OBJECTIVES: The cost-effectiveness of dapagliflozin in type 2 diabetes mellitus (T2DM) has been demonstrated using the Cardiff T2DM Model, using UK Prospective Diabetes Study (UKPDS) risk equations (RE). The utility of the UKPDS RE within contemporary analyses has been questioned, as some newer drugs have beneficial effects that are partly independent of changes in classical risk factors. This study assessed the impact upon predicted event incidence, costs and quality-adjusted life years (QALYs) of using UKPDS RE versus RE derived from a recent T2DM cardiovascular outcomes trial.

METHODS: The Cardiff T2DM Model was used to simulate lifetime outcomes for 1,000 patients using the published UKPDS82 RE and contrasted with RE derived from the Dapagliflozin Effect on Cardiovascular Events (DECLARE)-TIMI 58 trial. Baseline characteristics of the simulated cohorts were initialised to the DECLARE-TIMI 58 trial and the model used to evaluate outcomes associated with a 1% (11 mmol/mol) reduction in HbA1c employing UK 2018 costs.

RESULTS: Compared to UKPDS82, the DECLARE-TIMI 58 RE predicted higher incidence of hospitalisation for heart failure (20%) versus congestive heart failure (UKPDS: 7%) and longer life expectancy (+7 years). A 1% reduction in HbA1c was associated with undiscounted gains of 0.167 life years and 0.174 QALYs and £868 cost savings per person, when using UKPDS 82 RE. By comparison, gains of 0.274 life years, 0.293 QALYs and cost savings of £1,298 were estimated using DECLARE-TIMI 58 RE.

CONCLUSIONS: The use of DECLARE-TIMI 58 derived RE resulted in additional cost savings and gains in QALYs related to incidence of predicted vascular events compared to UKPDS82. Economic evaluations utilising UKPDS RE to predict the benefit of HbA1c reductions associated with dapagliflozin may underestimate its cost-effectiveness, due to reductions in heart failure that are potentially independent of HbA1c change.