Was STEP-8 double-blinded?

Only partially. The semaglutide vs. placebo comparison was double-blinded, but the semaglutide vs. liraglutide comparison was open-label. Patients and investigators knew which active drug was being used, which can introduce bias in subjective outcomes and adherence.

Why were patients with type 2 diabetes excluded from STEP-8?

The trial targeted the chronic weight management indication in adults without diabetes. This aligns with the FDA-approved labeling for Wegovy but limits generalizability to the large population of patients with obesity and coexisting diabetes.

How diverse was the STEP-8 study population?

The sample was approximately 75-84% White and 78% female. Representation of Black, Hispanic, Asian, and male participants was limited, raising questions about applicability across demographic groups.

Is 68 weeks long enough to evaluate obesity medications?

It is a standard regulatory timeframe, but it does not capture long-term weight maintenance, regain patterns after discontinuation, or cardiovascular event rates. Extension and outcome trials are needed for those questions.

Did Novo Nordisk fund the STEP-8 trial?

Yes. Novo Nordisk funded the trial and was involved in design, data collection, analysis, and manuscript preparation. The company manufactures both semaglutide (Wegovy) and liraglutide (Saxenda).

Were the doses of semaglutide and liraglutide fairly compared?

Both drugs were used at their maximum FDA-approved doses for weight management. However, critics note that maximum approved dose does not guarantee equivalent pharmacological exposure, given the drugs' different half-lives and receptor-binding properties.

Did STEP-8 measure cardiovascular outcomes?

No. The trial measured weight change and cardiometabolic biomarkers but did not adjudicate cardiovascular events. The SELECT trial later showed cardiovascular benefit for semaglutide 2.4 mg, but in a different population.

How much weight did participants regain after stopping semaglutide?

STEP-8 did not include an off-treatment follow-up period. Data from the STEP-1 extension study showed approximately two-thirds of lost weight was regained within one year of semaglutide discontinuation.

Could gastrointestinal side effects explain some of semaglutide's weight-loss advantage?

Possibly. GI side effects (nausea, vomiting, diarrhea) were numerically more common with semaglutide. Without detailed dietary intake data, the trial cannot fully separate appetite-mediated weight loss from caloric reduction driven by GI intolerance.

Should STEP-8 alone determine which GLP-1 a clinician prescribes for obesity?

No. STEP-8 provides one data point in a broader evidence base. Insurance coverage, patient preference for weekly vs. daily injections, individual tolerability, comorbid conditions, and cost should all factor into prescribing decisions.

Honest Criticisms and Limitations of the STEP-8 Trial

By HealthRX Medical Team

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

At a glance

| Parameter | Detail | |---|---| | N | 338 (semaglutide 2.4 mg: 126, liraglutide 3.0 mg: 127, placebo: 85) | | Intervention | Subcutaneous semaglutide 2.4 mg once weekly | | Comparator | Subcutaneous liraglutide 3.0 mg once daily + placebo | | Duration | 68 weeks (20-week dose escalation + 48-week maintenance) | | Primary endpoint | Percentage change in body weight from baseline to week 68 | | Key result | Semaglutide: −15.8% vs liraglutide: −6.4% (estimated treatment difference: −9.4 percentage points; P <0.001) |

Why a Limitations Deep-Dive Matters

STEP-8 was the first randomized trial to directly pit semaglutide 2.4 mg against liraglutide 3.0 mg for chronic weight management. The topline result, a roughly 2.5-fold greater weight loss with semaglutide, dominated headlines and shaped prescribing conversations. But headline numbers can obscure real methodological issues. This page catalogs each limitation, drawing on the published trial, editorial commentary, and related evidence so clinicians can form a more complete picture.

Limitation Framework for Evaluating STEP-8

The criticisms below are organized into six categories. Each one addresses a different threat to the trial's internal validity, external applicability, or clinical interpretation.

| Category | Core Question | |---|---| | Enrollment bias | Who was studied, and who was left out? | | Duration and design | Was 68 weeks long enough to answer the right questions? | | Open-label structure | Could knowledge of treatment assignment inflate differences? | | Statistical considerations | Do the analytical choices hold up under scrutiny? | | Sponsor and COI | How should industry funding color interpretation? | | Post-publication commentary | What did peer reviewers and letter-writers flag? |

1. Enrollment Bias and Narrow Demographics

STEP-8 enrolled adults aged 18 or older with a BMI of 30 kg/m² or higher (or 27 kg/m² with at least one weight-related comorbidity) who did not have diabetes. This mirrors the FDA-approved indication for semaglutide 2.4 mg, but it excludes the large population of patients with type 2 diabetes who represent a major segment of real-world GLP-1 prescribing.

