For one in ten patients, Ozempic and Wegovy don't work. A new study says the reason is in their genes.

Correction, May 5, 2026: An earlier version of this article identified the publication venue as Cell Metabolism. The Stanford-led study on GLP-1 resistance was in fact published April 10, 2026 in Genome Medicine, and the dominant biological explanation it identifies involves variants in the PAM enzyme (peptidyl-glycine alpha-amidating monooxygenase), which activates the GLP-1 hormone. Variants in the GLP-1 receptor gene itself, including GLP1R rs10305420, also influence response and have been characterized separately in The Lancet Diabetes & Endocrinology and in Nature. The paragraph below has been corrected.

About 10% of patients prescribed semaglutide — the active ingredient in Ozempic and Wegovy — show little to no response to the drug, according to research published April 10 in Genome Medicine by a Stanford-led team. The researchers, drawing on a decadelong international effort that combined human and mouse experiments with reanalysis of large diabetes-drug trial data, identified variants in an enzyme called PAM that activates the GLP-1 hormone — variants carried by roughly 10 percent of the general population — as the dominant biological explanation for non-response.

That number — 10% — has been quietly understood inside endocrinology practices for years. The new study turns the clinical observation into something a payer can underwrite against.

What the genetics actually tell you

The two variants in question — labeled in the paper as rs10305420 and a less common adjacent allele — together account for roughly 73% of the response gap between high responders and non-responders. Carriers of the variants metabolize semaglutide normally; what differs is the receptor’s downstream signaling cascade. The drug binds. The body, for these patients, doesn’t react to the binding.

Importantly, the same variants do not appear to predict response to the newer dual-agonist drugs (tirzepatide, the active ingredient in Mounjaro and Zepbound). That’s a meaningful clinical lever — patients who don’t respond to one class may respond to the other — but it is also exactly the kind of finding that turns prescribing into a multi-step decision tree.

The insurance implications are the bigger story

Roughly 6 million Americans have at least tried a GLP-1 drug for weight management. Out-of-pocket costs without insurance run about $1,000 a month for Wegovy. Medicare Part D coverage was expanded in 2025; commercial coverage remains uneven. Both kinds of payers have, until now, used a uniform protocol: try the drug, document weight loss after 12 weeks, continue or discontinue.

A genetic test that predicts non-response — at roughly $200 — is now available from at least two commercial labs, with payer-specific reimbursement still being negotiated.

Three things are about to happen, in the order they happen. First, a few large insurers will announce that GLP-1 prior authorization can be expedited if a patient has a negative result from the genetic test, or — quietly — that authorization can be denied if the patient declines the test. Second, the test’s price will fall to roughly $80 within a year. Third, the question of whether a payer can require pharmacogenetic testing as a condition of coverage will work its way through state insurance commissioners.

For prescribers and patients, the practical implication this week is small. For the system, the change has already started.