How was randomization stratified in the SCALE Obesity trial?

Randomization used a 2:1 ratio (liraglutide:placebo) stratified by prediabetes status and BMI category (≥30 or 27 to <30 kg/m²), with allocation concealment via interactive voice/web response system.

Why were patients with type 2 diabetes excluded from SCALE Obesity?

A separate trial, SCALE Diabetes, specifically studied liraglutide 3.0 mg in adults with T2D. Excluding T2D from the main trial prevented confounding from glucose-lowering medications and differing metabolic physiology.

What was the dose escalation schedule for liraglutide in SCALE?

Participants started at 0.6 mg daily, increasing by 0.6 mg per week over four weeks until reaching the target dose of 3.0 mg. This schedule was adopted directly into the Saxenda prescribing label.

How did SCALE handle missing data from dropouts?

The primary analysis used last observation carried forward (LOCF). Sensitivity analyses with mixed-effects repeated measures and multiple imputation confirmed consistent results, though LOCF in weight-loss trials tends to favor the active arm.

What dropout rate did the SCALE Obesity trial report?

Overall, 25.1% of liraglutide participants and 29.2% of placebo participants discontinued treatment. Adverse events (mainly gastrointestinal) drove 9.9% of liraglutide discontinuations versus 3.8% for placebo.

Did SCALE Obesity include an active comparator drug?

No. Both arms received placebo or liraglutide on top of identical lifestyle counseling (500 kcal/day deficit, ≥150 min/week exercise). There was no head-to-head comparison with other anti-obesity medications.

How does the completer analysis differ from the ITT result?

Completers who stayed on liraglutide for the full 56 weeks lost approximately 9.2% of body weight, compared with the ITT figure of 8.0%. The difference reflects dilution from early discontinuers in the ITT population.

Does SCALE Obesity provide evidence for cardiovascular benefit?

No. Weight change was the primary endpoint. Cardiovascular outcomes for liraglutide were evaluated separately in the LEADER trial, which used a lower dose (1.8 mg) in patients with type 2 diabetes.

What FDA benchmarks did the SCALE results meet?

FDA guidance requires either ≥5% placebo-subtracted weight loss or roughly double the proportion achieving ≥5% loss. SCALE met both: 5.4 percentage points mean difference and 63.2% vs 27.1% achieving ≥5% loss (a 2.3-fold ratio).

Can SCALE results be directly compared to semaglutide STEP trials?

Cross-trial comparisons are methodologically limited. STEP 1 and SCALE used similar placebo-controlled designs but enrolled different patient populations, ran at different sites, and tested different molecules at different doses.

Inside the SCALE Obesity Methodology: What Most Summaries Skip

Q: Why were patients with type 2 diabetes excluded from SCALE Obesity?

A separate trial, SCALE Diabetes, specifically studied liraglutide 3.0 mg in adults with T2D. Excluding T2D from the main trial prevented confounding from glucose-lowering medications and differing metabolic physiology.

Q: What was the dose escalation schedule for liraglutide in SCALE?

Participants started at 0.6 mg daily, increasing by 0.6 mg per week over four weeks until reaching the target dose of 3.0 mg. This schedule was adopted directly into the Saxenda prescribing label.

Q: How did SCALE handle missing data from dropouts?

The primary analysis used last observation carried forward (LOCF). Sensitivity analyses with mixed-effects repeated measures and multiple imputation confirmed consistent results, though LOCF in weight-loss trials tends to favor the active arm.

Q: What dropout rate did the SCALE Obesity trial report?

Overall, 25.1% of liraglutide participants and 29.2% of placebo participants discontinued treatment. Adverse events (mainly gastrointestinal) drove 9.9% of liraglutide discontinuations versus 3.8% for placebo.

Q: Did SCALE Obesity include an active comparator drug?

No. Both arms received placebo or liraglutide on top of identical lifestyle counseling (500 kcal/day deficit, ≥150 min/week exercise). There was no head-to-head comparison with other anti-obesity medications.

Q: How does the completer analysis differ from the ITT result?

Completers who stayed on liraglutide for the full 56 weeks lost approximately 9.2% of body weight, compared with the ITT figure of 8.0%. The difference reflects dilution from early discontinuers in the ITT population.

Q: Does SCALE Obesity provide evidence for cardiovascular benefit?

No. Weight change was the primary endpoint. Cardiovascular outcomes for liraglutide were evaluated separately in the LEADER trial, which used a lower dose (1.8 mg) in patients with type 2 diabetes.

Q: What FDA benchmarks did the SCALE results meet?

FDA guidance requires either ≥5% placebo-subtracted weight loss or roughly double the proportion achieving ≥5% loss. SCALE met both: 5.4 percentage points mean difference and 63.2% vs 27.1% achieving ≥5% loss (a 2.3-fold ratio).

