OBJECTIVES: Compare costs among patients with type 2 diabetes (T2D) treated with exenatide or insulin glargine.  These are injectable agents typically used after failure on oral antidiabetic agent(s) METHODS: Data from September 2004 to September 2007 were obtained from a large retrospective claims database.  Intent-to-treat cohorts of insulin-naïve adults diagnosed with T2D who initiated therapy on either exenatide (N=4090) or insulin glargine (N=1660).  Individuals were not allowed to use the other medication or other insulin in the one-year follow-up period.  Annual total medical costs and total diabetes related medical costs were estimated using stepwise multivariate regressions.  Major cost components were also examined using either stepwise multivariate regressions or a two-part model that controlled for the probability of using the service. Smearing estimates were used to transform estimated log costs into costs.  The analyses controlled for the potential impact of patient demographics, general health, prior resource use, comorbidities, and timing of treatment initiation. RESULTS: Initiation with exenatide compared to insulin glargine, was associated with significantly lower total direct medical costs ($19,293 vs $23,782, p<0.001) and total diabetes-related medical costs ($7,833 vs $8,536, p<0.0001).  Initiation of therapy with exenatide compared to insulin glargine was also associated with significantly lower inpatient ($4,212 vs $7,532, p<0.0001), outpatient ($9,501 vs $12,885, p<0.0001), and emergency room costs ($82 vs $131, p<0.0001) and significantly higher drug costs ($6,885 vs $5,936, p<0.0001).  Similarly, the use of exenatide compared to insulin glargine was also associated with significantly lower diabetes-related inpatient ($2172 vs $3538, p<0.0001) and outpatient costs ($2739 vs $3249, p<0.0001) and significantly higher diabetes-related drug costs ($3160 vs $2424, p<0.0001) CONCLUSIONS: Use of exenatide, compared to insulin glargine, was found to be associated with significantly lower annual total direct medical costs and total diabetes related medical costs even though diabetes related and total drug costs were higher.