SCALE Diabetes deep-dive: liraglutide 3.0 mg in T2D + obesity (Davies 2015 JAMA)
Last verified 2026-05-28 · Phase 3 · Completed (results published August 2015) · NCT01272232 ↗
SCALE Diabetes (NCT01272232; Novo Nordisk protocol NN8022-1922) is the pivotal phase-3 trial that established the efficacy and tolerability of liraglutide 3.0 mg in adults with type 2 diabetes and overweight or obesity — a population explicitly excluded from the parent SCALE Obesity and Prediabetes trial (Pi-Sunyer 2015 NEJM, PMID 26132939) that supported the original 2014 Saxenda approval. Novo Nordisk randomized 846 adults aged 18 or older with type 2 diabetes managed on diet, metformin, sulfonylurea, or combinations thereof, BMI at least 27, and HbA1c 7.0 to 10.0% in a 2:1:1 allocation to once-daily subcutaneous liraglutide 3.0 mg, liraglutide 1.8 mg (the approved Victoza dose for type 2 diabetes), or matching placebo, all on top of a 500 kcal/day deficit diet and exercise counseling, for 56 weeks. The three coprimary endpoints were percentage change in body weight from baseline to week 56 and the proportions achieving at least 5% and more than 10% weight loss. Results published by Davies and colleagues in the Journal of the American Medical Association on August 18, 2015 (PMID 26284720, DOI 10.1001/jama.2015.9676) showed a mean weight loss of 6.0% on liraglutide 3.0 mg versus 4.7% on liraglutide 1.8 mg versus 2.0% on placebo, with substantial concordant improvements in glycemic and cardiometabolic markers. The trial supported the FDA labeling that extends Saxenda's chronic weight management indication to adults with type 2 diabetes.
- Enrollment
- 846
- Duration
- 56 weeks of treatment plus a 12-week off-drug follow-up (week 68 observation)
- Drug
- Liraglutide 3.0 mg (Saxenda)
- Population
- Adults aged 18 or older with type 2 diabetes managed by diet, metformin, sulfonylurea, or combinations thereof (insulin and other GLP-1 agonists were excluded), BMI at least 27 kg/m squared, and HbA1c 7.0 to 10.0% at screening. Of 846 randomized: mean age 55 years, 50% female, 84% white, mean baseline body weight 105.7 kg, mean BMI 37.1 kg/m squared, mean HbA1c 7.9%, mean type 2 diabetes duration 7.3 years. Conducted at 126 sites across nine countries (United States, Belgium, Canada, Finland, Germany, Israel, Mexico, Russia, South Africa) from June 2011 through January 2013.
Primary endpoint
Percentage change in body weight from baseline to week 56 (coprimary)
Treatment arm
-6.0% (liraglutide 3.0 mg)
Comparator
-2.0% (placebo); -4.7% (liraglutide 1.8 mg)
Treatment difference: Estimated treatment difference liraglutide 3.0 mg vs placebo -4.00 percentage points (95% CI -5.10 to -2.90; P<0.001); liraglutide 3.0 mg vs 1.8 mg -1.34 percentage points (95% CI -2.10 to -0.57; P<0.001)
The coprimary endpoint also required proportions achieving at least 5% and more than 10% weight loss at week 56: 54.3% on liraglutide 3.0 mg vs 40.4% on 1.8 mg vs 21.4% on placebo achieved at least 5% loss (P<0.001 vs placebo); 25.2% vs 15.9% vs 6.7% achieved more than 10% loss (P<0.001 vs placebo).
