Study medications and variables

Insulin-receiving T2DM patients concomitantly treated with metformin, DPP-4 inhibitors, or SGLT-2 inhibitors, with the therapy duration of more than 30 days each, were assigned to insulin + DPP-4 inhibitor users or insulin + SGLT-2 inhibitor users versus insulin + metformin users (reference) according to oral antidiabetic combination history. To include T2DM patients with progressed or later-stage disease, only those already dependent on insulin therapy in outpatient settings were eligible for study inclusion. Use of other oral antidiabetic therapies were permitted in each comparison group. Prespecified variables included patient demographics (age and sex), and CCI, comorbid conditions (such as diabetes mellitus with complications, chronic lower respiratory disease, and renal disease), comedication patterns (sulfonylurea, thiazolidinedione, meglitinide, α-glucosidase inhibitor, GLP-1 analog, and glucocorticoid). Comedication patterns of corticosteroid were also identified in each patient in consideration of its prescription volume as well as its potential to suppress immunity, thereby increasing infection risks in patients. The index date was defined as the date when oral antidiabetic therapy (metformin, DPP-4 inhibitor, or SGLT-2 inhibitor) was initiated during the study period. The consecutive therapy was defined as when a study antidiabetic was re-prescribed within 1.5 times the days-supply added to the prior prescription date after its end date. Those patients who had been on both comparison drugs for an equivalent duration (the number of prescription days not differing by over 50% of each other’s) were excluded from the study, resulting in two mutually-exclusive comparison pairs.