The STEP Trials Compared: STEP-1, STEP-2, STEP-3, STEP-5, and STEP-8, A Cross-Trial Synthesis

By
Published Updated Last reviewed
GLP-1 medication and metabolic health image for The STEP Trials Compared: STEP-1, STEP-2, STEP-3, STEP-5, and STEP-8, A Cross-Trial Synthesis

At a Glance

At a glance

| Trial | N (sema / control) | Semaglutide dose | Follow-up | Primary endpoint | Primary result (sema vs control) | Dropout rate (sema) | Key adverse events | |---|---|---|---|---|---|---|---| | STEP-1 (Wilding 2021) | 1306 / 655 | 2.4 mg SC QW | 68 weeks | % change in body weight | −14.9% vs −2.4% | ~7% | Nausea (44%), diarrhea (30%), vomiting (24%) | | STEP-2 (Davies 2021) | 404 / 403 (1.0 mg) / 403 (placebo) | 2.4 mg SC QW | 68 weeks | % change in body weight | −9.6% vs −3.4% (placebo) | ~9% | Nausea (42%), diarrhea (29%) | | STEP-3 (Wadden 2021) | 407 / 204 | 2.4 mg SC QW + IBT | 68 weeks | % change in body weight | −16.0% vs −5.7% | ~9% | Nausea (49%), vomiting (30%) | | STEP-5 (Garvey 2022) | 304 / 152 | 2.4 mg SC QW | 104 weeks | % change in body weight | −15.2% vs −2.6% | ~14% | Nausea (43%), diarrhea (30%) | | STEP-8 (Rubino 2022) | 338 / 339 (liraglutide) | 2.4 mg SC QW | 68 weeks | % change in body weight | −15.8% vs −6.4% (liraglutide 3.0 mg) | ~4% | Nausea (44% sema vs 46% lira) |

Dropout rates are approximate discontinuation figures reported in each trial. Adverse events reflect the semaglutide arm unless noted. IBT = intensive behavioral therapy; QW = once weekly; SC = subcutaneous.

Population Differences

The five STEP trials share a common pharmacological agent and dose escalation scheme, but they enrolled meaningfully different patients. Understanding these differences is the first step toward reading the results accurately.

BMI entry criteria and diabetes status. STEP-1 required a BMI of 30 or greater, or 27 or greater with at least one weight-related comorbidity, and explicitly excluded patients with type 2 diabetes. STEP-2 (Davies 2021) required a BMI of 27 or greater and a confirmed diagnosis of type 2 diabetes with HbA1c between 7% and 10%, making it the only trial in the series to enroll a population with both obesity and established glucose dysregulation. STEP-3 (Wadden 2021) used the same BMI thresholds as STEP-1 and also excluded type 2 diabetes, but added the requirement that participants engage in intensive behavioral therapy throughout the trial. STEP-5 (Garvey 2022) mirrored STEP-1 population criteria at a 2-year horizon. STEP-8 (Rubino 2022) also excluded type 2 diabetes and matched STEP-1 inclusion criteria, with the distinguishing feature being an active comparator arm using liraglutide 3.0 mg rather than placebo.

The practical implication is that weight loss estimates from STEP-1, STEP-3, STEP-5, and STEP-8 cannot be straightforwardly applied to patients with type 2 diabetes. As STEP-2 demonstrated, the presence of type 2 diabetes attenuated the weight-loss response by roughly 5 percentage points relative to the non-diabetic STEP-1 population. This attenuation is biologically plausible: several antidiabetic agents that patients in STEP-2 continued (such as insulin or sulfonylureas) independently promote weight gain or resist weight loss, and the pathophysiology of type 2 diabetes itself alters energy homeostasis.

Baseline weight. Mean baseline body weight was approximately 105 kg in STEP-1 to 100 kg in STEP-2 to 106 kg in STEP-3 to 108 kg in STEP-5, and 104 kg in STEP-8. These figures are comparable enough that baseline weight alone is unlikely to explain outcome differences across the series. The more important population variable is diabetes status.

Sex distribution and racial composition. Across all five trials, women represented 70% to 78% of participants. Non-Hispanic white participants predominated in all trials, with Black and Asian representation remaining under 20% collectively in most arms. This limits the degree to which the series can speak to weight-loss outcomes in more diverse populations, a gap the Obesity Society has noted in the broader GLP-1 literature (TOS Clinical Practice Statement).

Methodology Differences

Background lifestyle intervention. All five trials included some form of counseling, but the intensity varied substantially, and this is the clearest methodological divergence across the series.

STEP-1, STEP-2, STEP-5, and STEP-8 all incorporated a standardized low-calorie diet (500 kcal/day deficit) and physical activity goal (150 minutes/week), delivered through brief counseling sessions at clinic visits. STEP-3 was structurally different: participants in both arms received an intensive behavioral therapy program of 30 sessions over 68 weeks, including a 1,000-to-1,200 kcal/day structured meal plan during the first 8 weeks and ongoing group and individual counseling. This design element makes STEP-3 the only trial in the series that answers the clinically distinct question of whether semaglutide adds benefit on top of a maximally intensive lifestyle program.

