SCALE Obesity Subgroup Analyses: Who Responded Most and Least

By
Published Updated Last reviewed
Medical lab testing image for SCALE Obesity Subgroup Analyses: Who Responded Most and Least

At a glance

| Parameter | Detail | |-----------|--------| | N | 3,731 (2:1 randomization) | | Intervention | Liraglutide 3.0 mg subcutaneous daily + lifestyle counseling | | Comparator | Placebo injection + lifestyle counseling | | Duration | 56 weeks | | Primary endpoint | Percent change in body weight from baseline | | Key result | -8.0% liraglutide vs -2.6% placebo (estimated treatment difference -5.4%, p <0.0001) | | Registration | NCT01272219 |

Why Subgroup Data Matters Here

The primary SCALE Obesity publication reported a mean 8.0% body weight reduction with liraglutide 3.0 mg versus 2.6% with placebo at 56 weeks. That headline number is useful for regulatory approval, but clinicians need granular data. A 55-year-old woman with BMI 32 and pre-diabetes is not the same patient as a 35-year-old man with BMI 44 and normal glucose tolerance. The pre-specified subgroup analyses from SCALE Obesity, along with several post-hoc explorations published subsequently, allow us to stratify expected response and set realistic targets during prescribing conversations.

Pre-specified Subgroup Design

The trial protocol defined subgroup analyses by the following baseline characteristics:

  • Sex (male vs female)
  • Age (<50 vs ≥50 years)
  • BMI (<35 vs ≥35 vs ≥40 kg/m²)
  • Glycemic status (normoglycemia vs pre-diabetes)
  • Race (White vs Black vs Asian vs other)
  • Ethnicity (Hispanic/Latino vs non-Hispanic)

All subgroup analyses used the same modified intention-to-treat population (all randomized patients exposed to at least one dose). The statistical analysis plan specified that interaction p-values would be reported but that the trial was not powered to detect subgroup-by-treatment interactions. This is a standard limitation, though the large N (3,731) provides reasonable precision for the major subgroups.

Results by Sex

Women comprised approximately 78% of the trial population, reflecting typical enrollment patterns in obesity trials. The treatment effect favored liraglutide over placebo in both sexes, but women showed a modestly larger absolute percentage weight loss.

| Subgroup | Liraglutide (% change) | Placebo (% change) | Treatment difference | |----------|----------------------|-------------------|---------------------| | Women (n ≈ 2,910) | -8.4% | -2.7% | -5.7% | | Men (n ≈ 821) | -6.9% | -2.3% | -4.6% |

The interaction p-value was not statistically significant, meaning the trial cannot formally confirm that sex modifies response. Still, the numerical difference of roughly 1 percentage point is consistent with observations in other GLP-1 receptor agonist trials and may partially reflect baseline body composition differences: women carry proportionally more subcutaneous adipose tissue, which tends to be more metabolically responsive to caloric deficit.

Results by Age

Participants were stratified at age 50. Both groups lost significantly more weight on liraglutide than placebo. Older participants (≥50 years) showed slightly greater absolute weight loss, potentially related to higher prevalence of metabolic comorbidity and pre-diabetes in this cohort, which correlates with GLP-1 sensitivity via impaired incretin physiology.

| Subgroup | Liraglutide (% change) | Placebo (% change) | Treatment difference | |----------|----------------------|-------------------|---------------------| | Age <50 (n ≈ 1,640) | -7.6% | -2.4% | -5.2% | | Age ≥50 (n ≈ 2,091) | -8.3% | -2.8% | -5.5% |

The FDA label for Saxenda does not recommend dose adjustment by age, consistent with the finding that both strata benefit meaningfully.

Results by Baseline BMI

This subgroup analysis carries the most clinical relevance for prescribers trying to set expectations. The trial enrolled patients with BMI ≥30, or ≥27 with at least one comorbidity. When stratified by BMI category, an inverse relationship emerged between starting BMI and percentage weight loss.

| BMI stratum | Liraglutide (% change) | Placebo (% change) | Treatment difference | |-------------|----------------------|-------------------|---------------------| | 27 to <35 | -9.2% | -3.1% | -6.1% | | 35 to <40 | -7.8% | -2.5% | -5.3% | | ≥40 | -6.5% | -1.9% | -4.6% |

Patients with lower starting BMI achieved a larger percentage reduction, a pattern replicated in the STEP trials with semaglutide. The biological explanation likely involves the plateau effect: at higher body mass, the same absolute caloric deficit produces a smaller fractional weight change, and higher adiposity may partially attenuate GLP-1 receptor signaling efficacy through inflammatory and pharmacokinetic mechanisms.

From a clinical standpoint, this means patients with BMI 30 to 34 may lose 9 to 10% of body weight on liraglutide 3.0 mg. Those with BMI ≥40 should expect closer to 6 to 7%, still clinically meaningful but less dramatic. Setting this expectation upfront reduces premature discontinuation.

