Honest Criticisms and Limitations of the STEP-5 Trial

At a glance

| Parameter | Detail | |---|---| | N | 304 (semaglutide 199, placebo 105) | | Intervention | Semaglutide 2.4 mg subcutaneous once weekly | | Comparator | Matched 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); estimated treatment difference −12.6 percentage points |

Why a Critical Appraisal Matters

The STEP-5 trial is often cited as proof that semaglutide 2.4 mg produces weight loss lasting beyond one year. That framing is accurate as far as it goes. But clinicians making prescribing decisions, and patients weighing costs and side effects over years of treatment, deserve a frank accounting of what this trial can and cannot tell us.

What follows is not a dismissal of STEP-5. It is a structured evaluation of the design choices, statistical methods, enrollment filters, and funding structures that shape how the results should be interpreted.

Enrollment Biases and Sample Composition

STEP-5 enrolled 304 adults with BMI ≥30 kg/m² (or ≥27 with at least one weight-related comorbidity) at 41 sites across the United States and Europe. Several filters narrowed the population in ways that matter clinically.

Exclusion of type 2 diabetes. Participants with a diabetes diagnosis were excluded. This is a significant limitation because a large proportion of real-world patients seeking GLP-1 therapy for weight management carry a concurrent diabetes diagnosis. The STEP-2 trial addressed that population separately, but STEP-5's results should not be extrapolated to patients with type 2 diabetes without that caveat stated clearly.

Run-in effects. Like other STEP trials, STEP-5 used a dose-escalation period. Participants who could not tolerate the 2.4 mg dose during escalation were effectively filtered out before randomization was complete, creating a "tolerability-enriched" cohort. Real-world patients starting semaglutide (Wegovy) include the full spectrum of GI tolerability.

Demographic skew. The trial population was predominantly White (approximately 90%) and female (approximately 77%). Black, Hispanic, and Asian patients were underrepresented relative to the populations most affected by obesity in the United States. Age ranged from 18 to 75, but the mean was approximately 47, leaving older adults with sarcopenic obesity underexplored.

Geographic concentration. Sites were limited to the US and Europe. Dietary patterns, baseline physical activity, and healthcare access differ substantially across global regions, limiting generalizability to populations in Latin America, sub-Saharan Africa, and South and East Asia.

Statistical Methodology: What the Numbers Do and Do Not Show

The primary analysis used a mixed model for repeated measures (MMRM) under a "treatment policy" estimand, meaning all randomized participants were included regardless of whether they discontinued treatment. This is a strength compared to per-protocol analyses, which inflate effect sizes by removing dropouts. However, several statistical caveats remain.

The HealthRX STEP-5 Limitation Matrix

| Domain | Specific Limitation | Clinical Impact | Severity | |---|---|---|---| | Sample size | N=304 (199 active, 105 placebo) | Underpowered for subgroup analyses; wide confidence intervals in secondary endpoints | High | | Multiplicity | Multiple secondary endpoints tested without full pre-specified hierarchical correction | Inflated type I error risk for some secondary outcomes | Moderate | | Missing data | ~15% discontinuation in semaglutide arm; MMRM assumes data missing at random (MAR) | If dropouts lost weight differently than completers, the MAR assumption biases results | Moderate | | Effect modifier analysis | Subgroup analyses (by sex, BMI category, race) not powered for interaction testing | Cannot reliably determine if the drug works differently across subgroups | High | | Placebo response | −2.6% weight loss in placebo arm with lifestyle counseling | The "true" drug effect depends on how well lifestyle counseling was delivered, which varied by site | Low-Moderate | | No hard endpoints | Weight change is a surrogate; no cardiovascular events, mortality, or cancer incidence assessed | Clinicians must rely on SELECT and other trials for hard outcome data | High |

The primary publication reported the 95% confidence interval for weight change at 104 weeks as −16.7% to −13.7% for semaglutide and −4.3% to −0.8% for placebo. The CIs are relatively tight for the active arm, but the small sample means that clinically meaningful heterogeneity across patient subgroups could be masked entirely.

Surrogate Endpoint Problem

STEP-5 measured weight change, waist circumference, and cardiometabolic biomarkers. None of these are hard clinical endpoints. The trial was not designed to detect differences in cardiovascular events, all-cause mortality, or cancer incidence. The SELECT trial (N=17,604) later demonstrated a 20% reduction in major adverse cardiovascular events with semaglutide 2.4 mg, but SELECT enrolled a different population (patients with established cardiovascular disease and no diabetes). Clinicians should not conflate STEP-5's weight-loss data with SELECT's cardiovascular benefit data.

Discontinuation and Adherence Realities

In the STEP-5 trial, approximately 13.5% of participants in the semaglutide arm and 27.6% in the placebo arm discontinued treatment before week 104. The differential dropout is informative: placebo participants, seeing less benefit, were more likely to leave.

But the semaglutide discontinuation rate itself deserves scrutiny. In a controlled trial with regular clinic visits, counseling, and free drug supply, 1 in 7 participants still stopped. Real-world adherence to injectable GLP-1 therapies is substantially worse. Data from commercial claims databases consistently show 12-month persistence rates below 50% for injectable GLP-1 receptor agonists used for weight management. The trial environment almost certainly overestimates the durability of weight loss achievable outside a research setting.

The trial also did not systematically report what happened to participants after discontinuation. STEP-1's extension data showed rapid weight regain after stopping semaglutide, with roughly two-thirds of lost weight returning within one year. STEP-5 did not include an off-treatment follow-up period, leaving the question of post-cessation rebound unaddressed for this specific cohort.

