Semaglutide in Obesity-Related HFpEF and Type 2 Diabetes (STEP-HFpEF DM)

Authors, Journal, Affiliations, Type, DOI

• Authors: Mikhail N. Kosiborod, Mark C. Petrie, Barry A. Borlaug, Javed Butler, Melanie J. Davies, G. Kees Hovingh, Dalane W. Kitzman, Daniel V. Møller, Morten B. Treppendahl, Subodh Verma, Thomas J. Jensen, Kristian Liisberg, Marie L. Lindegaard, Walter Abhayaratna, Fozia Z. Ahmed, Tuvia Ben-Gal, Vijay Chopra, Justin A. Ezekowitz, Michael Fu, Hiroshi Ito, Małgorzata Lelonek, Vojtěch Melenovský, Béla Merkely, Julio Núñez, Eduardo Perna, Morten Schou, Michele Senni, Kavita Sharma, Peter van der Meer, Dirk Von Lewinski, Dennis Wolf, Sanjiv J. Shah
• Journal: New England Journal of Medicine 2024;390(15):1394–1407 (published online April 6, 2024)
• Affiliations: Saint Luke's Mid America Heart Institute, Kansas City; University of Glasgow; Mayo Clinic; and 108 sites in 16 countries
• Type: Double-blind, randomized, placebo-controlled, event-driven-style multicentre trial (52 weeks)
• DOI: 10.1056/NEJMoa2313917

Overview

STEP-HFpEF DM extended the STEP-HFpEF findings to patients with obesity-related HFpEF and type 2 diabetes — a group with a more advanced phenotype, greater symptom burden and smaller weight loss on anti-obesity therapy. In 616 patients, semaglutide 2.4 mg weekly improved the KCCQ clinical summary score by +7.3 points more than placebo (13.7 vs 6.4) and reduced body weight by −6.4 percentage points more (−9.8% vs −3.4%) at 52 weeks, with consistent benefits in 6-minute walk distance, a hierarchical composite and CRP. Despite weight loss roughly 40% smaller than in patients without diabetes, symptom benefits were similar to STEP-HFpEF, suggesting mechanisms beyond weight loss. Serious adverse events were fewer with semaglutide.

Keywords

Semaglutide; GLP-1 receptor agonist; Heart failure with preserved ejection fraction; Obesity; Type 2 diabetes; KCCQ; C-reactive protein

Key Takeaways

Background

• HFpEF is now the predominant form of heart failure, partly because of rising obesity prevalence; excess adiposity and insulin resistance contribute to its development and progression.
• Type 2 diabetes is highly prevalent in HFpEF and marks a more severe phenotype — greater myocardial, microvascular, mitochondrial and skeletal-muscle dysfunction, more inflammation and insulin resistance, worse symptoms, exercise capacity and quality of life.
• Three reasons the effects of semaglutide might differ in diabetes: (1) weight loss is consistently smaller in patients with type 2 diabetes; (2) patients are more likely to receive SGLT2 inhibitors, now standard of care in HFpEF; (3) the phenotype is more advanced.
• In STEP-HFpEF (patients without diabetes), semaglutide improved HF symptoms, weight and exercise function.

Methods

Design and participants

• Double-blind, placebo-controlled trial at 108 sites in 16 countries in Asia, Europe and North and South America.
• Inclusion: age ≥18; documented HF; LVEF ≥45%; BMI ≥30; plus at least one of: elevated LV filling pressures, elevated natriuretic peptides with echocardiographic abnormalities, or HF hospitalization within 12 months with echocardiographic abnormalities/ongoing diuretics.
• Diabetes: diagnosis ≥90 days before screening; HbA1c ≤10%.
• Key exclusions: weight change >5 kg in 90 days; type 1 diabetes; GLP-1 receptor agonist within 90 days; uncontrolled diabetic retinopathy.
• Randomization: 1:1, stratified by BMI (<35 vs ≥35); semaglutide escalated from 0.25 mg over 16 weeks to 2.4 mg weekly, or placebo, for 52 weeks plus 5-week follow-up.
• Background glucose-lowering therapy could be any class except GLP-1 receptor agonists; adjustment was at investigator discretion.

Endpoints

• Dual primary: change in KCCQ-CSS (0–100, higher = better) and percentage change in body weight at week 52.
• Confirmatory secondary: 6-minute walk distance; a hierarchical composite (death, HF events, KCCQ-CSS differences of ≥15/≥10/≥5 points, 6MWD difference ≥30 m); and change in CRP.
• Safety: serious adverse events and adverse events of special interest (hypoglycaemia, diabetic retinopathy).
• Two estimands were used: treatment policy (regardless of discontinuation/rescue) and trial product (on-treatment). Results below are treatment policy unless stated.

Results

Population

• 616 randomized (310 semaglutide, 306 placebo); 94.2% and 95.1% completed.
• Median age 69 years; 44.3% female; median weight 102.7 kg; median BMI 36.9; 64.3% BMI ≥35.
• Median KCCQ-CSS 59.4; 6MWD 280 m; NT-proBNP 493 pg/mL; HbA1c 6.8%; diabetes duration 8.0 years.
• Background therapy: diuretics, RAS blockers and beta-blockers in most; MRAs 32.5%; SGLT2 inhibitors 32.8%; metformin 71.9%; insulin 20.8%.

