The same molecule, two very different odds of arriving
The active ingredient in the semaglutide tablet is the same acylated GLP-1 analogue that is in the weekly injection. Same 31-amino-acid backbone, same aminoisobutyric acid substitution at position 8 to blunt DPP-4 cleavage, same C18 fatty diacid chain that binds albumin and stretches the half-life to about a week. Nothing about the receptor pharmacology changes when you swallow it. What changes is the fraction that survives the trip.
Injected under the skin, semaglutide has a bioavailability of roughly 89 percent. Swallowed, the number collapses to somewhere around one percent, and depending on the study and the conditions it is often quoted lower. Everything odd about the oral form follows from that gap: why a 14 mg tablet is not fourteen times an injectable milligram, why the label wraps the dose in timing rules, and why two people taking an identical tablet can end up several-fold apart in plasma concentration.
It is worth stating the framing plainly, because a lot of consumer writing gets it backwards. The tablet is not a weaker or gentler version of the drug. It is the same drug pushed through a route that evolution built specifically to destroy peptides, using a chemical escort that works for a few minutes in one small patch of stomach lining.
Why peptides normally cannot be swallowed
An orally administered peptide faces three sequential problems. In the stomach it meets pepsin at a pH low enough to denature most proteins. If it reaches the small intestine it meets trypsin, chymotrypsin and a wall of brush-border peptidases. Whatever fragments survive still have to cross an epithelium that is selectively permeable to small, reasonably lipophilic molecules, and semaglutide at roughly 4,100 daltons is neither small nor passively permeable.
The default outcome is that essentially none of the dose is absorbed intact. That is why insulin, teriparatide, most GLP-1 analogues and nearly every therapeutic peptide on the market are injected. Attempts at oral peptide delivery have a long history of failure, and the handful of successes have all required an enhancer of some kind rather than a clever peptide design alone.
Semaglutide has two things going for it before any formulation work begins. Its albumin binding and protease resistance mean the small amount that does get in stays around for a week, so useful concentrations can accumulate from repeated small absorption events. And it is potent enough that milligram-scale delivery suffices, which makes throwing away 99 percent of the dose tolerable.
What SNAC actually does
The enabling component is SNAC, sodium N-(8-(2-hydroxybenzoyl)amino)caprylate, an amphiphilic salicylamide derivative developed as part of a carrier platform for oral macromolecules. Each tablet contains 300 mg of it, regardless of whether the semaglutide strength is 3, 7 or 14 mg. By mass the enhancer is the tablet; the peptide is a minority passenger.
SNAC does at least two things as the tablet erodes against the gastric mucosa. It raises the pH in the immediate microenvironment around the dissolving tablet, suppressing pepsin locally so the peptide is not cleaved before it can cross. It also increases epithelial membrane fluidity and favours the monomeric form of semaglutide, promoting transcellular passage rather than the paracellular route most permeation enhancers exploit. Work published in Science Translational Medicine tracing the absorption site identified the stomach, not the intestine, as where the drug gets in, and time to peak concentration after an oral dose is about an hour against one to three days for the injection.
The critical property is that the effect is local, brief and reversible. Absorption happens in a small patch of gastric wall touching the eroding tablet, over tens of minutes, after which the mucosa returns to baseline. That is reassuring on safety, since a permanently leaky gut wall would be a liability. It is also why the whole thing is fragile: anything that moves the tablet, dilutes the SNAC or changes the local pH degrades the absorption event.
Why the tablet numbers look so much bigger
Oral semaglutide is labelled at 3 mg, 7 mg and 14 mg once daily, with 3 mg used as a starting strength for the first month and not intended as a maintenance strength. Injectable semaglutide for type 2 diabetes is labelled at 0.25, 0.5, 1.0 and 2.0 mg once weekly, and the obesity formulation goes to 2.4 mg once weekly.
Put on the same footing, 14 mg daily is 98 mg of semaglutide swallowed per week against 1 or 2 mg injected. Roughly a hundredfold more drug is dispensed to produce a broadly comparable exposure. That ratio is not a safety margin or a potency difference. It is the reciprocal of the bioavailability, and it is the single most useful thing to hold in mind when a milligram figure for a tablet looks alarming next to a milligram figure for a pen.
