Background/Purpose: Periodontal disease (PD) and RA share the risk factors HLA DR B1 shared epitope (SE) and tobacco exposure (TE). PD may represent a risk factor for subsequent RA. Individuals with multiple ACPA have increased risk of subsequent RA. We developed a prospective cohort with moderate to severe PD but no clinical features of RA, determining: i) the prevalence of HLA-DRB1 alleles; ii) seropositivity to CCP-2 antibodies (CCP-2) and ACPA; iii) whether HLA-DRB1 alleles were associated with CCP-2 or ACPA; and iv)if HLA-DRB1 alleles modified the association between CCP-2 and ACPA.

Methods: CCP-2 and IgM RF were performed by CLIA–certified assays. CCP-2 or RF+ individuals were offered clinical assessment to ensure they were free of RA by 2010 ACR/EULAR criteria. Sera were analyzed by multiplex assay for distinct ACPA. Positive cutoffs for ACPA were established by ROC analysis of prior studies. HLA DRB1 SE alleles were assayed by PCR. Proportions and 95% confidence intervals (CI) were estimated using Fisher’s Exact Test. Mantel-Haenszel odds ratios (OR) were computed as appropriate. Interaction was assessed via the Breslow-Day test.

Results: Median age was 60 years. Key findings;

The most common ACPA was anti-citrullinated histone 2A; n=44(24%), all other ACPA occurred in < 7% of participants. There was a significant relationship between the presence of an indeterminate or low positive CCP-2 and the presence of ≥ 3 distinct ACPA (OR [95%CI] = 23.4[5.5, 100.1], p=0.0001). The presence of DRB1*0401 was marginally associated with an indeterminate or low positive CCP-2 (OR= 3.2[0.8, 12.3], p=0.09), and with the presence of ≥ 3 ACPA (OR= 5. 3[1.6, 17.5, p=0.01). RF was not associated with any specific ACPA, nor was it associated with indeterminate or low positive CCP-2. No relationships between tobacco exposure and any specific ACPA or RF were seen. Among DRB1*0401 negative individuals, prevalence estimates of ≥ 3 ACPA among those with an indeterminate or low positive CCP-2 vs. negative CCP-2 were 17% vs. 4%.This relationship was markedly stronger among DRB1*0401 positive individuals; prevalence estimates of ≥ 3 ACPA among those with an indeterminate or low positive CCP-2 vs. negative CCP-2 were 100% vs. 7% (p for interaction=0.06).

Conclusion: In this PD cohort, indeterminate or low positive CCP-2 is associated with the presence of ≥ 3 distinct ACPA. HLA DRB1*0401 alleles were associated with an increased risk of ≥ 3 ACPA, and with the presence of an indeterminate or higher CCP-2.  These results have implications for the follow-up of individuals with RA risk factors and indeterminate or low positive CCP-2 findings, and for the design of RA prevention trials.