BackgroundAnti-citrullinated protein antibodies (ACPA) predict increased disease activity and disability in patients with inflammatory arthritis such as rheumatoid arthritis. However, the absence of these antibodies does not confer universally good prognosis. Recently, a new set of antibodies, anti-carbamylated (anti-CarP) antibodies, have been identified in patients with rheumatoid arthritis. This study aimed to investigate the association between anti-CarP antibodies, disability, and disease activity in these patients. MethodsAdults with two or more swollen joints for at least 4 weeks were recruited from the Norfolk Arthritis Register (NOAR). At baseline patients completed the health assessment questionnaire (HAQ). The Disease Activity Score in 28 joints (DAS28) was calculated and rheumatoid arthritis classification criteria applied. ACPA and anti-CarP antibodies were measured on stored serum samples obtained within the first year of the study. The HAQ was repeated after 1, 2, 3, 5, 7, 10, 12, 15, and 20 years, and DAS28 scores done every 5 years. Generalised estimating equations (GEE) tested the association between anti-CarP antibodies and longitudinal HAQ and DAS28 scores. Findings1995 patients were included; 1310 (66%) were women and median age at onset was 55 years (IQR 43–66). Anti-CarP antibodies were positive in 460 patients (23%), and 1221 (61%) met rheumatoid arthritis classification criteria. Median follow-up was 7 years (IQR 5–11). Patients who were anti-CarP antibody positive had significantly more disability over time and higher levels of disease activity than those who were negative (multivariate GEE adjusted for age, sex, smoking status, ACPA, and year of recruitment to NOAR: β coefficient for HAQ 0·13, 95% CI 0·03–0·23, and for DAS28 0·31, 0·12–0·49). Statistically significant associations were also seen in a subanalysis of 1092 ACPA-negative patients (HAQ 0·15, 0·02–0·29; DAS28 0·37, 0·11–0·63). In ACPA-positive and rheumatoid arthritis subgroups, anti-CarP antibodies were significantly associated with DAS28 (0·30 [0·02–0·57] and 0·21 [0·04–0·37], respectively), and positive associations were also seen with HAQ scores, but these did not meet statistical significance. InterpretationIn this study the presence of anti-CarPA was associated with increased burden of disability as measured by the HAQ and higher disease activity in patients with inflammatory arthritis. Since GEE models include outcome data at all timepoints, these associations are long term. Our results suggest that anti-CarP antibodies might provide additional prognostic information to ACPA and in particular identify ACPA-negative patients with poor prognosis. FundingArthritis Research UK.