Do GLP-1 drugs work better for women than men?

TL;DR: the short version

• On average, women lost more weight than men on GLP-1 medicines. (GLP-1 medicines copy a gut hormone that helps you feel full.) A large 2026 review of 64 trials found this.
• The gap is small next to how much people vary. Women averaged about 11% weight loss and men about 7%, but plenty of men lost more than plenty of women. Your sex tells you very little about your own result.
• Age, ethnicity, starting weight and starting blood sugar made no big difference. That is the more reassuring finding, and it got less attention.
• Nobody knows yet why women did better on average. The researchers suggested three ideas (hormones, how the body handles the drug, and lower body weight) but did not test them, so the reason is not yet established.
• Comparing yourself with a partner tells you little. If your own progress worries you, talk to your prescriber. The options worth discussing with a physician are set out near the end of this page.

Educational content, not medical advice. GLP-1 receptor agonists are prescription-only medicines. No doses, titration schedules or regimen changes appear on this page. Nothing here is a reason to alter treatment; that conversation belongs with your prescriber.

---

This finding travelled fast, and it lost most of its content on the way. What was published as a careful subgroup analysis inside a large systematic review arrived in general coverage as "GLP-1s work better for women" - a headline that is defensible as a statement about two averages and misleading as a statement about any particular person.

It is worth taking apart properly, because the question underneath it - does this drug work as well for people like me - is one almost everyone starting treatment asks, and because the reasoning error involved is one that recurs across the whole of this field.

The finding: what the 2026 meta-analysis reported

A systematic review and meta-analysis led by the Johns Hopkins Bloomberg School of Public Health was published on 2 March 2026 in JAMA Internal Medicine. It examined 64 clinical trials of GLP-1 receptor agonists (medicines that copy GLP-1, a gut hormone that helps you feel full), including semaglutide and dulaglutide.

The sex analysis is a specific slice of that. It drew on 19,906 patients across six of the sixty-four trials, and reported mean weight loss of 10.88% of initial body weight in women versus 6.78% in men - a difference that reached statistical significance.

That is the whole of the headline finding, and it is a real result from a competently conducted analysis. Everything that follows is about what it means, which is a separate question from whether it is true.

What that figure is an average of

A mean is a summary of a distribution, and summarising discards information by design. The 10.88% figure is what you get if you take every woman in those six trials - the ones who lost a great deal, the ones who lost very little, the ones who discontinued early, the ones who responded unusually well - and collapse them into a single number. The same for the 6.78%.

Two distributions can have different means and still overlap almost entirely. That is not a hypothetical caveat here; it is the actual situation. Individual weight-loss response to GLP-1 receptor agonists is strikingly variable within any group you care to define, and the range within each sex in these trials substantially exceeds the roughly four-percentage-point gap between the two means.

So the correct reading of the finding is: if you took a large number of women and a large number of men and averaged each group, the women's average would be higher. That is a true statement about populations. It is not a statement about you, and it does not become one by being repeated.

What did not differ

Age, race, ethnicity, baseline BMI, baseline glycaemia

This is the part the coverage dropped, and it is the part more people should have heard.

Across the analysis, effectiveness was comparable across age, race, ethnicity, baseline BMI (starting weight relative to height) and baseline glycaemia (blood sugar at the start). No meaningful differences emerged on any of those axes.

That null result deserves to lead, because it answers the question most people are actually asking. Will this work as well for me at sixty as it would at thirty-five? The data say broadly yes. Will it work as well for someone of my ethnicity? Broadly yes. Does starting at a higher or lower BMI change how well it works? Broadly no. Does it matter whether my baseline glucose control is good or poor? Broadly no.

Null results are undersold generally, and they are badly undersold here. A drug class whose effectiveness holds up across age, ethnicity and baseline metabolic state is unusual and reassuring, and it is more clinically useful than a modest average difference between two groups that overlap heavily. The coverage inverted the priority because a difference makes a better headline than a similarity.

One honest caveat on the null findings themselves: "no significant difference found" is not identical to "no difference exists". Subgroup analyses are generally less well powered than the main analysis, and a small real difference could be present without being detected. The reasonable conclusion is that no large differences on those axes showed up, which is worth knowing and is not the same as proof of exact equivalence.

Plausible explanations

The authors proposed three explanations for the sex difference. All three are hypotheses generated to fit an observation, not mechanisms that have been demonstrated, and they should be held that way.

Body composition and fat distribution

The first is a possible synergistic interaction with oestrogen. Oestrogen influences appetite regulation, energy expenditure and fat distribution, and there is a plausible biological story in which it interacts with GLP-1 signalling to produce a larger effect. Plausible is the operative word: this analysis did not test it, and demonstrating it would require studies designed for the purpose - comparisons across menopausal status, for instance, which this pooled dataset was not set up to provide.