It also has consequences that have nothing to do with the patient. Manufacturing a hundredfold more active pharmaceutical ingredient for the same clinical effect is a substantial industrial burden, and semaglutide supply constraints during the peak of GLP-1 demand were partly a story about how much peptide the oral route consumes per treated person.
The fasting and water rules are the mechanism, not caution
The prescribing information directs that the tablet be taken on an empty stomach when the person first wakes, with no more than 4 ounces, about 120 millilitres, of plain water, and that no food, other drink or other oral medication follow for at least 30 minutes. Read against the mechanism, none of that is a generic instruction to be careful. Each clause protects one variable in the absorption event.
Food is the most consequential. Anything in the stomach raises the volume the SNAC has to work in, changes the pH, and speeds the tablet along on its way out. Studies of dosing with food found exposure reduced to a fraction of the fasted value, close enough to nothing that the label treats the fasted state as a condition of the drug working rather than an optimisation.
The water volume is counter-intuitive and is the clause people most often get wrong. More water is worse. Larger volumes dilute the SNAC around the tablet and can wash it out of the stomach before absorption is complete, and pharmacokinetic studies showed lower exposure at higher water volumes. The 30-minute wait was chosen as a workable compromise; longer post-dose fasting increased exposure further in dose-timing studies, but at some point a rule stops being followed.
What PIONEER actually measured
The oral formulation was developed through the PIONEER programme in type 2 diabetes. PIONEER 1 was the placebo-controlled monotherapy trial over 26 weeks and established dose-dependence: HbA1c fell progressively across the 3, 7 and 14 mg strengths, with the largest reduction a little above one percentage point at 14 mg. Other trials placed the tablet against active comparators, which is the more informative test.
PIONEER 2 compared oral semaglutide 14 mg with empagliflozin 25 mg over 52 weeks, and PIONEER 3 with sitagliptin 100 mg. PIONEER 4 is the most quoted, because it put the 14 mg tablet against subcutaneous liraglutide 1.8 mg and placebo over 52 weeks, and the tablet was at least as good as the daily injection on HbA1c and better on weight. Other trials covered renal impairment, add-on to insulin, and Japanese populations.
PIONEER 6 was the cardiovascular safety trial, roughly 3,200 people at high cardiovascular risk followed for a median of a little over a year. It met its non-inferiority objective against placebo for major adverse cardiovascular events. It was not powered to demonstrate superiority, and the point estimate favouring the drug was frequently over-read at the time.
PIONEER PLUS later asked whether the ceiling was the molecule or the dose, comparing 14 mg with 25 mg and 50 mg daily in type 2 diabetes. Higher oral doses produced greater HbA1c and weight reductions, which is the clearest available evidence that 14 mg was an absorption-limited dose rather than a pharmacologically maximal one.
SUSTAIN, STEP and the injectable benchmark
The injectable was built on the SUSTAIN programme in type 2 diabetes. SUSTAIN 6 is the trial that changed practice: about 3,300 people at high cardiovascular risk over 104 weeks, with a significant reduction in major adverse cardiovascular events against placebo, alongside an increase in diabetic retinopathy complications attributed largely to rapid glucose lowering in people with pre-existing retinopathy. SUSTAIN 7 beat dulaglutide head to head on HbA1c and weight.
The obesity indication came from the STEP programme using 2.4 mg weekly, where 68 weeks of treatment produced mean weight loss around 15 percent of body weight against roughly 2 to 3 percent on placebo. SELECT then showed a reduction in cardiovascular events in people with overweight or obesity and established cardiovascular disease but without diabetes, which is the result that moved the drug from a metabolic treatment to a cardiovascular one.
The oral route has been pushed toward the same territory. The OASIS trials tested higher-strength oral semaglutide for weight management, with the 50 mg dose producing weight loss in the range of the 2.4 mg injection over 68 weeks and a 25 mg strength studied subsequently. Readers should check the current label for what is approved in their jurisdiction, because the approved oral indications are narrower than the trialled ones and this picture has moved repeatedly.