Dual primary endpoints

• KCCQ-CSS: +13.7 vs +6.4 points; difference +7.3 (95% CI 4.1–10.4); P<0.001 (trial product estimand +8.6).
• Body weight: −9.8% vs −3.4%; difference −6.4 percentage points (95% CI −7.6 to −5.2); P<0.001 (trial product −7.9).
• Prespecified SGLT2 inhibitor interaction: KCCQ-CSS difference 5.3 points (95% CI −0.2 to 10.7) with SGLT2i at baseline vs 8.3 (4.5–12.1) without; weight difference −4.7 (−6.7 to −2.8) with vs −7.2 (−8.7 to −5.8) without. Benefits were directionally consistent but attenuated in SGLT2i users.

Confirmatory secondary endpoints

• 6MWD: +12.7 vs −1.6 m; difference +14.3 m (95% CI 3.7–24.9); P=0.008 (trial product +18.1).
• Hierarchical composite: win ratio 1.58 (95% CI 1.29–1.94); P<0.001; the KCCQ-CSS ≥15-point improvement contributed the most wins (trial product 1.86).
• CRP: −42.0% vs −12.8%; treatment ratio 0.67 (95% CI 0.55–0.80); P<0.001.

Supportive and safety

• NT-proBNP: treatment ratio 0.8 (95% CI 0.7–0.9).
• Adjudicated HF events: 7 vs 18 (HR 0.40; 95% CI 0.15–0.92) — a supportive, non-hierarchically tested endpoint.
• HbA1c: −0.8 percentage points (95% CI −1.0 to −0.6).
• Serious adverse events: 17.7% vs 28.8% (P=0.002); cardiac disorders 19 vs 40 (P=0.004).
• Discontinuation for serious AEs 1.9% vs 3.6%; for any AE 10.6% vs 8.2%.
• No increase in clinically significant hypoglycaemia and no increase in diabetic retinopathy events.
• Deaths: 16 total (6 semaglutide, 10 placebo); 5 cardiovascular (1 vs 4).

Discussion

• Semaglutide improved symptoms, physical limitations, exercise function and inflammation in obesity-related HFpEF with type 2 diabetes — extending the STEP-HFpEF findings to a higher-risk group.
• The weight loss was ~40% smaller than in STEP-HFpEF, yet symptom benefits were comparable. The authors argue this indicates mechanisms beyond weight loss: decongestion, vascular and skeletal-muscle function, mitochondrial function, epicardial adipose tissue, inflammation and insulin resistance.
• Benefits were consistent in patients taking and not taking SGLT2 inhibitors, supporting add-on use.
• The safety profile was favourable: reduced HbA1c without hypoglycaemia, and no retinopathy signal.
• Adjudicated HF events were few but favoured semaglutide, alongside substantial NT-proBNP reduction — a signal for disease modification requiring confirmation in outcome trials.
• The trial was not designed or powered for clinical events.

Limitations of the document

• Not powered for hard outcomes. The HF-event finding (7 vs 18; HR 0.40) is a supportive, non-hierarchically tested endpoint with wide confidence intervals and few events; it cannot establish a mortality or hospitalization benefit.
• 1-year follow-up only — durability, weight regain after discontinuation and long-term cardiovascular effects are unknown.
• Predominantly White population. Although 26% of US participants were Black (higher than national HFpEF data), the overall number of non-White participants was low, limiting generalizability.
• Open questions about background therapy: only 32.8% were on SGLT2 inhibitors, below contemporary guideline-directed use; the attenuation of benefit in SGLT2i users is hypothesis-generating and the interaction was not powered.
• No blinding of the weight outcome in practice — participants and investigators could infer assignment from weight loss and GI effects, which may inflate patient-reported KCCQ differences.
• The trial excludes HbA1c >10%, type 1 diabetes, recent GLP-1 use and uncontrolled retinopathy, so the sickest and most obese-diabetic patients are under-represented.
• Industry sponsor: Novo Nordisk designed and conducted the trial and performed the statistical analysis (with independent validation of the testing hierarchy); the first author had full data access.
• The 5-week follow-up is too short to assess rebound weight gain or withdrawal effects.

Key Concepts Mentioned

• entities/Semaglutide — the drug under test
• concepts/GLP-1-Receptor-Agonists — drug class and cardiovascular outcome data
• entities/HFpEF — the target phenotype
• entities/Obesity — the obesity-HFpEF phenotype and weight-loss mechanism
• entities/Heart-Failure — HF events and NT-proBNP
• concepts/Visceral-Adiposity — proposed mechanistic link
• concepts/Obesity-Paradox — intentional versus unintentional weight loss

Key Entities Mentioned

• Novo Nordisk — sponsor and manufacturer of semaglutide
• STEP-HFpEF, STEP-HFpEF DM, SUMMIT, SELECT, FLOW, SOUL — trial programme
• Kansas City Cardiomyopathy Questionnaire (KCCQ) — primary patient-reported instrument
• Statogen Consulting — independent statistician validating the testing hierarchy

Wiki Pages Updated

• Created: wiki/sources/semaglutide-stephfpef-dm-nejm-2024.md
• Updated: wiki/entities/Semaglutide.md
• Updated: wiki/entities/HFpEF.md
• Updated: wiki/concepts/GLP-1-Receptor-Agonists.md
• Updated: wiki/sourceindex.md
• Updated: log.md