Recent reports have described different risks of acute myocardial infarction (AMI) in association with specific oral antidiabetic medications. The present study compared the AMI incidence rates in new users of traditional neutral protamine Hagedorn (NPH) insulin and a long-acting synthetic insulin analog for basal insulin therapy. We retrospectively examined in-patient medical claims for AMI in a cohort of oral agent-treated patients with type 2 diabetes mellitus after the initiation of basal insulin therapy with either NPH (n = 5,461) or insulin glargine (n = 14,730) in a national administrative claims database comprising >30 managed healthcare plans in the United States. Poisson regression and Cox proportional hazards regression models, as well as the propensity score methods, were used to compare the subsequent AMI incidence rates after the initiation of NPH or glargine. At a mean follow-up of 2 years, the unadjusted AMI incidence was 17.6/1,000 person-years after the initiation of NPH versus 11.5/1,000 person-years after initiation of glargine (rate ratio 1.53, 95% confidence interval 1.29 to 1.81). The Cox regression model (hazard ratio 1.39, 95% confidence interval 1.14 to 1.69) and sensitivity analyses (hazard ratio range 1.30 to 1.56) showed a greater risk of AMI in the NPH group than in the glargine group. Propensity matched (1:1) analysis yielded similar results (odds ratio 1.55, 95% confidence interval 1.23 to 1.96 for NPH vs glargine). In conclusion, these results suggest that the initiation of basal insulin therapy with NPH rather than glargine in patients with type 2 diabetes mellitus is associated with a greater risk of AMI.