Inside the STEP-TEENS Methodology: What Most Summaries Skip

At a glance

| Parameter | Detail | |-----------|--------| | N | 201 (semaglutide 134, placebo 67) | | Intervention | Semaglutide 2.4 mg subcutaneous once weekly | | Comparator | Placebo injection + lifestyle intervention | | Duration | 68 weeks (16-week dose escalation + 52 weeks maintenance) | | Primary endpoint | Percentage change in BMI from baseline to week 68 | | Key result | −16.1% BMI change vs. −0.6% placebo (estimated treatment difference −14.0%; P <0.001) |

Why the trial population matters more than you think

Most summaries state that STEP-TEENS enrolled adolescents aged 12 to 17 with obesity. The actual inclusion criteria were more specific and clinically consequential. Participants needed a BMI at or above the 95th percentile for age and sex (the CDC definition of obesity in pediatrics), or a BMI at or above the 85th percentile with at least one weight-related comorbidity (Weghuber et al., 2022). They also required at least one documented unsuccessful weight-management effort.

This matters because the population skewed toward severe obesity. The mean baseline BMI was approximately 37 kg/m², well above the 95th-percentile threshold. Tanner stage had to be at least II (confirmed pubertal onset), which excluded younger children still in early growth phases. By requiring prior failed intervention, the trial selected for treatment-resistant adolescents, a population where placebo responses tend to be low.

Key exclusions: type 1 diabetes, prior bariatric surgery, and use of medications known to affect body weight (including other GLP-1 receptor agonists). Adolescents with type 2 diabetes were eligible, though they made up a small subset (FDA prescribing information, Wegovy).

Randomization and blinding architecture

STEP-TEENS used a 2:1 randomization ratio (semaglutide to placebo), stratified by Tanner stage (II-III vs. IV-V) and by the presence or absence of type 2 diabetes (Weghuber et al., 2022). The 2:1 ratio deserves attention. It generates more data on the active arm while reducing placebo exposure in minors. The tradeoff is reduced statistical power for a given total sample size compared with 1:1 allocation. At N = 201, this design still achieved significance because the effect size was large.

Both groups received identical lifestyle counseling (dietary guidance and physical activity recommendations), making this a placebo-controlled superiority trial with a lifestyle-intervention floor. The injections were matched in appearance. Site staff, participants, and outcome assessors were all blinded (Weghuber et al., 2022).

The estimand framework: treatment-policy vs. trial-product

This is where STEP-TEENS (and the broader STEP program) diverges from older obesity-trial methodology. The trial prespecified two estimands (Kelly et al., 2020; ICH E9 R1 addendum):

Treatment-policy estimand. The primary analysis. It evaluates the effect of initiating semaglutide regardless of whether participants actually continued treatment. If someone discontinued the drug at week 20 and was measured at week 68, their data still counted under the semaglutide arm. This mirrors real clinical decision-making: a physician prescribes; some patients stop; the population-level outcome is what matters for coverage and guideline decisions.

Trial-product estimand. A supplementary analysis. It estimates the effect assuming all participants stayed on treatment for the entire 68 weeks. This analysis censored data after treatment discontinuation and used a mixed model for repeated measures (MMRM). Results under this estimand were even larger (approximately −17% BMI change) because it excludes the dilution from early dropouts (Weghuber et al., 2022).

The distinction has regulatory and clinical implications. The FDA's 2022 expanded indication for Wegovy in adolescents aged 12 and older relied on the treatment-policy result, which is more conservative (FDA approval supplement).

Statistical methods: what drove the numbers

The primary analysis used an ANCOVA model with treatment group as a factor and baseline BMI, age, sex, and Tanner-stage stratum as covariates. Missing data were handled via multiple imputation under a "retrieved-dropout" assumption. This means missing week-68 values were imputed from observed data of participants who had the same reason for discontinuation and similar baseline characteristics (Weghuber et al., 2022).

The trial had co-primary endpoints: percentage change in BMI and the proportion of participants achieving at least 5% BMI reduction. A hierarchical testing procedure controlled family-wise type I error. The BMI-change endpoint was tested first; only if significant did the categorical endpoint proceed to formal testing.

Sample size was calculated assuming a 7% treatment difference in BMI change with 90% power at alpha = 0.025 (one-sided, equivalent to 0.05 two-sided). The observed difference of 14% was double the assumption, which explains the highly significant P value despite the modest sample (Weghuber et al., 2022).

