The Body That Won't Cooperate
For the better part of four years, the pharmaceutical industry has been operating on a simple premise: obesity is a drug problem, and it now has a drug solution. Ozempic. Wegovy. Mounjaro. Zepbound. The names have become cultural shorthand for a new era in medicine — the era in which a weekly injection could do what decades of diet culture, surgical intervention, and behavioral modification had failed to do at scale. The market agreed. Analysts project the GLP-1 receptor agonist sector will grow from roughly $73 billion in 2026 to more than $254 billion by 2034. Novo Nordisk and Eli Lilly have reordered the global pharmaceutical hierarchy. The revolution, everyone agreed, was here.
Then came a study that quietly complicated the story.
Published in April 2026 in the journal Genome Medicine, a decade-long international effort led by researchers at Stanford Medicine, Adelaide University, ETH Zürich, and the University of Oxford identified something the industry had not been forced to seriously confront: a significant portion of the population may be biologically incapable of responding to GLP-1 drugs the way the clinical trials promised. Not because they aren't trying. Not because they're non-compliant. Because their genetics won't allow it.
The condition has a name now. Researchers are calling it GLP-1 resistance.
What the Body Is Doing Wrong
To understand the finding, it helps to understand what GLP-1 drugs actually do. GLP-1 — glucagon-like peptide-1 — is a hormone produced naturally in the gut after eating. It tells the pancreas to release insulin, slows the emptying of the stomach, and signals to the brain that you're full. Drugs like semaglutide (the active ingredient in both Ozempic and Wegovy) work by mimicking this hormone, amplifying a system the body already uses to regulate blood sugar and appetite.
The Stanford-led team focused on a different part of that system: an enzyme called PAM, short for peptidyl-glycine alpha-amidating monooxygenase. PAM is, in the words of the study's senior author Anna Gloyn of Stanford Medicine, "a truly fascinating enzyme" because it's the only one in the human body capable of a chemical process called amidation — a modification that increases the potency and half-life of biologically active peptides. Without it, those peptides, including GLP-1, don't reach their full biological effect.
The researchers identified two specific variants in the PAM gene — labeled p.S539W and p.D563G — that handicap this enzyme. In people carrying these variants, GLP-1 levels in the blood are paradoxically higher than in non-carriers after eating. But the hormone isn't working properly. The signal is there; the response isn't. More GLP-1 circulating in the blood, less biological result.
This is the counterintuitive core of the finding: the problem isn't a shortage of the hormone. The problem is that the hormone has become, in some fundamental sense, illegible to the body.
The Scale of the Problem
Those PAM variants are carried by roughly 10% of the general population, according to the research team. That is not a marginal rounding error. It is one in ten people.
To confirm that this genetic quirk had real clinical consequences, the researchers turned to data from three clinical trials encompassing 1,119 participants with Type 2 diabetes. The results were unambiguous in one direction and unsettled in another. On blood sugar control — measured through HbA1c levels, a standard marker of long-term glucose regulation — the gap was stark. After six months of GLP-1 treatment, roughly 25% of non-carriers reached recommended HbA1c targets. Among carriers of the p.S539W variant, only 11.5% did. Among carriers of the p.D563G variant, the figure was 18.5%. In one analysis, the glucose-lowering effect of GLP-1 drugs was reduced by as much as 44% in people carrying PAM variants.
Notably, these variants did not blunt the effectiveness of other common diabetes medications — metformin, sulfonylureas, DPP-4 inhibitors all worked normally. The resistance appears specific to the GLP-1 pathway.
The weight loss question, however, is where the science gets more complicated — and where the study's authors are careful not to overclaim. The study was designed around blood sugar regulation, not obesity treatment. Only two of the clinical trials analyzed contained weight data, and those showed no significant difference in weight loss between carriers and non-carriers. Gloyn herself acknowledged the data is too limited to draw conclusions on that front. The weight loss doses of these drugs are substantially higher than the diabetes doses, which may matter. Whether PAM variants impair the appetite-suppression mechanism the same way they impair glucose regulation remains, as of this writing, an open question.
That open question matters enormously, because it sits at the center of the industry's largest growth market.
A Gold Rush Built on Averages
The GLP-1 revolution was built on clinical trial data, and clinical trial data is built on averages. The landmark STEP trials for semaglutide showed an average weight loss of around 15% of body weight over 68 weeks — an outcome that hadn't been seen in obesity pharmacology before. The SURMOUNT trials for tirzepatide went further, with some cohorts averaging over 20%. These numbers reshaped how doctors, insurers, and investors understood the treatable horizon for obesity.
