Patients with rheumatoid arthritis (RA) receiving treatment with biologic/targeted synthetic disease-modifying antirheumatic drugs (b/tsDMARDs) have a lower risk for incident dementia compared with those receiving conventional synthetic DMARDs (csDMARDs) alone, according to study results published study Seminars in Arthritis & Rheumatism.
Researchers conducted a retrospective cohort study using data from the Centers for Medicare & Medicaid Claims (CMS) national claims database collected between 2006 and 2017. Patients with RA were identified as individuals aged at least 40 years with 2 RA diagnoses by a rheumatologist diagnosed between 7 days and 365 days apart. All patients with a previous diagnosis of dementia were excluded from the study.
Person-time was defined as either (1) b/tsDMARD-exposed, which included tumor necrosis factor inhibitors (TNFis) bDMARDs, nonTNFi-bDMARDs, or tsDMARDs with or without csDMARDs, or (2) csDMARD-exposed, which included any csDMARD without a b/tsDMARD. All participants could contribute their time to different exposure groups if they switched medications. Incident dementia was defined as either (1) 1 inpatient or 2 outpatient ICD-9-CM or ICD-10 claims for dementia, or (2) prescription of a dementia-specific medication (ie, rivastigmine, memantine, donepezil, galantamine, or tacrine).
A total of 141,326 eligible patients with RA were identified. The median participant age at cs/b/tsDMARD initiation was 67 years. Overall, 80.0% of the patients were women and 75.3% were White. The median patient follow-up was 3.1 years (range, 1.4-5.6 years). Participants who initiated csDMARDs were older and more likely to be men. Additionally, patients who initiated b/ts/DMARDs were more likely to have dual eligibility status.
During the study period, a total of 3794 events of incident dementia were reported. In the csDMARD exposure group, 1711 events were identified over 85,268.0 person-years (PYs), which were associated with an age-adjusted incident rate (IR) of 20.1 (95% CI, 19.1-21.0) per 1000 PYs.
A total of 2083 events were reported over 159,809.3 PYs in the b/ts/DMARDs exposure group, yielding an age-adjusted IR of 17.0 (95% CI, 16.2-17.7) per 100 PYs. Following multivariable adjustment, patients with RA receiving b/tsDMARDs had a 19% lower risk for incident dementia (hazard ratio [HR], 0.81; 95% CI, 0.76-0.87) than those taking csDMARDs. A similar direction and magnitude of the decreased risk was seen in the competing risk model (HR, 0.85; 95% CI, 0.79-0.91) and when stratifying according to age groups. Subgroup analysis showed comparable risk reductions among TNFis, nonTNFis, and tsDMARDs.
There are several limitations to the study. Because the data were dependent on coding, there exists the potential risk for misclassification. As with other claims-based studies on dementia, the researchers could not identify different subtypes of dementia. Additionally, because inflammation appears to play a role in other forms of dementia, this renders coding for specific forms of dementia less relevant.
The study authors conclude that since no differences were observed between different classes of b/tsDMARDs, this suggests that decreased risk for dementia is possibly explained by an overall decrease in inflammation, rather than by a specific mechanism of action. They note, “Our results provide data to support another benefit of the treatment of systemic inflammation in patients with RA.”
Disclosure: Some of the study authors have declared affiliations with biotech, pharmaceutical, and/or device companies. Please see the original reference for a full list of authors’ disclosures.
Sattui SE, Navarro-Millan I, Xie F, Rajan M, Yun H, Curtis JR. Incidence of dementia in patients with rheumatoid arthritis and association with disease modifying anti-rheumatic drugs – analysis of a national claims database. Semin Arthritis Rheum. Published online August 17, 2022. doi:10.1016/j.semarthrit.2022.152083