Secondary endpoints
| Endpoint | Treatment | Comparator | Difference |
|---|---|---|---|
| Proportion achieving at least 5% weight loss at week 56 More than half of liraglutide 3.0 mg participants met the conventional clinically meaningful threshold despite the headwind that T2D consistently produces smaller weight responses than non-diabetic populations. | 54.3% (liraglutide 3.0 mg) | 21.4% (placebo); 40.4% (liraglutide 1.8 mg) | Estimated odds ratio liraglutide 3.0 mg vs placebo 4.8 (95% CI 3.4 to 6.8; P<0.001) |
| Proportion achieving more than 10% weight loss at week 56 Roughly one in four liraglutide 3.0 mg participants reached more than 10% loss, nearly four times the placebo rate. | 25.2% (liraglutide 3.0 mg) | 6.7% (placebo); 15.9% (liraglutide 1.8 mg) | Estimated odds ratio liraglutide 3.0 mg vs placebo 4.7 (95% CI 2.9 to 7.6; P<0.001) |
| Change in HbA1c from baseline to week 56 (percentage points) Glycemic response was larger than the 0.45 percentage-point drop seen in non-diabetic STEP-1 participants, reflecting the higher baseline HbA1c (mean 7.9%) and active beta-cell pathology in T2D. | -1.3 (liraglutide 3.0 mg) | -0.3 (placebo); -1.1 (liraglutide 1.8 mg) | Estimated treatment difference liraglutide 3.0 mg vs placebo -0.9 percentage points (95% CI -1.1 to -0.8; P<0.001) |
| Change in fasting plasma glucose from baseline to week 56 (mg/dL) Marked reduction concordant with the HbA1c drop, with most benefit captured at the 1.8 mg dose. | -38.0 mg/dL (liraglutide 3.0 mg) | -7.7 mg/dL (placebo); -34.7 mg/dL (liraglutide 1.8 mg) | Estimated treatment difference liraglutide 3.0 mg vs placebo -30.3 mg/dL (95% CI -36.1 to -24.6; P<0.001) |
| Proportion achieving HbA1c below 7.0% at week 56 More than two-thirds of liraglutide 3.0 mg participants reached the American Diabetes Association glycemic target, more than double the placebo rate. | 69.2% (liraglutide 3.0 mg) | 27.2% (placebo); 60.5% (liraglutide 1.8 mg) | Estimated odds ratio liraglutide 3.0 mg vs placebo 6.4 (95% CI 4.5 to 9.1; P<0.001) |
| Change in waist circumference from baseline to week 56 (cm) | -6.0 cm (liraglutide 3.0 mg) | -2.8 cm (placebo); -4.9 cm (liraglutide 1.8 mg) | Estimated treatment difference liraglutide 3.0 mg vs placebo -3.2 cm (95% CI -4.2 to -2.2; P<0.001) |
| Change in systolic blood pressure from baseline to week 56 (mmHg) Modest but statistically significant additional reduction; magnitude consistent with what is observed across SCALE program trials. | -3.0 mmHg (liraglutide 3.0 mg) | -0.4 mmHg (placebo); -2.6 mmHg (liraglutide 1.8 mg) | Estimated treatment difference liraglutide 3.0 mg vs placebo -2.6 mmHg (95% CI -4.6 to -0.6; P=0.01) |
| Change in fasting lipid profile from baseline to week 56 (triglycerides, % change) LDL and total cholesterol shifts were small and largely non-significant; triglyceride reduction tracked weight loss magnitude. | -13.4% (liraglutide 3.0 mg) | -3.3% (placebo); -8.7% (liraglutide 1.8 mg) | Estimated treatment ratio favoring liraglutide 3.0 mg vs placebo (P<0.001) |
| Change in patient-reported quality of life — IWQOL-Lite total score at week 56 Weight-specific quality-of-life instrument; larger improvement on liraglutide 3.0 mg consistent with greater weight loss in that arm. | +9.0 points (liraglutide 3.0 mg) | +5.3 points (placebo); +7.5 points (liraglutide 1.8 mg) | Estimated treatment difference liraglutide 3.0 mg vs placebo +3.7 points (95% CI +1.6 to +5.8; P<0.001) |
| Change in SF-36 physical-component summary score at week 56 | +3.3 points (liraglutide 3.0 mg) | +1.4 points (placebo); +2.6 points (liraglutide 1.8 mg) | Estimated treatment difference liraglutide 3.0 mg vs placebo +1.9 points; nominal P<0.001 |
Adverse events
| Event | Treatment rate | Comparator rate |
|---|---|---|
| Any gastrointestinal disorder GI events clustered during the 4-week dose-escalation phase (0.6 mg up to 3.0 mg). Most were mild-to-moderate and transient. | 59.4% (liraglutide 3.0 mg); 53.2% (liraglutide 1.8 mg) | 33.2% (placebo) |
| Nausea Most common GI event; peaked during titration and attenuated over weeks. Lower in absolute terms than in non-diabetic SCALE populations because diabetic participants generally tolerate liraglutide better. | 33.1% (liraglutide 3.0 mg); 27.5% (liraglutide 1.8 mg) | 10.6% (placebo) |
