OBJECTIVES:  To evaluate real-world outcomes among elderly patients (≥65 years) with type 2 diabetes mellitus (T2DM) initiating insulin treatment with insulin glargine via disposable pen (IG-Pen) or conventional vial/syringe (IG-Vial). METHODS:  The MarketScan® Medicare database was used to identify T2DM patients ≥65 years who initiated insulin treatment with IG-Pen or IG-Vial from January 2007 through June 2009. All patients had continuous health plan coverage for >6 months before (baseline) and >1 year after the index date (follow-up), and were insulin-naïve but had ≥1 oral anti-diabetes drug (OAD) or glucagon-like peptide-1 (GLP-1) analog during the baseline period. Endpoints included 1-year treatment persistence (continuous study drug use without discontinuation) and adherence (adjusted medication possession ratio: aMPR), hypoglycemia-related events, healthcare utilization and costs. Stringent 1:1 propensity score matching was applied to remove observed baseline selection bias between the two cohorts. RESULTS:  A total of 5,860 patients were matched and analyzed (n=2,930 in each cohort; 44% women; median baseline age 74 years, number of OADs 2.1, Charlson comorbidity index 1.29). During the 1-year follow-up, those who initiated with IG-Pen were more persistent (58.2% vs. 50.8%; P<0.0001), and adherent (aMPR 0.69 vs. 0.64; P<0.0001), had lower daily average consumption of insulin (28.6 U/day vs. 32.0 U/day; P=0.0002), were less likely to have hospitalization (all-cause 33.0% vs. 37.5%, P=0.0002; diabetes-related 16.7% vs. 18.8%, P=0.037), and had similar total healthcare costs ($22,265 vs. $21,669; P=0.5085), despite higher diabetes drug costs ($2,166 vs. $1,907; P<0.0001). Hypoglycemia-related event rates were 8.6% with IG-Pen vs. 10.4% with IG-Vial (P=0.0164). CONCLUSIONS:  This real-world study showed that for elderly T2DM patients initiating insulin treatment, using a pen rather than vial/syringe was associated with better treatment persistence and compliance, without increasing healthcare costs during the first year after initiation. These results may assist with clinical decision making and help optimize T2DM management in elderly patients.