Was the STEP-1 Extension designed as a discontinuation study?

No. The original STEP-1 trial was a 68-week treatment RCT. The extension was an observational off-treatment follow-up added afterward. There was no re-randomization at the point of discontinuation, which limits causal conclusions about drug withdrawal specifically.

How representative was the extension cohort?

Only 327 of 1,306 semaglutide-treated participants (25%) entered the extension. The cohort was predominantly white, female, and free of type 2 diabetes. This limits generalizability to the broader population prescribed semaglutide for weight management.

Did all participants regain weight after stopping semaglutide?

The group average showed about two-thirds of weight loss regained, but individual responses varied. Some participants maintained clinically meaningful weight loss at week 120, while others returned to baseline or higher. The trial did not extensively characterize predictors of maintained loss.

Could the lifestyle intervention ending explain part of the regain?

Possibly. Structured lifestyle counseling ended alongside semaglutide discontinuation. The trial design cannot separate the drug withdrawal effect from the loss of behavioral support.

Was Novo Nordisk involved in the study design?

Yes. Novo Nordisk funded the trial, employed several authors, and participated in design, data collection, analysis, and manuscript preparation. This is standard for industry-sponsored trials but means the framing and emphasis of results reflect sponsor interests.

Did the trial test gradual tapering or dose reduction?

No. All participants stopped semaglutide abruptly at week 68. The trial did not evaluate tapering, dose reduction, intermittent dosing, or transition to alternative medications.

How does the STEP-1 Extension compare to STEP-4?

STEP-4 used a withdrawal design with re-randomization at week 20 (continue semaglutide vs. switch to placebo), providing stronger evidence about continued treatment benefit. The two trials complement each other but address different questions.

Is one year of follow-up enough to understand weight regain?

One year captures the acute rebound phase but does not reveal whether weight stabilizes at a new plateau or continues rising. Longer-term (2-5 year) post-cessation data are needed to fully characterize discontinuation trajectories.

Does regained weight have the same body composition as the original weight?

Preliminary evidence suggests weight loss on GLP-1 agonists includes both fat and lean mass, while regain is disproportionately fat. This body composition shift was not a primary focus of the STEP-1 Extension but has clinical implications for metabolic health.

Should patients stop semaglutide based on this trial?

This trial does not provide a recommendation for or against stopping. It quantifies what happens on average when treatment stops abruptly without support. Clinical decisions about discontinuation should be individualized and ideally guided by ongoing monitoring.

Honest Criticisms and Limitations of the STEP-1 Extension Trial

By HealthRX Medical Team

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

At a glance

| Detail | Value | |---|---| | N (extension cohort) | 327 | | Intervention | Semaglutide 2.4 mg weekly for 68 weeks, then discontinuation | | Comparator | Off-treatment follow-up (no placebo re-randomization) | | Follow-up after stopping | 1 year (week 68 to week 120) | | Primary endpoint | Change in body weight from week 68 to week 120 | | Key result | ~two-thirds of weight loss regained by week 120 | | Sponsor | Novo Nordisk |

What the Extension Actually Measured

The original STEP-1 trial randomized 1,961 adults with obesity (or overweight with at least one weight-related comorbidity) to semaglutide 2.4 mg or placebo for 68 weeks. After the treatment period ended, a subset of 327 participants from the semaglutide arm entered an off-treatment extension that tracked weight, cardiometabolic markers, and adverse events for an additional 52 weeks.

The headline finding was stark: participants who had lost an average of 17.3% of body weight by week 68 regained about 11.6 percentage points of that loss by week 120. Cardiometabolic improvements in waist circumference, blood pressure, and lipids also reversed substantially.

But the trial was not designed as a discontinuation study from the outset. It was an observational extension grafted onto a treatment RCT, and that distinction matters for every conclusion drawn from it.

Enrollment and Selection Bias

Who Entered the Extension

Of the 1,306 participants originally randomized to semaglutide, only 327 (25%) continued into the off-treatment extension. The published data do not fully characterize what distinguished these 327 from the 979 who did not continue.

This creates a classic volunteer bias problem. Participants who agree to extended monitoring may differ systematically from those who decline. They could be more motivated, healthier, more engaged with clinical follow-up, or conversely, more concerned about weight regain. Either direction of bias skews the generalizability of the regain estimate.

Baseline Population Concerns

The original STEP-1 cohort was 74% female, 75% white, with a mean BMI of 37.9 kg/m². Mean age was 46 years. These demographics reflect a narrow slice of the global obesity population. Male patients, older adults, and individuals from non-white racial and ethnic backgrounds were underrepresented, a pattern FDA reviewers have flagged across GLP-1 obesity trials.

Exclusion criteria removed anyone with type 2 diabetes, prior bariatric surgery, or recent use of weight-loss medications. Real-world semaglutide discontinuation often happens in patients with these exact characteristics. The trial population was, by design, a healthier-than-typical cohort.

Methodological Gaps

No Re-Randomization at Discontinuation

A rigorous discontinuation study would re-randomize participants at the point of stopping: some continue treatment, others stop, and outcomes are compared prospectively. The STEP-1 Extension did not do this. Everyone stopped semaglutide at week 68. Without a continued-treatment comparator arm during the extension, the study cannot distinguish between weight regain due to drug withdrawal and weight regain that might have occurred even on continued therapy (due to metabolic adaptation or behavioral drift).

The STEP-4 trial, published in JAMA in 2022, partially addressed this gap by randomizing participants at week 20 to either continue semaglutide or switch to placebo. That trial confirmed the regain signal, but with a different design and timeline.

Lifestyle Intervention Confounding

During the initial 68-week treatment period, all participants received monthly lifestyle counseling. After discontinuation, this structured support also ended. The extension therefore measured the combined effect of stopping both the drug and the behavioral intervention simultaneously.

