Background/Purpose: DMARD treatment for Rheumatoid Arthritis (RA) has been shown to improve the glucose control and reduce the incidence of diabetes, usually assumed to be type 2 (T2D) in adults. While rare in adults, type 1 diabetes (T1D) can improve with DMARD treatment. We sought to differentiate the prevalence of T1D from T2D and examine the association of diabetes type with DMARDs, and other factors in patients with RA.

Methods: Participants in Forward, The National Databank for Rheumatic Diseases, during 1999-2018 who reported having diabetes and completed new questions added in 2017 regarding the type of diabetes, age of diagnosis, and treatments. Participants were characterized at the earliest time point available after diabetes diagnosis (baseline): either at enrollment into Forward or at the time of diagnosis during followup. Differences by diabetes type were assessed by statistical tests (Chi², T-test). Logistic generalized estimating equation (GEE) models were used in both univariate and multivariable manner to investigate associations between diabetes type and demographics, clinical measures, and RA treatment. Best models were found by QIC criterion.

Results: Of the included 700 diabetic RA patients, 8.6% (60) were of T1D, while the remaining 640 were T2D (91.4%). The onset age for T1D was 36 years (SD 22) vs 54 years (SD 13) for T2D. The majority of T1D was diagnosed before RA diagnosis (63.3%) whereas T2D was diagnosed mostly after RA diagnosis (71.6%). Patients with T1D tended to be younger, have higher education, lower BMI, worse HAQ, and more likely to smoke than patients with T2D (Table 1). These factors were significant in longitudinal models, although no RA severity measures were associated with diabetes type. Patients with T1D were slightly more likely to receive nonTNFi bDMARDs compared to patients with T2D although this was attenuated in the multivariable model (Table 2).

Conclusion: Surprisingly, the percentage of patients with RA reporting T1D was about double that expected in the population (~4% in the general population¹). These patients were younger and tended toward greater use of nonTNFi bDMARDs. Future studies should account for type of diabetes due to important differences in impact of auto-immune diseases by treatments.

¹American American Diabetes Association

Table 1 –Patients characteristics by diabetes type at baseline.

Table 2 –Association of T1D with RA disease characteristics, outcomes and DMARDs compared to T2D (OR (95%CI) from logistic GEE models)