Why GLP-1 Drugs Work Better for Some People Than Others

If you are wondering why GLP-1 drugs work better for some people, the short answer is this: genetics probably plays a part, but it is not the main practical explanation in most cases. A new Nature paper suggests common variants in GLP1R and GIPR influence weight loss and nausea or vomiting, but the effects are modest, and non-genetic factors such as drug choice, dose, duration, sex, diabetes status, and tolerability still matter greatly. [1] This article covers three things: what the paper actually found, what most coverage gets wrong, and what patients should do if treatment seems to be underperforming. This applies to people using semaglutide or tirzepatide for weight management. It is not a substitute for individual prescribing advice.

The short answer

The paper analysed 27,885 people who reported using semaglutide or tirzepatide-based drugs and found a significant association between a GLP1R variant and greater weight-loss efficacy, roughly 0.76 kg more weight loss per effect allele. It also found associations between receptor-gene variation and nausea or vomiting, with a GIPR signal that appeared particularly relevant to tirzepatide. [1]

The paper itself says the genetic effect sizes are modest. [1] Independent expert reaction has been similarly cautious: current evidence does not justify routine clinical decision-making based on these variants, because non-genetic factors still explain more of the observed variation than the common variants identified so far. [2]

What the new Nature paper found

The cohort was large, but not broadly representative. Participants were mostly female (82.4%), median age was 52, and 78.3% were of European ancestry. [1] The main efficacy analysis was conducted in a European-ancestry subset for GWAS purposes. [1]

The study found that treatment efficacy depended strongly on sex, drug type, time on treatment, and dose, with age having a weaker but measurable effect. [1] Women tended to lose more than men before correction, tirzepatide performed better than semaglutide on average, and type 2 diabetes status predicted lower weight-loss efficacy. [1]

The study also compared self-reported data with electronic health record data in a smaller subgroup. Self-reported weight loss looked materially larger than EHR-recorded weight loss, with a median BMI change of -11.8% in surveys versus -5.79% in EHR data. [1]

Why the paper matters

It matters because the biology is plausible. A receptor-gene signal for a drug acting on that receptor is much more interesting than a remote association with no obvious mechanistic relationship. It matters because the sample size is unusually large for this kind of obesity pharmacogenetic work. [1] And it matters because clinicians already know, from ordinary practice and major trials, that response to GLP-1-based therapy varies widely. [3,4,5,6]

In STEP 1, semaglutide 2.4 mg produced large mean weight loss, but with obvious inter-individual spread. [3] In SURMOUNT-1, tirzepatide produced larger mean weight loss again, but not uniformly. [4] None of that means every person will behave like the mean trial participant.

The paper addresses a real problem. It just does not solve it. In obesity medicine, the field likes single elegant explanations more than it likes awkward clinical reality.

What most articles miss

Response is not just a receptor story. The study's own modelling shows response depends strongly on sex, drug type, dose, and time on treatment. [1] Even within the paper, the answer is already wider than DNA.

The phenotype is noisy. Much of the study depends on self-report. The EHR comparison showed materially lower recorded BMI change than the survey data. [1] Sophisticated genetics does not cancel out soft outcome measurement.

Effect size is not the same as clinical usefulness. An extra 0.76 kg per allele is interesting scientifically. [1] It is not yet enough to change routine prescribing in a clinic tomorrow morning.

The question patients ask is different. Patients are rarely asking, "What is my pharmacogenomic profile?" They are asking, "Why is this not working for me?" That question may reflect inadequate titration, shortened treatment due to side effects, unrealistic time expectations, type 2 diabetes, or a switch decision made too late.

Maintenance and discontinuation are barely addressed. The study is about on-treatment efficacy and side effects. It says very little about what happens when appetite returns, what determines durable maintenance, or how to manage people whose initial success is followed by a plateau or rebound. That is a major clinical omission.

What the study can and cannot tell us

What it can tell us

• Common genetic variation in receptor pathways contributes to inter-individual variability in efficacy and side effects. [1]
• Non-genetic factors matter a great deal. The paper's linear model explained about 21.4% of variance in percentage BMI loss using non-genetic variables such as age, sex, starting BMI, drug type, dose, and time on treatment. [1]
• Type 2 diabetes status materially reduced expected weight-loss response. [1]

What it cannot tell us

• That genetic testing should now be routine before starting semaglutide or tirzepatide.
• That a disappointing response is "in your genes".
• That severe nausea is a useful sign of success.
• What the right clinician should do for the individual patient in front of them next week.

What to do if your GLP-1 drug is not working well

The first step is not panic, and it is not buying a gene test. The first step is to ask a more disciplined set of questions:

• Are you on the intended drug and the intended dose?
• Have you been on that dose for long enough?
• Have side effects limited escalation?
• Was the treatment started for obesity, diabetes, or both?
• Are you comparing your progress with trial averages that do not fit your starting point?
• Has nausea produced chaotic eating rather than therapeutic appetite control?
• Have you plateaued after initial benefit, or never meaningfully responded at all?
• Is there a clinical case for switching from semaglutide to tirzepatide?

A practical timeline if treatment seems to be underperforming

First 2 weeks

Focus on tolerance. It is usually too early to decide you are a non-responder. Early questions are about side effects, hydration, meal size, constipation, nausea, vomiting, and whether the patient can remain on treatment safely. [7,8]

Weeks 3 to 6

This is the period in which poor tolerance can quietly become poor treatment. If dose escalation is repeatedly interrupted by nausea or vomiting, the issue may be tolerability rather than lack of biological effect. Some people stop progressing because they cannot stay on a useful dose long enough. [7,8]

Beyond 6 weeks

If a patient has had enough time at an effective dose with tolerable side effects and reasonable adherence, but still shows weak response, it becomes more reasonable to think about genuine low responsiveness, type 2 diabetes-related attenuation of weight-loss effect, or whether another agent may be better. [1,4,6]

When to involve your clinician

Speak to your clinician promptly if: you cannot escalate because nausea or vomiting is persistent; you have minimal benefit despite an apparently adequate dose and duration; you have type 2 diabetes and your expectations were based on obesity-trial averages; you are considering switching drugs; or you are thinking about stopping because of poor tolerance or poor response.

When to seek urgent help

Seek urgent medical assessment for severe or persistent vomiting with inability to keep fluids down; signs of dehydration; severe abdominal pain especially if persistent or associated with vomiting; symptoms concerning for gallbladder disease or pancreatitis; or allergic symptoms such as facial swelling or breathing difficulty. These red flags matter more than genomics. [7,8]

Bottom line

This is a useful paper and it deserves attention. It shows that common variants in GLP1R and GIPR are associated with differences in GLP-1 efficacy and side effects. [1] But it does not support the louder conclusion many people will want to draw: that obesity prescribing is now becoming a straightforward genomic sorting exercise. The better conclusion is narrower and more clinically honest. Genetics probably explains some of why GLP-1 drugs work better for some people. But most of the practical work still lies elsewhere: in dose, duration, tolerability, drug selection, diabetes status, and proper clinical interpretation.