Separating pharmaceutical from behavioral contributions is not possible with this design. A patient who discontinues semaglutide but maintains intensive dietary counseling and exercise coaching may experience different weight trajectories than the extension data suggest.

Duration of Follow-Up

One year of off-treatment observation captures the acute rebound phase but tells us nothing about longer-term weight trajectories. Several questions remain open: Does weight stabilize at some new set point below baseline? Does regain continue past 12 months until all lost weight returns? Does a subset of patients maintain clinically significant weight loss long-term without the drug?

The AGA clinical practice guideline on pharmacotherapy for obesity (2022) acknowledges that long-term discontinuation data for GLP-1 agonists remain sparse. Two- and three-year post-cessation data would substantially change how clinicians counsel patients about stopping treatment.

Statistical Considerations

Handling of Missing Data

By week 120, a portion of extension participants had incomplete follow-up data. The primary analysis used a treatment-policy estimand (intention-to-treat principle), which includes all participants regardless of adherence to the monitoring schedule. Missing weight measurements were handled with multiple imputation.

This approach is standard but carries assumptions. If participants who regained the most weight were also more likely to drop out of follow-up (a plausible scenario given the psychological burden of weight regain), then the imputed values may underestimate the true magnitude of regain.

Percentage Points vs. Clinical Thresholds

The headline "two-thirds regained" is mathematically accurate but obscures important individual variation reported in the STEP-1 Extension data. Some participants maintained substantial weight loss at week 120, while others returned to or exceeded their baseline weight. Reporting group means smooths over this heterogeneity.

Clinically, the question that matters to patients is not "what happened on average" but "what is the probability that I specifically will maintain a 5% or 10% weight loss after stopping." The trial provides limited data on responder subgroup analyses at the individual level.

Cardiometabolic Reversals Reported Without Context

The extension reported that improvements in waist circumference, C-reactive protein, and lipid panels reversed after discontinuation. These reversal statistics were presented alongside weight data but without adjustment for the degree of weight regain in each participant. Whether cardiometabolic deterioration tracked proportionally with weight regain or followed a distinct pattern was not well characterized.

Novo Nordisk Involvement

Novo Nordisk funded the trial, employed several of the authors, and was involved in study design, data collection, data analysis, and manuscript preparation. This level of sponsor integration is common in phase 3 and extension studies but warrants scrutiny.

The trial's conclusion, that obesity requires chronic pharmacotherapy, directly supports the commercial interests of the company selling that pharmacotherapy. This does not mean the finding is wrong. The biology of weight regulation does support the chronic-disease model. But the framing of results, the choice of endpoints, and the narrative emphasis all passed through a sponsor filter.

What Was Not Studied

The trial did not compare discontinuation outcomes against dose reduction, intermittent dosing, or transition to lower-cost alternatives. From a public health perspective, these are the questions that matter most. A trial designed by an independent academic consortium might have prioritized these comparisons. A sponsor-funded trial had no commercial incentive to test whether patients could get by with less product.

External Commentary and Subsequent Debate

After publication, several letters to the editor and editorials raised points that the original manuscript did not emphasize.

The "chronic disease" framing debate. Critics noted that while the data clearly show weight regain after stopping semaglutide, using this to argue obesity must be treated chronically with medication conflates one specific intervention's rebound effect with the disease's inherent biology. Smoking cessation medications also show relapse after discontinuation, but nobody argues smokers must take varenicline for life. The analogy is imperfect, but the logical leap from "weight returns after stopping this drug" to "this drug must be taken indefinitely" deserved more cautious treatment.

Cost and access implications. At a list price exceeding $1,300 per month for Wegovy (semaglutide 2.4 mg), indefinite treatment creates enormous cost barriers. The extension data were cited in insurance coverage debates, with payers using the regain finding to justify both positions: some argued it proved the drug works and should be covered long-term, while others argued the temporary benefit did not justify permanent expenditure.

Body composition concerns. Weight regain after GLP-1 discontinuation may not restore the same ratio of lean mass to fat mass that existed before treatment. Preliminary data from body composition substudies suggest that loss includes both fat and lean tissue, while regain is disproportionately fat. This metabolic ratchet effect was not a focus of the STEP-1 Extension analysis but has become a significant concern in subsequent clinical discussion.

What Clinicians Should Take From This

The STEP-1 Extension confirmed something most obesity researchers already suspected: stopping an effective anti-obesity medication leads to weight regain. The trial quantified the magnitude and timeline of that regain with reasonable precision.

But the study cannot tell clinicians how to manage discontinuation. It cannot identify which patients might maintain weight loss off-drug. It cannot compare abrupt cessation against gradual tapering. And it cannot address whether intermittent or cyclical dosing strategies might preserve some benefit at lower cost and exposure.

These are the questions that matter for clinical practice, and they remain largely unanswered. The STEP-1 Extension is a useful data point, not a treatment algorithm.

Frequently asked questions

›

References

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
Rubino D, Abrahamsson N, Davies M, et al. Effect of continued weekly subcutaneous semaglutide vs placebo on weight loss maintenance in adults with overweight or obesity: The STEP 4 randomized clinical trial. JAMA. 2021;325(14):1414-1425. PubMed
FDA. Wegovy (semaglutide) prescribing information. 2021. FDA Label
Garvey WT, Mechanick JI, Brett EM, et al. American Association of Clinical Endocrinologists and American College of Endocrinology comprehensive clinical practice guidelines for medical care of patients with obesity. Endocr Pract. 2016;22(Suppl 3):1-203. PubMed
Amaro A, Sugimoto D, Crissman M, et al. AGA clinical practice guideline on pharmacological interventions for adults with obesity. Gastroenterology. 2022;163(5):1198-1225. PubMed