Quality-by-Design Oral Incretin Tablets

Evidence

Oral peptide delivery is both a biological and a manufacturing problem. Digestive conditions can destabilize a peptide, epithelial membranes limit absorption, and the large amount of an absorption enhancer can make a tablet difficult to produce consistently. Two independent studies used quality-by-design, or QbD, workflows to develop immediate-release tablets around sodium caprate (C10), an intestinal permeation enhancer. The newer study formulated tirzepatide; the second formulated semaglutide and provides commercially reusable full text.

The tirzepatide study, first published July 5, 2026, began by screening C10 through acid-neutralizing capacity, Caco-2 permeability experiments, and preliminary pharmacokinetic testing in rats. The researchers then defined a quality target product profile and critical quality attributes. Preliminary hazard analysis and failure-mode-and-effects analysis identified the binder-to-disintegrant balance as a critical material attribute and hammer-milling conditions as a critical process parameter. Face-centered central composite designs and multiple-response optimization were used to balance dissolution, powder flow, and tablet strength.

The optimized tirzepatide tablets had 0.58% friability and a Carr's index of 24, which the authors described as low friability and acceptable flowability. In dissolution media at pH 1.2, 4.0, and 6.8, formulations containing either 300 or 500 milligrams of C10 produced similarity-factor values above 50 against benchmark oral semaglutide tablets. This supports comparable laboratory release profiles under those test conditions; it does not establish equivalent absorption or therapeutic performance.

The animal result complicated the assumption that more enhancer necessarily produces more exposure. In beagle dogs, the 300-milligram C10 tirzepatide tablet produced higher systemic exposure than the 500-milligram formulation. For the 300-milligram tablet, mean peak concentration was 46.49 ± 23.79 nanograms per milliliter and mean area under the concentration-time curve to the last measurement was 1261.03 ± 690.44 hour-nanograms per milliliter. The abstract does not report group size, statistical significance, or absolute bioavailability.

The semaglutide study used a related QbD sequence: characterize the peptide and excipients, define target attributes, rank risks, and optimize tablet composition with a central composite design. Its full text reports that three semaglutide sources were poorly soluble from pH 1.2 through 5.0 but freely soluble at pH 6.8 or above. Four-week compatibility screening found that assay values generally remained between 95% and 105% with most tested excipients. Semaglutide mixed with sodium lauryl sulfate fell to about 0.2% assay at four weeks, while salicylic-acid mixtures had no detectable semaglutide. The authors therefore excluded those combinations, although they acknowledged that an assay cannot identify every degradation product.

Before tablet optimization, the semaglutide team orally gavaged four groups of five rats with semaglutide alone, semaglutide plus C10 at 100 or 200 milligrams per kilogram, or semaglutide plus SNAC at 200 milligrams per kilogram. Mean exposure to the last measured point rose from 58.9 hour-nanograms per milliliter for semaglutide alone to 230.5 with lower-dose C10 and 346.9 with higher-dose C10. The standard deviations were 76.4, 431.6, and 719.8, respectively, showing substantial variability around those means.

The final semaglutide formulation met the study's predefined criteria for hardness, disintegration, friability, and content uniformity. Its dissolution similarity factors against the reference product were 74.4 at pH 1.2, 74.7 at pH 4.0, and 71.3 at pH 6.8. Those are formulation-quality results, not evidence from a clinical comparison.

Analysis — Optimization Is Peptide-Specific

The cross-study pattern is that QbD can organize oral peptide development, but it cannot turn an enhancer into a universal recipe. Both teams translated a broad delivery problem into measurable attributes, ranked formulation risks, and used designed experiments rather than changing one ingredient at a time. That convergence supports an engineering inference: tablet mechanics, release, peptide stability, and permeation need to be optimized as a linked system.

The animal findings also point to a candidate-specific and potentially nonlinear relationship between C10 amount and exposure. Semaglutide exposure means increased across the two C10 doses tested in rats, whereas the lower-C10 tirzepatide tablet outperformed the higher-C10 tablet in dogs. This contrast is analysis, not a head-to-head result: the peptides, species, dosing formats, and studies differed. Still, it argues against using enhancer dose alone as a performance proxy. A stronger next experiment would vary C10 and matrix composition together, quantify intact peptide at the absorption site and in plasma, and test whether dissolution, local enhancer concentration, and systemic exposure move together. The useful emerging direction is a peptide-specific design space that includes both manufacturability and pharmacokinetics.

Limitations

Neither study demonstrates effectiveness or safety in people. The tirzepatide source was ingested only as a PubMed abstract, so this briefing cannot inspect animal numbers, variability for the 500-milligram group, statistical tests, detailed tablet composition, tissue findings, or the full set of process controls. Its benchmark-comparable dissolution result cannot substitute for bioequivalence, and greater exposure from one formulation does not by itself establish an optimal dose or safety margin.

The semaglutide paper provides fuller methods, but its animal screening used five rats per group and produced standard deviations larger than several mean exposure values. The compatibility study relied on retained assay rather than a validated related-substances method, so low-level or structurally altered degradants could have gone undetected. Dissolution testing used surfactant at two pH values to maintain sink conditions and was designed for formulation comparison rather than direct simulation of the gastrointestinal tract.

Finally, the studies used different peptides, species, enhancer amounts, and formulation details. Their exposure values cannot be pooled, and the opposite dose patterns cannot be assigned to peptide identity without controlled replication. Repeated-dose toxicology, local intestinal assessments, stability-indicating analytics, absolute bioavailability, and matched formulation comparisons would be needed before the design logic could be treated as translational evidence.