The trial's demographic breakdown further narrows generalizability:

| Characteristic | Semaglutide group | Liraglutide group | |---|---|---| | Female (%) | 78 | 78 | | White (%) | 75 | 84 | | Mean age (years) | 49 | 49 | | Mean BMI (kg/m²) | 37.2 | 38.0 | | Mean body weight (kg) | 104.5 | 108.8 |

The sample was predominantly White women in their late 40s. Representation of Black, Hispanic, and Asian participants was limited. Given documented differences in GLP-1 pharmacokinetics and body composition across racial and ethnic groups, STEP-8 cannot speak confidently to outcomes in populations that were underrepresented. The exclusion of patients with type 2 diabetes also means the trial says nothing about the interaction between glycemic control and weight-loss magnitude, an interaction that other STEP-program trials (STEP-2, specifically) showed can shift expected results.

Patients with a history of bariatric surgery, severe psychiatric illness, or active eating disorders were excluded. While these exclusions protect internal validity, they remove exactly the complex patients who present most frequently in obesity medicine clinics.

2. Duration and Design Constraints

Sixty-eight weeks is a standard timeframe for obesity pharmacotherapy trials, but it is too short to answer the questions that matter most: weight regain, cardiovascular events, and long-term tolerability.

Data from the STEP-1 extension study demonstrated that participants regained approximately two-thirds of lost weight within one year of stopping semaglutide. STEP-8 included no off-treatment follow-up phase. Clinicians therefore have no data from this trial on what happens when patients discontinue either drug or switch between them.

The 20-week dose-escalation period is another design element worth scrutinizing. Semaglutide was titrated from 0.25 mg weekly up to the full 2.4 mg dose over 16 weeks. Liraglutide was titrated from 0.6 mg daily to 3.0 mg over roughly 4 weeks, following its approved labeling. This means semaglutide spent a longer proportion of the trial at sub-therapeutic doses. Some commentators argued this asymmetry might actually have understated semaglutide's advantage, while others noted it simply reflects real-world dosing protocols and is therefore a reasonable design choice.

The trial also used a run-in period with lifestyle counseling before randomization. Participants who could not comply during this phase were excluded, which likely enriched the study population for motivated, adherent individuals.

3. Open-Label Comparator and Blinding Concerns

STEP-8 used a partially blinded design. The semaglutide versus placebo comparison was double-blinded. The semaglutide versus liraglutide comparison was open-label.

This is a significant limitation. Patients and investigators knew which active drug was being administered. Awareness of treatment assignment can influence:

Patient-reported outcomes and subjective well-being
Adherence to the lifestyle intervention component
Investigator enthusiasm during counseling visits
Threshold for reporting adverse events

The open-label design was a practical concession. Semaglutide is injected once weekly; liraglutide requires daily injections. Maintaining a true double-blind would have required daily placebo injections in the semaglutide group and weekly placebo injections in the liraglutide group, a significant burden. The investigators acknowledged this tradeoff in the published paper, but it remains a vulnerability. Any trial where the comparator requires more frequent self-injection inherently disadvantages that arm through differential burden and reduced blinding.

4. Statistical and Analytical Caveats

The primary analysis used a treatment-policy estimand (intention-to-treat principle), which accounts for treatment discontinuation. Sensitivity analyses using a trial-product estimand (on-treatment data only) showed even larger differences favoring semaglutide. Both approaches have merit. The treatment-policy approach is more conservative and clinically relevant, since it reflects what happens in practice when some patients stop therapy.

Several statistical points deserve attention:

Sample size. At 338 participants across three arms, STEP-8 was modestly powered. The trial was designed to detect a difference of 7 percentage points in weight change with 90% power, and it achieved this comfortably. But the sample was not powered for subgroup analyses by sex, race, baseline BMI, or comorbidity status. Any subgroup findings should be treated as hypothesis-generating only.

Multiple comparisons. The trial had two primary comparisons (semaglutide vs. liraglutide and semaglutide vs. placebo) and multiple secondary endpoints. A hierarchical testing procedure was used to control the family-wise error rate. This is appropriate methodology, but clinicians should note that some secondary endpoints could have been tested only if prior endpoints in the hierarchy reached significance.