Sex differences in fat distribution are real and relevant to how weight loss presents, but a difference in where fat sits is not automatically a difference in how much is lost.

Pharmacokinetics and body weight

The second and third explanations are related, and together they are the least exotic reading of the finding.

Sex differences in drug handling - absorption, distribution, clearance - are documented for many medicines, and could contribute here. More straightforwardly, women in these trials had lower median body weight than men. For a medicine given at a fixed amount rather than scaled to body weight, a lighter person receives more drug relative to their body mass than a heavier person. If response tracks with exposure per kilogram, some or all of an apparent sex effect could be a body-weight effect wearing a sex-shaped disguise.

This matters for interpretation. A weight-based exposure difference is not really a statement about being female; it is a statement about being lighter, which correlates with being female but is not the same thing. Distinguishing the two requires analyses that account for body weight directly, and that is a question for the next study rather than something this one resolves.

Trial-population composition

The last explanation is the least interesting and possibly the most important: the six trials contributing to the sex analysis were not designed to answer it. Their sex ratios, recruitment criteria and populations varied, and pooled subgroup comparisons can pick up differences between trials rather than between sexes when the groups are unevenly distributed across studies.

None of this means the finding is wrong. It means that "women respond better to GLP-1s" is one of several readings the data permit, and the analysis was not built to choose between them.

Why an average is a poor prediction of your own result

This is the spine of the piece, and it generalises well beyond GLP-1 medicines.

Statistically significant means one specific thing: that a difference this large is unlikely to have arisen by chance given the sample size. With nearly twenty thousand patients, quite small differences clear that bar comfortably. Significance is a statement about confidence that a difference exists in the population sampled.

It does not mean the difference is large. It does not mean it is clinically important. And it emphatically does not mean it applies to any given individual. Those are four separate claims, and the word "significant" - which in ordinary English means "important" - smuggles them together. Most misreading of medical statistics happens in that gap.

The spread inside each group

Picture the two distributions rather than the two means. Both are wide. Both are centred at different points, and both cover almost the same territory. Substantial numbers of men in these trials lost more weight than the average woman. Substantial numbers of women lost less than the average man. The between-group gap is modest compared with the within-group range, which is the technical way of saying that knowing someone's sex tells you very little about what their result will be.

An average is a property of a population. It is not a prediction about a person, and it is not a target. This is the same reasoning error that leads people to read a trial's headline percentage as a promise about their own body, and it is why we treat dose-response curves as descriptions of average behaviour rather than personal forecasts.

Two further limits belong here. This is a meta-analysis of trial populations, and trial populations are selected - enrolment criteria exclude people, participants receive adherence support and structured follow-up that routine care does not reliably provide, and discontinuation is handled differently. Effect sizes from trials generally exceed what the same medicine produces in ordinary practice. And the sex analysis rests on six of the sixty-four trials, so it is a smaller and more particular evidence base than the headline "64 trials" suggests.

What this does and does not change for someone already on treatment

For almost everyone: nothing.

If you are a woman on a GLP-1 receptor agonist, this finding does not mean you should expect a specific result, and it does not mean a slower response indicates something is wrong. If you are a man, it does not mean the medicine is working less well for you than it should be. Both statements follow from the same point - the group average is not your expected value.

The scenario that prompts most of the searching is comparative. Two people in the same household start the same medicine at the same time, and one loses weight faster. That comparison feels like evidence and is almost useless as evidence, because sex is one input among many that differ between any two people: starting weight, body composition, what they eat, how they sleep, how much they move, other medications, thyroid and metabolic conditions, alcohol, stress, and the substantial unexplained individual variation that remains after accounting for all of it. A four-point difference in group means is a small contributor sitting inside that noise.

What it emphatically is not is a reason to change anything on your own. This page contains no doses and no regimen suggestions by design. If your response is not what you or your prescriber expected, the useful step is a conversation with the clinician managing your treatment, who can look at the actual inputs. If weight loss has stalled after an initial period, that is a distinct and well-described pattern with its own explanations - covered in why GLP-1 weight loss plateaus.

When a difference in response is worth raising with a clinician

Some situations do warrant a conversation, and the distinction is between slower than a comparison you have invented and different from what your own clinical picture predicts.

Worth raising: no meaningful change over a sustained period despite consistent use; side effects severe enough to affect eating, hydration or daily function; a response that changes direction after being stable; new symptoms you cannot attribute; or a concurrent condition or medication that might be interacting. Also worth raising is the situation where the goal has shifted - where preserving muscle or managing a specific metabolic marker has become more relevant than the number on the scale.

Not worth acting on alone: a partner, sibling or colleague losing weight faster. That comparison is not a clinical signal.

