Introduction: Liraglutide is a glucagon-like peptide-1 analogue used to treat type 2 diabetes mellitus (T2DM). To date, limited long-term data ([ 2 years) exist comparing real-world diabetes-related effectiveness and costs for liraglutide versus insulin treatment. Methods: This retrospective claims data analysis covered the period from 1 January 2010 to 31 December 2017 and included continuously insured patients with T2DM who initiated insulin or liraglutide and had 3.5 or 5 years' follow-up data, identified using the German AOK PLUS dataset. Propensity score matching (PSM) was used to adjust for patient characteristics. Results: After PSM, there were 825 and 436 patients in the liraglutide and insulin groups at 3.5 and 5 years' follow-up, respectively. Baseline characteristics were similar between compared cohorts. The respective change from baseline to follow-up in mean glycated haemoglobin for liraglutide and insulin patients was-0.88% and-0.81% (p [ 0.100) after 3.5 years and-1.15%/-1.02% (p [ 0.100) after 5 years. Mean respective changes in body mass index (kg/m 2) were-1.21/? 1.14 (p \ 0.001) after 3.5 years and-1.29/? 1.13 after 5 years (p \ 0.001). Liraglutide-versus insulin-treated patients were less likely to have an early T2DMrelated hospitalisation (3.5-year hazard ratio [HR]: 0.414 [95% confidence interval (CI) 0.263-0.651]; 5-year HR: 0.448 [95% CI 0.286-0.701]). At 5 years' follow-up, there was no statistically significant difference in total direct costs between treatment groups (cost ratio: 1.069 [95% CI 0.98-1.13]; p [ 0.100).