But averages conceal distributions. Behind every impressive mean is a population of hyper-responders pulling the number up and non-responders pulling it down. Clinical experience has always reflected this: physicians treating patients with GLP-1 drugs have long observed what lead study author Mahesh Umapathysivam described as "a huge variation in response to these GLP-1-based medications." Until now, that variation had no genetic anchor. The Stanford team's work begins to provide one.
What makes the finding structurally significant is not just the PAM discovery in isolation. It is part of a broader, accelerating field. In early April 2026, a separate large-scale genome-wide association study conducted with 23andMe data and published in Nature identified additional genetic predictors of GLP-1 response and side-effect profiles, including variants in the GLP-1 receptor gene itself and the GIP receptor gene. The picture emerging from pharmacogenomics is not of a miracle drug that works for everyone — it is of a drug class whose effects are meaningfully modulated by the genome.
The industry has not been particularly eager to engage with this complexity. A market built on the idea of a universal solution has limited commercial incentive to publicize its exceptions.
Access Was Already Unequal
Before the genetics, the access problem was already severe. The GLP-1 gold rush happened largely for people who could afford it. Without insurance, injectable GLP-1 drugs cost anywhere from $900 to $1,300 per month at standard retail pricing. The Trump administration's Most Favored Nation pricing push in late 2025 nudged some cash-pay prices toward $350 per month through manufacturer programs — still a figure that forecloses the treatment for a substantial portion of the population.
Insurance coverage has, if anything, been moving in the wrong direction. By 2026, over 41 million Americans had no commercial insurance coverage for Wegovy. California's Medi-Cal program ended coverage for GLP-1 drugs prescribed solely for weight loss effective January 2026. Only 19% of firms with 200 or more employees included weight-loss GLP-1 coverage in their health plans in 2025; the figure rose to 43% only at firms with 5,000 or more employees — meaning smaller employers, who cover a disproportionate share of lower-wage workers, largely aren't offering it.
The people most likely to be obese in America — those in lower-income brackets, those with less consistent access to healthcare, those working for small employers — are also the people least likely to have access to the drugs marketed as the solution to their condition. The GLP-1 revolution has, at its current trajectory, been a revolution primarily for people who were already relatively well-positioned within the healthcare system.
The genetic finding does not create this inequity. But it deepens its stakes. If genetic testing eventually becomes part of the prescription workflow for GLP-1 drugs — identifying likely non-responders so clinicians can try other approaches sooner — that testing infrastructure will not be evenly distributed. The patients most likely to be failed by the drug, and least likely to get the genetic clarity that might redirect their care, may be the same people.
What Comes Next
The Stanford team is clear-eyed about the limits of their work. The mechanism by which PAM variants produce GLP-1 resistance has not been fully established. "That is the million-dollar question," Gloyn said. The study showed that the GLP-1 receptor's ability to bind the hormone appears normal in PAM variant carriers — the problem emerges somewhere downstream in the signaling cascade, but its precise location has not been pinned down. Future research will need to identify the mechanism before pharmacological workarounds can be rationally designed.
What the study does definitively accomplish is something arguably more valuable in the short term: it identifies a genetic marker that predicts treatment failure. That is the first step in what researchers call precision medicine — not a single treatment for everyone, but a system that matches patients to therapies based on biological reality rather than statistical probability.
Umapathysivam put it plainly: "This is the first step in being able to use someone's genetic make-up to help us improve that decision-making process." For the estimated 10% of patients who have been or will be prescribed GLP-1 drugs for diabetes and found them failing to meet blood sugar targets, that genetic explanation matters. It converts a frustrating clinical mystery into an actionable data point.
The broader obesity question — whether PAM variants affect weight loss at the doses used to treat obesity — will require a new wave of studies, including prospective trials with genetic stratification built in from the start. A trove of clinical trial data on this question likely already exists inside pharmaceutical companies that have run the large GLP-1 obesity trials. Whether that data gets analyzed through a pharmacogenomics lens is partly a scientific question and partly a commercial one.
The GLP-1 market is now large enough that even a 10% non-response rate represents millions of patients and billions of dollars in therapeutic dead-ends. The incentive to understand that failure — and to develop next-generation approaches for those patients, whether through alternative drug classes, combination therapies, or pharmacogenomically guided dosing — is not purely altruistic. It is a market opportunity dressed up as a scientific problem.
That framing is neither cynical nor reassuring. It is simply the world in which obesity medicine now operates: one where the science is genuinely advancing, the money is genuinely enormous, and the distribution of both has never been particularly fair.
The body that won't cooperate with a $100-billion industry's best drug is, in the end, just telling the truth about itself. The more interesting question is whether the system around it will listen.
