Semaglutide Not Working? What a Gene Study Found

For about 1 in 400 people, semaglutide may not be working because of their genes. People carrying rare damaging changes in GLP1R, the gene for the receptor semaglutide acts on, lost 2.4% of their body weight on the drug against 8.8% for closely matched people without them.¹

Tirzepatide mostly held up. Carriers who took it instead lost 13.3% compared with 15.2%, a gap too small for the study to tell apart from chance.¹ Nature Metabolism published the study of 13,725 patients on November 11, 2026.² Its semaglutide result is large and rests on very few people. Only 17 carriers took semaglutide, and 19 took tirzepatide.¹

What did the study test?

The study tested whether rare changes in two receptor genes alter weight loss on semaglutide or tirzepatide. It used the health records of people who started either drug in routine care.¹ Nobody was assigned a drug. The records came from the Helix Research Network, a program linking clinical-grade exome sequencing to electronic health records and insurance claims across 13 North American health systems and covering 351,818 people.¹

From that network the authors drew 13,725 adults with a BMI of 25 or more who began injected semaglutide or tirzepatide. Injected semaglutide is the drug sold as Ozempic and Wegovy, and tirzepatide is sold as Mounjaro and Zepbound, though the paper uses no brand names. Their outcome was the lowest weight recorded 6 to 12 months after starting, as a percentage of starting weight. People with earlier bariatric surgery, a pregnancy in the previous nine months or a cancer diagnosis in the previous year were left out.¹

Each carrier was compared with as many as 10 non-carriers on the same drug and the same highest dose category, matched for sex, ancestry group, type 2 diabetes and earlier GLP-1 drug use. Starting BMI had to fall within 3 points and follow-up within 30 days.¹ Carriers were always measured next to people on the same drug at the same dose level, so a lower dose cannot explain the gap.

What is GLP1R, and why would a change in it matter?

GLP1R is the gene carrying the instructions for the GLP-1 receptor, the protein on cells that semaglutide binds to. Tirzepatide binds the same receptor and a second one, the GIP receptor, whose gene is GIPR.¹ If GLP1R is damaged, semaglutide could in principle have less to work on. Tirzepatide would still have its second route.

The study sorted variants by how much harm they were predicted to do. Loss-of-function changes ranked as the most severe, expected to stop the gene from making a working receptor. Next came missense changes that swap a single building block in the receptor, scored by two prediction tools, AlphaMissense and REVEL, as damaging or possibly damaging.¹ Missense changes predicted to be harmless served as a control.¹ People carrying those showed no loss of response.¹

How much less weight did carriers lose?

Carriers lost about a quarter of the usual weight on semaglutide and close to the usual amount on tirzepatide.¹

Weight lost at 6 to 12 months	Carriers	Matched non-carriers	Response ratio (95% CI)	P value
Both drugs	8.2%	12.5%	0.65 (0.46 to 0.85)	0.0007
Semaglutide, 17 carriers	2.4%	8.8%	0.27 (0.01 to 0.68)	0.0007
Tirzepatide, 19 carriers	13.3%	15.2%	0.88 (0.68 to 1.08)	0.23

A response ratio of 0.27 means carriers on semaglutide lost 27% as much as similar people without the variants. In the authors' words, carriers achieved "roughly one-quarter of the typical weight loss."¹ Seventeen people make for a wide confidence interval, and this one runs from 0.01 to 0.68. Even at the top of that range, carriers would have lost about a third less. When the authors tested whether the carrier shortfall differed between the two drugs, they found it 4.4 percentage points larger on semaglutide, with an interval from 1.1 to 7.7 and P = 0.009.¹

For a person starting at 250 pounds, 2.4% is 6 pounds and 8.8% is 22 pounds, while on tirzepatide the same arithmetic gives 33 pounds versus 38. Those pounds are this page's conversion, not figures from the paper. The non-carriers' 8.8% comes from routine care in the first year of treatment, so it is not a trial result.

How common are these GLP1R variants?

