Semaglutide tablets vs injections: how a peptide survives being swallowed

You've probably seen headlines about "the semaglutide pill" and wondered how that works, given that peptides are famously not taken by mouth. The short answer: the tablet is not a different molecule. It is the same peptide as the injection, packed with a second ingredient whose whole job is to get it across the stomach wall before it is destroyed. This page explains that chemistry and how the oral and injected trial programmes were designed. It does not cover what the medicine does in a person.

Why a peptide normally can't be a tablet

Peptides are chains of amino acids joined by peptide bonds, and breaking those bonds is exactly what digestion is for. Acid and the enzyme pepsin attack them in the stomach, more proteases wait in the small intestine, and anything that survives still has to cross a gut lining that lets very little of its size through. We go through each barrier in why most peptides are not taken orally.

Semaglutide is a 31-residue peptide of about 4,114 Da, roughly four times the 1,000 Da mark above which, as one review of the oral formulation puts it, several barriers must be overcome: enzymatic breakdown, pH-driven changes in shape, poor permeability and variable absorption. So a semaglutide tablet needs help. That help is SNAC.

What SNAC is

SNAC is short for sodium N-[8-(2-hydroxybenzoyl) aminocaprylate]. Its non-proprietary name is salcaprozate sodium. It is a small fatty-acid derivative and it is an absorption enhancer, sometimes called a permeation enhancer: an ingredient added to a formulation to help another molecule get across a biological barrier. It is not new. Reviews describe it being studied with other molecules, vitamin B12 among them, before semaglutide.

The tablet is semaglutide and SNAC pressed together. The interesting part is where and how they work.

How the tablet gets through the stomach

The key paper is Buckley and colleagues, Science Translational Medicine, 2018. Using imaging in people and studies in dogs, they found three things that surprised people who assumed any oral peptide would be absorbed in the intestine, as small-molecule drugs usually are:

1. Absorption happens in the stomach, and only in a small area right next to the surface of the dissolving tablet.
2. It requires SNAC. Without the enhancer co-formulated in the tablet, the peptide is not absorbed.
3. It goes through the cells, not between them. The authors describe the route as transcellular, with no evidence of any effect on the tight junctions that seal the gaps between cells. And the effect of SNAC is transient: it enhances absorption only briefly.

The reviews that followed spell out the mechanism in more detail:

• Local buffering. As the tablet erodes, SNAC raises the pH of the small pocket of fluid around it. In-vitro evidence suggests that a less acidic microenvironment reduces the conversion of pepsinogen into pepsin, the enzyme that would otherwise chop the peptide up.
• Monomerisation. Semaglutide molecules tend to clump together (oligomerise). SNAC changes the polarity of the surrounding solution so the molecules stay as single units, which is the form that can be absorbed.
• A brief opening. SNAC transiently increases transcellular permeability of the stomach lining, so a fraction of the peptide passes through the cells into the circulation.

One honest aside: this does not work for every peptide. The Aroda review cited below notes that liraglutide, a close relative, was a poor fit for SNAC because it binds membranes more strongly and clumps more readily. The formulation is matched to the molecule. SNAC is not a general key for oral peptides.

Why most of the peptide in a tablet never arrives

Even with SNAC, most of the peptide in a tablet never reaches the bloodstream. The review by Aroda and colleagues (2022) reports an oral bioavailability of 0.8 per cent under the conditions studied, and describes oral absorption as faster but lower and more variable than with subcutaneous injection. Bioavailability means the share of a substance that reaches the circulation intact.

An injection bypasses the gut altogether. A tablet loses more than 99 per cent of its peptide along the way, which is why formulation, not the molecule, is the hard part of an oral peptide.

Once it is absorbed, though, the route stops mattering. The same review reports that semaglutide is distributed, metabolised and eliminated in the same way whether it arrived by vein, by injection or by mouth, and that its half-life of around a week is not determined by how it was given. That long half-life, which comes from the same albumin-binding design described in tirzepatide, explained, is one reason semaglutide was considered a good candidate for a tablet in the first place.

For contrast, the other GLP-1 tablet licensed in the UK takes the opposite route. Orforglipron is not a peptide at all but a small molecule, so it has no peptide bonds for digestion to break and needs no SNAC-type formulation.

How the oral and injected trials were designed

Semaglutide has four main phase 3 programmes. SUSTAIN tested the injection in type 2 diabetes, STEP the injection in obesity, PIONEER the tablet in type 2 diabetes and OASIS the tablet in obesity. How semaglutide compares with tirzepatide and retatrutide at the receptor level is covered in semaglutide vs tirzepatide vs retatrutide.

Here are five of those trials by design only. If any term is unfamiliar, how to read a peptide study explains them all.

Tablet trials

Injection trials

What's worth noticing:

• The programmes mirror each other. PIONEER 1 and SUSTAIN 1 ask the same placebo question in a similar diabetes population, one for each form. OASIS 1 and STEP 1 do the same in obesity, with the same 68-week length. That parallel design is deliberate, and it's why the two forms are often discussed side by side, but they are still separate trials with separate participants.
• PIONEER 4 is the one that crosses routes. It compared the tablet with an injected drug, liraglutide. You cannot hide a tablet as an injection, so it used a double-dummy design: every participant received both a tablet and an injection, with an inactive dummy standing in for whichever they had not been assigned. That keeps the trial blind even though the two treatments look nothing alike. Compare SURPASS-2 in tirzepatide, explained, which was open-label.
• The primary endpoint and the trial length are not always the same. PIONEER 4 ran for 52 weeks but its primary endpoint was measured at week 26. The registration tells you which time point the trial was built to answer.
• No trial in this table compares the tablet with the injection of semaglutide directly. If a post claims one form "beats" the other, ask which trial measured them against each other.

Is oral semaglutide licensed in the UK?

Yes, as a prescription-only medicine. On 11 June 2026 the MHRA announced the UK's first GLP-1 receptor agonist tablet for weight management, a semaglutide tablet, describing the decision as authorisation of a new indication for the tablet. The MHRA's own statement stresses that, as with all GLP-1 receptor agonists, it is a prescription-only medication. In the United States, an oral semaglutide formulation for type 2 diabetes was approved by the FDA in 2019.

Injectable semaglutide is also licensed in the UK and also prescription only. Neither form is a research compound, and Biovanta does not supply either.

What this means for research peptides

The tablet is a showcase for how much engineering an oral peptide needs: a peptide chosen for its size, half-life and potency, an enhancer matched to that one molecule, and a tablet designed to release both at the same spot in the stomach. A capsule of an unmodified research peptide has none of that. For laboratory research, the question of route doesn't arise in the same way, but it's the reason claims of "oral" versions of other peptides deserve a sceptical read.