OBJECTIVES To assess the cost-effectiveness of saxagliptin, a dipeptidyl peptidase-4 inhibitor (DPP4i), compared to the glucagon-like peptide-1 (GLP-1) analogues, exenatide twice-daily or lixisenatide, when added to insulin (±metformin) for the treatment of patients with type 2 diabetes mellitus (T2DM) who are inadequately controlled on insulin. METHODS The validated CARDIFF diabetes model was used to conduct the analyses. Clinical inputs for the model were sourced from a systematic review and network meta-analysis (NMA), which found that the treatments were similarly effective at lowering HbA1c and that the GLP-1 analogues resulted in greater and clinically relevant weight loss. A UK healthcare perspective was used for costs, and quality-adjusted life years (QALYs) were calculated from utility data sourced from the published literature. To assess uncertainty, univariate deterministic and multivariate probabilistic sensitivity analyses (PSA) were performed. RESULTS Saxagliptin was found to dominate lixisenatide in the base case, being associated with lower costs (-£472 [95%CI: -1,378, -223]) and slightly more QALYs (0.010 [95%CI: -0.040, 0.076]). Saxagliptin was less costly (-£1,402 [95%CI: -£1,932, -£1,218]) and slightly less effective (-0.012 QALYs [95%CI: -0.054, 0.035]) than exenatide twice-daily. The lower costs associated with saxagliptin were primarily due to the lower drug acquisition costs. The estimated QALY differences were small and associated with the impact of weight reduction on health-related quality of life. Univariate sensitivity analyses demonstrated that the results were most sensitive to varying the HbA1c treatment effect and the effect of saxagliptin on weight change. The PSA estimated that saxagliptin had a 70.4% and 34.3% probability of being dominant, and a 99.4% and 100% probability of being cost-saving compared to lixisenatide and exenatide twice-daily, respectively. CONCLUSIONS Saxagliptin as add-on to insulin was shown to be a cost-saving treatment option from a UK healthcare perspective for patients with T2DM who are inadequately controlled on insulin.