#3807 heart failure in ascend-nd and ascend-d

Abstract Background and Aims Heart failure (HF) is a common cardiovascular (CV) complication in chronic kidney disease (CKD) [1]. Two CV outcome trials (ASCEND-ND [NCT02876835][2] and ASCEND-D [NCT02879305][3]) have provided the principal assessment of CV risk with daprodustat in CKD: both studies showed noninferiority of daprodustat to erythropoiesis-stimulating agent (ESA). Here, we examined the risk of HF in daprodustat-treated patients in the ASCEND-ND and ASCEND-D trials. Method The Phase 3, open-label, randomised, event-driven ASCEND trials investigated oral, once-daily daprodustat vs injectable ESA in patients with anaemia of CKD not on (ASCEND-ND) or on (ASCEND-D) dialysis. Time-to-first adjudicated major adverse CV event (MACE) was the primary CV endpoint. Time-to-first MACE or hospitalisation for heart failure (HHF) was a key secondary endpoint; the composite of time-to-first adjudicated MACE + HHF was assessed as a principal secondary endpoint, and time-to-first HHF alone as an additional secondary endpoint. To better understand HF-specific outcomes, post-hoc analyses accounting for competing risk of death were conducted via the composite endpoint of all-cause mortality (ACM) + HHF – more typically employed to evaluate HF-specific outcomes and to assess “hospitalisation-free survival”. Post-hoc subgroup analyses of patients identified as having a baseline HF history were also conducted. Results Overall, 3872 and 2964 patients were randomised in ASCEND-ND and ASCEND-D, respectively. All key prespecified and post-hoc CV endpoints in ASCEND-ND and ASCEND-D are shown in the Table, and the composite endpoint of ACM + HHF is shown in the Figure. In summary, for components of MACE for daprodustat vs ESA, respectively, ACM events were similar between treatment arms in both studies: ASCEND ND: 11.6% vs 12.2%; ASCEND-D: 14.8% vs 14.6%. For HHF alone in both studies, event rates for daprodustat vs ESA, respectively, were: 7.2% vs 5.9% (hazard ratio [HR] 95% confidence interval [CI] = 1.22 [0.95, 1.56]) in ASCEND ND; and 7.5% vs 6.8% (HR [95% CI] = 1.10 [0.84, 1.45]) in ASCEND-D. In post-hoc analyses using the composite endpoint of ACM + HHF in ASCEND-ND, the event rate for daprodustat was 20.3% (n = 393/1937), and the event rate for ESA was 19.0% (n = 368/1935) (HR [95% CI] = 1.09 [0.94, 1.26]). More HHF events were observed in patients with a history of HF (daprodustat: 20.4% [n = 54/265]; ESA: 13.4% [n = 34/254]). In ASCEND-D, ACM + HHF events were similar for daprodustat (24.4% [n = 363/1487]) and ESA (24.8% [n = 366/1477]; HR [95% CI] = 0.98 [0.85, 1.14]). In this composite endpoint in patients with a history of HF in ASCEND-D, there were also more HHF events in those treated with daprodustat (18% [n = 47/267]) vs ESA (13% [n = 32/254]), but there were fewer deaths with daprodustat (21% [n = 56/267]) than ESA (28% [n = 71/254]). For subgroup analyses of daprodustat vs ESA, the proportion of patients with a history of HF at baseline was greater in ASCEND-D (18.0% vs 17.2%) than ASCEND-ND (13.7% vs 13.1%). Incidence of MACE + HHF was similar between treatment arms in both studies. Conclusion Our findings showed that there was a higher HHF risk in daprodustat-treated patients who had a history of HF compared to ESA-treated patients. While HF risk was similar between treatments in patients on dialysis (ASCEND-D) when combining ACM + HHF in a post-hoc analysis, the risk remained numerically higher for daprodustat vs ESA in patients not on dialysis (ASCEND-ND). The underlying mechanism responsible for these observations and differences is not known, although patients with advanced CKD and comorbid HF who are not on dialysis are recognised to be at underlying risk of HF decompensation.