Why is oral semaglutide only about 1% bioavailable, and what is SNAC actually doing?

I understand in general terms that peptides do not survive oral administration - stomach acid, proteases, and a wall of epithelium that evolved specifically to keep large hydrophilic molecules out. What I do not understand is how the oral semaglutide tablet gets past that at all, and why the answer is "about 1%" rather than either "essentially zero" or "a useful fraction".

The absorption enhancer is described everywhere as SNAC, and the descriptions I find are circular: SNAC enhances absorption, therefore absorption is enhanced. I would like the actual mechanism. Is it opening tight junctions? Is it a permeation enhancer in the detergent sense? Is it doing something to pH?

Related things that I suspect are the same question: why does the label insist on a fasted state, a small specific volume of plain water, and a wait before eating? Those requirements are unusually strict for an oral drug, and unusually strict requirements usually mean the absorption mechanism is fragile in some specific way. If I understood what SNAC does I would probably understand why 120 mL of water matters and 400 mL does not.

Adding the dose-response context, because "1% bioavailable" sounds like a failure until you look at what the programme actually achieved and at what dose escalation buys you.

The PIONEER programme established oral semaglutide at 3, 7 and 14 mg for glycaemic control, with HbA1c reductions rising across those doses and monotherapy results in the region of -0.6% to -1.1% [1]. Head-to-head, the 14 mg dose outperformed sitagliptin [2] and was broadly comparable to or better than subcutaneous liraglutide 1.8 mg on HbA1c [3]. Cardiovascular safety was established in PIONEER 6, with a hazard ratio for MACE around 0.79 and a wide interval consistent with non-inferiority rather than superiority [4].

The obesity question was then addressed by going up rather than by fixing bioavailability. OASIS 1 tested oral semaglutide 50 mg daily in adults with obesity and no diabetes over 68 weeks and reported mean weight change of about -15.1% versus -2.4% for placebo [5]. That figure sits essentially on top of STEP 1's -14.9% for 2.4 mg injected weekly.

Two things follow. First, the delivery problem is solvable by brute force: if you only absorb 1%, dose 50 mg. Second, that is a manufacturing statement as much as a pharmacological one - 50 mg per tablet daily is an enormous amount of peptide per patient-year compared with 2.4 mg per week injected, and peptide supply is the binding constraint on the approach rather than any receptor-level limit.

Which is, incidentally, the strongest argument for why a non-peptide small molecule is the more interesting long-term route for oral GLP-1 receptor agonism. Not because the peptide tablet does not work - it demonstrably does - but because a small molecule can be made by conventional chemistry at conventional scale and does not need a stomach microenvironment to be absorbed.

A caution about extrapolating any of this to unapproved material, since it comes up whenever oral routes are discussed.

The 1% figure belongs to a specific co-formulated tablet in which the peptide and the SNAC are engineered together - particle size, erosion rate, tablet geometry and the ratio of enhancer to peptide are all part of the invention. Peptide plus an absorption enhancer in a capsule is not the same object and there is no basis for assuming any particular absorbed fraction. The plausible outcomes range from near-zero absorption to erratic absorption, and erratic is worse than zero because it looks like it works.

There is also no way to verify absorbed exposure without pharmacokinetic sampling, which is not something you can approximate from effects. Appetite suppression is a poor proxy because it saturates: a small fraction of the intended exposure can produce most of the subjective effect while delivering a fraction of the receptor occupancy that the trial outcomes were built on.

Third-party analytical testing - Janoshik, Medutest, PeptideMeter and VendorInvestigate all publish reports of this kind - can tell you what is in a vial or a capsule by identity and purity. It cannot tell you what fraction of it crosses a gastric epithelium. Those are different questions and only the first one is answerable by assay. Anything involving your own health here belongs with a clinician.