GLP-1s: what is behind the variation in efficacy?

Since the approval of semaglutide (Wegovy; Novo Nordisk) as a weight-loss injection in the United States in 2021, the effect of GLP-1 medicines on the treatment of obesity has been undeniable. In the UK alone, it is estimated that 8 million adults have taken or are considering taking a GLP-1^​1​.

Following the arrival of Wegovy tablets in June 2026 — and more oral GLP-1s on the horizon — this impact could be even greater.

However, studies have shown variation in the efficacy of GLP-1s across sex, age and ethnicity, among other factors.

Research published in June 2026 showed patients taking the medicines experienced dramatically different weight-loss outcomes: while some lost more than 25% of their body weight, others lost less than 5%^​2​.

The study results also revealed that female and white patients taking semaglutide or tirzepatide (Mounjaro; Eli Lilly) were more likely to achieve significant weight loss than male, black and Hispanic patients^​2​.

Findings from the 2024 SELECT trial, which included more than 17,000 people who were overweight or obese, showed that younger participants taking semaglutide experienced progressively greater mean weight loss than older participants; although, the differences between age groups were small^​3​.

Meanwhile, exclusive data shared with The Pharmaceutical Journal by Patientric, a platform that captures data from private in-person weight management consultation records, suggests that differences between groups may be more to do with those who remain on treatment, rather than weight-loss outcomes.

The figures showed that patients from the most deprived fifth of the country are half as likely to still be on treatment at a year compared with those in the least deprived fifth (11.5% versus 21.5%). Some 20.2% of white patients are still on treatment at a year compared with 11.4% of black patients, and only 13.2% of under-40s are still on treatment at a year, compared with 21.8% of those aged over 40 years.

Those with three or more health conditions were also found to be the most persistent group, with 23.2% staying on the treatment for at least a year compared with 16.5% of those with no health conditions.

Do GLP-1s work differently in different populations, and to what extent are these differences because of access to and experience of treatment?

Sex differences

In the June 2026 study, which included 20,678 patients taking tirzepatide or semaglutide, researchers found differences in weight-loss response by sex^​2​. Patients were divided into five groups according to their weight-loss outcomes, with ‘high responders’ losing more than 15% of their body weight and the ‘minimal weight-loss’ group losing less than 5% over the two-year study period^​2​.

Among patients taking tirzepatide, women accounted for 74.9% of the high-response group, compared with 59.2% of the minimal-weight-loss group, while men accounted for 25.1% and 40.8%, respectively^​2. A similar pattern was seen with semaglutide, where women accounted for 80.3% of high responders and 58.5% of minimal responders^​2​. However, the reasons for this remain unclear.

Katrina Bicknell, professor of pharmacology and pharmacy education at the University of Reading, says the differences may be owing to body composition, and the effects of oestrogen and testosterone on fat and muscle.

“Generally, men tend to lose weight faster than women and it is really interesting that women are seen to have greater weight loss outcomes when they use GLP-1 medicines,” she adds.

Bicknell explains that a man — on average — tends to have more muscle and less body fat than a woman of the same size. “GLP-1 medicines seem to preferentially reduce fat mass over lean body mass; if women have a higher percentage of their body weight as fat mass, greater weight losses might be seen compared to men who have less fat mass to lose,” she continues.

However, the results of a 2025 meta-analysis of 14 randomised controlled trials highlighted that the difference in weight loss between the sexes was relatively small. Across the studies, women taking GLP-1 receptor agonists lost an average of 1.04kg more than men, equivalent to a 1.69 percentage-point greater reduction in body weight^​4​.

Referring to the analysis, Amira Guirguis, chief scientist at the Royal College of Pharmacy, describes the differences as “modest”, adding that there is “substantial overlap between individuals”.

“Sex tells us something about differences at a population level, but very little about how a particular person will respond,” she says.

Age

As noted, there was some evidence in the SELECT trial that suggested that age may have an impact on weight-loss outcomes with GLP-1s, with younger people potentially more likely to lose more weight.

Bicknell says hormones could be responsible for differences in weight-loss responses at different ages, such as menopause.

“Reduced oestrogen levels seen in post-menopausal women may reduce the weight-loss response, whereas this is not observed in women taking hormone replacement therapy,” she says.

Although Guirguis notes: “Age may have some influence on average weight loss but it is not a particularly useful predictor on its own.”

Indeed, studies have suggested age does not have a great bearing on the effectiveness of GLP-1s. The results of a study published in JAMA Internal Medicine in March 2026, which examined 64 clinical trials involving semaglutide and dulaglutide, showed that the medicines have similar effectiveness in people aged under 65 years and those aged 65 years and over.

