Do GLP-1 Drugs Work Differently by Race?

GLP-1 Patient Populations

No significant interaction by race

A 2026 JAMA Internal Medicine meta-analysis of 64 randomized trials found no significant heterogeneity of GLP-1 treatment effects by race or ethnicity for weight loss.

Alexander et al., JAMA Internal Medicine, 2026

Alexander et al., JAMA Internal Medicine, 2026

View as image

GLP-1 receptor agonists have become the most prescribed weight loss drugs in the United States, but the clinical trial populations that established their efficacy were overwhelmingly White. This raises two distinct questions. First, do the drugs work differently across racial and ethnic groups? Second, do all groups have equal access to the drugs? The answer to the first question, based on the subgroup data available, is largely no. The answer to the second is emphatically no. The efficacy story is reassuring. The access story is not. Both are part of understanding who these drugs are actually reaching and who they are missing.

What the clinical trial subgroup data shows

The most comprehensive analysis comes from Alexander et al. (2026), published in JAMA Internal Medicine. This systematic review and meta-analysis examined 64 randomized controlled trials of GLP-1 receptor agonists and specifically analyzed heterogeneity of treatment effects across patient subgroups including race and ethnicity.^[1]

The findings were clear: among 9 trials with 25,229 patients analyzed by race and 7 trials with 8,328 patients analyzed by ethnicity, there was no significant heterogeneity of treatment effect. GLP-1 drugs produced consistent weight loss across racial and ethnic subgroups. The only demographic variable that showed a significant difference was sex: women lost approximately 10.9% of body weight (95% CI, 7.0%-14.8%) compared to 6.8% for men (95% CI, 4.6%-9.0%).

Rubino et al. (2024) performed a post hoc analysis of the STEP 1, 2, and 3 trials specifically examining semaglutide 2.4 mg by race and ethnicity. Participants in the pooled STEP 1 and 3 analysis were 75.3% White, 8.8% Black, 10.6% Asian, and 5.3% other racial groups. Hispanic or Latino participants represented 13.9%.^[2]

The estimated treatment difference for semaglutide versus placebo ranged from 9.3% to 12.5% across all racial and ethnic subgroups in the STEP 1 and 3 sample and 4.5% to 7.3% in STEP 2. There were no significant interactions between treatment effect and race (STEP 1 and 3: p at least 0.07; STEP 2: p at least 0.15) or ethnicity (p at least 0.40 and p at least 0.85, respectively). Safety was also consistent across subgroups.

East Asian populations: the STEP 6 data

The STEP 6 trial was specifically designed for East Asian populations, enrolling Japanese and Korean adults with overweight or obesity. Kadowaki et al. (2024) published a subgroup analysis examining how baseline characteristics affected weight loss outcomes.^[3]

Semaglutide 2.4 mg produced clinically meaningful weight loss across all subgroups, ranging from 9.40% to 16.42%. The treatment differences favored semaglutide over placebo regardless of baseline body weight, BMI, age, glycemic status, dyslipidemia, or hypertension status.

Sex was the only significant modifier: women and men responded differently to semaglutide in this East Asian population (p = 0.0005 for semaglutide 2.4 mg versus placebo), consistent with the sex-based difference seen in the broader Alexander meta-analysis. Presence of type 2 diabetes at baseline also reached significance (p = 0.0381), with patients without diabetes losing more weight.

The STEP 6 data is important because it confirms that GLP-1 drugs work effectively in a population with different baseline BMI distributions, body composition profiles, and metabolic characteristics than the predominantly White Western populations in earlier trials. Asian populations tend to develop metabolic complications at lower BMI thresholds than White populations, and the effectiveness of semaglutide at lower starting weights had not been established in earlier trials.

The prescribing disparity problem

While the efficacy data is reassuring, the access data tells a different story. Multiple studies have documented substantial racial and ethnic disparities in who actually receives GLP-1 prescriptions.

