthritis (RA) presentations, delaying treatment and permitting structural damage. Anti-carbamylated protein (anti-CarP) antibodies target homocitrullinated proteins via an HLA-DRB1 shared-epitope-independent pathway biochemically distinct from anti-CCP, making them candidates to fill this diagnostic gap. Prospective Indian data evaluating this hypothesis are absent. This study determined whether anti-CarP identifies anti-CCP-seronegative early RA, independently predicts joint deformity, and extends 2010 ACR/EULAR classification, with comprehensive dose-response characterisation in a North Indian cohort. Methods Open-label, cross-sectional, comparative observational study at IMS-BHU, Varanasi enrolling 200 female RA patients: 115 Early RA (duration < 12 months; anti-CCP−/RF−) and 85 Established RA (> 12 months; anti-CCP+/RF+). Anti-CCP and RF served exclusively as group-defining criteria to prevent circular analysis. Anti-CarP IgG was measured by CE-marked ELISA (DKO149; positivity ≥ 10 AU/mL). Primary analyses included restricted cubic spline (RCS) dose-response modelling, bootstrap-validated multivariable logistic regression for joint deformity (B = 1000 resamples), and 2010 ACR/EULAR serology substitution analysis. Secondary analyses comprised Spearman correlations with disease activity and structural damage, six-antibody concordance-discordance phenotype mapping, and HLA-DRB1 shared-epitope independence testing. Results Anti-CarP identified 54/115 anti-CCP-negative patients as seropositive (47.0%; 95% CI 38.1–56.0%), a 47% relative diagnostic gain substantially exceeding the 16–30% reported from European cohorts. Substituting anti-CarP for anti-CCP in the ACR/EULAR serology domain extended Early RA classification from 56.5% to 70.4% (+ 13.9%; cohort-wide 75.0%→83.0%; κ = 0.761). Anti-CarP independently predicted joint deformity on multivariable analysis (adjusted OR 3.95, 95% CI 1.35–11.54; p = 0.012; bootstrap-corrected C-index 0.945; calibration slope 0.847). RCS modelling confirmed significant non-linearity with HAQ-DI, DAS28-ESR, Sharp/van der Heijde score, and joint deformity (all non-linearity p ≤ 0.049), supporting a data-driven threshold of 23.05 AU/mL over binary positivity. HLA-DRB1 shared epitope was not associated with anti-CarP titres (p = 0.20), confirming independence from the citrulline axis. Only 11.0% of patients remained truly seronegative across a six-antibody panel. Conclusions Anti-CarP identifies seronegative early RA missed by anti-CCP, independently predicts joint deformity with a fourfold odds ratio supported by rigorous bootstrap validation, and meaningfully extends 2010 ACR/EULAR classification in a North Indian cohort. Dose-response characterisation establishes 23.05 AU/mL as the clinically optimal threshold beyond conventional binary positivity.