Comparator arms. Four of the five trials used placebo as the control. STEP-8 is the exception: it used open-label liraglutide 3.0 mg (Saxenda), the previously approved GLP-1 receptor agonist for weight management, as the active comparator. The open-label design in STEP-8 introduces the possibility of differential adherence or reporting bias between arms, though the magnitude of the semaglutide advantage (approximately 9.4 percentage points on absolute weight loss) is unlikely to be explained by expectation effects alone. STEP-2 included a third arm of semaglutide 1.0 mg (the diabetes dose), providing a partial dose-response comparison within a diabetic population.

Primary endpoint definition. All five trials used percentage change in body weight from baseline to the end of the on-treatment observation period as the primary endpoint. This consistency is one of the series' genuine strengths and permits direct numerical comparison. Secondary endpoints diverged: STEP-3 emphasized body composition and cardiometabolic markers; STEP-5 added bone mineral density measurements given the 2-year duration; STEP-2 included HbA1c change as a key secondary endpoint.

Statistical handling of missing data. All trials used a treatment-policy estimand as the primary analysis, meaning dropouts were included using multiple imputation. A secondary "trial product estimand" (on-treatment analysis) was also reported and consistently showed larger effect sizes. Readers comparing raw numbers should confirm which estimand they are using, because the difference can be 2 to 3 percentage points.

Results, Matched

Mean weight loss across the series

The headline weight-loss figures, taken from the treatment-policy estimand, span a range of approximately 10 to 16 percent for semaglutide 2.4 mg across the five trials. (Wilding 2021) reported −14.9% for semaglutide versus −2.4% for placebo at 68 weeks in non-diabetic patients with obesity. STEP-3 (Wadden 2021) reached −16.0% in the semaglutide plus IBT arm, the highest figure in the series, though the IBT-only control arm itself achieved −5.7%, higher than the −2.4% placebo arm in STEP-1. STEP-5 (Garvey 2022) showed −15.2% at 104 weeks, confirming that the plateau observed at 68 weeks in STEP-1 is largely maintained rather than reversed over a longer horizon. STEP-8 (Rubino 2022) showed −15.8% for semaglutide and −6.4% for liraglutide at 68 weeks. STEP-2 (Davies 2021) showed −9.6% for semaglutide 2.4 mg versus −3.4% for placebo, the lowest absolute figure in the series.

The between-trial range of roughly 6 percentage points in the semaglutide arm is clinically meaningful. Attributing it entirely to diabetes status is tempting but probably incomplete. Differences in background behavioral support, baseline antidiabetic medication burden, and trial-specific dropout patterns all contribute.

Effect of background intervention intensity

STEP-3 adds approximately 1.1 percentage points of weight loss above STEP-1 (−16.0% vs −14.9%) when comparing semaglutide arms. That difference is modest. The more striking comparison is in the control arms: IBT-alone produced −5.7% versus lifestyle counseling alone producing −2.4% in STEP-1. This tells us that intensive behavioral therapy has a meaningful independent effect on weight, but it also tells us that semaglutide's therapeutic contribution is largely preserved on top of IBT. The drug roughly triples the weight-loss benefit regardless of whether the background program is minimal or intensive.

This finding has direct clinical significance. Patients who cannot access structured IBT programs should not be discouraged from using semaglutide on the assumption that results will be substantially inferior. The absolute weight loss with semaglutide plus standard counseling (−14.9% in STEP-1) is only marginally below the IBT combination (−16.0% in STEP-3).

Long-term durability: STEP-5

STEP-5 is the only trial in the series to extend to 104 weeks, and its findings address one of the most commonly raised clinical questions: does the weight loss achieved by 68 weeks hold? The answer, in the on-treatment population, is largely yes. The −15.2% figure at 104 weeks is numerically indistinguishable from the −14.9% at 68 weeks in STEP-1. Weight regain did not accelerate during weeks 68 through 104 in patients who continued therapy. Cardiometabolic risk markers, including waist circumference, blood pressure, and fasting lipids, also maintained improvement at 2 years (Garvey 2022).

What STEP-5 cannot tell us is what happens after discontinuation at the 2-year mark, or whether durability holds beyond 2 years of continuous use. The STEP-4 trial (not included in this comparison but worth noting) addressed the withdrawal question and found substantial weight regain within 52 weeks of stopping semaglutide, which contextualizes the long-term maintenance findings from STEP-5 as dependent on continued treatment.

Semaglutide versus liraglutide: STEP-8

STEP-8 is the only head-to-head comparison of the two approved GLP-1 receptor agonists for obesity in the STEP series. At 68 weeks, semaglutide 2.4 mg produced −15.8% versus −6.4% for liraglutide 3.0 mg, a difference of −9.4 percentage points (Rubino 2022). The proportion of participants achieving 5% or greater body weight reduction was 86.6% with semaglutide versus 58.1% with liraglutide. For the 10% threshold, the figures were 70.9% versus 25.6%.

