SCALE Obesity Results in Detail: Numbers, Subgroups, and Time Course

At a glance

| Parameter | Detail | |-----------|--------| | N | 3,731 (2:1 randomization; liraglutide n=2,487, placebo n=1,244) | | Intervention | Liraglutide 3.0 mg subcutaneous once daily + diet/exercise counseling | | Comparator | Matching placebo injection + identical diet/exercise counseling | | Duration | 56 weeks (with 12-week observational follow-up off drug) | | Primary endpoint | Percent change in body weight from baseline | | Co-primary endpoints | Proportion achieving ≥5% weight loss; proportion achieving >10% weight loss | | Key result | -8.0% vs -2.6% (p <0.001); 63.2% vs 27.1% hit ≥5%; 33.1% vs 10.6% hit >10% | | Population | Adults with BMI ≥30 or ≥27 with dyslipidemia/hypertension, without diabetes |

Primary Endpoint: Percent Weight Change at 56 Weeks

The trial's primary outcome was mean percent change in body weight from randomization to week 56. The intention-to-treat analysis with last observation carried forward (LOCF) showed:

• Liraglutide group: -8.0 ± 6.7% (mean ± SD)
• Placebo group: -2.6 ± 5.7%
• Estimated treatment difference: -5.4 percentage points (95% CI: -5.8 to -5.0; p <0.001)

The absolute weight loss in kilograms followed the same pattern: liraglutide participants lost a mean of 8.4 kg versus 2.8 kg with placebo. Baseline mean body weight was approximately 106 kg in both groups, confirming well-matched randomization across the 3,731-person cohort.

Co-Primary Categorical Responder Endpoints

The FDA required categorical responder thresholds as co-primary endpoints, not just mean change. Both were met with wide separation from placebo.

| Threshold | Liraglutide 3.0 mg | Placebo | Odds Ratio (95% CI) | p-value | |-----------|--------------------:|--------:|---------------------|---------| | ≥5% weight loss | 63.2% | 27.1% | 4.8 (4.1 to 5.6) | <0.001 | | >10% weight loss | 33.1% | 10.6% | 4.3 (3.5 to 5.3) | <0.001 | | >15% weight loss | 14.4% | 3.5% | NR | <0.001 |

These categorical numbers matter clinically because FDA labeling for Saxenda recommends discontinuation if a patient fails to achieve 4% weight loss by 16 weeks. The trial data show that responders who cleared the early threshold often went on to much larger losses, while non-responders at 16 weeks rarely caught up.

Time Course of Weight Loss

Weight loss was not linear. The published Figure 2 shows a characteristic rapid-then-plateau curve:

• Weeks 0-16: Steepest weight decline. Liraglutide participants lost roughly 5-6% of body weight during this initial phase. Placebo participants lost approximately 2%.
• Weeks 16-32: Continued but slower loss. The liraglutide curve flattened toward week 28-32.
• Weeks 32-56: Weight stabilized near the nadir. Minimal additional loss occurred.
• Weeks 56-68 (off-treatment follow-up): Participants regained approximately 2.9 kg in the 12 weeks after stopping liraglutide, confirming weight regain upon cessation, a consistent finding across GLP-1 receptor agonist trials.

The dose-escalation period (weeks 1-4, increasing from 0.6 mg to 3.0 mg) overlapped with early weight loss, making it difficult to separate pharmacologic effect from the behavioral activation that accompanies trial enrollment.

Response Distribution: Beyond the Mean

Means obscure important variability. The SCALE data reveal a wide distribution of individual responses:

| Weight loss category | Liraglutide (%) | Placebo (%) | |---------------------|----------------:|------------:| | <5% loss or gained weight | 36.8 | 72.9 | | 5% to <10% | 30.1 | 16.5 | | 10% to <15% | 18.7 | 7.1 | | ≥15% | 14.4 | 3.5 |

About one in seven liraglutide-treated participants achieved what would now be considered "high-response" territory (≥15% loss). On the other end, over a third did not reach the 5% clinical significance threshold despite 56 weeks of treatment. This bimodal tendency supports the early stopping rule in current prescribing guidance.

Secondary Endpoints: Cardiometabolic Parameters

Weight loss translated to measurable metabolic improvements. Key secondary outcomes at 56 weeks:

| Endpoint | Liraglutide (change from baseline) | Placebo | Treatment difference (95% CI) | |----------|------------------------------------:|--------:|-------------------------------| | Waist circumference | -8.2 cm | -3.9 cm | -4.2 cm (-4.7 to -3.8) | | Systolic blood pressure | -4.2 mmHg | -1.5 mmHg | -2.8 mmHg (-3.5 to -2.1) | | Diastolic blood pressure | -2.6 mmHg | -1.9 mmHg | -0.9 mmHg (-1.4 to -0.4) | | Fasting glucose | -7.1 mg/dL | -0.8 mg/dL | -6.3 mg/dL (-7.0 to -5.6) | | HbA1c | -0.3% | -0.1% | -0.23% (-0.25 to -0.21) | | Fasting insulin | -15.0 pmol/L | -3.3 pmol/L | Significant reduction | | hsCRP | -0.75 mg/L | -0.11 mg/L | Significant reduction | | Total cholesterol | Modest improvement |, | NS for LDL specifically |

The prediabetes conversion endpoint deserves separate attention: among participants with prediabetes at baseline (n=2,254), liraglutide reduced the rate of progression to type 2 diabetes by 80% over 56 weeks compared with placebo (0.2% vs 1.1% annualized incidence).

