Objectives: With new classes of T2DM medications offering weight reduction in addition to better glycaemic control, HTA agencies expect fuller accounting of the impact of post-trial weight trajectory assumptions on cost-effectiveness. Four alternative scenarios were examined using the example of the novel SGLT-2 inhibitor canagliflozin (CANA) in the UK treatment setting. MethOds: The importance of alternative assumptions was illustrated using CANA 300mg versus glimepiride (GLIM; 1mg titrated to 6mg or 8mg) in dual therapy with metformin simulated over 40 years using the ECHO-T2DM model loaded with patient characteristics, treatment effects, and adverse event rates from the DIA3009 trial, in which CANA 300mg reduced body weight by-5.7% versus GLIM over 52 weeks. HbA1c was assumed to drift annually by 0.14% for CANA (similar to metformin in ADOPT), 0.24% for GLIM (as sulphonylurea in ADOPT), and 0.15% for rescue therapy with insulin (initiated when HbA1c > 7.5%). Upon treatment discontinuation, four alternative weight-trajectory assumptions were applied: (A) weight change maintained permanently; (B) CANA weight reduction disappears fully at treatment discontinuation, GLIM weight-gain permanent; (C) GLIM weight-gain permanent, forced convergence of weight upon CANA discontinuation; (D) weight changes disappear fully at discontinuation for both treatments. A weight increase was applied when insulin was initiated and proportional weight changes were applied when insulin dose was titrated upwards. Results: CANA 300mg generated more QALYs at modest incremental cost, resulting in ICERs of £2,766 to £4,317 in the scenarios. Maintaining the benefits permanently (A) generated the largest QALY gain (0.243); complete elimination of benefits at discontinuation (D) offered the smallest (0.198). The proportions of incremental QALYs attributable to weight differences were 34.4%, 19.5%, 18.9% and 17.4% for Scenarios A to D, respectively. cOnclusiOns: CANA 300mg was cost-effective in each of four weight scenarios following discontinuation. Further work is required to define the most clinically plausible scenarios. PDB54 The CosT-effeCTiveness of DaPagliflozin (forxiga®) versus a DPP-4 inhiBiTor in The TreaTmenT of TyPe 2 DiaBeTes melliTus (T2Dm) in englanD anD Wales

A441combination with insulin versus insulin alone for patients who are inadequately controlled despite high doses of insulin. Methods: The published and validated CARDIFF diabetes model was used to conduct the analysis. Clinical inputs were derived from a randomized clinical trial comparing dapagliflozin add-on to insulin with insulin regimens. Based on clinical inputs and the United Kingdom Prospective Diabetes Study (UKPDS) equations, the model predicts disease progression and the number of micro-and macro-vascular complications, along with diabetes-specific and all-cause mortality. The perspective of the National Health Service in England and Wales was adopted over a lifetime horizon. Local unit costs and utility data were assigned to the appropriate model parameters to calculate total Quality-Adjusted-Life-Years (QALYs) and total costs. Univariate and probabilistic sensitivity analyses (PSA) were conducted. Results: Compared to insulin, dapagliflozin added to insulin was associated with 0.342 incremental QALYs (95%CI: 0.288; 0.480) at an additional cost of £1,813 (95%CI: £1,165; £2,381), resulting in an incremental cost-effectiveness ratio (ICER) point estimate of £5,295 per QALY gained. The univariate analyses showed that no input parameter change inflated the ICER above £15,000 per QALY. At a willingness-topay threshold of £20,000 per QALY gained, the dapagliflozin treatment strategy was estimated to have a 100% probability of being cost-effective when compared to the insulin treatment strategy. These findings were shown to be robust with all sensitivity analyses. ConClusions: Dapagliflozin was shown to be a cost-effective treatment option in combination with insulin for patients who are inadequately controlled with insulin alone within established UK cost-effectiveness thresholds. PDB59Assessment of the Key Drivers of Cost-effeCtiveness in the eConomiC moDelling of CAnAgliflozin (CAnA) versus glimePiriDe (glim) in the treAtment of tyPe 2 DiABetes mellitus (t2Dm) in the uK setting