BACKGROUND: Type 2 diabetes patients often initiate treatment with a sulfonylurea and subsequently intensify their therapy with insulin. However, information on optimal treatment regimens for these patients is limited. OBJECTIVE: To compare risk of cardiovascular disease (CVD) and hypoglycemia between sulfonylurea initiators who switch to or add insulin. DESIGN: This was a retrospective cohort assembled using national Veterans Health Administration (VHA), Medicare, and National Death Index databases. PARTICIPANTS: Veterans who initiated diabetes treatment with a sulfonylurea between 2001 and 2008 and intensified their regimen with insulin were followed through 2011.
MAIN MEASURES:The association between insulin versus sulfonylurea + insulin and time to CVD or hypoglycemia were evaluated using Cox proportional hazard models in a 1:1 propensity score-matched cohort. CVD included hospitalization for acute myocardial infarction or stroke, or cardiovascular mortality. Hypoglycemia included hospitalizations or emergency visits for hypoglycemia, or outpatient blood glucose measurements <60 mg/dL. Subgroups included age < 65 and ≥ 65 years and estimated glomerular filtration rate ≥ 60 and < 60 ml/min. KEY FINDINGS: There were 1646 and 3728 sulfonylurea monotherapy initiators who switched to insulin monotherapy or added insulin, respectively. The 1596 propensity score-matched patients in each group had similar baseline characteristics at insulin initiation. The rate of CVD per 1000 person-years among insulin versus sulfonylurea + insulin users were 49.3 and 56.0, respectively [hazard ratio (HR) 0.85, 95 % confidence interval (CI) 0.64, 1.12]. Rates of first and recurrent hypoglycemia events per 1000 person-years were 74.0 and 100.0 among insulin users compared to 78.9 and 116.8 among sulfonylurea plus insulin users, yielding HR (95 % CI) of 0.94 (0.76, 1.16) and 0.87 (0.69, 1.10), respectively. Subgroup analysis results were consistent with the main findings. CONCLUSIONS: Compared to sulfonylurea users who added insulin, those who switched to insulin alone had numerically lower CVD and hypoglycemia events, but these differences in risk were not statistically significant.