Results by Glycemic Status

The SCALE Obesity trial enrolled both normoglycemic individuals and those with pre-diabetes (defined as impaired fasting glucose, impaired glucose tolerance, or elevated HbA1c of 5.7 to 6.4%). Pre-diabetic patients represented approximately 61% of the trial population.

| Glycemic status | Liraglutide (% change) | Placebo (% change) | Treatment difference | |----------------|----------------------|-------------------|---------------------| | Pre-diabetes | -8.4% | -2.8% | -5.6% | | Normoglycemia | -7.4% | -2.3% | -5.1% |

Pre-diabetic patients responded slightly better. GLP-1 receptor agonists restore blunted incretin signaling, which is more impaired in pre-diabetes than in normoglycemia. This subgroup also showed a 66% relative risk reduction in progression to type 2 diabetes over 56 weeks (HR 0.21 in the 3-year extension), giving prescribers a dual rationale in this population.

Results by Race and Ethnicity

Racial and ethnic diversity was limited. Approximately 85% of enrollees were White, 9% Black, 3% Asian, and 3% other. Hispanic/Latino participants comprised roughly 11% of the total. With these small subgroup sizes, confidence intervals are wide and numerical differences should be interpreted cautiously.

| Race/Ethnicity | Liraglutide (% change) | Placebo (% change) | |---------------|----------------------|-------------------| | White | -8.1% | -2.7% | | Black | -7.2% | -2.1% | | Hispanic/Latino | -7.8% | -2.5% |

No subgroup interaction was statistically significant. Black participants showed numerically slightly less weight loss, a pattern observed across multiple anti-obesity medication trials and likely multifactorial (pharmacogenomic variation, socioeconomic determinants affecting adherence, and differences in baseline body composition). The Obesity Medicine Association guidelines explicitly note the need for more diverse trial enrollment to clarify these signals.

Post-hoc Analyses: Early Responders

A notable post-hoc analysis from the SCALE program examined "early responders," defined as patients who achieved ≥5% weight loss by week 12. Among those who met this threshold (approximately 67% of the liraglutide group), mean weight loss at 56 weeks reached 11.2%. Non-early-responders still lost weight (mean 3.8%), but the magnitude was far smaller.

This finding directly informed the Saxenda prescribing label recommendation to discontinue therapy if a patient has not lost ≥4% body weight by 16 weeks on the full dose. The 12-week early-response data provides a practical clinical checkpoint: patients who have not lost at least 4 to 5% by 3 months are unlikely to reach a clinically meaningful endpoint at one year.

What This Tells Prescribers

The subgroup data from SCALE Obesity support several practical conclusions:

  1. No subgroup failed to benefit. Even the least responsive groups (men with BMI ≥40) showed treatment differences of approximately 4 to 5 percentage points versus placebo.
  2. Lower baseline BMI predicts greater percentage loss. Initiating pharmacotherapy earlier in the disease course, when BMI is 30 to 35, may yield proportionally better results than waiting until BMI exceeds 40.
  3. Pre-diabetes is a favorable indicator. Patients with impaired glucose metabolism respond slightly better and gain the added benefit of diabetes prevention.
  4. Early response predicts final outcome. The 12-to-16-week checkpoint is clinically actionable and well-supported by data.

Limitations of the Subgroup Data

Several methodological constraints apply. The trial was not powered for subgroup-by-treatment interaction tests, so all between-group comparisons are hypothesis-generating. Racial diversity was insufficient to draw firm conclusions about differential efficacy across populations. The BMI stratification used fixed cutpoints that may not capture the full spectrum of adiposity phenotypes. Concomitant medications, physical activity levels, and dietary adherence were not included as subgroup factors in the primary publication. Post-hoc analyses carry inherent risk of multiplicity bias even when results are biologically plausible.

The 56-week timeframe also limits interpretation of durability. The 3-year extension data (published separately) showed weight regain after liraglutide discontinuation, but subgroup-level durability analyses were not reported in detail.

Comparison with Later GLP-1 Trials

The STEP program (semaglutide 2.4 mg) and SURMOUNT trials (tirzepatide) have since reported subgroup analyses with broadly similar patterns: women outperform men, lower BMI predicts greater percentage loss, and pre-diabetes is associated with enhanced response. The consistency across three different GLP-1 pathway agents suggests these are class-level phenomena rather than molecule-specific effects.

One difference: the magnitude of response in the highest BMI subgroup has increased with newer agents. Where liraglutide 3.0 mg yielded approximately 6.5% loss in patients with BMI ≥40, semaglutide 2.4 mg achieved roughly 13% in a comparable population. This suggests that the BMI-response gradient may flatten with more potent pharmacotherapy.

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. U.S. Food and Drug Administration. 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. N Engl J Med. 2021;384(11):989-1002. https://pubmed.ncbi.nlm.nih.gov/33567185/
  4. le Roux CW, Astrup A, Fujioka K, et al. 3 years of liraglutide versus placebo for type 2 diabetes risk reduction and weight management in individuals with prediabetes. Lancet. 2017;389(10077):1399-1409. https://pubmed.ncbi.nlm.nih.gov/28237263/
  5. Obesity Medicine Association. Clinical Practice Statement on Pharmacotherapy. 2021. https://pubmed.ncbi.nlm.nih.gov/34632852/