Background: In a recent analysis of European electronic medical records, prior methotrexate use was associated with a lower risk of dementia (Alzheimer’s disease [AD] or vascular) among patients with rheumatoid arthritis (RA). Variations in prescribing practices may exist outside of Europe, therefore, it is important to understand the generalizability of this relationship among other healthcare systems as well as dementia subtypes, including AD alone.
Objectives: To understand the risk of AD associated with methotrexate use among commercially insured patients in the U.S.
Methods: Patients ≥ 50 years of age with RA were identified using Optum's de-identified Integrated Electronic Health Records and Claims database 01/01/2007 – 03/31/2022. Cases had evidence of AD after the observed RA diagnosis (first observed AD diagnosis deemed index). At least 12 months of continuous enrollment from RA diagnosis to index was required. Up to three controls with no evidence of AD were matched to cases based on age, sex, and the number of months from RA diagnosis to index date (a random index date with at least 12 months of continuous enrollment from date of RA diagnosis was chosen for controls). Exposure to methotrexate, as well as the time on treatment, was evaluated from the time of RA diagnosis to index. Conditional logistic regression models were fit to estimate the risk of AD, controlling for age, history of stroke, myocardial infarction, BMI, and either methotrexate treatment (binary variable), or duration of methotrexate treatment (continuous variable).
Results: 494 AD cases were matched to 1,393 controls. The average time from RA diagnosis to AD was 7.3 years (SD ± 3.8 years). Methotrexate use was observed in 17% of each group, with a mean time on treatment of 2.8 years (SD ± 3.3 years) and median of 1.4 years (IQR 30 days – 4.7 years). Both prior methotrexate exposure (binary variable, OR 0.88, 95% CI 0.66-1.18) and years of methotrexate treatment (continuous variable, OR 0.98, 95% CI 0.91-1.04) were not significantly associated with a lower risk of AD. Results were similar in separate regressions that evaluated the risk of AD among treatment duration subgroups relative to patients with no methotrexate exposure (methotrexate exposure < 1.4 years: OR 0.81, 95% CI 0.55-1.19; methotrexate exposure ≥1.4 years: OR 0.87, 95% CI 0.62-1.22).
Conclusions: Methotrexate use among a U.S. population with RA was not associated with a lower risk of AD and limited to no effect of treatment duration was observed. This analysis highlights the need for further research to clarify the relationship between antirheumatic drugs and risk of dementia in real-world data.
