Delayed Initiation of ARNI and SGLT2i in Heart Failure With Corrected Aetiology (DELAY-HF), Pilot Study

• Scientific Background and Rationale Contemporary heart failure (HF) guidelines recommend the early and simultaneous initiation of the four foundational pillars of guideline-directed medical therapy (GDMT)-ARNI (or ACEi/ARB), beta-blocker, mineralocorticoid receptor antagonist, and SGLT2 inhibitor-in patients with a reduced left ventricular ejection fraction (LVEF). The pivotal trials supporting these recommendations, however, were conducted predominantly in patients with chronic HFrEF in whom the underlying aetiology was either non-correctable or had not been definitively addressed.

A clinically distinct population is encountered when the underlying cause of heart failure is fully reversible: patients with ischaemic cardiomyopathy who undergo complete revascularization, and patients with left-sided valvular disease causing chronic pressure or volume overload (severe aortic stenosis, aortic regurgitation, primary mitral regurgitation) who undergo valvular surgery or transcatheter intervention. In these patients, removal of the haemodynamic insult itself drives a substantial portion of subsequent LVEF recovery, independent of neurohormonal blockade. Consequently, universal and immediate initiation of ARNI and SGLT2i in this population may expose patients who would have recovered spontaneously to drug-related adverse effects (symptomatic hypotension, hyperkalaemia, renal dysfunction, genitourinary infection, ketoacidosis), to the burden of polypharmacy, and to avoidable cost, with uncertain incremental benefit on cardiac function or clinical outcome.

The optimal timing of ARNI and SGLT2i initiation in patients whose underlying cause of HF has been definitively corrected has not been prospectively examined.

• Study Hypothesis The central hypothesis of DELAY-HF is that, in patients in whom the underlying cause of heart failure has been surgically or procedurally corrected, the LVEF recovery attributable to treatment of the underlying cause is substantially greater than the incremental recovery attributable to ARNI and SGLT2i. Under this hypothesis, the 12-month LVEF change in the delayed-initiation arm is expected to be no more than 5 percentage points worse than in the immediate-initiation arm. As a pilot trial, DELAY-HF estimates this difference with a 95% confidence interval and evaluates it against the pre-specified decision criterion; a formal non-inferiority hypothesis test will be performed in a subsequent, separately designed confirmatory trial.
• Pre-specified Analyses

Four pre-specified analyses are planned to fully characterise the contribution of ARNI/SGLT2i in this population:

• Effect of early ARNI/SGLT2i on LVEF recovery. Patients in the delayed-initiation arm who never received ARNI/SGLT2i during the 12-month follow-up (those without recovery at 6 months who were ultimately not started, plus those who recovered spontaneously) will be compared with the immediate-initiation arm to estimate the contribution of early ARNI/SGLT2i to LVEF recovery beyond the recovery driven by treatment of the underlying cause.
• Safety of delayed initiation. Within the delayed-initiation arm, patients who received delayed ARNI/SGLT2i and recovered will be compared with those who received delayed therapy without recovery, to evaluate whether postponing therapy by 6 months in non-recovering patients is associated with an attenuation of subsequent LVEF response.
• Twelve-month head-to-head comparison. The proportion of patients achieving LVEF recovery and the magnitude of LVEF improvement at the 12-month assessment will be compared directly between the delayed-initiation and immediate-initiation arms.
• Spontaneous versus pharmacologic recovery. Among patients in the delayed-initiation arm whose LVEF recovered spontaneously after correction of the underlying cause (without medication), the magnitude of LVEF recovery will be compared with that achieved in patients in the immediate-initiation arm whose recovery occurred while on ARNI/SGLT2i.