| Diarrhea | 21.4% (liraglutide 3.0 mg); 19.7% (liraglutide 1.8 mg) | 8.5% (placebo) |
| Vomiting | 12.9% (liraglutide 3.0 mg); 9.2% (liraglutide 1.8 mg) | 3.7% (placebo) |
| Constipation | 11.6% (liraglutide 3.0 mg); 8.0% (liraglutide 1.8 mg) | 8.5% (placebo) |
| Dyspepsia | 8.5% (liraglutide 3.0 mg); 6.6% (liraglutide 1.8 mg) | 1.9% (placebo) |
| Headache No treatment-related signal. | 10.0% (liraglutide 3.0 mg); 7.8% (liraglutide 1.8 mg) | 10.1% (placebo) |
| Documented symptomatic hypoglycemia (plasma glucose at or below 70 mg/dL with symptoms) Higher than in non-diabetic SCALE trials because background sulfonylureas were permitted; the protocol pre-specified sulfonylurea dose reduction at randomization for participants on those agents. No severe hypoglycemia (requiring third-party assistance) was reported in any arm. | 21.8% (liraglutide 3.0 mg); 27.0% (liraglutide 1.8 mg) | 16.9% (placebo) |
| Cholelithiasis (gallstones) Consistent with the gallstone signal observed across rapid weight-loss interventions and in later GLP-1 trials. | 1.4% (liraglutide 3.0 mg); 1.4% (liraglutide 1.8 mg) | 0.5% (placebo) |
| Acute pancreatitis (confirmed) Single confirmed case on liraglutide 3.0 mg; numerically rare but contributed to the class-level pancreatitis warning maintained on the Saxenda label. | 0.2% (1/423, liraglutide 3.0 mg); 0% (liraglutide 1.8 mg) | 0% (placebo) |
| Permanent treatment discontinuation due to adverse events Most discontinuations were for gastrointestinal intolerance during or shortly after the dose-escalation phase. | 9.7% (liraglutide 3.0 mg); 4.9% (liraglutide 1.8 mg) | 4.7% (placebo) |
| Any serious adverse event No specific SAE class was concentrated in the liraglutide arms beyond the gallbladder and isolated pancreatitis signals. | 9.7% (liraglutide 3.0 mg); 6.6% (liraglutide 1.8 mg) | 7.5% (placebo) |
Subgroup analyses
- Background sulfonylurea use at randomization: Mandatory sulfonylurea dose reduction at randomization mitigated hypoglycemia risk; effect direction on weight and HbA1c was preserved.
Pre-specified protocol amendment to halve the sulfonylurea dose at study start once liraglutide was added.
- Baseline HbA1c above 8.5% vs at or below 8.5%: HbA1c reduction was larger in participants with higher baseline values, as expected for any antihyperglycemic agent; weight loss did not differ meaningfully by baseline HbA1c.
Pre-specified subgroup; consistent with broader GLP-1 class evidence.
- Sex (female vs male): Weight loss was directionally larger in female participants (typical of obesity trials); HbA1c effects were similar by sex.
The trial was 50% female, atypical of obesity trials and reflecting the T2D population mix.
Clinical significance
SCALE Diabetes is the regulatory document that brought liraglutide 3.0 mg into the T2D + obesity subset, an indication explicitly excluded from the parent SCALE Obesity and Prediabetes trial. The 6.0% mean weight loss on liraglutide 3.0 mg is smaller than the 8.0% reported in the non-diabetic SCALE Obesity trial (Pi-Sunyer 2015 NEJM, PMID 26132939) — a consistent pattern across the GLP-1 class where T2D populations lose 30 to 40% less weight than non-diabetic populations for any given drug and dose. What SCALE Diabetes added to the maintenance-dose argument is the demonstration that the higher-dose Saxenda still produces incremental weight and glycemic benefit over the Victoza dose (1.8 mg) in T2D, with a tolerable safety profile when background sulfonylureas are dose-reduced. The 1.3 percentage-point HbA1c drop is clinically substantial and concordant with the magnitude observed in cardiovascular outcome trials of liraglutide 1.8 mg (LEADER, PMID 27295427). In contemporary practice the trial's relevance has narrowed: with semaglutide 2.4 mg (Wegovy) approved for chronic weight management and tirzepatide (Zepbound) producing weight loss of 20% or more in obesity, liraglutide 3.0 mg is rarely the first choice for new starts in T2D + obesity. It remains a second-line option for patients who do not tolerate weekly dosing or who have a strong preference for daily injection.