Eberly et al. (2021), published in JAMA Health Forum, analyzed 1,180,260 commercially insured patients with type 2 diabetes in the US. Overall, only 7.7% were treated with a GLP-1 RA during the study period (2015-2019). In multivariable analyses controlling for age, sex, comorbidities, and socioeconomic factors, lower rates of GLP-1 RA use were found among Asian patients (adjusted odds ratio 0.59), Black patients (aOR 0.81), and Hispanic patients (aOR 0.91) compared to White patients. Higher household income was independently associated with higher GLP-1 RA use.^[5]

Kukhareva et al. (2024) found similar patterns in a more recent analysis of 57,320 patients with type 2 diabetes using the TriNetX electronic health record network. The disparities were even more pronounced for newer, more effective agents. For tirzepatide, adjusted odds of prescriptions compared to White patients were 0.3 for Asian patients, 0.4 for Hispanic patients, 0.6 for American Indian/Alaska Native patients, and 0.7 for Black patients. For semaglutide, the gaps were somewhat narrower but still significant: 0.5 for Asian patients, 0.6 for Hispanic patients, and 0.8 for Black patients.^[4]

The older GLP-1 RA dulaglutide showed smaller disparities, with Black patients having equivalent prescribing rates to White patients (aOR 1.0). This pattern suggests that cost may be a primary driver: newer, more expensive agents show larger disparities, while older, less expensive ones are prescribed more equitably.

Why the trials are not representative

The clinical trials that established GLP-1 drug efficacy enrolled populations that do not reflect the demographics of obesity and type 2 diabetes in the United States. Black Americans have a higher prevalence of obesity (49.9%) than White Americans (41.4%), according to CDC data. Hispanic Americans also have higher obesity rates (45.6%). Yet these populations were consistently underrepresented in the pivotal trials.

In the STEP program, Black participants represented only 8.8% of the pooled STEP 1 and 3 sample. Asian participants were 10.6%. Hispanic participants were 13.9%. The SURMOUNT program for tirzepatide had similar enrollment patterns. Meta-analyses of GLP-1 obesity trials have found that approximately 79% of participants across all major RCTs were White or Caucasian.

This underrepresentation matters even when subgroup analyses show no significant differences, because statistical power to detect differences in small subgroups is limited. The absence of a significant interaction does not prove equivalence. It means the trials were not large enough in their minority subgroups to detect modest differences if they existed. The subgroup data is reassuring but not definitive.

Cost and insurance as structural barriers

The prescribing disparities are not primarily explained by differences in clinical need or patient preferences. They track closely with socioeconomic gradients and insurance coverage patterns.

Semaglutide (Wegovy) has a list price exceeding $1,300 per month. Tirzepatide (Zepbound) is similarly priced. Even with insurance, copays can exceed $200-500 per month depending on plan structure. Eberly et al. found that higher zip code-linked median household income was independently associated with higher GLP-1 RA use, even after controlling for race and ethnicity.^[5]

These cost barriers interact with racial wealth disparities. The median Black household in the US holds approximately one-sixth the wealth of the median White household. Hispanic households hold approximately one-fifth. This means that even when insurance coverage is equivalent, out-of-pocket costs represent a larger share of disposable income for non-White households.

The rise of compounded semaglutide has partially addressed cost barriers, with compounded versions costing $100-400 per month. However, compounded products carry their own quality and safety concerns, and their availability has been subject to ongoing regulatory changes.

Genetic variation and GLP-1 response

Beyond race as a demographic category, there is the separate scientific question of whether genetic variation affecting GLP-1 receptor signaling influences drug response. The GLP-1 receptor gene (GLP1R) has known variants that differ in frequency across populations. Some of these variants affect receptor function, ligand binding, or downstream signaling.

Population-based pharmacogenomic studies on GLP-1 drugs remain limited. Most of the published pharmacogenomic data comes from type 2 diabetes glycemic response studies rather than weight loss studies. A few variants in GLP1R and related genes (TCF7L2, CTRB1/2, ARRB1) have been associated with differential glycemic response to GLP-1 RAs, but these associations have not been consistently replicated and are not strong enough to guide clinical prescribing decisions.

The distinction between race (a social construct) and genetic ancestry (a biological variable) matters here. Racial categories used in clinical trials are imprecise proxies for genetic variation. Two patients who both identify as "Black" may have very different genetic ancestry profiles. The subgroup analyses by race in the STEP and SURMOUNT trials are testing social categories, not genetic clusters. The consistent efficacy across racial subgroups is reassuring, but it does not rule out the possibility that specific genetic variants, which may vary in frequency across ancestry groups, could modify individual response.