What does not exist anywhere in this literature is a phase 3 trial of oral semaglutide against subcutaneous semaglutide. The comparators were liraglutide, sitagliptin, empagliflozin, dulaglutide and placebo. Every statement of the form the tablet is equivalent to a particular injectable dose is a cross-trial inference across different populations, baselines and years, not a measured result.
Variability is the part the averages hide
Trial means conceal how differently the tablet behaves between people. Reported inter-individual variability in oral semaglutide exposure is on the order of a hundred percent as a coefficient of variation, against something closer to twenty percent for the subcutaneous route. In plain terms, two people on the same tablet strength, both following the instructions, can sit several-fold apart in steady-state concentration.
The reasons are all downstream of the mechanism. Gastric emptying rate, resting stomach volume, gastric pH, mucosal surface in contact with the tablet and the precision of the fasting window all vary between people and within the same person day to day. None of those inputs matter to a subcutaneous depot, which is why the injection is comparatively predictable.
This is also the real reason the tablet is taken daily. Semaglutide has a terminal half-life of about a week whichever way it enters, so clearance alone would permit weekly oral dosing. Daily administration exists because each absorption event is small and unreliable, and averaging thirty of them a month produces a far steadier exposure than betting the month on one. Steady state still takes roughly four to five weeks to establish, as it does with the injection.
Side effects and what actually differs
Because it is the same molecule acting at the same receptor, the adverse effect profile is the profile of semaglutide: nausea, vomiting, diarrhoea and constipation dominating early, largely dose-related, largely attenuating with time and titration. The class boxed warning for thyroid C-cell tumours, derived from rodent studies, applies to the oral and injectable products alike, as do the cautions around pancreatitis, gallbladder disease and severe gastrointestinal reactions.
The differences that do exist are route-specific rather than pharmacological. Injection-site reactions belong to the pen and not the tablet. The tablet in turn carries an interaction concern that the injection does not, because a drug absorbed through a narrow gastric window over half an hour sits in direct competition with everything else taken orally in the morning, and levothyroxine in particular has been flagged for altered exposure.
What orforglipron changes about the question
Orforglipron is a non-peptide, small-molecule GLP-1 receptor agonist taken orally once daily. Because it is not a peptide it is not a substrate for digestive proteases, and it needs no permeation enhancer to cross the gut wall. Its phase 3 programme reported in 2025, with obesity trials showing mean weight loss in the region of a tenth of body weight over about 72 weeks.
The interesting fact for this comparison is not the efficacy number but the absence of a food and water rule. Orforglipron was studied without the fasting requirements that the semaglutide tablet cannot function without. That is the cleanest possible demonstration that the constraints on oral semaglutide are a peptide delivery problem, not something intrinsic to oral GLP-1 pharmacology.
Regulatory status should be stated carefully: semaglutide is approved in oral and injectable forms, whereas orforglipron completed phase 3 and entered regulatory review, and its approval status varies by jurisdiction and has changed recently. It is not a compounded or grey-market compound, but nor should it be described as an established treatment on the strength of press releases.
Where the comparison breaks down
The honest summary is that oral and injectable semaglutide are the same drug at different and less certain delivered doses, tested in overlapping but non-identical programmes, and never against each other. Anyone converting between them with a fixed ratio is inventing a number. The literature supports the statement that the tablet at higher strengths lands in the same neighbourhood as mid-range injectable doses, and nothing more precise.
The route choice is also not purely pharmacological. A weekly injection requires cold-chain handling, device familiarity and tolerance of needles, and delivers a predictable dose. A daily tablet requires a rigid morning routine, competes with other oral medication, and delivers an unpredictable one. Which trade is better is a question about a particular person's circumstances, and it is not one this article can answer for anybody.
What the absorption problem does settle is how to read the numbers. A tablet strength and an injection strength are not on the same scale and never were, and a claim that compares them directly, in either direction, is the fastest way to identify writing that has not understood the drug.