However, there is still a lack of research specifically examining the impact of obesity medications in older adults. Authors of another review, published in 2025, concluded that relatively few studies have focused on adults aged 65 years and over. It also found that older adults had a higher prevalence of side effects on the stomach and intestines when taking the medications.

There are also particular considerations when prescribing weight-loss medication to older adults, Bicknell says. This is because muscle mass naturally declines with age, while further weight loss with GLP-1s could increase the risk of frailty.

However, Guirguis says age should not necessarily be viewed as a barrier to treatment: “Older adults can still achieve clinically meaningful weight loss, although for older or frail people the focus should not be on the number on the scales alone. Maintaining muscle mass, nutrition and physical function also matters.”

Diabetes

Evidence suggests that people with type 2 diabetes mellitus (T2DM) lose less weight on GLP-1s than those who do not have the condition.

This pattern has been observed with both semaglutide and tirzepatide. In the STEP 1 trial, which included 1,961 people without T2DM, participants taking semaglutide lost an average of 14.9% of their body weight after 68 weeks^​5​.

In the STEP 2 trial, which included 1,210 people with T2DM, participants taking the same dose of semaglutide lost an average of 9.6% after 68 weeks^​6​. Similar differences were noted in the SURMOUNT trials of tirzepatide^​7​.

The reasons for this difference are not fully understood. People with T2DM may have insulin resistance, hyperinsulinemia (i.e. higher than normal levels of insulin in the blood) and a longer history of obesity and metabolic disease, while some diabetes medications, including insulin and sulfonylureas, can promote weight gain.

T2DM is also associated with a reduced incretin effect, meaning that the body’s natural hormonal response to food is impaired. These factors may contribute to a metabolic environment in which weight loss is more difficult to achieve.

Bicknell says a reduction in insulin-producing beta cells in the pancreas could also play a role. GLP-1 medicines stimulate these cells to release insulin, alongside reducing appetite and slowing gastric emptying. In T2DM, reduced beta-cell function could therefore affect how the medicines influence insulin production and insulin-dependent glucose metabolism. Although it is unclear how much this contributes to the difference in weight loss seen in clinical trials.

“It is feasible that this limits the effects that GLP-1 medicines have on the levels of insulin and insulin-dependent glucose metabolism, even if appetite is suppressed and absorption of nutrients is slowed,” she says.

Bicknell adds that this is not to say GLP-1s do not produce clinically meaningful weight loss in people with T2DM.

Andy Kenneth Edwards

Race and ethnicity

Studies suggest there is a significant difference in weight loss between those from diverse racial and ethnic groups.

In the June 2026 study, among patients taking tirzepatide who lost more than 15% of their body weight, 91.2% were white and 6.0% were black. By comparison, among those who lost less than 5% of their body weight, 79.8% were white and 10.9% were black^​2​.

However, Guirguis warns: “These findings need to be interpreted very carefully. Race and ethnicity should not be treated as straightforward proxies for biology or genetic ancestry. In real-world studies, they can be closely linked to factors, such as access and affordability of treatment, prescribing, dose escalation, adherence and persistence, as well as wider socioeconomic circumstances.

“This makes it very difficult to separate a potential biological difference from differences in people’s experience of treatment. Current evidence does not support using race or ethnicity to predict an individual’s response to a GLP-1 medicine or to determine access to treatment.”

Bicknell adds: “We really don’t have enough evidence to say that genetic ancestry from race or ethnicity might lead to different outcomes for GLP-1 medicines. The clinical trials did not recruit sufficient numbers of ethnically diverse participants and real-world studies also fail to show whether there are differences in responses because ethnic minority groups are less likely to be prescribed GLP-1 medicines.

“It is feasible that genetic ancestry might lead to differences in body composition, increased susceptibility to T2DM, metabolic conditions or certain genetic variants, but there is insufficient evidence to take this beyond the hypothetical.” 

There are early indications that genetic variants could play a role. The results of a study published in March 2026 suggest a link between variants of the PTPRU gene and the response to GLP-1 drugs, with patients of mixed races experiencing lower weight loss when taking semaglutide^​8​.

However, Bicknell says that researchers are only beginning to understand the genetics behind variations in response.

She adds that researchers are investigating genetic variants that could influence how people respond to GLP-1s, including those involved in regulating appetite, the brain’s reward response to food, insulin secretion and how the body uses energy. However, she says there is currently “no front runner” that can explain differences in weight-loss response.

“Just as there is no single gene that predicts an individual’s likelihood of developing obesity, it is very unlikely that a single gene variant will be identified that will identify who would most benefit from a particular GLP-1 medicine,” Bicknell says.