Frequently asked questions
How does SCALE Diabetes compare with STEP-2 (semaglutide 2.4 mg in T2D + obesity)?
STEP-2 (Davies 2021 Lancet, PMID 33667417) is the semaglutide analog of SCALE Diabetes and was designed almost identically: 1,210 adults with T2D and BMI at least 27 randomized to once-weekly semaglutide 2.4 mg, semaglutide 1.0 mg (the diabetes-dose comparator, equivalent to Ozempic), or placebo for 68 weeks. Mean weight loss in STEP-2 was 9.6% on semaglutide 2.4 mg vs 6.0% on semaglutide 1.0 mg vs 3.4% on placebo; HbA1c fell by 1.6 percentage points on 2.4 mg vs 0.4 on placebo. Compared with SCALE Diabetes (6.0% loss on liraglutide 3.0 mg vs 2.0% on placebo, 1.3 HbA1c drop), STEP-2 produced roughly 1.6 times the weight loss and a slightly larger HbA1c reduction. The trials are not head-to-head, but the indirect comparison aligns with the broader pattern that semaglutide 2.4 mg outperforms liraglutide 3.0 mg in both non-diabetic and T2D populations.
Why was liraglutide 1.8 mg included as a lower comparator?
Liraglutide 1.8 mg is the maximum approved Victoza dose for type 2 diabetes and was already in widespread clinical use when SCALE Diabetes was designed. Including it as a comparator answered the regulatory question of whether the higher 3.0 mg Saxenda dose produced incremental weight and glycemic benefit over what patients with T2D would already receive on the standard diabetes dose. The result — an additional 1.3 percentage points of weight loss (6.0% vs 4.7%) and a small additional HbA1c reduction (1.3 vs 1.1 percentage points) — supported the labeling claim that the 3.0 mg dose adds meaningful weight benefit on top of the glycemic benefit available from the 1.8 mg dose. The 1.8 mg arm also served as an internal active control to confirm the trial's assay sensitivity, demonstrating that liraglutide at either dose substantially outperformed placebo on HbA1c.
Was hypoglycemia common in SCALE Diabetes?
Documented symptomatic hypoglycemia was reported in 21.8% of participants on liraglutide 3.0 mg vs 16.9% on placebo over 56 weeks — higher than in non-diabetic SCALE trials but expected given that the trial permitted background sulfonylureas, which are insulin secretagogues that drive hypoglycemia risk when combined with any GLP-1 agonist. The protocol pre-specified halving the sulfonylurea dose at randomization for participants on those agents, which mitigated but did not eliminate the risk. No severe hypoglycemia (events requiring third-party assistance) was reported in any arm. The 1.8 mg comparator arm actually showed numerically higher hypoglycemia (27.0%) than the 3.0 mg arm, reflecting the random allocation and small numeric differences rather than a dose-response signal. In contemporary practice, sulfonylurea dose reduction is standard when starting any GLP-1 agonist, and the trial's protocol matches current guideline recommendations.
Did the HbA1c improvement persist over the full 56 weeks?
Yes. HbA1c trajectories in the Davies 2015 publication showed rapid reduction during the first 12 to 16 weeks of treatment (concordant with the dose-escalation period) followed by sustained low values through week 56, with no meaningful drift back toward baseline. Mean HbA1c at week 56 was 6.6% on liraglutide 3.0 mg vs 7.0% on liraglutide 1.8 mg vs 7.6% on placebo, with 69.2% of liraglutide 3.0 mg participants meeting the American Diabetes Association glycemic target of below 7.0% vs 27.2% on placebo. Fasting plasma glucose followed an identical pattern. The 12-week off-drug follow-up phase showed partial regression of glycemic benefit, consistent with what was later demonstrated for semaglutide in STEP-1 extension (PMID 35441470) — GLP-1 effects on both weight and glycemia are largely contingent on continued therapy.
Is liraglutide 3.0 mg still preferred for T2D + obesity given Wegovy availability?