Q: Can SCALE results be directly compared to semaglutide STEP trials?

Cross-trial comparisons are methodologically limited. STEP 1 and SCALE used similar placebo-controlled designs but enrolled different patient populations, ran at different sites, and tested different molecules at different doses.

By HealthRX Medical Team

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

At a glance

| Parameter | Detail | |---|---| | Trial | SCALE Obesity and Prediabetes (NCT01272219) | | N | 3,731 randomized (2,487 liraglutide, 1,244 placebo) | | Intervention | Liraglutide 3.0 mg subcutaneous daily + diet/exercise counseling | | Comparator | Matched placebo injection + identical diet/exercise counseling | | Duration | 56 weeks (with 4-week dose escalation) | | Primary endpoint | Percent change in body weight from baseline at week 56 | | Key result | −8.0% liraglutide vs −2.6% placebo (difference −5.4 percentage points, p < 0.0001) | | Publication | Pi-Sunyer et al., NEJM 2015 |

The 2:1 Randomization and Why It Matters

Most trial summaries state the result. Few explain why the SCALE investigators chose a 2:1 randomization ratio favoring active drug over placebo. The rationale was straightforward: liraglutide's side-effect profile (primarily nausea and gastrointestinal symptoms) was already well-characterized from type 2 diabetes trials. Assigning more participants to the active arm maximized statistical power to detect weight-loss differences without meaningfully increasing safety risk.

Randomization was stratified by prediabetes status (yes or no) and BMI (≥30 or 27 to <30 kg/m²). This stratification ensured balanced subgroups across arms, which mattered because the FDA's 2014 approval review for Saxenda later examined efficacy by glycemic status. Without stratification, imbalanced prediabetes prevalence could have confounded weight-loss estimates in either direction.

Allocation concealment used an interactive voice/web response system. Participants, investigators, and sponsor personnel remained blinded throughout the 56-week treatment period.

Inclusion and Exclusion Criteria: Selecting the Right Population

The trial enrolled adults aged ≥18 with BMI ≥30 kg/m² or BMI ≥27 kg/m² with at least one weight-related comorbidity (treated or untreated dyslipidemia or hypertension). This definition tracked the FDA's longstanding obesity-drug indication framework, which requires either frank obesity or overweight plus a comorbid condition.

Key exclusions limited generalizability in specific ways:

Type 2 diabetes: Excluded entirely. A separate SCALE trial (SCALE Diabetes) addressed this population. Clinicians extrapolating results to patients with T2D should reference that trial's data instead.
Prior bariatric surgery: Excluded. Altered GI anatomy could confound GLP-1 pharmacokinetics.
Recent weight change: Participants who lost or gained >5 kg within 90 days before screening were excluded, reducing regression-to-the-mean effects.
GLP-1 receptor agonist use within 90 days: Prevented carryover effects from prior exposure.

The population was 78.7% female and 85.8% white. Mean baseline BMI was 38.3 kg/m². These demographics matter when applying results to male or ethnically diverse populations, a limitation the authors acknowledged directly.

The Dose Escalation Protocol

Liraglutide was not started at 3.0 mg. Participants underwent a standardized 4-week dose escalation: 0.6 mg daily for week 1, increasing by 0.6 mg each week until reaching 3.0 mg by week 4. This protocol aimed to reduce the gastrointestinal side effects (nausea, vomiting, diarrhea) that peak during GLP-1 receptor agonist initiation.

This detail is clinically relevant because the Saxenda prescribing information carried this escalation schedule directly from the trial protocol into the approved label. Clinicians who skip or compress the escalation are departing from the evidence base that demonstrated the 8.0% weight-loss result.

Co-Primary Endpoints and the FDA's Bar

SCALE used three co-primary endpoints evaluated hierarchically:

Percent change in body weight from baseline
Proportion of participants losing ≥5% of baseline weight
Proportion of participants losing >10% of baseline weight

The FDA guidance on obesity drug development specifies that an anti-obesity medication should demonstrate either (a) a mean placebo-subtracted weight loss ≥5% with statistical significance, or (b) a proportion achieving ≥5% loss that is roughly double the placebo proportion. SCALE met both bars.

| Endpoint | Liraglutide 3.0 mg | Placebo | Difference | |---|---|---|---| | Mean weight change | −8.0% | −2.6% | −5.4 pp (p < 0.0001) | | ≥5% weight loss | 63.2% | 27.1% | 2.3× (p < 0.0001) | | >10% weight loss | 33.1% | 10.6% | 3.1× (p < 0.0001) |

The hierarchical testing procedure controlled for type I error inflation across the three endpoints. If the first failed, the second and third would not be formally tested. All three cleared their thresholds.

Statistical Approach and Missing Data

The primary analysis used a modified intention-to-treat population: all randomized participants who received at least one dose of the study drug and had at least one post-baseline body-weight assessment. This excluded 3.4% of randomized participants.