BMI vs. body weight: why the primary endpoint choice matters

Adult STEP trials used percentage change in body weight. STEP-TEENS used percentage change in BMI. In growing adolescents, body weight is expected to increase with linear growth. A stable weight in a 14-year-old who grows 3 cm taller actually represents a meaningful reduction in adiposity. BMI accounts for height changes, making it the more appropriate measure in pediatrics (Styne et al., 2017).

The American Academy of Pediatrics 2023 clinical practice guideline for pediatric obesity endorsed BMI-based metrics for evaluating treatment response, consistent with the STEP-TEENS design (Hampl et al., 2023).

The comparator: lifestyle intervention alone

Both arms received counseling targeting 60 minutes of daily physical activity and a calorie-reduced diet. Critics might argue this does not reflect usual care, where many adolescents receive no structured counseling. However, the 2023 AAP guideline now recommends intensive health behavior and lifestyle treatment (IHBLT) as the baseline for all adolescents with obesity (Hampl et al., 2023). In that context, the comparator arm actually represents guideline-concordant care, making the placebo result (−0.6% BMI change) a fair baseline.

The near-zero placebo response also confirms what prior trials demonstrated: lifestyle intervention alone produces minimal BMI reduction in adolescents with established severe obesity, supporting pharmacotherapy as an adjunct rather than a failure of willpower (Kelly et al., 2020).

Limitations the authors disclosed (and one they didn't)

The published paper acknowledged several limitations (Weghuber et al., 2022):

• Short duration relative to a chronic disease. 68 weeks provides no data on long-term growth, bone density, or weight regain after discontinuation.
• Limited racial and ethnic diversity. The trial enrolled participants from 5 countries, with underrepresentation of Black and Hispanic adolescents relative to obesity prevalence in those groups.
• No data on younger children. Enrollment started at age 12; nothing in STEP-TEENS speaks to safety or efficacy before puberty.
• Small subgroups. Only 5% of participants had type 2 diabetes, making subgroup conclusions about glycemic outcomes unreliable.

A limitation not explicitly highlighted: the 7-week follow-up after treatment cessation showed rapid BMI regain, consistent with adult STEP 1 extension data (Wilding et al., 2022). This suggests that stopping semaglutide in adolescents leads to weight rebound, raising questions about indefinite treatment duration in a still-developing population.

How these choices shape clinical translation

The methodology decisions in STEP-TEENS directly affect how clinicians should interpret and apply the results:

| Design choice | Clinical implication | |---------------|---------------------| | Treatment-policy estimand | The 16.1% result is achievable in real practice, not just under perfect adherence | | 2:1 randomization | More safety data on semaglutide, fewer teens exposed to placebo | | BMI primary endpoint | Accounts for adolescent growth; don't compare directly to adult weight-loss percentages | | 68-week duration | Long enough for efficacy signal, too short for growth-plate or bone-density conclusions | | Lifestyle-intervention floor | Both arms got counseling; the drug effect is additive to behavioral treatment |

The Endocrine Society's 2023 guideline on pediatric obesity pharmacotherapy cited STEP-TEENS as the basis for recommending semaglutide in adolescents aged 12 and older with BMI at or above the 95th percentile who have not responded to lifestyle intervention alone (Styne et al., 2017; updated guidance).

Frequently asked questions

›

›

›

›

›

›

›

›

›

›

References

1. Weghuber D, Barrett T, Barrientos-Pérez M, et al. Once-weekly semaglutide in adolescents with obesity. N Engl J Med. 2022;387(24):2245-2257. PubMed
2. Hampl SE, Hassink SG, Skinner AC, et al. Clinical practice guideline for the evaluation and treatment of children and adolescents with obesity. Pediatrics. 2023;151(2):e2022060640. PubMed
3. Wilding JPH, Batterham RL, Davies M, et al. Weight regain and cardiometabolic effects after withdrawal of semaglutide. Diabetes Obes Metab. 2022;24(8):1553-1564. PubMed
4. Styne DM, Arslanian SA, Connor EL, et al. Pediatric obesity: assessment, treatment, and prevention. J Clin Endocrinol Metab. 2017;102(3):709-757. PubMed
5. Kelly AS, Auerbach P, Barrientos-Pérez M, et al. A randomized, controlled trial of liraglutide for adolescents with obesity. N Engl J Med. 2020;382(22):2117-2128. PubMed
6. FDA. Wegovy (semaglutide) prescribing information. Revised 2023. FDA Label