Two adjacent questions come up often enough to separate out. If the underlying context is PCOS or fertility, that is a distinct indication with its own evidence, covered in GLP-1s for women with PCOS and fertility - it is not the same question as the one on this page. If the context is the menopausal transition, where changing body composition and weight distribution complicate the picture, see GLP-1s and menopause weight. And there is a broader piece, peptides in men vs women, on how peptides in general are studied and discussed differently in men and women, which is worth reading if this article's statistical point interests you more than its specific subject. For the underlying pharmacology, how GLP-1 agonists work is the primer.

Where Vivre Labs fits: options to discuss with a physician

If this article has you thinking about GLP-1 medicines for yourself, the next step is a conversation with a doctor, not a decision based on a group average. These are the compounds most often raised in that conversation. None of this is a prescription, and which option (if any) suits you depends on your health, your history and your blood tests.

• Semaglutide. A GLP-1 medicine, and one of the drugs studied in the review above. FDA-approved for type 2 diabetes and for chronic weight management. Semaglutide at Vivre Labs.
• Tirzepatide. Works on two hormone signals: GLP-1 and GIP (a second gut hormone involved in blood sugar and fat handling). FDA-approved for type 2 diabetes and for chronic weight management. Tirzepatide at Vivre Labs.
• Retatrutide. Works on three signals: GLP-1, GIP and glucagon (a hormone that affects how the body uses energy). It is investigational: still in clinical trials and not approved anywhere. Eli Lilly has said it plans to apply for US approval in early 2027. Its page explains where it stands. Retatrutide at Vivre Labs.

One honest note: the review on this page did not look at tirzepatide or retatrutide by sex, so nothing here suggests either one works differently for women or for men.

Vivre Labs offers access, and we recommend starting with baseline bloodwork so any decision rests on your own numbers. See what to test before you start. For a plain side-by-side of these medicines, read tirzepatide vs semaglutide vs retatrutide vs CagriSema.

Frequently asked questions

Does this mean GLP-1s work better for women? On average, in this pooled trial data, women lost more weight than men. But the distributions overlap heavily - many men lost more than many women - so the average does not predict an individual outcome. It is a real population-level finding and a poor personal forecast.

Why did the women in the analysis lose more? The authors offered three possibilities: a synergistic interaction with oestrogen, sex differences in how the drug is handled, and women's lower median body weight producing greater exposure relative to body mass. All three are hypotheses proposed to explain an observation. None was tested in this analysis, and the third would make the finding partly about body weight rather than sex.

I'm losing less than my partner on the same medicine. Is something wrong? Probably not, and sex would be a small part of the explanation even if it contributed. Two people differ on many inputs that affect response, and individual variation is wide even after accounting for the measurable ones. If you are concerned about your own response, raise it with your prescriber rather than adjusting anything independently.

Does age, ethnicity or starting BMI affect how well GLP-1s work? In this analysis, effectiveness was comparable across all of those. That is the more reassuring finding and the one that got less attention. The usual caution applies - subgroup analyses are less well powered than main analyses, so "no large difference detected" is the accurate version rather than "proven identical".

---

Related reading

• GLP-1s for women with PCOS and fertility - a distinct indication, not the same question
• Why GLP-1 weight loss plateaus - when a response changes over time
• GLP-1s and menopause weight - body composition through the transition
• The GLP-1 dose-response curve - average behaviour, not a personal forecast
• How GLP-1 agonists work - the underlying pharmacology

Part of GLP-1 & Metabolic Health.

Explore it with verified supply. Every Vivre Labs lot is independently tested and the certificate of analysis is published, so you know exactly what is in the vial. Explore semaglutide.

Educational content, not medical advice. GLP-1 receptor agonists are prescription-only medicines and should be used only under clinical supervision. No doses, titration schedules, administration methods or regimen changes appear anywhere on this page. Nothing here is a basis for altering treatment. Claims for the Malaysian market are subject to MAB requirements. Position stated as of 13 September 2026.

References

• Systematic review and meta-analysis led by the Johns Hopkins Bloomberg School of Public Health, published 2 March 2026 in JAMA Internal Medicine; 64 clinical trials of GLP-1 receptor agonists including semaglutide and dulaglutide.
• Sex analysis within that review: 19,906 patients across six trials; mean weight loss 10.88% of initial body weight in women versus 6.78% in men; difference statistically significant.
• Same analysis: effectiveness comparable across age, race, ethnicity, baseline BMI and baseline glycaemia.
• Authors' proposed explanations for the sex difference - possible synergistic interaction with oestrogen; sex differences in drug handling; women's lower median body weight producing greater exposure relative to body mass. Presented by the authors as hypotheses; none tested within the analysis.
• Interpretive note: statistical significance indicates a difference unlikely to be attributable to chance given the sample; it does not indicate effect magnitude, clinical importance, or applicability to an individual. Within-group variation in these trials exceeds the between-group difference in means.
• Methodological note: meta-analysis of clinical-trial populations, which are selected by enrolment criteria and supported by trial-specific adherence and follow-up procedures that differ from routine care. The sex analysis draws on six of the sixty-four included trials.