Glucocorticoid dose-dependent risk of type 2 diabetes in six immune-mediated inflammatory diseases: a population-based cohort analysis

INTRODUCTION: In immune-mediated inflammatory diseases, there is a lack of -estimates of glucocorticoid dose–response diabetes risk that consider changes in prescribed dose over time and disease activity. RESEARCH DESIGN AND METHODS: Population-based longitudinal analysis of electronic health records from the UK Clinical Practice Research Datalink, linked to hospital admissions and the mortality registry (1998–2017). We included 100 722 adult patients without diabetes history, diagnosed with giant cell arteritis or polymyalgia rheumatica (n=32 593), inflammatory bowel disease (n=29 272), rheumatoid arthritis (n=28 365), vasculitis (n=6082), or systemic lupus erythematosus (n=4410). We estimated risks and HRs of type 2 diabetes associated with time-variant daily and total cumulative prednisolone-equivalent glucocorticoid dose using Cox regression methods. RESULTS: Average patient age was 58.6 years, 65 469 (65.0%) were women and 8858 (22.6%) had a body mass index (BMI) ≥30 kg/m(2). Overall, 8137 (8.1%) people developed type 2 diabetes after a median follow-up of 4.9 years. At 1 year, the cumulative risk of diabetes increased from 0.9% during periods of non-use to 5.0% when the daily prednisolone-equivalent dose was ≥25.0 mg. We found strong dose-dependent associations for all immune-mediated diseases, BMI levels and underlying disease duration, even after controlling for periods of active systemic inflammation. Adjusted HR for a <5.0 mg daily dose versus non-use was 1.90, 95% CI 1.44 to 2.50; range 1.70 for rheumatoid arthritis to 2.93 for inflammatory bowel disease. CONCLUSIONS: We report dose-dependent risks of type 2 diabetes associated with glucocorticoid use for six common immune-mediated inflammatory diseases. These results underline the need for regular diabetic risk assessment and testing during glucocorticoid therapy in these patients.