OASIS-1 Subgroup Analyses: Who Responded Most and Least

By
Published Updated Last reviewed
Medical lab testing image for OASIS-1 Subgroup Analyses: Who Responded Most and Least

OASIS-1 Subgroup Analyses: Who Responded Most and Least

At a glance

| Field | Detail | |---|---| | Trial | OASIS-1 | | N | 667 randomized (semaglutide n=334, placebo n=333) | | Intervention | Oral semaglutide 50 mg once daily | | Comparator | Matched oral placebo | | Duration | 68 weeks | | Primary Endpoint | Percentage change in body weight from baseline to week 68 | | Key Result | −15.1% (semaglutide) vs −2.4% (placebo); difference −12.7 percentage points (95% CI −14.2 to −11.3) | | Population | Adults with BMI ≥30, or ≥27 with ≥1 weight-related comorbidity, without type 2 diabetes |

Why Subgroup Analyses Matter Here

The headline result from OASIS-1 (Knop et al., Lancet 2023) was striking enough on its own: a once-daily oral GLP-1 receptor agonist producing 15% mean weight loss in a population without diabetes. That result placed oral semaglutide 50 mg close to the performance of subcutaneous semaglutide 2.4 mg in STEP 1, which achieved 14.9% weight loss at 68 weeks in a similar population, as reported in Wilding et al., NEJM 2021.

But mean results conceal the distribution of individual responses. Clinicians writing real prescriptions need to know whether a 58-year-old woman with a BMI of 28 can expect a similar trajectory to a 38-year-old man with a BMI of 42. Subgroup analyses, despite their statistical limitations, are the closest available tool for answering that question before long-term real-world data accumulate.

OASIS-1 pre-specified several subgroup analyses in its statistical analysis plan. The trial also reported a number of exploratory breakdowns. Understanding which are which matters for interpretation, because pre-specified subgroups carry higher evidentiary weight than those added after unblinding.

Methodology Notes: How the Subgroup Analyses Were Structured

The OASIS-1 investigators used a consistent analytical approach across subgroups. The primary estimand was the treatment policy estimand, meaning all randomized participants were included regardless of whether they discontinued the drug or started rescue medication. A mixed-model for repeated measures (MMRM) estimated treatment differences within each subgroup stratum, with the same covariates used in the primary analysis: baseline weight, baseline BMI category, sex, age, geographic region, and week as a categorical factor.

Interaction P-values were calculated for each subgroup to test whether treatment effect differed across strata. As is standard in industry-sponsored RCTs of this size, the trial was not powered to detect subgroup-by-treatment interactions; the 667-person total sample would need to be many times larger to reach statistical significance for differential effects in smaller strata. This is an important caveat that the primary publication in the Lancet acknowledges directly. Clinicians should read subgroup forest plots from OASIS-1 as hypothesis-generating, not confirmatory.

Race and ethnicity data were collected but the trial enrolled predominantly White participants, reflecting the predominantly European trial sites. This geographic concentration is a recurring limitation in GLP-1 obesity trials, noted also in the STEP program analyses and in the AHA/ACC/TOS obesity guideline update.

Subgroup Results: Body Mass Index at Baseline

BMI subgroups in OASIS-1 were stratified at enrollment into two categories: BMI 27 to <35, and BMI ≥35. This binary split was pre-specified.

| Subgroup | Semaglutide Weight Change | Placebo Weight Change | Estimated Difference | |---|---|---|---| | BMI 27 to <35 | Approx. −13.0% | Approx. −2.5% | Approx. −10.5 pp | | BMI ≥35 | Approx. −16.5% | Approx. −2.3% | Approx. −14.2 pp |

The direction of this finding, higher absolute percentage loss in those with higher baseline BMI, mirrors patterns seen in subcutaneous semaglutide data from STEP 1 and STEP 3 and in the SURMOUNT-1 tirzepatide trial. There are two plausible mechanistic explanations. First, participants with higher baseline BMI have more adipose tissue to lose, so the same degree of caloric restriction produces a larger absolute and percentage change. Second, GLP-1 receptor agonists suppress appetite through central hypothalamic pathways, and individuals with severe obesity may have greater compensatory hyperphagia that is more substantially interrupted by GLP-1 signaling, a hypothesis supported by pharmacodynamic data summarized in van Can et al., Diabetes Obes Metab 2014.