Just 33 of the 13,725 people in the study carried one of these variants, about 1 in 400, and the paper rounds that to approximately 0.2%.¹ The drug groups of 17 and 19 add to more than 33 because three carriers took both drugs in turn and count in each.¹ Six of the 33 had a loss-of-function change and five a missense change scored as damaging. The other 22 had changes scored only as possibly damaging, and 15 of those 22 shared one variant, named rs146868158.¹

The paper found the difference between the drugs most pronounced for rs146868158 itself, with a semaglutide response ratio of 0.38 and a tirzepatide ratio of 1.13.¹ Nearly half the carriers behind the headline figure therefore share a single change that software rates only as possibly damaging.

What could this mean if semaglutide is not working for you?

For most people who lose little on semaglutide, often called non-responders, this study offers no explanation, since 13,692 of its 13,725 participants carried none of the qualifying GLP1R variants. Weight loss on these drugs "is heterogeneous across individuals," as the paper's first sentence puts it, and a separate April 2026 study from 23andMe researchers found sex, age and the choice of drug strongly associated with outcomes.¹³

The few carriers raise a different question, about switching. The three carriers who took semaglutide first and tirzepatide later are the closest the study comes to a switch. Their results are not reported on their own.¹ Switching between the two drugs is a prescribing decision, and our semaglutide vs tirzepatide comparison covers how they differ.

Everyone in the study started treatment through a health system, and the paper does not mention compounded products.¹ The study had no reason to ask whether a vial held what its label said. Someone using a gray-market vial does.

Why did tirzepatide still work for carriers?

The paper does not show why tirzepatide still worked, and the authors call their own explanation speculative. They suggest "partial compensation through GIPR co-agonism," meaning tirzepatide's action on the GIP receptor may make up for some of the lost GLP-1 signal.¹ Proving it would take direct experiments, they add. Carriers of the most severe changes did lose somewhat less on tirzepatide too. The paper calls it "a trend towards attenuation" and gives a response ratio of 0.67 with P = 0.015. The authors treat that as a hypothesis-generating observation in a small subgroup.¹

People with rare changes in GIPR itself, tirzepatide's second target, lost weight normally on both drugs by the study's tests. On semaglutide, 33 GIPR carriers lost 7.9% versus 9.5%, and on tirzepatide, 30 carriers lost 12.1% compared with 14.1%, so neither drug showed a gap large enough to reach statistical significance.¹

Did the semaglutide result hold up in other groups?

The semaglutide shortfall pointed the same way in two smaller groups, and it reached statistical significance in one. A later set of Helix patients, with records through March 31, 2026, included seven semaglutide carriers. They lost 4.1% while matched non-carriers lost 7.5%, a ratio of 0.54 with P = 0.004. The 11 tirzepatide carriers in that set showed no shortfall, at a ratio of 1.19.¹

All of Us, the NIH research program, supplied the second check. Its eight semaglutide carriers lost 2.6% compared with 5.7%, a ratio of 0.46. At P = 0.12, that gap could be chance. This group also counted oral semaglutide users, and only one tirzepatide carrier turned up, too few to test.¹

Do common GLP1R variants matter too?

Common GLP1R variants matter much less, according to the same paper. A widespread one called P7L was linked to slightly more weight loss, a response ratio of 1.05 per copy. Another, Ala316Thr, went the other way at 0.87 per copy. Both results come from the European ancestry group.¹

The April 2026 study in Nature from 23andMe researchers reached a similar view of common variants. Using self-reported data from 27,885 people taking GLP-1 drugs, they linked P7L, listed there as rs10305420, to about 0.76 kg of extra weight lost per copy, roughly 1.7 pounds, and tied a GIPR variant to nausea and vomiting on tirzepatide with no link to how much weight people lost.³ Nature's press summary calls the genetic effects modest.³

Should anyone get a genetic test before starting semaglutide?

The Helix paper does not call for testing before semaglutide, and nothing in it shows that choosing a drug by genotype improves results. The furthest the authors go is that carriers "may be preferentially suited to tirzepatide or other agents engaging non-GLP-1R pathways."¹

Several limits stand between this paper and a clinical test. Prediction software classified the variants, and the authors note such tools can be less reliable for membrane proteins in this receptor family. Carriers were few. The records captured adherence and weight imperfectly, and most participants were of European ancestry.¹ The Helix team found these variants with clinical-grade exome sequencing, a method that reads every protein-coding gene, while the 23andMe study analyzed common variants in self-reported 23andMe data.¹³

Who ran the study, and who paid for it?