This is an area where larger, more diverse pharmacogenomic studies are needed. The current evidence base cannot distinguish between "GLP-1 drugs genuinely work the same for everyone" and "our studies are too small and too genetically homogeneous to detect meaningful differences."

What this means for the evidence base

The combination of consistent efficacy data and inconsistent access creates a specific pattern: the drugs appear to work similarly across racial groups, but the populations benefiting from them are disproportionately White and affluent.

This has implications for real-world effectiveness data. As more post-marketing studies and registry analyses are published, they will predominantly reflect the experience of White patients because that is who received the prescriptions. Any safety signals, long-term efficacy patterns, or rare adverse events that differ by race could be missed because non-White patients are underrepresented in both the clinical trial data and the real-world treatment population.

For the broader question of who qualifies for these drugs, BMI thresholds themselves have been criticized for not accounting for racial and ethnic differences in body composition and metabolic risk. Asian populations develop type 2 diabetes and cardiovascular disease at lower BMIs than White populations, which has led some guidelines to recommend lower BMI cutoffs for treatment initiation in Asian patients. Whether these adjusted thresholds are consistently applied in clinical practice is unclear.

For the question of how semaglutide performs for weight loss without diabetes, the same racial composition caveat applies: the evidence comes predominantly from White populations. For patients wondering about GLP-1 drugs in older adults or after bariatric surgery, the pattern is the same: broadly reassuring trial data, but populations studied do not reflect the full diversity of patients who could benefit.

The Bottom Line

The clinical trial evidence consistently shows that GLP-1 receptor agonists produce similar weight loss across racial and ethnic subgroups, with sex as the only significant demographic modifier. However, substantial prescribing disparities persist: non-White patients are significantly less likely to receive GLP-1 prescriptions, with the gaps widest for newer, more expensive agents. The trials themselves underrepresent minority populations, limiting the power to detect modest differences. The result is a class of drugs that appears to work for everyone but reaches predominantly White, higher-income patients.

Frequently Asked Questions

Do GLP-1 weight loss drugs work the same for Black patients?

Available subgroup data indicates yes. Rubino et al. (2024) found that semaglutide 2.4 mg produced statistically significant and clinically relevant weight loss across all racial subgroups including Black patients, with no significant treatment-by-race interaction in the STEP trials.^[2] However, Black patients represented only 8.8% of participants, limiting the statistical power to detect modest differences.

Are GLP-1 drugs effective for Asian patients?

Yes. The STEP 6 trial, designed specifically for East Asian populations, showed semaglutide 2.4 mg produced 9.4% to 16.4% weight loss across all demographic subgroups in Japanese and Korean adults.^[3] Asian patients tend to have different body composition profiles and metabolic risk thresholds, but the drug's efficacy was consistent.

Why are fewer Black and Hispanic patients prescribed GLP-1 drugs?

Multiple factors contribute. Cost is a major barrier: semaglutide and tirzepatide exceed $1,300 per month at list price. Eberly et al. (2021) found that lower household income was independently associated with lower GLP-1 use, and racial wealth disparities amplify this effect.^[5] Insurance coverage patterns, physician prescribing practices, and structural healthcare access barriers also play roles.

Do men and women respond differently to GLP-1 drugs?

Yes. Alexander et al. (2026) found that women lost approximately 10.9% of body weight compared to 6.8% for men on GLP-1 drugs, the only demographic subgroup with a statistically significant difference in treatment effect.^[1] The STEP 6 East Asian analysis also found sex as a significant treatment modifier. The biological mechanisms behind this difference are not fully understood.

Were GLP-1 clinical trials diverse enough?

No. Across major GLP-1 obesity trials, approximately 79% of participants were White. Black, Asian, and Hispanic populations were consistently underrepresented relative to their prevalence of obesity in the US. While available subgroup analyses show no significant efficacy differences, the small minority subgroup sizes limit the power to detect modest differences.

Should BMI thresholds for GLP-1 drugs be different for Asian patients?

Some guidelines already recommend lower BMI cutoffs for metabolic intervention in Asian populations because type 2 diabetes and cardiovascular disease develop at lower BMI thresholds in this group. Whether prescribers consistently apply adjusted thresholds when considering GLP-1 treatment for Asian patients is unclear from available data.