Skip to content
GLP Loss
← Research
Evidence

SOUL Explained: Oral Semaglutide and Heart Outcomes

SOUL cut major cardiovascular events from 13.8% to 12.0% over four years in adults with type 2 diabetes and established disease. The benefit tracked blood sugar, not weight, and the tablet it tested is 0.8% bioavailable.

Owen Castellanos10 min read
Hazard ratios for major cardiac eventsOral semaglutide against placebo. Lower favors the drug.SOUL, all participants0.86 (0.77 to 0.96)PIONEER 6, all patients0.79 (0.57 to 1.11)SOUL, prior heart failure0.78 (0.63 to 0.96)SOUL, no heart failure1.01 (0.84 to 1.20)0.50.751.01.25The vertical line at 1.0 is no difference from placebo.The two heart-failure rows are the same trial, split in two.

A tablet version of semaglutide has a cardiovascular outcome trial behind it, and sellers of compounded oral preparations have started to gesture at it. The trial is SOUL, the result is real, and almost nothing about the population or the product transfers to a compounded tablet. How the two formats compare on weight is a separate question, handled in the oral-versus-injectable article. This page is about what SOUL measured, in whom, and what its endpoint means.

The endpoint, and who it was measured in

SOUL was a double-blind, placebo-controlled, event-driven superiority trial. Participants had to be 50 years of age or older, have type 2 diabetes with a glycated hemoglobin of 6.5% to 10.0%, and have known atherosclerotic cardiovascular disease, chronic kidney disease, or both. They were randomly assigned to once-daily oral semaglutide at a maximal dose of 14 mg or to placebo, in addition to standard care.[1]

Among the 9,650 participants randomized, mean follow-up was 47.5 months (SD 10.9) and median follow-up 49.5 months. The primary outcome was major adverse cardiovascular events — a composite of death from cardiovascular causes, nonfatal myocardial infarction or nonfatal stroke — in a time-to-first-event analysis. It occurred in 579 of 4,825 participants (12.0%; 3.1 events per 100 person-years) on oral semaglutide against 668 of 4,825 (13.8%; 3.7 per 100 person-years) on placebo: hazard ratio 0.86 (95% CI, 0.77 to 0.96; P = 0.006).[1]

Two figures deserve to travel together. The relative reduction is 14%; the absolute difference across roughly four years is 1.8 percentage points. Serious adverse events occurred in 47.9% of the semaglutide group and 50.3% of the placebo group, and the incidence of gastrointestinal disorders was 5.0% and 4.4%.[1]

The confirmatory secondary outcomes, including a five-point composite of major kidney disease events, did not differ significantly between the two groups.[1] The kidney evidence for this class rests on a different trial, described in the FLOW article.

Why the earlier oral trial proved less with a lower ratio

PIONEER 6 came first, in 3,183 patients at high cardiovascular risk, with a median time in the trial of 15.9 months. Major adverse cardiovascular events occurred in 61 of 1,591 patients (3.8%) on oral semaglutide and 76 of 1,592 (4.8%) on placebo: hazard ratio 0.79 (95% CI, 0.57 to 1.11).[2]

That point estimate is lower than SOUL's, and it established far less. The trial was designed to rule out 80% excess cardiovascular risk against a noninferiority margin of 1.8, and its confidence interval runs from a 43% reduction to an 11% increase. The components pointed in different directions: cardiovascular death at 0.49 (0.27 to 0.92), nonfatal myocardial infarction at 1.18 (0.73 to 1.90), nonfatal stroke at 0.74 (0.35 to 1.57), and death from any cause at 0.51 (0.31 to 0.84).[2]

The lesson generalizes past this drug. What a trial establishes is set by its design and its interval, not by how favorable its central number looks, and the same discipline applies to the injectable outcome trial described in the SELECT article.

The benefit did not follow the weight

A post hoc analysis of cardiovascular risk factors in SOUL found early improvements at 13 weeks in glycated hemoglobin (−0.87 percentage points), body weight (−2.54%), systolic blood pressure (−3.84 mm Hg) and high-sensitivity C-reactive protein (−18.08%), sustained across the trial. At week 156 the estimated treatment differences were −0.47 percentage points for glycated hemoglobin (95% CI, −0.52 to −0.42) and −3.26 percentage points for body weight (−3.55 to −2.98), with no significant difference for low-density lipoprotein cholesterol or diastolic blood pressure.[5]