Helix researchers led the study. The authors report no specific funding for it. Eight of the 21 authors, including the first and last, are Helix employees, and the others declare no competing interests.¹ Helix also markets a GLP-1 dataset from the same network, combining exome sequencing with health records, to drug developers.⁴ That does not make the result wrong, but the finding comes from a company with a commercial interest in the data behind it. The study data are not public, though qualified researchers at network sites or approved institutions can request access.¹

What is known and what is not?

One association in a small group is known, and most of what a person would need before acting on it is not.

Established by the study: in 13,725 treated adults, the 17 semaglutide users carrying a qualifying GLP1R variant lost about a quarter of the usual weight; tirzepatide users with the same kinds of variants lost close to normal; two smaller groups pointed the same way for semaglutide, one significantly; common GLP1R variants shift weight loss only slightly.¹

Not established:

Whether these variants damage the receptor at all. The classifications are computer predictions, and the authors flag them as less reliable for this protein family.¹
Why tirzepatide escapes. The GIPR explanation is the authors' speculation, and they say it needs direct experiments.¹
Whether choosing a drug by genotype leads to more weight loss. The paper makes no such comparison.¹
Whether carriers have more or fewer side effects. The paper reports no side effect data for them.¹
How the findings apply to people of other ancestries, since most participants were of European ancestry.¹
What happens past two years. The longest window checked ran 12 to 24 months, and its results sit in a supplement this page has not read.¹

Common questions

Can your genes keep semaglutide from working?

Rare changes in the GLP1R gene were linked to much smaller weight loss on semaglutide, the drug in Ozempic and Wegovy, in a November 2026 Nature Metabolism study. Carriers lost 2.4% of body weight against 8.8% for matched non-carriers, but only about 1 in 400 people carried one.

Does this apply to Ozempic, Wegovy, Mounjaro and Zepbound?

Yes, Ozempic and Wegovy are brand names for injected semaglutide, and Mounjaro and Zepbound for tirzepatide, the two drugs the Helix study tested. The paper itself uses no brand names.

What is GLP1R?

GLP1R is the gene that encodes the GLP-1 receptor, the protein semaglutide binds to. Tirzepatide binds that receptor and the GIP receptor, encoded by a separate gene, GIPR.

How common are GLP1R variants that weaken semaglutide?

GLP1R variants of this kind turned up in 33 of the 13,725 people in the Helix study, about 1 in 400. The paper rounds this to approximately 0.2%, and 15 of the 33 shared a single variant, rs146868158.

How much weight did GLP1R carriers lose on semaglutide?

The 17 carriers on semaglutide lost 2.4% of body weight at 6 to 12 months, against 8.8% for matched non-carriers. The authors describe it as roughly one-quarter of the typical weight loss.

Did tirzepatide work for people with GLP1R variants?

Tirzepatide worked nearly as well as usual, with 19 carriers losing 13.3% against 15.2% for non-carriers, a gap the study could not separate from chance. Carriers of the most severe changes showed a smaller shortfall the authors call hypothesis-generating.

Why would tirzepatide work when semaglutide does not?

The Helix authors suggest that tirzepatide partly makes up for a weakened GLP-1 receptor by also acting on a second target, the GIP receptor. They call the idea speculative and say it needs direct experimental evidence.

Do GIPR variants affect tirzepatide weight loss?

Rare GIPR variants did not significantly change weight loss on either drug in the Helix study. In the 23andMe study, a common GIPR variant was linked to nausea and vomiting on tirzepatide but not to weight lost.

How many people were in the semaglutide gene study?

The study included 13,725 adults on semaglutide or tirzepatide, drawn from a research network of 351,818 people across 13 North American health systems.

Was the semaglutide gene study a clinical trial?

No, the Helix study was observational, built from routine health records and insurance claims. Nobody was assigned a drug, and carriers were matched with similar non-carriers after the fact.

Did the semaglutide gene study account for dose?

