STEP-3 Subgroup Analyses: Who Responded Most and Least

At a glance

Detail	Value
Trial	STEP-3 (Semaglutide Treatment Effect in People with Obesity, trial 3)
N	611 randomized (407 semaglutide, 204 placebo)
Intervention	Semaglutide 2.4 mg subcutaneous once weekly + IBT (30 sessions)
Comparator	Placebo injection once weekly + IBT (30 sessions)
Duration	68 weeks (16-week initial low-calorie diet phase, then standard diet)
Primary endpoint	Percentage change in body weight from baseline to week 68
Key result	−16.0% semaglutide+IBT vs −5.7% placebo+IBT (estimated treatment difference −10.3 percentage points; p < 0.001)

Why Subgroup Data From STEP-3 Matters More Than Usual

The STEP program comprised four phase 3 trials, each testing semaglutide 2.4 mg in a slightly different population or alongside a different co-intervention. STEP-3 is unique because both arms received an identical 30-session IBT program plus an initial 8-week low-calorie diet (1,000 to 1,200 kcal/day) followed by a conventional reduced-calorie plan. This design isolates the pharmacologic contribution of semaglutide on top of the most rigorous behavioral intervention used in any of the four trials.

Because IBT itself is a potent intervention (producing 5% to 7% weight loss in the placebo arm), subgroup analyses here reveal something the other STEP trials cannot: which patient characteristics predict an additive drug benefit beyond what structured behavioral change alone delivers. For prescribers deciding between lifestyle-only programs and pharmacotherapy layered on top, these data directly inform the conversation.

Pre-Specified Subgroup Analyses: The Forest Plot

STEP-3 investigators pre-specified subgroup analyses for the primary endpoint (percentage body weight change at 68 weeks) across several baseline characteristics. The primary publication reports forest plots showing treatment effect consistency. Below is a reconstruction of the key subgroup findings.

By Sex

Women comprised roughly 77% of participants and showed a larger absolute treatment difference than men. This sex-based pattern appeared across the entire STEP program. Possible explanations include differences in body composition (women carry proportionally more adipose tissue responsive to GLP-1-mediated appetite suppression), hormonal interactions with incretin signaling, and potentially greater adherence to the dietary components of IBT. The FDA-approved Wegovy label does not recommend sex-based dose adjustments, and both sexes exceeded the 5% clinically meaningful threshold by a wide margin.

By Baseline BMI

Participants with lower baseline BMI lost a greater percentage of body weight. This is a consistent finding across GLP-1 receptor agonist trials and does not mean the drug is less effective in severe obesity. Patients with BMI ≥40 still lost nearly 14% of body weight on average, a result that rivals outcomes from some bariatric procedures at 12 months. The narrower treatment difference in the highest BMI stratum (−9.3 pp vs −11.1 pp) may reflect the greater total caloric expenditure and proportionally smaller impact of fixed-dose appetite suppression in patients with higher body mass.

By Age

Younger participants showed modestly larger responses in both arms. Age-related differences in metabolic rate, lean mass preservation, and physical activity capacity likely contributed. The 2022 AGA guideline on pharmacologic management of obesity does not recommend age-stratified dosing but notes that older adults warrant closer monitoring for lean mass loss and sarcopenia risk during rapid weight reduction.

By Race and Ethnicity

The STEP-3 cohort was 75% White, 17% Black or African American, and 22% Hispanic or Latino. The trial publication reported subgroup forest plots by race and ethnicity with confidence intervals that crossed zero for some smaller subgroups due to limited statistical power. Key observations:

• Black participants showed treatment differences broadly consistent with the overall cohort, though with wider confidence intervals.
• Hispanic/Latino participants responded at least as well as the overall population.
• No racial or ethnic subgroup showed evidence of a null treatment effect, but the trial was not powered to detect statistically significant interactions by race.

This limitation matters. Underrepresentation of non-White populations in the STEP trials has been a persistent criticism. Real-world data from the OSMO registry and post-marketing studies suggest similar efficacy across racial groups, but prospective evidence remains thin.

Glycemic Status: Prediabetes and the STEP-3 Wrinkle

STEP-3 excluded participants with established type 2 diabetes (those patients were studied in STEP-2). Roughly 50% of STEP-3 participants had prediabetes (HbA1c 5.7% to 6.4%) at baseline. This created a natural subgroup comparison.

