Most people assume that swallowing a drug is the simplest way to take it. For small molecules like aspirin or caffeine, that is largely true. For peptides, which are chains of amino acids, the digestive system is essentially a wrecking yard. Enzymes break them apart, the gut wall blocks them, and the liver clears whatever slips through. The result is that very little of an orally swallowed peptide ever reaches general circulation.
GLP-1 analogues are a prominent example. These peptides mimic a natural gut hormone involved in regulating blood sugar and appetite. Injected versions have been studied extensively. Getting them to work as a pill is a different challenge entirely. A 2026 study published in Xenobiotica used mathematical modelling to map exactly how one GLP-1 analogue travels through the body after oral dosing, and what variables most influence how much of it survives the journey.
The bioavailability problem
Bioavailability is the fraction of a drug that makes it from the dose site into the bloodstream intact and active. For most oral small-molecule drugs, bioavailability ranges from around 30 percent to nearly 100 percent. For the GLP-1 analogue examined in this study, the published figure is somewhere between 0.4 percent and 1 percent under fasting conditions.
That is not a typo. Of every milligram swallowed on an empty stomach, researchers measure less than one hundredth of a milligram reaching the systemic circulation. The reasons are layered. The peptide has a large, complex molecular structure that the gut lining was not designed to transport. Gastric acid and digestive enzymes begin degrading it almost immediately. Even molecules that survive digestion face a dense cellular barrier in the stomach and intestinal wall, and then a first pass through the liver that filters out more.
Understanding those layers individually is not just an academic exercise. It informs how formulators design tablets, how clinicians interpret dosing studies, and how researchers predict what happens in people with different body weights, kidney function, or gastrointestinal conditions.
How a permeation enhancer changes the picture
The oral formulation studied in this paper includes a co-ingredient called SNAC, short for sodium N-[8-(2-hydroxybenzoyl) amino] caprylate. SNAC is not the active peptide. It is a small molecule that travels alongside the peptide and temporarily makes the stomach lining more permeable, allowing more of the peptide to slip through before degradation catches up.
The researchers built their pharmacokinetic model to capture this interaction directly. They found that the amount of SNAC present in the formulation, not just the peptide dose itself, meaningfully shifts how much peptide is absorbed at the gastric level. This is an important mechanistic insight because it suggests that formulation design, specifically the ratio of SNAC to peptide, is a key lever in oral peptide delivery.
The study also confirmed that absorption for this peptide happens primarily in the stomach rather than the small intestine. That is unusual for most drugs and has practical implications. Stomach emptying rate, food in the stomach, and gastric pH all become critical variables. Taking the tablet with food, for example, dilutes the SNAC concentration and reduces absorption, which is why the fasting-state bioavailability figure is specified separately.
The semi-mechanistic modelling approach
Rather than simply fitting curves to observed blood-level data, the researchers developed what they call a semi-mechanistic pharmacokinetic model. This kind of model incorporates known biological processes, such as gastric emptying, intestinal transit, hepatic extraction, and enzyme degradation, as structural compartments. Data from the literature and from clinical trials is then used to estimate the rates operating within those compartments.
The team built separate models for intravenous dosing, single oral doses across a range of peptide amounts and SNAC concentrations, and repeated daily oral dosing. They validated the steady-state model, meaning the predictions were tested against clinical data collected after subjects had been taking the oral dose long enough for blood levels to stabilize at a consistent rhythm.
The fact that a model developed on single-dose data could accurately predict steady-state concentrations is meaningful. It suggests the model captures the underlying biology rather than just memorizing one dataset. That kind of validated model can then be interrogated to answer questions that clinical trials alone cannot easily address.
Key variables the model identified
The study highlighted four main factors that drive variability in how much of the oral peptide reaches systemic circulation: the dose of the peptide itself, the concentration of SNAC in the formulation, gastrointestinal permeability at the gastric level, and the intestinal first-pass effect.
Gastrointestinal permeability is partly an individual biological trait and partly a situational one. People differ in how permeable their gut lining is, and the same person may have different permeability depending on hydration, inflammation, or other medications. The intestinal first-pass effect refers to metabolism that happens in the gut wall and liver before the drug reaches general circulation. Both factors were shown to be meaningful sources of variability in predicted blood levels.
These findings have direct relevance for population simulations. When the model was used to simulate outcomes across a diverse group of virtual subjects rather than just one average individual, the spread of predicted drug exposures was wide. This kind of population-level insight helps researchers understand why some individuals in clinical trials show quite different responses to the same dose.
Implications for future formulation research
The authors note that the semi-mechanistic model developed here is a stepping stone toward more complex physiologically based pharmacokinetic models, often called PBPK models. PBPK models incorporate even more granular anatomy and physiology, mapping drug movement through specific tissues and organs rather than abstracted compartments.
A validated PBPK model built on this foundation could help researchers predict how the oral formulation behaves in people with gastric conditions, altered liver function, or different body composition without running separate clinical trials for every subgroup. It could also help predict drug-drug interactions, since the SNAC-mediated absorption pathway might be sensitive to other compounds taken at the same time.
The broader lesson from this work is that oral peptide delivery is not simply a matter of swallowing what was previously injected. The formulation chemistry, the timing relative to food, the individual biology of the person taking it, and the dose-concentration relationship of the absorption enhancer all interact in ways that require careful mathematical modelling to understand and optimize.
What this means for peptide science broadly
GLP-1 analogues are among the most studied peptides in current clinical literature, but the pharmacokinetic principles this study explores apply far beyond that single class. Any research peptide with a large molecular structure faces similar barriers to oral absorption. The development of permeation enhancers, protective coatings, and other formulation strategies is an active area of research across many peptide families.
For researchers and informed readers following the peptide space, this study is a useful reminder that the route of administration matters as much as the molecule itself. A peptide delivered subcutaneously, intranasally, or orally can produce very different blood-level profiles even at the same nominal dose, because the absorption barriers and first-pass effects are completely different by route.
The Xenobiotica paper adds a rigorous quantitative framework to that principle, giving future researchers a validated model they can build on. Early data from this kind of mechanistic modelling tends to inform not just drug development but also the design of more informative clinical trials, where dosing strategies can be tested against model predictions before committing to large-scale studies.
