Background/Purpose:   Rheumatoid arthritis (RA) and psoriatic arthritis (PsA) are chronic inflammatory conditions that are associated with increased risk of insulin resistance and cardiovascular disease (CVD). However, the evidence for the association of these conditions with incident diabetes mellitus (DM) has been inconsistent. The purpose of this study was to evaluate this association in a cohort of patients with physician entered diagnosis and laboratory values.

Methods:   We studied a retrospective cohort of patients with RA or PsA from 1/1/2001 to 12/31/2012 with age- (± 2.5 years) and gender-matched non-rheumatic controls (NRC) in a tertiary health center.  Data were extracted from electronic health records. RA and PsA were defined using ICD-9 codes 714.0 and 696.0, respectively, entered twice by a rheumatologist; with a primary care physician within the health system were included. Patients with prevalent DM were excluded. Patients were followed over time for development of DM, defined by ICD-9 250, random blood glucose >200 mg/dl, HgA1c ≥6.5% or use of anti-diabetic medications; a more stringent DM diagnosis, with DM defined as HgA1c ≥6.5%, was also examined.  A Poisson regression model was used to estimate the incidence rate (IR) of DM per 1000 person-years for the RA and PsA groups separately compared to the NRC. Cox proportional hazard regression (HR) models, accounting for the correlation induced by matching , were used to estimate the association on developing incident DM after adjusting for age, gender, race, body mass index ( BMI), hypertension, hyperlipidemia, ESR, number of office visits, number of drug classes, glucocorticoid,  immunosuppressive agent and statin use in the year prior to cohort entry. For the primary outcome, our study had 80% power to detect a 25% difference in the risk for incident diabetes between the RA/PsA and NRC groups.

Results:  1502 RA patients, 359 PsA patients and 7887 NRC were included.  Overall, the mean age of the cohort was 55 years; 70% were female, and 98% were Caucasian. RA patients on average were older, more likely to be female, taking steroids and immunosuppressive drugs, and had higher ESR and lower median BMI than PsA patients.  Statin use was not different across all groups. During observation, there were 1195 incident diabetes cases: 199 in the RA, 47 in the PSA and 859 in the NRC, with IR 23.2, 23.0 and 17.7 respectively, with an incidence rate ratio (IRR) of 1.32 (1.13-1.54) and 1.30 (0.97-1.75) for RA and PsA respectively compared to NRC. The adjusted HR (95% CI) for DM was 1.14 (0.94-1.38) and 1.28 (0.94-1.75) for RA and PsA respectively compared to NRC. Defining DM as HgA1c ≥ 6.5%, there were 708 incident cases: 122 in the RA, 25 in the PsA and 561 in the NRC, with IR 13.8, 11.8 and 11.2 respectively and IRR 1.05 (0.70-1.57) and 1.22 (1.01-1.49) for RA and PsA respectively compared to NRC. The HR (95% CI) was 1.06 (0.87-1.30) and 1.12 (0.82-1.52) for RA and PsA respectively compared to NRC.

Conclusion: In this cohort, RA and PsA were not associated with increased risk of incident DM, defined by both clinical and stricter laboratory criteria.  Our results suggest that an excessive burden of DM does not account for the high risk of CVD in these patients.