Participants with prediabetes lost less weight in both arms. This aligns with findings from STEP-2 (which enrolled only patients with type 2 diabetes) showing a clear inverse relationship between baseline insulin resistance and weight loss response. The biological rationale is straightforward: hyperinsulinemia promotes lipogenesis and opposes lipolysis, creating a metabolic headwind against weight loss regardless of caloric intake. GLP-1 receptor agonists improve insulin sensitivity over time, which may explain why some patients with prediabetes showed accelerated weight loss after week 20 in the STEP-3 time-course data.

Post-Hoc Analyses and Categorical Response Rates

Beyond mean percentage weight loss, categorical responder analyses reveal the distribution of outcomes.

Proportion Achieving Clinically Meaningful Thresholds

Over one-third of semaglutide-treated participants lost at least 20% of their body weight. This threshold is significant because bariatric surgery benchmarks for sleeve gastrectomy report approximately 25% total body weight loss at one year. The semaglutide+IBT combination put a meaningful proportion of patients into a range previously achievable only through surgery.

Non-Responders

Approximately 13% of semaglutide-treated participants did not achieve the ≥5% weight loss threshold. These "non-responders" likely include patients with poor medication adherence (nausea-driven discontinuation), genetic variation in GLP-1 receptor expression, and those whose caloric intake patterns were resistant to appetite suppression. The STEP-3 protocol did not collect granular adherence data for the IBT sessions, making it impossible to separate pharmacologic non-response from behavioral non-adherence.

The IBT Contribution: What the Placebo Arm Tells Us

The placebo arm lost 5.7% of body weight, a result that exceeded the placebo arms in STEP-1 (−2.4%) and STEP-4 (limited comparability due to design). This 3-percentage-point gap between STEP-1 and STEP-3 placebo arms quantifies the independent contribution of IBT. For the semaglutide arms, the gap was smaller: −16.0% in STEP-3 vs −14.9% in STEP-1 (a difference of about 1.1 percentage points). This suggests diminishing marginal returns when stacking IBT on top of pharmacotherapy. The drug already suppresses appetite so effectively that the additional behavioral coaching contributes proportionally less than it does alone.

Limitations of the Subgroup Data

Several constraints limit how far clinicians can push these findings:

1. Multiplicity. No corrections for multiple comparisons were applied to subgroup analyses. Treatment-by-subgroup interaction p-values were reported, and none reached statistical significance, meaning the observed differences could reflect chance variation.

2. Sample size. With 611 total participants, individual subgroups (especially men, racial minorities, and the ≥40 BMI stratum) had limited statistical power. Confidence intervals for treatment differences in smaller subgroups were wide.

3. Short duration. 68 weeks is clinically informative but does not address durability. STEP-4 and the STEP-5 extension showed weight regain after discontinuation, raising questions about whether subgroup differences persist, converge, or diverge over longer horizons.

4. Homogeneous population. Participants were overwhelmingly White, female, and without type 2 diabetes. Generalizability to men, non-White populations, and patients with established diabetes requires cross-referencing with STEP-2 data and real-world evidence.

5. IBT standardization. All 611 participants were assigned to the same 30-session IBT program. In practice, access to structured behavioral therapy varies enormously. Whether the subgroup patterns hold under less intensive behavioral support is unknown.

Clinical Translation: What Prescribers Should Take Away

For patients considering semaglutide with or without behavioral support, the STEP-3 subgroup data suggest the following practical points:

• Expect larger percentage weight loss in patients with lower baseline BMI. A patient starting at BMI 33 will likely see a higher percentage reduction than one starting at BMI 45, though both will benefit substantially.
• Women may respond more robustly than men, but the difference is modest (about 2 to 3 percentage points) and should not influence prescribing decisions.
• Patients with prediabetes or insulin resistance should be counseled that their trajectory may be slower but still clinically meaningful. Monitoring HbA1c and fasting insulin alongside weight can help frame progress.
• Adding structured behavioral therapy to semaglutide provides a real but incremental benefit (roughly 1 percentage point of additional weight loss). For patients already on semaglutide who cannot access IBT, the drug alone delivers the majority of the effect.
• Non-response (failure to lose ≥5%) occurred in about 1 in 8 patients. Early weight loss trajectory (by week 16) may help identify non-responders who need dose escalation, adherence support, or alternative strategies.

