How long before incretins help in HFpEF?

A July 2026 systematic review and meta-analysis in the Journal of Cardiac Failure found that semaglutide and tirzepatide reduce worsening heart-failure events and cardiovascular death in obese patients with heart failure with mildly reduced or preserved ejection fraction (HFmrEF/HFpEF), but that the benefit does not become statistically sustained until roughly the six-month mark^[1].

For patients and clinicians watching for early signs that a drug is working, that timeline matters as much as the headline hazard ratio.

What the meta-analysis found

The review, authored by Bernardo F. Spiazzi, Carolina P. Zingano, and Amanda R. Vest, pooled four randomized controlled trials with 4,149 participants and a weighted average median follow-up of 2.47 years^[1]. Using iterative landmark analysis, the researchers recalculated hazard ratios day by day, truncating the dataset at each point to identify when drug effects first became statistically significant.

For the composite outcome of worsening heart failure or cardiovascular death, the hazard ratio was 0.59 (95% CI 0.45-0.78; moderate certainty). Statistical significance was first reached at day 126 (4.1 months) and was sustained from day 162 (5.3 months) onward^[1].

For worsening heart-failure events alone, the effect was larger: hazard ratio 0.33 (95% CI 0.20-0.54; high certainty). Sustained significance for that endpoint came somewhat later, after day 183 (6.0 months)^[1].

The underlying trials

The four trials pooled in the meta-analysis are the STEP-HFpEF and STEP-HFpEF DM studies of once-weekly semaglutide 2.4 mg, and SELECT and FLOW (which contributed participants with a history of heart failure). The primary STEP-HFpEF trial, published in the New England Journal of Medicine in September 2023, enrolled 529 patients with ejection fraction above 45% and a body-mass index of 30 or higher^[2]. It reported clinically meaningful improvements in the Kansas City Cardiomyopathy Questionnaire Clinical Summary Score (KCCQ-CSS) and six-minute walk test distance, alongside a reduced rate of worsening heart failure.

The SUMMIT trial of tirzepatide, also published in the New England Journal of Medicine, enrolled 731 patients with ejection fraction at or above 50% and BMI at or above 30, and followed them for a median of 104 weeks^[3]. The primary composite endpoint (cardiovascular death or worsening heart failure) occurred in 9.9% of the tirzepatide group versus 15.3% on placebo (hazard ratio 0.62; 95% CI 0.41-0.95; P=0.026). The KCCQ-CSS score improved by a mean of 19.5 points on tirzepatide versus 12.7 points on placebo at 52 weeks.

Why 'time to benefit' is a separate question

A drug can produce a statistically significant pooled hazard ratio over two years while providing little protection in the first few months. The landmark-analysis method used in the Spiazzi review tests whether the separation between treatment and placebo curves is real at each point in time, not just at the end of follow-up.

From a clinical standpoint, this shapes how both patients and prescribers interpret the early phase of treatment. A lack of evident benefit at three months is not a signal of failure; the data suggest the protection accrues from roughly the five-to-six-month window and holds from there.

The mechanism is not fully settled. Both GLP-1 receptor agonism (semaglutide) and dual GIP/GLP-1 receptor agonism (tirzepatide) reduce body weight and adipose-related inflammation, lower filling pressures in the left ventricle, and may have direct cardiac effects, though the relative contribution of each pathway in HFpEF is still an area of active investigation. Weight reduction alone does not explain the full magnitude of the effect, because benefit emerges before meaningful weight loss would be expected to translate into structural cardiac changes.

What this does not tell us

The meta-analysis does not establish which specific drug, dose, or patient subgroup drives most of the benefit. The four included trials enrolled different populations (some with type-2 diabetes, some without), used different primary endpoints, and were not individually powered for time-to-benefit analyses. Spiazzi and colleagues are clear that this is a study-level analysis; individual participant data would allow more precise subgroup estimates.

The analysis also does not address durability beyond the follow-up windows of the contributing trials, or what happens to HF event rates if treatment is discontinued before six months. That question remains open.

Where this sits in the incretin and heart-failure picture

The STEP-HFpEF and SUMMIT results have already reshaped how cardiology guidelines are approaching GLP-1 and dual agonist use in HFpEF with obesity. The July 2026 meta-analysis adds a clinically practical answer to a question guidelines do not usually address: not whether these agents work, but when a patient should expect to see protection. The answer, at least for the composite endpoint, appears to be around five months rather than immediately. For the full regulatory and prescribing picture on each agent, see the semaglutide overview and the tirzepatide overview on this site.