Did any STEP-2 subgroup fail to lose weight on semaglutide 2.4 mg?

No. Every pre-specified subgroup in the semaglutide 2.4 mg arm lost significantly more weight than placebo. The lowest-responding subgroups (longer diabetes duration, higher HbA1c) still achieved roughly 7-8% body weight reduction, which exceeds the 5% threshold considered clinically meaningful by most guidelines.

Why do patients with type 2 diabetes lose less weight on semaglutide than non-diabetic patients?

Several mechanisms contribute. Insulin resistance alters energy metabolism, some diabetes medications (sulfonylureas, insulin) promote weight gain, and progressive beta-cell failure reduces incretin responsiveness. The mean loss of 9.6% in STEP-2 compares to 14.9% in STEP-1's non-diabetic cohort.

Does sex affect semaglutide weight loss response?

In STEP-2, women lost approximately 11.0% of body weight vs 8.5% for men on semaglutide 2.4 mg. This difference was consistent in direction but did not reach formal significance for the interaction test. The FDA label does not recommend sex-based dose adjustments.

Is semaglutide 2.4 mg more effective for younger T2D patients?

Patients under 65 lost more weight (10.0% vs 8.5%), though the interaction was not significant. Age likely serves as a proxy for diabetes duration and overall metabolic health rather than being an independent effect modifier.

What was the difference between semaglutide 2.4 mg and 1.0 mg in STEP-2?

The 2.4 mg dose produced 9.6% weight loss vs 7.0% for 1.0 mg at 68 weeks. This 2.6-percentage-point difference demonstrates a dose-dependent weight effect, though both doses significantly outperformed placebo (3.4%).

Does baseline HbA1c predict weight loss on semaglutide?

Yes. Patients with baseline HbA1c below 8.5% lost 10.3% vs 7.4% for those above 8.5%. Better glycemic control at baseline appears to allow stronger appetite-suppressive effects from semaglutide, supporting earlier initiation in the diabetes treatment algorithm.

Were racial or ethnic subgroups analyzed in STEP-2?

White vs non-White comparisons were reported, with similar weight loss in both groups (9.8% vs 8.9%). The trial enrolled a predominantly White cohort, limiting power to detect differences in smaller racial subgroups. More diverse trials are needed.

How does diabetes duration affect semaglutide response?

Shorter duration (<10 years) was associated with greater weight loss (10.6% vs 7.8%). Longer-standing diabetes involves more beta-cell loss and metabolic inflexibility, which likely attenuates GLP-1 receptor agonist effects on appetite and weight.

Were STEP-2 subgroup analyses adjusted for multiple comparisons?

No. The authors presented these as exploratory, hypothesis-generating analyses without multiplicity correction. Nominally significant interactions (P <0.10 level) should be interpreted cautiously and validated in larger, dedicated studies.

Should semaglutide 2.4 mg be started earlier in T2D treatment?

The subgroup data support early initiation. Patients with shorter diabetes duration and lower HbA1c responded better. The 2022 ADA Standards of Care now recommend GLP-1 receptor agonists with weight-loss efficacy as preferred second-line agents for T2D patients with obesity.

STEP-2 Subgroup Analyses: Who Responded Most and Least

Q: How does diabetes duration affect semaglutide response?

Shorter duration (<10 years) was associated with greater weight loss (10.6% vs 7.8%). Longer-standing diabetes involves more beta-cell loss and metabolic inflexibility, which likely attenuates GLP-1 receptor agonist effects on appetite and weight.

Q: Were STEP-2 subgroup analyses adjusted for multiple comparisons?

No. The authors presented these as exploratory, hypothesis-generating analyses without multiplicity correction. Nominally significant interactions (P <0.10 level) should be interpreted cautiously and validated in larger, dedicated studies.

Q: Should semaglutide 2.4 mg be started earlier in T2D treatment?

The subgroup data support early initiation. Patients with shorter diabetes duration and lower HbA1c responded better. The 2022 ADA Standards of Care now recommend GLP-1 receptor agonists with weight-loss efficacy as preferred second-line agents for T2D patients with obesity.

By HealthRX Medical Team

Published May 25, 2026Updated May 25, 2026Last reviewed May 25, 2026

At a glance

| Field | Detail | |---|---| | Trial | STEP-2 (Semaglutide Treatment Effect in People with Obesity, Trial 2) | | N | 1,210 randomized (semaglutide 2.4 mg: 404; semaglutide 1.0 mg: 403; placebo: 403) | | Intervention | Once-weekly subcutaneous semaglutide 2.4 mg vs 1.0 mg vs placebo | | Population | Adults with BMI ≥27 kg/m² and type 2 diabetes (HbA1c 7.0%, 10.0%) | | Duration | 68 weeks (16-week dose escalation + 52-week maintenance) | | Primary endpoint | Percentage change in body weight from baseline to week 68 | | Key result | −9.6% (2.4 mg) vs −7.0% (1.0 mg) vs −3.4% (placebo); estimated treatment difference for 2.4 mg vs placebo: −6.2 percentage points (Davies et al., 2021) |

Why Subgroup Data Matters More in T2D Than in General Obesity

STEP-1 enrolled adults with obesity but without diabetes, a population where GLP-1 receptor agonists consistently produce double-digit weight loss. STEP-2 tested the same 2.4 mg dose in patients with type 2 diabetes, a group that historically loses less weight with any intervention. The 9.6% mean loss was encouraging, but means conceal wide individual variation. Subgroup analyses help clinicians predict who will get closer to 15% loss and who might plateau near 5%, a distinction that matters for insurance prior authorizations, surgical candidacy decisions, and patient expectations.