Access to GLP-1s

As Guirguis highlights, socioeconomic factors have an impact on people’s access to GLP-1s.

In the June 2026 study, the researchers acknowledged that while the data show demographic differences mirrored across both semaglutide and tirzepatide, these are prone to bias and cannot account for factors such as concurrent use of other medications, variation in adherence and other social determinants of health^​2​.

An analysis, published in 2024 by the Health Foundation, of 113,000 patients who received a private prescription for a GLP-1 between November 2024 and October 2025, found that people in the most deprived areas were accessing 32% fewer GLP-1 prescriptions than those in the least deprived areas, despite much higher obesity levels in more deprived areas^​9​.

The analysis also found that almost 8 in 10 prescriptions were for women, while the highest uptake was among people aged 30–49 years, falling sharply after the age of 60 years^​9​.

In addition, it found people in more deprived areas tended to start treatment at a higher BMI. In the most deprived areas, around 45% of patients aged 30-49 years started treatment with a BMI of 35 or above, compared with around 30% in the least deprived areas^​9​, the analysis highlighted. In contrast, it showed that 11.1% of patients in the most deprived areas started treatment with a BMI below 30, compared with 15.7% in the least deprived areas^​9​. The Health Foundation said the findings suggest people in more deprived areas may be accessing treatment later^​9​.

Commenting on the findings, co-author Samantha Field, senior Fellow in prevention at the Health Foundation, said: “With an estimated 2.4 million people in the UK already prescribed weight-loss medications, our findings reveal a stark divide. The groups bearing the greatest burden of obesity are seeking GLP-1 treatments less frequently, and often at higher BMIs.

“The NHS should be taking these findings into account as the roll-out of these medications progresses, to ensure they are reaching the people who are most in need of them.”

The NHS began a phased 12-year rollout of GLP-1s in June 2025. Over the first three years, it is prioritising a cohort of 220,000 patients, which is based on clinical need^​10​.

Treatment persistence

The exclusive analysis shared with The Pharmaceutical Journal by Patientric suggests factors, such as deprivation and ethnicity, have an impact on treatment persistence, rather than weight-loss outcomes. Among patients in the most deprived one-fifth of England, 11.5% remained on GLP-1 treatment at one year, compared with 21.5% in the least deprived one-fifth of the country. It appears people still lose weight if they continue with treatment irrespective of whether they came from a deprived area or not.  

A similar pattern emerged with ethnicity: 20.2% of white patients remained on treatment at one year compared with 11.4% of black patients, a difference that persisted after adjustment for deprivation, age, sex and starting BMI. However, Patientric said there were too few patients remaining at one year across ethnic groups to draw meaningful conclusions about differences in weight-loss efficacy.

Patientric noted that it cannot track patients that leave its network, meaning some patients who are classified as having stopped treatment may have continued through another provider.

More research needed

Guirguis says more studies are needed on underrepresented groups in society.

“The issue is not that we have evidence showing GLP-1 medicines are less effective in underrepresented populations. Rather, when some groups are poorly represented, we have less statistical confidence in identifying or ruling out modest differences in efficacy, safety or tolerability,” she adds.

“Trials therefore need to do more than simply recruit diverse participants. They need enough people from different groups to properly analyse those differences, and should also capture the social and healthcare factors that can influence someone’s treatment and outcomes.”

Ultimately, even defining who responds ‘best’ to GLP-1s may be more complicated than comparing percentage weight loss.

“Firstly, and most importantly, how do we define a ‘good response’?” says Bicknell. “Even very modest weight loss, which might be considered a poor response, can improve insulin sensitivity and glycaemic control in a person living with T2DM, or might reduce the risk of someone going on to develop obesity-related complications.

She adds: “Secondly, would the cost of developing a robust genetic test be justified if the focus was just on weight-loss response rather than safety and identifying who might have serious, potentially life-threatening side effects?

“We have a way to go to identify a genetic test that can predict whether a person is at risk of developing the serious adverse effects associated with GLP-1 medicines.”  

Guirguis says: “We need to move beyond simply asking whether one demographic group loses more weight than another and understand why individuals respond differently. That means larger and more representative studies, which capture treatment exposure, dose escalation, adherence, morbidities, body composition, socioeconomic factors and genetic information together.

“The ultimate goal would be to combine these factors with someone’s response after treatment starts to make genuinely useful predictions. But any personalised approach would need to be tested prospectively and shown to improve patient outcomes before it could be used routinely. We’re not there yet, but this is where the field is heading.”