The open-label design is the primary caveat here. Daily liraglutide injection versus weekly semaglutide injection may influence adherence independently of drug effect, and patients aware they are on the "older" agent might be less motivated. Nonetheless, the magnitude of the difference aligns with network meta-analyses comparing the two agents in separate placebo-controlled trials. The FDA label for semaglutide 2.4 mg (Wegovy) does not make a comparative efficacy claim, but the STEP-8 data provide the most direct available evidence.

What the Trials Together Do and Do Not Establish

What they establish together. The STEP series collectively confirms that once-weekly subcutaneous semaglutide 2.4 mg produces clinically meaningful, statistically significant weight loss compared to placebo across three distinct patient contexts: obesity without diabetes, obesity with type 2 diabetes, and obesity with co-administered intensive behavioral therapy. The effect size is consistent enough across sites (trials ran in Europe, North America, Asia, and South America) to suggest it is not a population-specific phenomenon. Weight loss is sustained at 2 years without dose escalation beyond the 2.4 mg maintenance level. Semaglutide 2.4 mg demonstrably outperforms liraglutide 3.0 mg on the same primary endpoint.

What they do not establish. The series does not include a cardiovascular outcomes trial at the 2.4 mg dose in a purely obese population (SELECT addressed this question subsequently, though that trial is outside the present comparison). The trials do not include patients with class I obesity and no comorbidities, meaning the BMI 27-to-30 range is represented only by patients who had at least one comorbidity. The series offers no data on patients with severe renal impairment, active eating disorders, or prior bariatric surgery. The racial and ethnic composition limits generalizability to non-white populations. No trial in the STEP series compared semaglutide 2.4 mg to a surgical intervention or to combination pharmacotherapy.

The series also does not resolve the question of optimal treatment duration. STEP-5 ends at 2 years, and the STEP-4 withdrawal findings suggest that stopping is associated with weight regain, but neither the minimum effective duration nor the long-term safety beyond 2 years has been established within the STEP program itself.

Outstanding Questions for the Next Trial

  1. Cardiovascular outcomes at 2.4 mg in obesity without diabetes. The SELECT trial examined semaglutide 2.4 mg in patients with established cardiovascular disease, but the series lacks a dedicated outcomes trial in patients with obesity and no prior cardiovascular event or diabetes.

  2. Comparative effectiveness against tirzepatide. SURMOUNT-1 and STEP-1 enrolled comparable non-diabetic populations but were not designed as head-to-head comparisons. A prospective randomized comparison of semaglutide 2.4 mg versus tirzepatide 10 mg or 15 mg is the clearest evidence gap in the GLP-1 receptor agonist obesity literature.

  3. Response in racially and ethnically diverse populations. Given the predominantly white trial populations, dedicated trials or pre-specified subgroup analyses in Black, Latino, and East Asian patients are needed to determine whether dose adjustments or population-specific endpoints are warranted.

  4. Optimal duration and discontinuation strategy. Is a planned drug holiday or dose reduction feasible without full weight regain? At what point, if any, can a maintenance dose lower than 2.4 mg sustain the treatment effect?

  5. Additive effects with structured exercise. STEP-3 used structured diet and behavioral therapy as the background intervention. A trial pairing semaglutide with a supervised resistance or aerobic exercise program of equivalent rigor would clarify the drug-exercise interaction, particularly for outcomes such as lean mass preservation.

  6. Type 2 diabetes with BMI below 27. STEP-2 required a BMI of 27 or greater. The weight-loss and glycemic benefits of semaglutide 2.4 mg in patients with type 2 diabetes who have a BMI in the normal or mildly overweight range remain unstudied in the STEP framework.

Frequently asked questions

›
›
›
›
›

References

  • Wilding JPH, et al. Once-Weekly Semaglutide in Adults with Overweight or Obesity. N Engl J Med. 2021;384(11):989-1002. https://pubmed.ncbi.nlm.nih.gov/33567185/
  • Davies M, et al. Semaglutide 2.4 mg once a week in adults with overweight or obesity, and type 2 diabetes (STEP 2). Lancet. 2021;397(10278):971-984. https://pubmed.ncbi.nlm.nih.gov/33667417/
  • Wadden TA, et al. Effect of Subcutaneous Semaglutide vs Placebo as an Adjunct to Intensive Behavioral Therapy on Body Weight in Adults With Overweight or Obesity (STEP 3). JAMA. 2021;325(14):1403-1413. https://pubmed.ncbi.nlm.nih.gov/33625476/
  • Garvey WT, et al. Two-year effects of semaglutide in adults with overweight or obesity: the STEP 5 trial. Nat Med. 2022;28(10):2083-2091. https://pubmed.ncbi.nlm.nih.gov/36280822/
  • Rubino DM, et al. Effect of Weekly Subcutaneous Semaglutide vs Daily Liraglutide on Body Weight in Adults With Overweight or Obesity Without Diabetes (STEP 8). JAMA. 2022;327(2):138-150. https://pubmed.ncbi.nlm.nih.gov/35089258/
  • The Obesity Society. Clinical Practice Statement. Obesity. 2021;29(3):E1. https://pubmed.ncbi.nlm.nih.gov/33555847/
  • FDA. Wegovy (semaglutide) Prescribing Information. 2021. https://www.accessdata.fda.gov/drugsatfda_docs/label/2021/215256s000lbl.pdf