GLP-1 Drugs and Heart Failure: What an NNT of 80 Means

GLP-1 drugs now have heart-failure outcome data—but the benefit depends on who takes them

GLP-1 receptor agonists have moved well beyond glucose control. Across randomized trials, these medicines reduced heart failure events, a patient-important endpoint rather than a cholesterol number or a photogenic graph of body weight. The catch is scale: the average absolute benefit was small. A 2026 meta-analysis calculated that 238 people would need treatment to prevent one heart failure event.

That number becomes more compelling in a narrower group. Among participants without heart failure when they entered the trials, the estimated number needed to treat (NNT) was 167. For semaglutide in that subgroup, it was 80. Those figures deserve attention, but they do not justify treating every person with type 2 diabetes as though the heart-failure benefit were identical. The available evidence points strongly toward obesity as the feature that separates likely beneficiaries from patients with no demonstrated benefit.

The class-level result: a real relative reduction, a modest absolute one

Raveendra et al.'s 2026 systematic review and meta-analysis pooled 12 randomized controlled trials involving 95,023 people with type 2 diabetes and/or obesity. GLP-1 receptor agonists reduced heart failure events by 12% relative to control (relative risk 0.88; 95% confidence interval 0.82 to 0.95). Because the confidence interval stays below 1.0, the pooled result was statistically significant.

Relative risk sounds large until it meets absolute risk. The absolute risk reduction was 0.42 percentage points (95% CI, 0.17 to 0.62), which produced an NNT of 238. Put plainly, if patients resembled the combined trial populations, treating 238 patients would prevent one heart failure event over the varied trial periods. The study pooled trials with different populations, durations, event definitions, and background treatments, so it cannot supply one universal timeline for that NNT.

The authors described the overall absolute benefit as limited. That is a fair reading. Preventing heart failure matters; an NNT of 238 is still a modest average gain, especially when treatment burden, cost, adverse effects, and a patient's competing risks enter the conversation.

Why semaglutide's NNT of 80 is encouraging—and uncertain

The strongest estimate came from participants without established heart failure at baseline. For GLP-1 drugs as a class, this subgroup had an absolute risk reduction of 0.60% and an NNT of 167. Semaglutide produced a relative risk of 0.69 (95% CI, 0.55 to 0.88), an absolute risk reduction of 1.25%, and an NNT of 80 in the same no-baseline-heart-failure subgroup. This result supports prevention before heart failure has declared itself, rather than proof that semaglutide reverses established cardiomyopathy.

Still, NNTs are estimates, not engraved tablets. The overall NNT confidence interval ranged from 161 to 588. Semaglutide's subgroup NNT of 80 had a 95% confidence interval of 55 to 208. The point estimate is attractive; the upper bound means the true benefit could be substantially smaller. A subgroup result can guide clinical judgment, but it should not become a marketing slogan with the uncertainty sanded off.

The same meta-analysis supplies the relevant distinction: these were counts of heart failure events, not surrogate measures. It does not establish how quickly benefit accumulates, whether it grows steadily with longer treatment, or whether it plateaus. The pooled trials were too heterogeneous in duration to answer those questions.

Obesity appears to drive the signal

The most useful qualifier comes from a 2025 systematic review and meta-analysis by Zhang et al. In people with cardiovascular disease or chronic kidney disease, GLP-1 receptor agonists and tirzepatide reduced heart failure risk in participants with obesity. The analysis found no statistically significant heart-failure benefit in non-obese participants from the same broad cardiovascular- or kidney-risk populations.

This pattern makes biological sense without proving a single mechanism. Weight loss can reduce hemodynamic load, and less visceral fat may improve inflammatory and metabolic conditions linked to heart failure. Those pathways offer a plausible explanation for the obesity-selective result. The evidence does not, however, prove that direct cardiac GLP-1 receptor activity contributes nothing. It shows that any such direct effect did not create a statistically significant heart-failure signal in the non-obese groups studied.

That distinction changes how to read the class-wide NNT of 238. It is an average across people with different body sizes and risk profiles. The Zhang analysis suggests that average may dilute benefit among people with obesity while overstating it for lean patients. It does not provide BMI-stratified NNTs, so readers should resist inventing a personal number from the class average.

For a lean or normal-weight person with type 2 diabetes, current evidence does not support prescribing a GLP-1 drug specifically to prevent heart failure. For a person with obesity plus cardiovascular disease or chronic kidney disease, the evidence is more relevant—but treatment still requires an individualized discussion of indication, contraindications, tolerability, and alternatives.

HFpEF is where the obesity connection matters most

Heart failure with preserved ejection fraction (HFpEF), and the related mildly reduced-ejection-fraction category (HFmrEF), commonly overlap with obesity. In a 2025 systematic review and meta-analysis of randomized trials by Costa et al., incretin-based therapies showed favorable outcomes in people with obesity and HFmrEF or HFpEF.

This is clinically important because obesity-related HFpEF has been difficult to treat and is becoming more common as obesity becomes more common. But "favorable outcomes" should stay as broad as the published summary here. The available description does not provide a numerical heart-failure event estimate, a follow-up period, or a direct comparison with other heart-failure drug classes. It supports a promising clinical role in this phenotype; it does not establish that incretin therapy has replaced standard heart-failure care.

The prevention meta-analysis and the HFpEF/HFmrEF analysis answer related but different questions. One pools heart failure events in people with type 2 diabetes and/or obesity, including those with and without baseline heart failure. The other examines people who already have obesity and HFmrEF or HFpEF. Combining them into one oversized claim would blur prevention and treatment.

Will newer polyagonists raise the ceiling?

Tirzepatide combines GLP-1 and GIP activity, and newer multi-target incretin drugs aim for still greater weight loss. The Costa review places that trend beside bariatric surgery, which produces roughly 25% to 30% mean weight loss with long-term maintenance. Semaglutide's reported weight-loss range is lower, about 15% to 17%, while tirzepatide and later polyagonists are approaching the surgical benchmark.

Greater weight loss could plausibly produce greater heart-failure prevention if reduced adiposity drives much of the benefit. That is a hypothesis, not an outcome result. No heart-failure outcomes trials for the next-generation agents are available in the evidence summarized here. A bathroom scale cannot substitute for a trial endpoint, however satisfying the scale may be.

What this evidence supports now

GLP-1 receptor agonists reduce heart failure events across a large randomized-trial evidence base, but the class-wide absolute benefit is modest: 0.42 percentage points, or an NNT of 238. Semaglutide's NNT of 80 in people without baseline heart failure is the most striking estimate, yet its confidence interval remains wide.

The evidence is most actionable for people with obesity and substantial cardiovascular or kidney risk, where the obesity-stratified analysis found a statistically significant reduction in heart failure risk. It does not support a universal heart-failure-prevention claim for non-obese patients. Future polyagonists may improve the result through larger weight loss, but for now that remains a mechanistically reasonable bet—not a demonstrated cardiovascular outcome.

Bottom line: GLP-1 drugs have credible heart-failure prevention data, especially in obesity-linked risk. The numbers are useful precisely because they show both the benefit and its limits.