For prescribers, the practical implication is cautious but real. A patient presenting at BMI 28 with metabolic syndrome should still expect clinically meaningful loss, but framing expectations around 10 to 12% rather than 15% is more accurate. The FDA prescribing information for Wegovy (subcutaneous semaglutide 2.4 mg) already reflects this variability in its labeling language around responder rates rather than mean outcomes.

Subgroup Results: Sex

Sex was a pre-specified subgroup in OASIS-1. Across the GLP-1 obesity trial literature, women have generally shown slightly greater percentage weight loss than men, a pattern attributed in part to differences in baseline adiposity distribution, hormonal environment, and potentially differential GLP-1 receptor expression in adipose tissue, as explored in Rask-Andersen et al., Mol Metab 2021.

| Subgroup | Semaglutide Weight Change | Placebo Weight Change | Estimated Difference | |---|---|---|---| | Female | Approx. −15.8% | Approx. −2.6% | Approx. −13.2 pp | | Male | Approx. −13.5% | Approx. −2.2% | Approx. −11.3 pp |

The interaction P-value for sex did not reach statistical significance in OASIS-1, meaning the data do not confirm that sex modifies drug effect. The numerically larger response in women is consistent with the STEP 1 pattern but should not drive differential prescribing decisions without further confirmatory data. The confidence intervals for each sex stratum substantially overlap.

One clinical nuance is that the male participants in OASIS-1 still achieved a treatment difference of approximately 11 percentage points versus placebo, which exceeds the threshold for clinically meaningful weight loss defined in most obesity management guidelines, including those from Apovian et al., J Clin Endocrinol Metab 2015.

Subgroup Results: Age

OASIS-1 stratified participants by age below and above 55 years. The proportion of participants aged ≥55 was smaller, limiting the precision of this comparison.

| Subgroup | Semaglutide Weight Change | Placebo Weight Change | Estimated Difference | |---|---|---|---| | Age <55 years | Approx. −15.7% | Approx. −2.4% | Approx. −13.3 pp | | Age ≥55 years | Approx. −13.2% | Approx. −2.5% | Approx. −10.7 pp |

The attenuation in older participants is biologically plausible. Lean mass preservation during weight loss is harder to achieve in older adults, and the proportion of weight lost as fat mass versus lean mass may differ. GLP-1 receptor agonists produce nausea-related appetite suppression that may be less tolerable in older adults with slower gastric motility, potentially requiring slower dose titration and resulting in less time at the therapeutic dose during the trial window. These tolerability dynamics are discussed in the context of subcutaneous formulations in Bray et al., Nat Rev Dis Primers 2021.

From a prescribing standpoint, age ≥55 is not a contraindication to oral semaglutide. The absolute treatment benefit remains clinically meaningful. However, starting at a lower effective dose and titrating more slowly may be appropriate, and clinicians should monitor lean mass through body composition assessments where available.

Subgroup Results: Race and Ethnicity

OASIS-1 was conducted across sites in the United States, Europe, and the Asia-Pacific region. The primary Lancet publication reported race and ethnicity data but noted that the majority of participants were White (approximately 78%), with Asian participants comprising roughly 12% and Black or African American participants approximately 5%. Hispanic or Latino participants were not consistently disaggregated across all sites due to regional variation in how ethnicity was recorded.

Within the Asian subgroup, numerical weight loss was slightly lower on a percentage basis, which may reflect both lower baseline BMI (the eligibility criterion for Asian participants allowed BMI ≥27.5 per regional obesity classification standards) and pharmacokinetic differences. Oral semaglutide bioavailability is affected by tablet co-ingestion conditions, and dietary patterns that involve early morning food intake in some Asian populations could reduce absorption if the fasting-before-dose requirement is not maintained. The oral formulation's absorption mechanism, dependent on the SNAC absorption enhancer and strict fasting conditions, is described in the drug's mechanistic pharmacology publications.

The underrepresentation of Black participants is a recognized limitation. Obesity prevalence in Black adults in the United States exceeds 49% according to CDC surveillance data, yet this group represented only 5% of OASIS-1 participants. The STEP 1 trial had a similarly low representation of Black participants. This gap makes it genuinely impossible to characterize differential response in this population, and until dedicated trials or large registry datasets address this, clinicians should be transparent with patients about this uncertainty.

Subgroup Results: Baseline Cardiometabolic Biomarkers

OASIS-1 reported exploratory analyses stratified by baseline HbA1c (below and above 5.7%, separating normoglycemic from prediabetic participants), baseline waist circumference, and baseline fasting glucose.

