OBJECTIVES: The NICE reference case for T2DM requires cost-utility analysis with a lifetime horizon. The denominator in cost-utility outcomes is the incremental QALY, which is driven in most T2DM economic models by several key features, including: life extension, micro- and macrovascular events, treatment and treatment-related adverse events (AEs), and excess bodyweight. This analysis aims to determine which factors most impact QALY gains using the example of CANA 300mg versus GLIM (titrated from 1mg to 6mg or 8mg) when combined with metformin. METHODS: The ECHO-T2DM model was used to simulate CANA versus GLIM over 40 years.  ECHO-T2DM was loaded with patient characteristics, treatment effects, and AE rates from the DIA3009 trial.  HbA1c was assumed to drift upwards at an annual rate of 0.14% for CANA (similar to metformin in the ADOPT trial) and 0.24% for GLIM (as sulphonylurea in ADOPT). CANA and GLIM were discontinued and insulin initiated (annual drift 0.15% as in UKPDS) when patients failed to maintain HbA1c under 58 mmol/mol (7.5%). RESULTS:  Hypothetical patients experienced 0.21 more QALYs when treated with CANA 300mg versus GLIM. Because GLIM was associated with greater use of rescue medication (and extra insulin-mediated HbA1c lowering) in the simulation, HbA1c values converged asymptotically limiting the differences in microvascular complications. However, lower blood pressure for patients on CANA versus GLIM was associated with reductions of 2.2% to 4.1% for the rates of macrovascular outcomes (although associated QALY gains were small due to discounting). Differences in weight and especially hypoglycaemic events, related both to GLIM and to earlier initiation of insulin, were associated with improvements in utility (0.04 and 0.16 QALYs, respectively).  CONCLUSIONS: Patients treated with CANA in dual therapy experienced an additional 0.21 QALYs over 40 years versus patients treated with GLIM.  The primary drivers were improved weight while on agent and fewer hypoglycaemic events.