Novo Nordisk, the company that makes Ozempic and Wegovy, took compounded and follow-on copies of its own weight-loss and diabetes drugs, ran them through immune-response assays, and reported that the copies carried different impurities than its branded products, some of which could provoke an immune reaction. The analysis, published online July 30 in Pharmaceutical Research ↗, covered semaglutide (the molecule in Ozempic and Wegovy) and liraglutide (Victoza and Saxenda).
Compounded versions are copies mixed by pharmacies rather than made on the originator's production line. In the United States they proliferated during the GLP-1 shortages, when the branded pens were hard to get and, for many people, unaffordable. The FDA allowed compounders to fill the gap while semaglutide and tirzepatide sat on its official shortage list, then began winding that allowance down once the shortages eased, which is the live regulatory fight this analysis lands in the middle of. A month of the branded drug can cost more than a full-time minimum-wage paycheck ↗, and telehealth companies still sell compounded semaglutide and tirzepatide at a fraction of the list price. Follow-on products are separate manufacturers' own versions made outside the originator's supply chain. Both sit next to the brand at the pharmacy counter, and both were the subject of this study.
The Novo team, led by Katharina Kopp and Morten Hach, used four lab techniques. A method called MAPPs (short for major histocompatibility complex II-associated peptide proteomics) exposes human immune cells to a drug and reads back which peptide fragments those cells display, a proxy for what the immune system might learn to attack. They added mass spectrometry to catalog impurities, a light-exposure test to see how the products degrade, and a fibrillation assay to check whether the peptide clumps into fibers. The copies came back with distinct impurity profiles: amino acids deleted or added to the peptide chain, plus impurities the analysis could not identify. Compounded semaglutide showed a significant jump in high-molecular-weight protein, meaning aggregated clumps, when exposed to light, and the liraglutide follow-ons were less physically stable than the originator.
The immune part is where the caution has to sit. In the MAPPs assay, dendritic cells taken from healthy donors and stimulated with the copies' impurities presented a different set and number of peptide fragments than the originator did. That is a laboratory signal that the impurities could, in principle, train the immune system to react. It is not a patient outcome. There were no patients in this study, no injected doses, and no measured immune reactions in a human body. The word the authors use is "potential."
The other thing to weigh is who ran the analysis. All ten authors work for Novo Nordisk, the maker of the originator drugs and the company with the most to lose from compounded and follow-on competition that undercuts a franchise worth tens of billions of dollars a year. That does not make the chemistry wrong. Deletions, additions, aggregation under light, and reduced stability are measurable facts, and a copy that degrades differently than the brand is a real reason to want independent testing. It does mean the interpretive leap, from a dendritic-cell readout to "could lead to an undesirable immune response," arrives from the party that benefits most if patients read it as a warning.
The impurity question is not about the molecule itself. On peptidemodel, the semaglutide ↗ and liraglutide ↗ cards describe the same peptide sequences the copies are trying to reproduce, both hitting the GLP-1 receptor ↗. The originator and the copy are aiming at the identical target with the identical active peptide. What this study says differs is everything riding alongside it: the deletion here, the addition there, the clump that forms under a lightbulb. Purity is a manufacturing property, not a property of the drug's design.
The honest bottom line is narrow. The study shows compounded and follow-on GLP-1 products are not chemically identical to the branded versions and can degrade in different ways, which is a genuine argument for scrutiny of who is mixing these copies and how. It does not show that any patient was harmed, and it comes from the one company that profits when patients hear the word immunogenic and go back to the pen. The missing pieces, independent laboratories running the same assays and real immunogenicity data from people actually taking the copies, are exactly what would turn a plausible concern into a proven one.