• Sample Size Justification Sample Size Justification. The sample size of 80 patients (40 per arm) was not derived from a power calculation for hypothesis testing, in keeping with the exploratory purpose of a pilot trial. It is based on: (1) the recommendation of Whitehead et al. (2016) of 25-40 participants per arm for a pilot trial preceding a main trial with a standardised effect size of 0.5 (decision margin Δ = 5 percentage points, assumed SD ≈ 10 percentage points); (2) consistency with prior randomised trials of GDMT withdrawal in recovered heart failure (ReReRe n = 80, TRED-HF n = 51, CATHEDRAL-HF n = 60); (3) detection of a safety signal-approximately 80% power (Fisher's exact test, α = 0.05) to detect a rescue-therapy rate of 25% in the delayed arm versus 5% in the immediate arm; (4) estimation of feasibility metrics (recruitment, adherence, 12-month follow-up completion) within a 95% confidence interval of approximately ±10 percentage points; and (5) an allowance of 10%

• Exploratory Analyses A pre-specified exploratory aim of the trial is to estimate the proportion of patients in the delayed-initiation arm in whom ARNI and SGLT2i can ultimately be deemed unnecessary-that is, those whose cardiac function has recovered sufficiently at 6 months that initiation is not warranted under the trial protocol. A formal cost-effectiveness analysis (incremental cost-effectiveness ratio) will compare the two strategies from the healthcare-system perspective.

• Expected Implications If the delayed-initiation strategy proves safe and feasible in this pilot, with no clinically important signal of harm on the LVEF decision criterion, the findings will support a subsequent confirmatory trial and, ultimately, a more individualised prescribing approach in patients with a corrected cause of heart failure-reserving ARNI and SGLT2i for those who fail to recover spontaneously-thereby reducing unnecessary drug exposure, adverse-effect burden, and cost in a population that is currently treated uniformly under generalised HFrEF guidelines.

• Safety Monitoring Safety Monitoring. An independent Data and Safety Monitoring Board (DSMB) will perform two interim safety reviews: after the first 40 patients complete 12-month follow-up and after enrolment of 80 patients. Safety data from the intensive 6-month observation period of the delayed-initiation arm (rescue-therapy rate, heart failure hospitalisation, death) will be reviewed separately by aetiology (valvular vs ischaemic). A rescue-therapy rate exceeding 20% in the delayed-initiation arm triggers a DSMB review, and stratum-specific discontinuation of enrolment may be recommended.

Inclusion criteria

• Age ≥ 18 years.
• Preoperative left ventricular ejection fraction (LVEF) ≤ 40% on echocardiography.
• Successful surgical or procedural correction of a correctable underlying cause of heart failure: Valvular heart disease: mitral valve surgery, aortic valve surgery, transcatheter aortic valve implantation (TAVI), or tricuspid valve surgery, OR Ischemic cardiomyopathy: complete revascularization by coronary artery bypass grafting (CABG) or percutaneous coronary intervention (PCI).
• Enrollment between 3 months before and 10 days after the corrective procedure.
• Hemodynamically stable post-operative state with NYHA class I, defined at the time of randomization as: Systolic blood pressure ≥ 90 mmHg sustained for at least 6 hours; No up-titration of intravenous diuretics within the preceding 6 hours; No use of intravenous vasodilators within the preceding 6 hours; No use of intravenous inotropes within the preceding 24 hours; Heart rate 50-110 bpm and no clinical signs of volume overload.
• Patients receiving ARNI or SGLT2 inhibitors prior to enrollment must complete a 1-week washout period before randomization.

Provision of written informed consent.

Exclusion criteria

• Prior history of sustained ventricular tachycardia or ventricular fibrillation.
• Greater-than-moderate paravalvular leak or residual mitral regurgitation after aortic or mitral valve surgery.
• Incomplete revascularization in patients with coronary artery disease (residual significant disease in any major coronary territory: LAD, LCX, or RCA).
• Graft occlusion documented on coronary CT angiography after CABG. Planned pregnancy during the study period.
• Uncontrolled hypertension on medical therapy (systolic blood pressure > 160 mmHg).
• Estimated glomerular filtration rate (eGFR) < 30 mL/min/1.73 m².
• Inability to tolerate ARNI, defined as inability to take sacubitril/valsartan 25 mg twice daily (e.g., due to symptomatic hypotension, history of angioedema, or bilateral renal artery stenosis).
• Inability to tolerate SGLT2 inhibitors (e.g., type 1 diabetes mellitus or history of recurrent diabetic ketoacidosis).
• Any other condition that, in the opinion of the investigator, would interfere with study participation or follow-up.