Subgroup Analyses

Pre-specified subgroup analyses showed consistent treatment effects across:

• Sex: Women (-8.3%) and men (-7.5%) both exceeded placebo by similar margins
• Age: Participants ≥55 years showed comparable efficacy to younger cohorts
• Baseline BMI: Those with BMI ≥40 lost similar percentages as those with BMI 27-35
• Prediabetes status: Participants with prediabetes lost slightly less weight (-7.4%) than those with normoglycemia (-8.5%), a pattern later confirmed in the STEP trials with semaglutide
• Race: Efficacy was consistent across racial subgroups, though the trial population was predominantly White (approximately 85%)

No subgroup showed a statistically significant interaction with treatment, meaning the drug effect was not meaningfully modified by any baseline characteristic tested.

Methodological Notes That Affect Interpretation

Several design features shape how these results should be read:

Run-in period: There was no placebo run-in. Randomization occurred after screening. This means early dropouts are captured in ITT analyses, which tends to dilute the treatment effect compared with completer analyses.

LOCF imputation: The primary analysis used last observation carried forward. This approach assumes dropouts would have maintained their last measured weight. Given that liraglutide dropouts (27.5% discontinued) were often doing so due to adverse events (nausea, vomiting), their trajectories may have been declining regardless.

Completer analysis: Among participants who completed 56 weeks on treatment, liraglutide produced -9.2% weight loss versus -3.0% for placebo completers. The gap between ITT (-8.0%) and completer (-9.2%) results indicates moderate dropout dilution.

Lifestyle counseling intensity: Both groups received 500 kcal/day deficit dietary counseling and ≥150 minutes/week physical activity guidance. The magnitude of placebo response (2.6%) was modest, suggesting the counseling alone had limited effect in this population.

Injection burden: The once-daily subcutaneous injection introduces an inherent unblinding risk. Participants experiencing GI side effects would reasonably suspect active drug assignment. The FDA review acknowledged this limitation but noted that the magnitude of treatment difference made functional unblinding unlikely to fully explain results.

Safety Signals Relevant to Efficacy Interpretation

Adverse events influenced the efficacy population. Key safety data that contextualize the results:

• Discontinuation for adverse events: 9.9% liraglutide vs 3.8% placebo
• GI events (nausea): 40.2% liraglutide vs 14.7% placebo, predominantly in weeks 1-8
• Gallbladder events: 2.5% vs 1.0% (consistent with rapid weight loss from any cause)
• Heart rate increase: Mean +2.4 bpm with liraglutide vs +0.1 bpm placebo

The higher discontinuation rate means the ITT population includes participants who took liraglutide for only a few weeks. Their minimal weight loss pulls down the group mean, making the drug's efficacy in adherent patients somewhat higher than the headline 8.0% figure.

Placing SCALE in the GLP-1 Efficacy Hierarchy

The SCALE result (8.0% placebo-subtracted: 5.4%) established a benchmark that subsequent GLP-1 trials have exceeded. For context within the class:

• Liraglutide 3.0 mg (SCALE, 2015): ~5.4% net
• Semaglutide 2.4 mg (STEP 1, 2021): ~12.4% net
• Tirzepatide 15 mg (SURMOUNT-1, 2022): ~17.8% net

SCALE's historical importance lies not in being the largest effect, but in being the trial that proved a GLP-1 receptor agonist could produce clinically meaningful weight loss in a large non-diabetic population with acceptable safety. It opened the regulatory pathway that semaglutide and tirzepatide later followed.

Frequently asked questions

›

›

›

›

›

›

›

›

›

›

References

1. 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. https://pubmed.ncbi.nlm.nih.gov/26132939/
2. FDA. Saxenda (liraglutide) Prescribing Information. 2014. https://www.accessdata.fda.gov/drugsatfda_docs/label/2014/206321Orig1s000lbl.pdf
3. 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. https://pubmed.ncbi.nlm.nih.gov/33567185/
4. Jastreboff AM, Aronne LJ, Ahmad NN, et al. Tirzepatide Once Weekly for the Treatment of Obesity (SURMOUNT-1). N Engl J Med. 2022;387(3):205-216. https://pubmed.ncbi.nlm.nih.gov/35658024/
5. Rubino DM, Greenway FL, Khalid U, et al. Effect of Continued Weekly Subcutaneous Semaglutide vs Placebo on Weight Loss Maintenance in Adults With Overweight or Obesity (STEP 4). JAMA. 2021;325(14):1414-1425. https://pubmed.ncbi.nlm.nih.gov/33755728/