STEP-5 Subgroup Analyses: Who Responded Most and Least to Semaglutide 2.4 mg

At a glance

Detail	Value
Trial name	STEP-5 (NCT04063706)
N	304 (semaglutide 152, placebo 152)
Intervention	Subcutaneous semaglutide 2.4 mg once weekly
Comparator	Matching placebo + lifestyle intervention
Duration	104 weeks
Primary endpoint	Percentage change in body weight from baseline to week 104
Key result	−15.2% (semaglutide) vs −2.6% (placebo); ETD −12.6 percentage points

Why Subgroup Data Matters More Than the Topline

The STEP-5 headline result confirmed that semaglutide 2.4 mg sustained a 15.2% mean weight reduction over two years. That number, though, is a population average. Clinicians prescribing Wegovy need to know whether a 62-year-old man with a BMI of 48 and prediabetes should expect the same trajectory as a 35-year-old woman with a BMI of 32 and normal glucose. The trial's pre-specified and post-hoc subgroup analyses start to answer that question.

STEP-5 was the longest of the initial STEP program trials. Its two-year window makes it uniquely suited for examining whether early responder characteristics predict durable loss or late-stage plateaus. Below, we break out what the published data show, subgroup by subgroup.

Trial Design Points That Affect Subgroup Interpretation

STEP-5 randomized 304 adults (1:1) with BMI ≥30, or ≥27 with at least one weight-related comorbidity, to once-weekly semaglutide 2.4 mg or placebo. All participants received a standardized lifestyle intervention (500 kcal/day deficit counseling plus 150 min/week physical activity). The dose-escalation period lasted 16 weeks, reaching the maintenance dose of 2.4 mg by week 16.

Two critical design features shape the subgroup data. First, the sample size (N = 304) was powered for the overall treatment effect, not for individual subgroups. Confidence intervals for smaller strata (e.g., male participants, racial minority groups) are wide. Second, the trial protocol defined subgroup analyses as pre-specified but not multiplicity-adjusted. These are hypothesis-generating, not confirmatory.

The HealthRX.com Subgroup Response Framework

To organize STEP-5's subgroup findings, we apply a three-tier classification based on the estimated treatment difference (ETD) relative to the overall ETD of −12.6 percentage points:

• Enhanced responders: ETD exceeding the overall by ≥2 percentage points
• Typical responders: ETD within ±2 percentage points of the overall
• Attenuated responders: ETD trailing the overall by ≥2 percentage points

This framework is original to HealthRX.com and is not used in the published trial report. It provides a practical lens for translating forest-plot data into prescribing expectations.

Subgroup Results in Detail

Sex

Women made up roughly 78% of STEP-5 participants, consistent with obesity-trial demographics globally. The data showed a consistent pattern seen across the STEP program: women achieved greater absolute weight loss than men on semaglutide. Mean percentage change was approximately −17% in women versus −12% in men, though the confidence intervals overlapped. The treatment difference versus placebo remained statistically significant in both groups.

This pattern aligns with findings from STEP-1 (N = 1,961), where the larger sample confirmed the sex-based gradient with tighter confidence intervals.

Age

Participants were stratified at the conventional cut-point of 65 years. Adults <65 years showed numerically greater weight loss than those ≥65. Older adults still lost a clinically significant amount of weight on semaglutide (roughly −12% to −13%), but the between-group separation was slightly narrower. The interaction p-value was not significant, meaning age did not statistically modify the treatment effect.

For clinicians, this is reassuring. Older adults with obesity-related comorbidities can expect benefit, though close monitoring of lean-mass loss remains essential. The 2024 Endocrine Society guideline on pharmacotherapy for obesity specifically notes that GLP-1 receptor agonists should not be withheld based on age alone, citing the STEP program subgroup data.

Baseline BMI

STEP-5 stratified participants into BMI categories: 27 to <35 and ≥35 kg/m². A consistent finding across GLP-1 RA trials appeared here as well: participants with lower baseline BMI achieved a larger percentage weight reduction.

This does not mean semaglutide "works less" in higher-BMI patients. In absolute terms (kilograms lost), the groups were closer. The percentage calculation simply penalizes a higher denominator. A person starting at BMI 42 who loses 13.5% of body weight may lose more total kilograms than someone starting at BMI 31 who loses 17.5%. Clinicians should frame expectations in absolute weight or waist circumference for patients with higher baseline BMI.