Missing data handling is where methodology separates useful trials from misleading ones. SCALE used last observation carried forward (LOCF) as the primary imputation method for its main analysis. LOCF assumes that a participant's last measured weight persists unchanged. For weight-loss trials, this typically favors the active drug because dropouts in the drug arm often leave during early weight loss, locking in a lower weight.

Dropout rates were substantial: 25.1% in the liraglutide arm and 29.2% in the placebo arm discontinued treatment prematurely. The most common reason for discontinuation in the liraglutide group was adverse events (9.9% vs 3.8% placebo), predominantly GI-related. Sensitivity analyses using mixed-effects repeated measures (MMRM) and multiple imputation produced consistent results, which strengthened the primary finding.

The Endocrine Society's 2015 guidelines on obesity pharmacotherapy noted that LOCF in obesity trials deserves scrutiny, though the SCALE investigators' sensitivity analyses addressed the most common criticisms.

The Comparator Choice: Placebo Plus Lifestyle

Both arms received a 500 kcal/day deficit diet and ≥150 minutes/week of physical activity counseling. This was not a drug-versus-nothing comparison. The placebo arm's 2.6% weight loss reflects real lifestyle intervention effects.

This design means the trial answers: "What does adding liraglutide 3.0 mg to structured lifestyle modification produce?" It does not answer: "How does liraglutide compare to other anti-obesity medications?" No active comparator arm existed. Head-to-head data against other agents (orlistat, phentermine-topiramate, semaglutide) requires separate trials. The later STEP 1 trial for semaglutide 2.4 mg used a similar placebo-controlled design, making cross-trial comparisons tempting but methodologically limited by different patient populations and sites.

The Estimand Framework: What Was Actually Being Estimated

Before the ICH E9(R1) addendum formalized estimand thinking, SCALE's protocol implicitly used a "treatment policy" estimand for its primary analysis. All participants were included regardless of adherence, and post-discontinuation weights (when available) contributed to the LOCF carry-forward.

A separate "on-treatment" analysis restricted to completers showed larger effects: approximately 9.2% weight loss in the liraglutide group. The gap between ITT and completer analyses tells clinicians something practical. Patients who tolerate liraglutide for a full year can expect somewhat better results than the headline 8.0% figure. Patients who discontinue early (roughly 1 in 4) contribute to the diluted ITT estimate.

Limitations the Authors Acknowledged

The published report listed several limitations directly:

Trial duration: 56 weeks provides no data on durability beyond one year. Weight regain after GLP-1 agonist discontinuation was later documented in a SCALE maintenance extension.
Population homogeneity: Predominantly white, female, and from North America or Europe.
No active comparator: Cannot rank liraglutide against other approved agents.
Cardiometabolic outcomes: Weight loss was the primary endpoint. Cardiovascular event reduction required the separate LEADER trial, which tested liraglutide 1.8 mg in T2D patients.

What This Means for Clinical Interpretation

The 5.4 percentage point placebo-subtracted difference is real but context-dependent. It was measured in a selected population (no T2D, predominantly female, mean BMI 38), using a specific dose escalation, over exactly 56 weeks, with LOCF imputation, in conjunction with structured lifestyle counseling. Each of those parameters defines the boundaries of valid extrapolation.

Clinicians prescribing Saxenda should set patient expectations based on these trial conditions, not on cross-trial comparisons with newer GLP-1 agents tested in different populations.

Frequently asked questions

›

References

Pi-Sunyer X, Astrup A, Fujioka K, et al. A Randomized, Controlled Trial of 3.0 mg of Liraglutide in Weight Management. N Engl J Med. 2015;373(1):11-22. PubMed
FDA. Saxenda (liraglutide) Prescribing Information. 2014. FDA Label
FDA. Developing Products for Weight Management: Guidance for Industry (Revision 1). FDA Guidance
Wadden TA, Hollander P, Klein S, et al. Weight maintenance and additional weight loss with liraglutide after low-calorie-diet-induced weight loss: the SCALE Maintenance randomized study. Int J Obes. 2013;37(11):1443-1451. PubMed
Marso SP, Daniels GH, Tanaka K, et al. Liraglutide and Cardiovascular Outcomes in Type 2 Diabetes (LEADER). N Engl J Med. 2016;375(4):311-322. PubMed
Apovian CM, Aronne LJ, Bessesen DH, et al. Pharmacological Management of Obesity: An Endocrine Society Clinical Practice Guideline. J Clin Endocrinol Metab. 2015;100(2):342-362. PubMed
Wilding JPH, Batterham RL, Calanna S, et al. Once-Weekly Semaglutide in Adults with Overweight or Obesity (STEP 1). N Engl J Med. 2021;384(11):989-1002. PubMed