Some studies suggest an association between onset and/or poor control of type 2 diabetes mellitus and inhaled corticosteroid (ICS) therapy for chronic obstructive pulmonary disease (COPD), and also between increased fracture risk and ICS therapy; however, study results are contradictory and these associations remain tentative and incompletely characterized. This matched cohort study used two large UK databases (1983-2016) to study patients (≥ 40 years old) initiating ICS or long-acting bronchodilator (LABD) for COPD from 1990-2015 in three study cohorts designed to assess the relation between ICS treatment and (1) diabetes onset (N = 17,970), (2) diabetes progression (N = 804), and (3) osteoporosis onset (N = 19,898). Patients had ≥ 1-year baseline and ≥ 2-year outcome data. Matching was via combined direct matching and propensity scores. Conditional proportional hazards regression, adjusting for residual confounding after matching, was used to compare ICS vs. LABD and to model ICS exposures. Median follow-up was 3.7-5.6 years/treatment group. For patients prescribed ICS, compared with LABD, the risk of diabetes onset was significantly increased (adjusted hazard ratio 1.27; 95% CI, 1.07-1.50), with overall no increase in risk of diabetes progression (adjusted hazard ratio 1.04; 0.87-1.25) or osteoporosis onset (adjusted hazard ratio 1.13; 0.93-1.39). However, the risks of diabetes onset, diabetes progression, and osteoporosis onset were all significantly increased, with evident dose-response relationships for all three outcomes, at mean ICS exposures of 500 µg/day or greater (vs. < 250 µg/day, fluticasone propionate-equivalent). Long-term ICS therapy for COPD at mean daily exposure of ≥ 500 µg is associated with an increased risk of diabetes, diabetes progression, and osteoporosis.