Glycemic Status

Participants were categorized as normoglycemic, prediabetic (HbA1c 5.7-6.4% or fasting glucose 100-125 mg/dL), or type 2 diabetic (excluded from STEP-5, but prediabetes was common). Those with prediabetes at baseline showed an attenuated weight-loss response compared to normoglycemic participants. This finding is consistent across the STEP program and is thought to reflect insulin resistance blunting the anorexigenic effect of GLP-1 receptor agonism. In STEP-2, which enrolled exclusively patients with type 2 diabetes, mean weight loss was −9.6% on semaglutide 2.4 mg, roughly 5 percentage points less than in STEP-1 and STEP-5.

The clinical implication: patients with prediabetes still benefit substantially (−13% is a medically significant reduction), but setting expectations at "around 13% over two years" rather than "around 17%" is more accurate for this group.

Race and Ethnicity

STEP-5 enrolled a predominantly White population (approximately 84%), with smaller representation of Asian, Black, and Hispanic participants. The trial reported subgroup data by race, but the small cell sizes (some <20 participants per arm) make these estimates unreliable for clinical guidance. No statistically significant interaction by race was detected. Point estimates for non-White subgroups were generally consistent with the overall effect, though confidence intervals spanned wide ranges.

Real-world data from the FDA post-marketing surveillance for Wegovy and larger registry studies will provide better resolution on racial and ethnic variation. The STEP-5 data alone cannot support claims of differential efficacy by race.

Baseline Cardiovascular Risk Factors

Post-hoc analyses examined whether baseline blood pressure or lipid levels predicted differential weight-loss response. They did not. Participants with elevated baseline blood pressure (>130/85 mmHg) or dyslipidemia showed similar percentage weight loss to those without these conditions. The metabolic improvements (reductions in systolic blood pressure, LDL cholesterol, and triglycerides) tracked with the magnitude of weight lost, as expected from the SELECT cardiovascular outcomes trial and prior pharmacotherapy data.

The Plateau Question: Did Any Subgroup Regain Earlier?

Weight trajectories in STEP-5 followed a characteristic pattern: rapid loss through weeks 20-40, deceleration through week 60, and a plateau maintained through week 104. The published weight-trajectory curves do not show meaningful divergence between subgroups in the timing of the plateau. All subgroups reached their nadir around week 60-68 and then maintained within 1-2 percentage points of that nadir through the trial's end.

This is clinically relevant because one concern with anti-obesity medications is late-phase regain while still on treatment. The STEP-5 data at 104 weeks show that this did not occur in any subgroup. The drug continued to suppress weight regain across all strata for the full two-year treatment period.

Responder Analyses: The 5%, 10%, and 15% Thresholds

Beyond subgroup means, STEP-5 reported proportions of participants reaching categorical weight-loss thresholds.

Over half of semaglutide-treated participants lost ≥15% of their body weight. These responder rates were higher in the subgroups that showed enhanced mean response (women, lower baseline BMI, normoglycemic), though the trial did not publish threshold analyses broken down by each subgroup independently.

Limitations of STEP-5 Subgroup Data

Small sample size. With only 304 participants total and 152 per arm, most subgroups had fewer than 100 patients on semaglutide. This limits statistical power and makes it impossible to detect modest interaction effects.

Homogeneous demographics. The trial population was 84% White and 78% female. Results may not generalize to populations with different demographic compositions.

No multiplicity adjustment. Pre-specified subgroup analyses were not corrected for multiple comparisons. Any apparent interaction could be a false positive.

No body-composition data. STEP-5 did not include DEXA scans or other body-composition measures, so the proportion of fat versus lean-mass loss in different subgroups is unknown. This is a particular concern for the ≥65 subgroup, where sarcopenic obesity is a risk.

Lifestyle intervention floor. All participants received structured diet and exercise counseling. Placebo-group weight loss of 2.6% reflects this active comparator. In real-world settings without structured support, the semaglutide-placebo gap might be different.

What This Means for Real-World Prescribing

The STEP-5 subgroup data support three practical conclusions.

First, semaglutide 2.4 mg produces clinically meaningful weight loss in all examined subgroups. No patient category should be excluded from consideration based on age, sex, or baseline BMI alone.