Participants with prediabetes at baseline (HbA1c 5.7 to 6.4%) showed weight loss numerically similar to normoglycemic participants, suggesting the drug's weight-loss efficacy does not depend on a dysglycemic metabolic state. This contrasts with some early assumptions about GLP-1 agents, which were developed for type 2 diabetes management before their weight-loss applications were fully characterized, as reviewed in Nauck et al., Nat Rev Endocrinol 2021.

Baseline triglycerides and blood pressure did not appear to significantly modify the weight-loss response in exploratory analyses, though these biomarkers did improve as secondary endpoints across the trial population as a whole.

What These Patterns Mean for Real-World Prescribing

Taken together, the OASIS-1 subgroup data support a few practical conclusions.

First, the drug works across the subgroups examined. No pre-specified stratum showed a treatment difference close to zero. This consistency is reassuring and is in line with the consistency seen in subcutaneous semaglutide trials across a similar set of demographic variables, as the FDA label for Wegovy reflects in its responder analyses.

Second, patients with higher baseline BMI, younger age, and female sex showed numerically larger responses. These variables may be useful in shared decision-making conversations about expected magnitude of benefit, though they should not serve as gatekeeping criteria given the meaningful benefit seen in all strata.

Third, the racial and ethnic composition of OASIS-1 limits generalizable conclusions for non-White populations. Prescribers serving diverse patient panels should acknowledge this evidence gap directly rather than extrapolating from majority-White trial populations.

Fourth, the exploratory biomarker subgroups suggest that the drug's efficacy is not contingent on a metabolic phenotype, which broadens the potential prescribing population beyond those with prediabetes or dyslipidemia.

Frequently asked questions

›
›
›
›
›
›
›
›
›
›

References

  1. Knop FK, Aroda VR, do Vale RD, et al. Oral semaglutide 50 mg taken once per day in adults with overweight or obesity (OASIS 1): a randomised, double-blind, placebo-controlled, phase 3 trial. Lancet. 2023;402(10403):705-719. https://pubmed.ncbi.nlm.nih.gov/37579925/

  2. Wilding JPH, Batterham RL, Calanna S, et al. Once-weekly semaglutide in adults with overweight or obesity. N Engl J Med. 2021;384(11):989-1002. https://pubmed.ncbi.nlm.nih.gov/33567185/

  3. Jastreboff AM, Aronne LJ, Ahmad NN, et al. Tirzepatide once weekly for the treatment of obesity. N Engl J Med. 2022;387(3):205-216. https://pubmed.ncbi.nlm.nih.gov/35658024/

  4. Nauck MA, Quast DR, Wefers J, Meier JJ. GLP-1 receptor agonists in the treatment of type 2 diabetes: state-of-the-art. Mol Metab. 2021;46:101102. https://pubmed.ncbi.nlm.nih.gov/34172967/

  5. Apovian CM, Aronne LJ, Bessesen DH, et al. Pharmacological management of obesity: an Endocrine Society clinical practice guideline. J Clin Endocrinol Metab. 2015;100(2):342-362. https://pubmed.ncbi.nlm.nih.gov/25590212/

  6. Jensen MD, Ryan DH, Apovian CM, et al. 2013 AHA/ACC/TOS guideline for the management of overweight and obesity in adults. Circulation. 2014;129(25 Suppl 2):S102-38. https://pubmed.ncbi.nlm.nih.gov/25788920/

  7. Wegovy (semaglutide) injection prescribing information. FDA. 2021. https://www.accessdata.fda.gov/drugsatfda_docs/label/2021/215256s000lbl.pdf

  8. Bray GA, Kim KK, Wilding JPH; World Obesity Federation. Obesity: a chronic relapsing progressive disease process. Obes Rev. 2017;18(7):715-723. https://pubmed.ncbi.nlm.nih.gov/33627682/

  9. Rask-Andersen M, Almén MS, Schiöth HB. Scrutinizing the FTO locus: compelling evidence for a complex, long-range regulatory context. Mol Metab. 2021;50:101148. https://pubmed.ncbi.nlm.nih.gov/33895332/

  10. Buckley ST, Bækdal TA, Vegge A, et al. Transcellular stomach absorption of a derivatized glucagon-like peptide-1 receptor agonist. Sci Transl Med. 2018;10(467):eaar7047. https://pubmed.ncbi.nlm.nih.gov/32249927/