Three percentage points of body weight over three years is not what a weight-loss buyer is shopping for. It is also not what produced the cardiovascular result. A separate analysis found the reduction in major adverse cardiovascular events differed across baseline glycated hemoglobin categories (P for interaction = .04, with greater reduction at higher baseline values) and tracked in-trial glycated hemoglobin reductions at both 13 and 52 weeks (P for interaction = .005 and < .001). Across baseline body-mass index categories, and across changes in body-mass index, the benefit was consistent (P for interaction = .88 and .64).[4]

Median age in that analysis was 66 years, mean glycated hemoglobin 8.0% (SD 1.1) and mean body-mass index 31.1 (SD 5.8).[4] This is a glycemic and vascular trial that happens to involve a molecule sold for weight loss, and the difference between those two indications is set out in the two-approvals article.

Where the result reverses

Of the 9,650 participants, 2,229 (23.1%) had a history of heart failure — 991 (10.3%) with preserved ejection fraction, 592 (6.1%) with reduced ejection fraction and 646 (6.7%) of unknown subtype. For the prespecified composite heart-failure outcome, the hazard ratio was 0.78 (95% CI, 0.63 to 0.96) in participants with a heart-failure history and 1.01 (0.84 to 1.20) in those without, with a P for interaction of .06.[3]

Inside the heart-failure group it splits again. The hazard ratio was 0.59 (0.39 to 0.86) with preserved ejection fraction and 0.98 (0.70 to 1.38) with reduced ejection fraction.[3] Meanwhile the major adverse cardiovascular event reduction showed no heterogeneity by heart-failure history at all, at 0.83 (0.68 to 1.01) and 0.86 (0.75 to 0.98), P for interaction .77.[3]

So one endpoint in this trial is close to uniform across the cohort and another is concentrated in about a tenth of it. A page quoting the 0.78 without the 1.01 beside it has reported a real number and described a false picture, and the broader cardiovascular record for the class is in the heart-benefits article.

The tablet is a delivery system, and a fragile one

Oral semaglutide is a peptide that would ordinarily be digested, carried through the stomach wall by an absorption enhancer in the tablet. A population pharmacokinetic analysis across six clinical pharmacology trials put bioavailability at 0.8% under the recommended dosing conditions — 30 minutes of post-dose fasting, taken with no more than 120 mL of water. Post-dose fasting time, water volume and body weight were the most important covariates affecting exposure. Within-subject variability in bioavailability was 137%, which with once-daily dosing and a long half-life corresponds to 33% within-subject variability in steady-state exposure.[6]

A dedicated dosing-conditions trial makes the dependency concrete. In the food-effect arm, subjects who ate 30 minutes before the dose showed limited or no measurable semaglutide exposure, while every subject in the fasting arm had measurable exposure; exposure increased with longer post-dose fasting (p < 0.001), and the conditions that worked became the ones used in the phase 3 program and written into the label.[7] What else a person swallows around that window matters, which is covered in the oral medications article.

Read those two papers together and the 14 mg in SOUL stops looking like a dose. It is a tablet specification. Roughly ninety-nine parts in a hundred never reach the bloodstream, and the fraction that does depends on an empty stomach, a small glass of water and half an hour of waiting.

What does not transfer to a compounded tablet

No compounded oral or sublingual semaglutide preparation was in SOUL, in PIONEER 6, or in any cardiovascular outcome trial. Compounded drugs are not FDA-approved and are not reviewed by the FDA for safety, efficacy or quality before they are dispensed, so there is no published figure for how much semaglutide a given compounded tablet or dissolvable delivers into the blood. For a molecule whose manufactured tablet lands at 0.8% bioavailability and 137% within-subject variability, that gap is the whole question rather than a technicality.

A milligram figure printed on a compounded oral product is a statement about what went into it, not about what reaches a receptor. The general version of that problem is in the compounded vial article, the regulatory difference is in the compounded-versus-brand article, and the sellers offering an oral route are listed on the oral semaglutide board.