Missing data. Approximately 15-20% of participants in each group discontinued treatment before week 68. The primary analysis used a mixed model for repeated measures, which handles missing data under a missing-at-random assumption. If dropout was related to unmeasured factors (such as privately experienced side effects not captured in the data), the results could be biased.

STEP-8 was funded by Novo Nordisk, the manufacturer of both semaglutide (Wegovy) and liraglutide (Saxenda). The company was involved in trial design, data collection, statistical analysis, and manuscript preparation.

This creates an unusual conflict dynamic. Novo Nordisk manufactures both drugs being compared, which means a positive result for semaglutide comes partly at the expense of liraglutide sales. In practice, the financial incentive clearly favored semaglutide: Wegovy carries a higher price point and represents the company's growth strategy, while Saxenda (liraglutide 3.0 mg) was already facing generic competition timelines.

Multiple authors disclosed financial relationships with Novo Nordisk, including consulting fees and research grants. This does not invalidate the data, but it is standard practice to consider sponsor involvement when interpreting effect sizes, particularly in an open-label comparison. An independently funded replication would meaningfully strengthen confidence in the magnitude of difference observed.

6. Post-Publication Commentary and Peer Criticism

Several points were raised in editorials and letters-to-the-editor following the trial's publication in JAMA:

Dose equivalence. Critics questioned whether 2.4 mg of once-weekly semaglutide and 3.0 mg of once-daily liraglutide represent pharmacologically "fair" comparisons. Semaglutide 2.4 mg is the maximum approved obesity dose. Liraglutide 3.0 mg is also its approved ceiling. But the GLP-1 receptor engagement profiles differ substantially due to differences in half-life (approximately 7 days for semaglutide vs. 13 hours for liraglutide), albumin binding, and steady-state pharmacokinetics. Some argued that comparing maximum approved doses does not necessarily mean comparing maximum pharmacological potential.

Gastrointestinal tolerability. Nausea, vomiting, and diarrhea were common in both groups but numerically more frequent with semaglutide. Whether the greater weight loss is partly mediated by reduced caloric intake due to GI side effects (rather than central appetite suppression alone) remains debated. The trial did not include validated dietary recall data sufficient to disentangle these mechanisms.

Absence of hard outcomes. STEP-8 measured weight change, waist circumference, and cardiometabolic biomarkers. It did not assess cardiovascular events, quality of life using validated instruments, or patient-centered functional outcomes. The SELECT trial subsequently demonstrated cardiovascular benefit for semaglutide 2.4 mg, but that was a different population (established cardiovascular disease) and a different question. STEP-8 alone cannot support claims about cardiovascular superiority of semaglutide over liraglutide.

Lifestyle intervention intensity. Both arms received monthly counseling sessions encouraging a 500 kcal/day deficit and 150 minutes/week of physical activity. The real-world intensity of lifestyle support varies enormously. Patients receiving less structured counseling may see different absolute weight-loss numbers with either drug.

What This Means for Clinical Decision-Making

STEP-8 provides strong evidence that semaglutide 2.4 mg produces greater short-term weight loss than liraglutide 3.0 mg in a selected population. That finding is internally valid and consistent with the broader STEP program. But clinicians choosing between these agents should recognize what the trial does not tell them: whether the difference persists beyond 68 weeks, whether it holds in diverse populations, whether it translates to differences in hard clinical endpoints when the two drugs are compared directly, and how much of the effect-size gap is amplified by open-label awareness.

For patients with type 2 diabetes, prior bariatric surgery, or significant psychiatric comorbidity, STEP-8's results require extrapolation rather than direct application. Cost, injection frequency preference, insurance formulary constraints, and individual tolerability often matter more in practice than the 9.4-percentage-point difference observed under controlled trial conditions.

Frequently asked questions

›

References

Rubino DM, Greenway FL, Khalid U, et al. Effect of weekly subcutaneous semaglutide vs daily liraglutide on body weight in adults with overweight or obesity without diabetes: the STEP 8 randomized clinical trial. JAMA. 2022;327(2):138-150. PubMed
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. PubMed
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
Lincoff AM, Brown-Frandsen K, Colhoun HM, et al. Semaglutide and cardiovascular outcomes in obesity without diabetes. N Engl J Med. 2023;389(24):2221-2232. PubMed
Wegovy (semaglutide) prescribing information. U.S. Food and Drug Administration. FDA Label
Saxenda (liraglutide) prescribing information. U.S. Food and Drug Administration. FDA Label