POS0309\u2005SEROPOSITIVITY INCREASES RISK OF INCIDENT DEMENTIA IN INDIVIDUALS WITH RHEUMATOID ARTHRITIS: A POPULATION-BASED COHORT STUDY

Abstract

Several studies show increased risk of dementia among individuals with rheumatoid arthritis (RA), while others show no association. One reason for this discrepancy might be differential association by serostatus. No prior studies have investigated the association between RA and dementia by serostatus.We aimed to evaluate the risk of incident dementia among individuals with RA, stratified by serostatus.This population-based cohort study included all cases of incident RA within Olmsted County, Minnesota with index date of RA onset between 1 January 1999 and 31 December 2013. We matched RA cases to non-RA comparators 1:1 on age and sex. All RA cases met 1987 ACR criteria for RA. We defined seropositivity as positive rheumatoid factor or anti-cyclic citrullinated peptide antibodies. We defined incident dementia as having two ICD-9/10 codes for dementia at least 30 days apart, with the date of the second code representing the time of dementia onset. We excluded individuals with dementia prior to index date. We estimated the cumulative incidence of dementia adjusting for the competing risk of death. For the main analysis, cox proportional hazard models estimated adjusted hazard ratios (aHR) with 95% confidence intervals (CI) for incident dementia, adjusting for age, sex, index year, body mass index, and smoking status (never, former, current). These models compared the incidence of dementia for RA versus non-RA, seropositive and seronegative RA versus their matches, and seropositive versus seronegative RA. To validate these results, we also performed sensitivity analyses using groups matched via inverse probability weighting on age, sex, index year, obesity, smoking status, and race.We identified 597 RA cases (mean age 56, 70% female), and 594 non-RA comparators. Of the RA cases, 388 (65%) were seropositive, and 209 (35%) were seronegative. The ten-year cumulative incidence of dementia in patients with RA was 3.3 (95%CI 2.0,5.5) per 100,000 compared to 2.4 (95%CI 1.3,4.2) in non-RA comparators, for aHR of 1.26 (95%CI 0.7,2.3). When stratifying by serostatus, the ten-year incidence of dementia for seropositive RA cases was 3.6 (95%CI 2.0,6.5), corresponding to aHR of 1.45 (95%CI 0.73,2.9) compared to matched non-RA cases. In contrast, the incidence of dementia in seronegative RA cases was 2.6 (95%CI 1.0,7.0), for an aHR of 1.0 (95%CI 0.29,3.5). Overall, the incidence of dementia in seropositive RA cases was significantly higher than seronegative cases (aHR 3.2, 95%CI 1.2,8.5). Indeed, sensitivity analysis using inverse probability weighting confirmed that among participants aged 50 and older, dementia incidence was higher for seropositive than seronegative RA (aHR 2.9, 95%CI1.1,7.8) (Figure 1).Individuals with seropositive RA have an increased risk for incident dementia than those with seronegative RA. Future studies should replicate these findings and investigate the mechanism for this association.Figure 1.Cumulative incidence of dementia for individuals with seropositive versus seronegative RA aged 50 and older, balanced by inverse probability weightingThis work was funded by grants from the National Institutes of Health, NIAMS (R01 AR46849) and NIA (R01 AG068192, R01 AG034676).Vanessa Kronzer: None declared, Tina Gunderson: None declared, Cynthia S. Crowson: None declared, John M Davis III Grant/research support from: Pfizer, Maria Vassilaki Shareholder of: equity ownership in Abbott Laboratories, Johnson and Johnson, Medronic and Amgen;, Grant/research support from: Roche and Biogen, Michelle Mielke Consultant of: Biogen and Brain Protection Company, Elena Myasoedova: None declared