Yes, every carrier in the Helix study was matched with non-carriers who reached the same highest dose category of the same drug. Matching also covered sex, ancestry group, type 2 diabetes, earlier GLP-1 use, starting BMI and follow-up time.

Did the semaglutide finding replicate?

The semaglutide finding partly replicated in two small groups, one from Helix and one from the All of Us program. The Helix group of seven carriers showed a significant shortfall, and the All of Us group of eight pointed the same way without reaching significance.

Is there a common gene variant that helps GLP-1 weight loss?

A common GLP1R variant called P7L was linked to a response ratio of 1.05 per copy in the Helix paper, about 5% more weight loss than in people without it. The 23andMe study put it at roughly 1.7 pounds of extra loss per copy.

Do GLP1R variants cause more side effects?

The Helix paper reports no side effect data for GLP1R carriers, so whether the variants change nausea or other effects is not known from it. Our GLP-1 side effects guide covers what the trials recorded.

Should I switch from semaglutide to tirzepatide if it is not working?

Not on the strength of this study alone, since the Helix paper compared groups and found a genetic explanation in about 1 in 400 people. A switch is a prescribing decision, and our semaglutide and tirzepatide pages carry each drug's trial record.

Who paid for the semaglutide gene study?

The Helix gene study had no specific funding, according to its authors, and eight of its 21 authors work for Helix, the genomics company whose research network supplied the data.

When was the semaglutide gene study published?

Nature Metabolism published it online on November 11, 2026, under the title "Rare GLP1R variants attenuate weight loss with semaglutide but not tirzepatide".

Sources

Levy ME, Schiabor Barrett KM, Telis N, Lagarde WH, Phatak PD, Maher JF, et al. Rare GLP1R variants attenuate weight loss with semaglutide but not tirzepatide. Nature Metabolism, 2026. https://doi.org/10.1038/s42255-026-01656-z. No specific funding; eight authors employed by Helix.
Read at source: the open-access PDF on nature.com, read in fragments through a fetch tool that caps quotation length. Every figure on this page was checked against a verbatim fragment of the Results, Methods, Discussion, Table 1 or Figure 1. The PDF retrieved prints its online date as a placeholder, so the date comes from source 2. Supplementary tables and figures were not opened, so the 12 to 24 month results are unavailable to this page. ↩ ↩ ↩ ↩ ↩ ↩ ↩ ↩ ↩ ↩ ↩ ↩ ↩ ↩ ↩ ↩ ↩ ↩ ↩ ↩ ↩ ↩ ↩ ↩ ↩ ↩ ↩ ↩ ↩ ↩ ↩ ↩ ↩ ↩ ↩ ↩ ↩ ↩ ↩
Rare GLP1R variants diminish the weight loss effect of semaglutide but not of tirzepatide. Research Briefing, Nature Metabolism, November 13, 2026. https://www.nature.com/articles/s42255-026-01661-2.
Read at source. A plain-language summary of source 1 written with its authors. Gives no figures and is used only for the paper's online date of November 11, 2026. ↩
Nature press release. Medicine: Genetic differences may influence GLP1 treatment outcomes (Nature). April 9, 2026. https://www.natureasia.com/en/info/press-releases/detail/9297. Summarizes Su QJ, Ashenhurst JR, Xu W, et al., Nature, April 8, 2026, https://doi.org/10.1038/s41586-026-10330-z.
Read via a reproduction. The press release was read at source; the paper itself was not, because nature.com refused both the article page and the PDF. The 0.76 kg figure, the GIPR nausea finding and the description of genetic effects as modest are the press office's, and the conversion to pounds is this page's arithmetic. ↩ ↩ ↩ ↩
Helix. GLP-1 focus area, Life Sciences. https://www.helix.com/life-sciences/focus_areas/glp-1.
Read at source, November 14, 2026. A company page, cited only to show that Helix markets a GLP-1 dataset from the same network to drug developers. ↩

Citing this page: Peptide Decoding. Is Semaglutide Not Working Because of Your Genes? For About 1 in 400. Figures from Levy et al., Nature Metabolism, published November 11, 2026, with the April 2026 23andMe study read via Nature's press release. Reviewed November 14, 2026. https://peptidedecoding.com/news/semaglutide-glp1r-gene-variants