Initiation of insulin among veterans with type 2 diabetes and sustained elevation of A1c

Aim: To examine if the relationship between insulin initiation and glycemic control among veterans with poorly controlled type 2 diabetes (T2DM) varies by the number of oral-glucose-lowering-medication classes used prior to insulin initiation. Methods: The cohort consisted of veterans with T2DM with at least two glycosolated hemoglobins (A1c) ≥8.0% within a 12-month period but without prior insulin use. The study period was October 1998 until May 2006. Cox regression analyses were used to assess the predictors of the rate of insulin initiation. Results: Among 40,537 who met the inclusion criteria, 17,519 (43.2%) had insulin initiated over a median follow-up period of 58.6 months. The rate of insulin initiation due to 1% increase in A1c increased by 33.6%, 28.8%, 24.2%, 19.7%, 15.4% for patients exposed to 0, 1, 2, 3, 4 classes of oral-glucose-lowering agents. A higher insulin initiation rate was also associated with younger age, more comorbidities, non-Hispanic white race/ethnicity, obesity, longer diabetes duration, and attending endocrinology clinics. Conclusions: Poor glycemic control is associated with increased rates of insulin initiation. This relationship is attenuated by the number of distinct oral-glucose-lowering-medication classes used prior to insulin initiation.