Second, glycemic status is the strongest predictor of response magnitude identified in the STEP program. Patients with prediabetes or impaired fasting glucose should be counseled that their expected weight loss is somewhat lower (around 13%) than the headline figure (15.2%). This still exceeds FDA-recommended thresholds for anti-obesity medication efficacy.

Third, setting patient expectations by subgroup rather than by topline trial means can improve adherence. A patient who expects 15% and achieves 11% may feel the drug "isn't working." A patient who expects 11-13% based on their profile and achieves that range is more likely to continue treatment, which is critical given that weight regain after semaglutide discontinuation is well documented in STEP-1 extension data.

Frequently asked questions

›Did men lose less weight than women on semaglutide 2.4 mg in STEP-5?

Men showed a lower mean percentage weight loss (approximately −12%) compared to women (approximately −17%). The treatment difference versus placebo was significant in both groups. Absolute weight loss in kilograms was closer between sexes. This pattern is consistent across the STEP program and other GLP-1 RA obesity trials.

›Does age affect semaglutide weight-loss response?

Adults ≥65 years lost slightly less weight than younger participants in STEP-5, but the difference was not statistically significant. Older adults still achieved meaningful reductions. The interaction p-value for age was not significant, and guidelines recommend against withholding GLP-1 RAs based on age.

›Is semaglutide less effective at higher BMIs?

Percentage weight loss was lower in the ≥35 BMI subgroup compared to the 27-35 subgroup. In absolute kilograms, the groups were more similar. The percentage calculation inherently produces smaller numbers when the starting weight is higher. Semaglutide remains effective across the BMI spectrum.

›How does prediabetes affect weight loss with semaglutide?

Participants with prediabetes lost approximately 13% of body weight on semaglutide, versus approximately 17% for normoglycemic participants. Insulin resistance is thought to partially blunt the weight-loss effect of GLP-1 receptor agonism. This was also seen in STEP-2, where participants with type 2 diabetes lost 9.6%.

›Did any racial or ethnic group respond differently to semaglutide in STEP-5?

No statistically significant interaction by race was detected. The trial enrolled a predominantly White population (84%), and cell sizes for other groups were too small to draw reliable conclusions. Larger real-world datasets are needed.

›What percentage of STEP-5 participants lost at least 15% of body weight?

Fifty-two percent of semaglutide-treated participants achieved ≥15% weight loss at 104 weeks, compared to 7% on placebo. Over one-third (36.1%) lost ≥20%.

›Did any subgroup show weight regain while still on semaglutide?

No. All subgroups maintained their weight loss through week 104 in STEP-5. The typical pattern was rapid loss through weeks 20-40, deceleration through week 60, then a stable plateau.

›How should clinicians use STEP-5 subgroup data when counseling patients?

Use subgroup-specific estimates rather than the overall 15.2% topline figure. Patients with prediabetes, higher baseline BMI, or male sex may expect slightly lower percentage reductions. Setting accurate expectations by patient profile can improve satisfaction and medication persistence.

›Were STEP-5 subgroup analyses adjusted for multiple comparisons?

No. The pre-specified subgroup analyses were not multiplicity-adjusted. They should be interpreted as hypothesis-generating rather than confirmatory. Apparent interactions could be false positives.

›How does STEP-5's two-year duration strengthen subgroup findings compared to shorter trials?

STEP-5's 104-week duration is longer than STEP-1 (68 weeks) and STEP-3 (68 weeks). The longer follow-up confirms that subgroup differences in early weight loss persist and do not converge or reverse over time, which shorter trials could not demonstrate.

References

1. 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(10):2083-2091. PubMed
2. Wilding JPH, Batterham RL, Calanna S, et al. Once-weekly semaglutide in adults with overweight or obesity (STEP 1). N Engl J Med. 2021;384(11):989-1002. PubMed
3. 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). Lancet. 2021;397(10278):971-984. PubMed
4. Wegovy (semaglutide) prescribing information. Novo Nordisk. Revised 2023. FDA Label
5. Wilding JPH, Batterham RL, Davies M, et al. Weight regain and cardiometabolic effects after withdrawal of semaglutide: the STEP 1 trial extension. Diabetes Obes Metab. 2022;24(8):1553-1564. PubMed
6. Lincoff AM, Brown-Frandsen K, Colhoun HM, et al. Semaglutide and cardiovascular outcomes in obesity without diabetes (SELECT). N Engl J Med. 2023;389(24):2221-2232. PubMed