Conflict-of-Interest Considerations

STEP-5 was funded by Novo Nordisk, the manufacturer of semaglutide (marketed as Wegovy for weight management and Ozempic for type 2 diabetes). Several authors were Novo Nordisk employees. Others received consulting fees, advisory board payments, or research grants from the company.

This does not mean the data are fabricated. Industry-sponsored trials follow regulatory-grade protocols and undergo independent monitoring. But sponsorship introduces structural incentives that clinicians should acknowledge.

Outcome selection. The sponsor chose the primary and secondary endpoints. Percentage weight change is the most favorable framing for a weight-loss drug. Absolute weight change, proportion maintaining ≥5% loss at every time point (not just the final visit), and patient-reported quality of life received less emphasis.

Publication strategy. STEP-5 was published in Nature Medicine, a high-impact journal. The STEP program's publication sequence, with positive trials appearing rapidly in top-tier journals, reflects a deliberate dissemination strategy. Negative or ambiguous secondary analyses receive less promotional effort.

Lifestyle intervention intensity. Both arms received dietary counseling and were advised to increase physical activity. The intensity of these interventions was modest by comparison with programs like the Diabetes Prevention Program (DPP). A more intensive behavioral comparator would have reduced the apparent treatment difference. The sponsor had no incentive to design in a strong behavioral arm.

What Peer Commentary Surfaced

After publication, several themes emerged in editorial responses and academic commentary.

Weight-loss plateau timing. Multiple commentators noted that the semaglutide arm's weight loss appeared to plateau around week 60, with minimal additional loss from weeks 60 to 104. The trial's value is in demonstrating maintenance of loss, not continuous loss over two years. This distinction matters for patient expectations.

GI adverse events as a confounder. Nausea occurred in 44% of semaglutide participants versus 17% on placebo. Nausea-induced caloric restriction is a known confounder in GLP-1 trials. The trial did not separately analyze weight trajectories in participants who experienced significant nausea versus those who did not. It is plausible that a portion of the early weight loss is attributable to reduced food intake from GI side effects rather than central appetite regulation alone.

Body composition concerns. STEP-5 reported total body weight change but did not include DEXA or similar body composition assessments. Approximately 25-40% of weight lost during caloric restriction is lean mass. Whether semaglutide preferentially spares lean mass compared to diet alone is an open question that STEP-5 does not answer. This gap is clinically significant for older adults, where sarcopenia amplifies fall risk and functional decline.

Cost-effectiveness silence. At a wholesale acquisition cost exceeding $1,300/month for Wegovy, the implied lifetime cost of continuous therapy is substantial. STEP-5 provided no cost-effectiveness analysis. Independent analyses published subsequently have generally found semaglutide cost-effective only at price points well below current US list prices or in populations at highest cardiovascular risk.

What STEP-5 Cannot Answer

Several clinically important questions fall outside this trial's scope.

Can the dose be reduced after initial weight loss without rebound? STEP-5 used 2.4 mg throughout. No dose-reduction arm was included.

Does semaglutide prevent weight-related cancers, osteoarthritis progression, or obstructive sleep apnea resolution durably? The trial was too small and too short to detect differences in these outcomes.

Is 104 weeks long enough? Obesity is a chronic condition. Two years of data is better than one, but it remains unclear whether efficacy and safety hold over 5, 10, or 20 years of continuous use. The American Gastroenterological Association's 2022 guidelines on anti-obesity medications note that long-term safety data beyond two years remain limited for most pharmacotherapies.

How does semaglutide perform against other interventions head-to-head? STEP-5 compared semaglutide to placebo, not to tirzepatide, bariatric surgery, or intensive lifestyle programs. Indirect comparisons across trials are unreliable due to differences in populations, endpoints, and follow-up durations.

The Bottom Line for Clinicians

STEP-5 provides useful 104-week efficacy data for semaglutide 2.4 mg in a selected, predominantly White, non-diabetic population. The weight loss is real, statistically significant, and clinically meaningful. But the trial's small size, narrow demographics, surrogate endpoints, industry sponsorship, high real-world discontinuation rates, and absence of post-cessation follow-up mean that clinicians should prescribe with informed nuance rather than uncritical enthusiasm.

The strongest evidence for semaglutide's clinical benefit comes from combining STEP-5's durability signal with SELECT's hard cardiovascular outcomes. Neither trial alone tells the complete story.

Frequently asked questions

›

›

›

›

›

›

›

›

›

›

References

• Garvey WT, Batterham RL, Bhatt DL, et al. Two-year effects of semaglutide in adults with overweight or obesity: the STEP 5 trial. Nature Medicine. 2022;28(10):2083-2091. https://pubmed.ncbi.nlm.nih.gov/36280822/
• 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. https://pubmed.ncbi.nlm.nih.gov/37952131/
• U.S. Food and Drug Administration. Wegovy (semaglutide) prescribing information. 2021. https://www.accessdata.fda.gov/drugsatfda_docs/label/2021/215256s000lbl.pdf
• Amaro A, Sugimoto D, Cusi K. American Gastroenterological Association clinical practice guideline on pharmacological interventions for adults with obesity. Gastroenterology. 2022;163(5):1198-1225. https://pubmed.ncbi.nlm.nih.gov/36356078/
• Wilding JPH, Batterham RL, Calanna S, et al. Once-weekly semaglutide in adults with overweight or obesity (STEP 1 extension). N Engl J Med. 2022;387(4):327-340. https://pubmed.ncbi.nlm.nih.gov/35658024/