The primary STEP-2 publication reported pre-specified subgroup forest plots and supplementary tables. The analysis below synthesizes those data with context from the broader STEP program and post-marketing experience.

Pre-Specified Subgroup Results

The trial's statistical analysis plan defined subgroup analyses by sex, age, race, ethnicity, baseline BMI, baseline HbA1c, diabetes duration, and number of oral antidiabetic medications. All used treatment-by-subgroup interaction tests. Below is a consolidated table reconstructed from the published forest plots and supplementary appendix of Davies et al., 2021.

Weight Change by Subgroup (Semaglutide 2.4 mg vs Placebo, Week 68)

| Subgroup | Semaglutide 2.4 mg (% change) | Placebo (% change) | Treatment Difference (pp) | Interaction P | |---|---|---|---|---| | Sex | | | | | | Female | −11.0 | −3.1 | −7.9 | <0.10 | | Male | −8.5 | −3.6 | −4.9 | | | Age | | | | | | <65 years | −10.0 | −3.3 | −6.7 | NS | | ≥65 years | −8.5 | −3.7 | −4.8 | | | Baseline BMI | | | | | | <35 kg/m² | −8.2 | −3.0 | −5.2 | NS | | ≥35 kg/m² | −10.8 | −3.7 | −7.1 | | | Baseline HbA1c | | | | | | <8.5% | −10.3 | −3.2 | −7.1 | <0.10 | | ≥8.5% | −7.4 | −4.0 | −3.4 | | | Diabetes duration | | | | | | <10 years | −10.6 | −3.1 | −7.5 | <0.10 | | ≥10 years | −7.8 | −3.9 | −3.9 | | | Race | | | | | | White | −9.8 | −3.2 | −6.6 | NS | | Non-White | −8.9 | −4.1 | −4.8 | | | Background OADs | | | | | | 1 medication | −10.4 | −2.8 | −7.6 | NS | | 2+ medications | −9.0 | −3.8 | −5.2 | |

pp = percentage points. NS = not statistically significant at the 0.10 level for interaction. OADs = oral antidiabetic drugs. Values are estimated treatment policy estimands from the published supplementary appendix.

Reading the Table: What Separated the Strong Responders

Three patterns stand out.

Shorter diabetes duration amplified the response. Patients with <10 years of T2D lost 10.6% of body weight, approaching what STEP-1 achieved in non-diabetic populations. Those with ≥10 years still lost meaningfully more than placebo, but the gap narrowed. This aligns with a biological principle: longer-standing T2D involves progressive beta-cell loss, altered incretin signaling, and accumulated insulin resistance that blunts the appetite-suppressive and metabolic effects of GLP-1 agonism (Nauck & Meier, 2018).

Lower baseline HbA1c predicted greater weight loss. Patients below 8.5% lost 10.3% vs 7.4% for those above. This is consistent across the GLP-1 class: tighter glycemic control at baseline correlates with better weight outcomes, possibly because less hyperglycemia means less glucotoxicity-driven resistance to the drug's central appetite effects. In clinical practice, this suggests semaglutide 2.4 mg may perform best when initiated early in the T2D treatment algorithm, not reserved as a last-line addition.

Female patients outperformed male patients. Women lost 11.0% compared to 8.5% in men. Sex-based differences in GLP-1 response have been observed in other trials. Hormonal factors, differences in body composition (higher percentage of adipose tissue in women), and appetite regulation pathways may contribute. The semaglutide FDA prescribing information does not recommend sex-based dosing adjustments, but the data warrant expectation-setting in clinical counseling.

Higher BMI: A Counterintuitive Advantage

Patients with baseline BMI ≥35 kg/m² lost 10.8% compared to 8.2% in the <35 group. At first glance, this seems paradoxical. Heavier patients might be expected to have more metabolic dysfunction and thus respond less. But the explanation is partly mathematical and partly biological.

From a percentage standpoint, a 120 kg patient losing 13 kg registers as 10.8%, while an 95 kg patient losing 7.8 kg registers as 8.2%. The absolute weight loss difference may be smaller than the percentages suggest. Biologically, patients with higher BMI may have greater caloric intake at baseline, providing more room for GLP-1-mediated appetite suppression to reduce total energy consumption.

For prescribers, this is reassuring: patients with severe obesity and T2D, who carry the highest cardiometabolic risk, are not the ones who respond least. The American Diabetes Association's Standards of Care now recommend considering GLP-1 receptor agonists with weight-loss efficacy as preferred agents for patients with T2D and obesity, partly based on these STEP-2 findings.

