Authors:

B Kartman¹ , P Mcewan² , H Bennett² , C Edmonds³ , I Gause-Nilsson¹ , ¹AstraZeneca - Gothenburg - Sweden , ²Health Economics and Outcomes Research Ltd - Cardiff - United Kingdom of Great Britain & Northern Ireland , ³AstraZeneca - Cambridge - United Kingdom of Great Britain & Northern Ireland ,

Citation:

Background: Type 2 diabetes mellitus (T2DM) and heart failure (HF) represent a considerable burden to patients, healthcare systems and society globally. The Dapagliflozin Effect on Cardiovascular Events (DECLARE)-TIMI 58 trial evaluated the effects of dapagliflozin on cardiovascular (CV) and renal outcomes in patients with T2DM and either established CV disease or multiple risk factors (MRF) for CV disease. The trial demonstrated that treatment with dapagliflozin was associated with a lower rate in the composite of CV death or hospitalisation for heart failure (HHF) compared to placebo (hazard ratio: 0.83; 95% confidence interval: 0.73 to 0.95); a finding driven by a lower rate of HHF in dapagliflozin-treated patients (hazard ratio: 0.73; 95% confidence interval: 0.61 to 0.88).

Purpose: To estimate the direct healthcare costs associated with HHF event rates reported from DECLARE-TIMI 58, from a US payer perspective.

Methods: DECLARE-TIMI 58 event rates were used to predict HHF incidence in a hypothetical cohort of 1,000 people with T2DM, over a 4-year time horizon. A range of published HHF event costs ($26,893–$37,076, 2017 $) were applied to predicted HHF events. Long-term costs associated with managing HF were not considered. Future costs were discounted at 3% per annum. Sensitivity analyses were conducted for pre-defined DECLARE-TIMI 58 subgroups.

Results: Over a 4-year modelled time horizon, 24 HHF events were predicted per 1,000 patients treated with dapagliflozin versus 33 HHF events for placebo (difference: -9 events). When modelled by prior HF status, more events were avoided in those with prior HF due to higher underlying event risk in this subgroup (difference: -33 events per 1,000 patients). Total estimated costs associated with HHF were $630,237–$868,875 per 1,000 patients treated with dapagliflozin versus $860,164–$1,185,864 per 1,000 patients treated with placebo (difference: $229,927–$316,989). For the overall population, a 27% reduction in HHF costs was estimated over 4 years, compared to 21% and 35% reductions estimated in the established CV disease and MRF subgroups, respectively.

Conclusions: The reduction in HHF events demonstrated for dapagliflozin versus placebo in the DECLARE-TIMI 58 trial will translate to savings within healthcare systems in clinical practice. Further to hospitalisation costs estimated in this study, additional savings associated with HF management may be attained in those without prior disease. These short-term cost savings complement other previously shown benefits of dapagliflozin related to improved glycaemic control and weight loss.