Rarely. Once-weekly semaglutide 2.4 mg (Wegovy, approved June 2021) produces substantially larger weight loss than daily liraglutide 3.0 mg in T2D + obesity — STEP-2 reported 9.6% mean loss vs SCALE Diabetes 6.0% in similar populations — with a more convenient dosing schedule. Tirzepatide (Zepbound, approved November 2023), the GIP/GLP-1 dual agonist, produces weight loss of more than 15% in T2D populations per SURMOUNT-2 (Garvey 2023 Lancet, PMID 37385275). For new starts in T2D + obesity, most obesity-medicine specialists now select tirzepatide first, then semaglutide 2.4 mg, with liraglutide 3.0 mg reserved for patients who cannot tolerate weekly dosing, who have a strong preference for daily injection (some patients find daily routines easier than weekly), or whose insurance covers only liraglutide. The SCALE Diabetes data remain on the Saxenda label and define the FDA-approved use, but real-world utilization has shifted decisively toward weekly agents over the past five years.
Why was SCALE Diabetes conducted as a separate trial rather than as a T2D subgroup of SCALE Obesity?
The parent SCALE Obesity and Prediabetes trial (Pi-Sunyer 2015 NEJM, PMID 26132939) explicitly excluded participants with type 2 diabetes — both to keep the population homogeneous for the primary weight endpoint and because background antihyperglycemic agents (especially sulfonylureas and insulin) would have complicated the safety profile. The FDA's general approach to obesity-drug development requires separate adequate-and-well-controlled trials in the T2D + obesity subset before labeling can extend there, because T2D populations consistently lose less weight on any anti-obesity agent and because background diabetes medications create unique drug-drug interactions and hypoglycemia risks. SCALE Diabetes provided that dedicated dataset and supported the labeling that lets clinicians prescribe Saxenda for chronic weight management in adults with T2D and BMI at least 27. The same pattern was followed by Novo Nordisk for semaglutide (STEP-1 in non-diabetic adults, STEP-2 in T2D + obesity) and by Eli Lilly for tirzepatide (SURMOUNT-1 in non-diabetic adults, SURMOUNT-2 in T2D + obesity).
References
- 1.Davies MJ, Bergenstal R, Bode B, et al.; NN8022-1922 Study Group. Efficacy of Liraglutide for Weight Loss Among Patients With Type 2 Diabetes: The SCALE Diabetes Randomized Clinical Trial. JAMA. 2015. PMID: 26284720.
- 2.Novo Nordisk A/S. Effect of Liraglutide on Body Weight in Overweight or Obese Subjects With Type 2 Diabetes: SCALE - Diabetes (NCT01272232). ClinicalTrials.gov. 2015. https://clinicaltrials.gov/study/NCT01272232
- 3.Pi-Sunyer X, Astrup A, Fujioka K, et al.; SCALE Obesity and Prediabetes NN8022-1839 Study Group. A Randomized, Controlled Trial of 3.0 mg of Liraglutide in Weight Management. N Engl J Med. 2015. PMID: 26132939.
- 4.Davies M, Færch L, Jeppesen OK, et al.; STEP 2 Study Group. Semaglutide 2.4 mg once a week in adults with overweight or obesity, and type 2 diabetes (STEP 2): a randomised, double-blind, double-dummy, placebo-controlled, phase 3 trial. Lancet. 2021. PMID: 33667417.
- 5.Garvey WT, Frias JP, Jastreboff AM, et al.; SURMOUNT-2 Investigators. Tirzepatide once weekly for the treatment of obesity in people with type 2 diabetes (SURMOUNT-2): a double-blind, randomised, multicentre, placebo-controlled, phase 3 trial. Lancet. 2023. PMID: 37385275.
- 6.Marso SP, Daniels GH, Brown-Frandsen K, et al.; LEADER Steering Committee. Liraglutide and Cardiovascular Outcomes in Type 2 Diabetes. N Engl J Med. 2016. PMID: 27295427.
- 7.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. PMID: 35441470.
Related trial deep-dives
- SCALE Teens deep-dive: liraglutide 3.0 mg in adolescents (Kelly 2020 NEJM)
- SCALE Maintenance deep-dive: liraglutide 3.0 mg after diet-induced weight loss (Wadden 2013)
- STEP-2 trial deep-dive: semaglutide 2.4 mg in obesity + type 2 diabetes (Davies 2021 Lancet)
- STEP-1 trial deep-dive: semaglutide 2.4 mg for obesity (Wilding 2021 NEJM)