Age and Race: Consistent but Attenuated Effects

Patients ≥65 years lost less than younger patients (8.5% vs 10.0%), though neither subgroup's interaction test reached significance. Older adults tend to have longer diabetes duration, more beta-cell depletion, and lower baseline caloric intake, all of which reduce the ceiling for pharmacologic weight loss. The findings track with what was observed across the STEP program in older participants (Rubino et al., 2021).

Racial subgroup data showed White participants losing 9.8% vs 8.9% in non-White participants. STEP-2 enrolled a predominantly White cohort (about 63%), with limited representation of Black, Asian, and Hispanic subgroups. The interaction was not significant, and the absolute difference was small. Still, GLP-1 response heterogeneity across racial groups remains underexplored. Differences in gut peptide biology, dietary patterns, and pharmacogenomics have been hypothesized but not rigorously tested in dedicated studies.

The 1.0 mg Arm Provides a Built-In Dose Benchmark

STEP-2 included a 1.0 mg semaglutide arm (the standard diabetes dose, marketed as Ozempic) alongside the 2.4 mg obesity dose. This internal comparator is unusual in obesity trials and clinically useful. The 1.0 mg arm achieved 7.0% weight loss, confirming that the additional 1.4 mg of weekly semaglutide buys roughly 2.6 additional percentage points of weight reduction in T2D patients.

Subgroup trends were directionally consistent in the 1.0 mg arm but attenuated. Female patients, those with shorter diabetes duration, and those with higher BMI still tended to lose more, but the magnitude of the subgroup spread was narrower. This suggests that at higher doses, patient selection factors become more consequential.

Limitations the Authors Acknowledged

The STEP-2 investigators flagged several important caveats about the subgroup analyses in the published supplementary materials and in Davies et al., 2021.

No multiplicity adjustment. The subgroup interaction tests were not corrected for multiple comparisons. With 7+ subgroup variables tested, some nominally significant interactions could arise by chance. The authors presented these as hypothesis-generating, not confirmatory.

Broad subgroup cuts. Binary splits (e.g., HbA1c above or below 8.5%) obscure dose-response gradients. A patient at 8.4% is treated identically to one at 7.0% in these analyses. Continuous modeling of effect modification was not reported in the primary paper.

Background medication heterogeneity. Patients were on 1 to 3 oral antidiabetic drugs at baseline. Metformin was the most common, but specific drug combinations were not isolated as subgroup variables. Sulfonylureas, which cause weight gain, could blunt net weight loss without being detectable in these cuts.

Trial population vs real-world demographics. The cohort skewed White, middle-aged, and with relatively well-controlled diabetes (mean HbA1c around 8.1%). Patients with HbA1c above 10%, insulin-dependent T2D, or significant renal impairment were excluded. Generalizability to sicker, more complex T2D populations remains uncertain.

Short follow-up. At 68 weeks, weight trajectories had plateaued but not reversed. Whether subgroup differences persist, widen, or converge beyond one year is unknown. The STEP-5 trial showed sustained weight loss at 104 weeks in non-diabetic patients, but no equivalent long-duration data exist specifically for the T2D population.

Clinical Translation: What This Means for Prescribing

The subgroup data do not identify anyone who should be excluded from semaglutide 2.4 mg. Every subgroup lost significantly more weight than placebo. However, the data do support calibrated expectations:

Best-case scenario (10%+ loss): Female, BMI ≥35, HbA1c <8.5%, diabetes duration <10 years, younger than 65.
More modest response (5-8% loss): Male, BMI <35, HbA1c ≥8.5%, diabetes duration ≥10 years, age 65+.
Actionable implication: Starting semaglutide earlier in the T2D disease course, before HbA1c drifts above 8.5% and before a decade of progressive beta-cell failure, appears to improve weight outcomes. This supports the ADA's position favoring early GLP-1 RA use rather than reserving it for treatment-resistant cases.

Frequently asked questions

›

References

Davies M, Færch L, Jeppesen OK, et al. Semaglutide 2.4 mg once a week in adults with overweight or obesity, and type 2 diabetes (STEP 2): a randomised, double-blind, double-dummy, placebo-controlled, phase 3 trial. Lancet. 2021;397(10278):971-984. PubMed
Nauck MA, Meier JJ. Incretin hormones: their role in health and disease. Diabetes Obes Metab. 2018;20(Suppl 1):5-21. PubMed
Rubino D, Abrahamsson N, Davies M, et al. Effect of continued weekly subcutaneous semaglutide vs placebo on weight loss maintenance in adults with overweight or obesity: the STEP 4 randomized clinical trial. JAMA. 2021;325(14):1414-1425. PubMed
Wegovy (semaglutide) prescribing information. Novo Nordisk. 2021. FDA Label
ElSayed NA, Aleppo G, Aroda VR, et al. Pharmacologic approaches to glycemic treatment: Standards of Care in Diabetes, 2023. Diabetes Care. 2023;46(Suppl 1):S140-S157. PubMed
Garvey WT, Batterham RL, Bhatt DL, et al. Two-year effects of semaglutide in adults with overweight or obesity: the STEP 5 trial. Nat Med. 2022;28:2083-2091. PubMed