What SOUL settles, and what it leaves open

It settles a narrow and valuable thing. In adults aged 50 and over with type 2 diabetes and established atherosclerotic cardiovascular disease, chronic kidney disease or both, already receiving standard care, four years of once-daily manufactured oral semaglutide lowered the rate of major adverse cardiovascular events from 13.8% to 12.0%.[1]

It leaves open everything outside that description: adults without diabetes, adults without established disease, anyone under 50, anyone taking a preparation other than the tablet the trial used, and anyone whose reason for taking it is the scale rather than the arteries. On that last group the trial is not silent but discouraging, since the weight difference it produced was 3.26 percentage points and the cardiovascular benefit did not vary with it.[4][5]

Frequently asked

What did the SOUL trial find?
In 9,650 adults aged 50 and over with type 2 diabetes and established atherosclerotic cardiovascular disease, chronic kidney disease or both, once-daily oral semaglutide reduced major adverse cardiovascular events from 13.8% to 12.0% over a mean 47.5 months, a hazard ratio of 0.86 (95% CI 0.77 to 0.96). The confirmatory secondary kidney outcomes did not differ significantly.
Why did PIONEER 6 report a lower hazard ratio but prove less?
PIONEER 6 was a noninferiority trial designed to rule out 80% excess risk, with 3,183 patients and a median 15.9 months in the trial. Its hazard ratio of 0.79 carried a confidence interval from 0.57 to 1.11, spanning both benefit and harm, and its components disagreed — cardiovascular death at 0.49 but nonfatal myocardial infarction at 1.18. Design and interval decide what a trial establishes, not the point estimate.
Does the SOUL result depend on losing weight?
No. The estimated treatment difference in body weight at week 156 was 3.26 percentage points, and a post hoc analysis found the cardiovascular benefit consistent across baseline body-mass index categories and across changes in body-mass index, with interaction P values of .88 and .64. It did vary with glycated hemoglobin, being larger at higher baseline values and with larger in-trial reductions.
Did oral semaglutide help with heart failure?
Only in the participants who already had it. The composite heart-failure outcome had a hazard ratio of 0.78 in the 23.1% with a heart-failure history and 1.01 in those without. Within the heart-failure group the ratio was 0.59 for preserved ejection fraction and 0.98 for reduced ejection fraction.
Does SOUL apply to a compounded oral semaglutide tablet?
No compounded oral or sublingual preparation was in SOUL or in any cardiovascular outcome trial. Compounded drugs are not FDA-approved and are not reviewed by the FDA for safety, efficacy or quality before dispensing, and no published figure describes how much semaglutide a given compounded tablet delivers into the blood.
How much of an oral semaglutide tablet is actually absorbed?
Bioavailability is about 0.8% under the labeled conditions of 30 minutes of post-dose fasting with no more than 120 mL of water, with within-subject variability in bioavailability of 137%. In a dedicated trial, subjects who ate 30 minutes before dosing showed limited or no measurable exposure at all, while every subject in the fasting arm had measurable exposure.

Sources

  1. [1] McGuire DK, Marx N, Mulvagh SL, et al. (2025). Oral Semaglutide and Cardiovascular Outcomes in High-Risk Type 2 Diabetes. N Engl J Med. PMID 40162642
  2. [2] Husain M, Birkenfeld AL, Donsmark M, et al. (2019). Oral Semaglutide and Cardiovascular Outcomes in Patients with Type 2 Diabetes. N Engl J Med. PMID 31185157
  3. [3] Pop-Busui R, Rasmussen S, Deanfield JE, et al. (2026). Oral Semaglutide and Heart Failure Outcomes in Persons With Type 2 Diabetes: A Secondary Analysis of the SOUL Randomized Clinical Trial. JAMA Intern Med. PMID 41627802
  4. [4] Inzucchi SE, Abdul Ghani R, Deanfield J, et al. (2026). Oral Semaglutide and CV Benefits in the SOUL Trial: How Do Baseline or Changes in HbA1c or BMI Affect Clinical Outcomes? J Clin Endocrinol Metab. PMID 42478869
  5. [5] Mulvagh SL, Inzucchi SE, Marx N, et al. (2026). Oral Semaglutide and Change in Cardiovascular Risk Factors in High-Risk Type 2 Diabetes: A Post Hoc Secondary Analysis of the SOUL Randomized Clinical Trial. JAMA Cardiol. PMID 41879791
  6. [6] Overgaard RV, Navarria A, Ingwersen SH, et al. (2021). Clinical Pharmacokinetics of Oral Semaglutide: Analyses of Data from Clinical Pharmacology Trials. Clin Pharmacokinet. PMID 33969456
  7. [7] Bækdal TA, Breitschaft A, Donsmark M, et al. (2021). Effect of Various Dosing Conditions on the Pharmacokinetics of Oral Semaglutide, a Human Glucagon-Like Peptide-1 Analogue in a Tablet Formulation. Diabetes Ther. PMID 34080123

Where to get it

Best GLP-1 injections

Every injectable seller we can verify, with the price each one publishes and an honest read of what the trials measured.

Compare providers →

More in Evidence