OBJECTIVES: To compare the direct diabetes treatment costs (DTC) in type 2 diabetes (T2D) patients treated with insulin Glargine (GLA) to those with NPH insulin-based regimens (NPH) in Germany. METHODS: A cost-minimization analysis from the statutory health insurance (SHI) perspective was conducted. Resource use data were collected within a cross-sectional, retrospective study performed between February and May 2007 in 199 primary care centres. Consecutive T2D patients with SHI status treated with either GLA- or NPH-based regimens for at least 6 months prior to study were enrolled. For costing public price lists were used. DTC were calculated as summarized costs of antidiabetic medications, blood glucose self-monitoring (test strips, lancets), glucagon use and needles for a 6 months period. Sensitivity analyses for cost variables were performed. RESULTS: A total of 1602 (982 GLA and 620 NPH) patients were included. Mean DTC were €658±258 and €685±242 per patient during 6 months in GLA and NPH patients, respectively. NPH was mainly used in a basal-bolus (ICT) (79%) whereas GLA was more frequently prescribed in a basal-oral (BOT) antidiabetic regimen (43%). Higher basal insulin costs for GLA vs. NPH (€194±97 vs. €116±74) during 6 months were compensated by lower costs for co-prescribed short-acting insulins (€96±133 vs. €158±133). Further cost compensations were due to a lower consumption of test strips (375±249 vs. 447±251 units) and needles (159±142 vs. 185±176 units) per 6 month in GLA- vs. NPH-based regimens, respectively. Hypoglycemia with consecutive glucagon use was only reported for NPH (4 patients). Within the sensitivity analyses the DTC for GLA remained lower vs. NPH. CONCLUSIONS: Under real-life conditions direct diabetes treatment costs are similar in GLA and NPH treated T2D patients. As GLA is advantageous for a patient due to reduced hypoglycemic risk, less injection frequency and flexible dosing compared to NPH, GLA can be regarded as first-line insulin approach in BOT and ICT.