Frequently asked questions

›Were the STEP-3 subgroup analyses pre-specified or post-hoc?

Most subgroup analyses (by sex, age, baseline BMI, race/ethnicity) were pre-specified in the statistical analysis plan. Categorical responder analyses (≥5%, ≥10%, ≥15%, ≥20% thresholds) were also pre-specified secondary endpoints. Some exploratory analyses by glycemic status and time-course trajectories were characterized as post-hoc or descriptive.

›Did any subgroup fail to benefit from semaglutide in STEP-3?

No. Every pre-specified subgroup showed statistically significant weight loss superiority for semaglutide+IBT over placebo+IBT. Treatment-by-subgroup interaction tests were non-significant across all subgroups, indicating consistent efficacy.

›Why did participants with higher baseline BMI lose less percentage body weight?

Higher baseline body mass means a larger denominator for percentage calculations. Absolute weight loss (in kg) was actually similar or slightly greater in the higher BMI subgroups. The percentage gap also reflects the metabolic reality that fixed-dose appetite suppression has proportionally less impact relative to total energy expenditure at higher body weights.

›How did STEP-3 define intensive behavioral therapy?

IBT consisted of 30 individual counseling sessions over 68 weeks, delivered by registered dietitians or trained counselors. Sessions covered dietary modification, physical activity (target 100 min/week increasing to 200 min/week), and behavioral strategies. All participants also followed an initial 8-week low-calorie diet (1,000 to 1,200 kcal/day) using meal replacements.

›Was STEP-3 powered to detect subgroup differences by race?

No. The trial enrolled 75% White participants. Black (17%) and Hispanic/Latino (22%) subgroups were too small for definitive conclusions about differential efficacy. The forest plots showed overlapping confidence intervals across racial subgroups, consistent with similar treatment effects but not definitive.

›How does the STEP-3 subgroup data compare to STEP-1 subgroups?

Patterns were broadly consistent. Both trials showed larger percentage weight loss in women, younger participants, and those with lower baseline BMI. The main difference is that STEP-3's placebo arm performed better due to IBT, narrowing the absolute treatment difference in some subgroups compared to STEP-1.

›Should prescribers adjust semaglutide dosing based on subgroup membership?

The FDA-approved Wegovy label recommends the same dose escalation schedule regardless of sex, age, BMI, or race. Subgroup data from STEP-3 inform expectations and counseling but do not support dose modifications. The 2.4 mg maintenance dose was used uniformly across all subgroups.

›What happened to subgroup weight loss after semaglutide was stopped?

STEP-3 did not include a structured post-treatment follow-up period. STEP-4 and STEP-5 data show that weight regain occurs in most patients after discontinuation, regardless of subgroup. Whether certain subgroups maintain weight loss better than others remains an open research question.

›Did the low-calorie diet phase in STEP-3 affect subgroup outcomes differently?

The 8-week low-calorie diet phase produced similar early weight loss across subgroups. The divergence between semaglutide and placebo arms became apparent primarily after week 8, during the maintenance phase when appetite suppression from semaglutide had its greatest differential impact.

›Is intensive behavioral therapy necessary when prescribing semaglutide for obesity?

Not strictly necessary. STEP-1 demonstrated strong weight loss (−14.9%) with semaglutide plus standard lifestyle counseling (not IBT). STEP-3 showed that adding IBT contributed roughly 1 additional percentage point of weight loss on top of semaglutide. For patients without access to structured IBT, semaglutide alone with basic dietary guidance remains highly effective.

References

1. Wadden TA, Bailey TS, Billings LK, et al. Effect of subcutaneous semaglutide vs placebo as an adjunct to intensive behavioral therapy on body weight in adults with overweight or obesity: the STEP 3 randomized clinical trial. JAMA. 2021;325(14):1403-1413. PubMed

2. 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

3. 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

4. Amaro A, Sugimoto D, Wharton S, et al. Efficacy and safety of semaglutide for weight management: evidence from the STEP program. Postgrad Med. 2023;135(sup1):56-67. PubMed

5. U.S. Food and Drug Administration. Wegovy (semaglutide) injection prescribing information. Revised 2023. FDA Label

6. Grunvald E, Shah R, Herber-Gast R, et al. AGA clinical practice guideline on pharmacological interventions for adults with obesity. Gastroenterology. 2022;163(5):1198-1225. PubMed