Honest Criticisms and Limitations of the SCALE Obesity Trial

At a glance

| Parameter | Detail | |-----------|--------| | N | 3,731 randomized (2:1 liraglutide:placebo) | | Intervention | Liraglutide 3.0 mg subcutaneous daily + diet/exercise counseling | | Comparator | Placebo injection + identical lifestyle counseling | | Duration | 56 weeks | | Primary endpoint | Percent change in body weight from baseline | | Key result | −8.0% liraglutide vs −2.6% placebo (treatment difference −5.4 percentage points, p < 0.001) | | Registration | NCT01272219 |

The Run-In Problem: Who Actually Entered Randomization?

SCALE used a 2-week screening period rather than a classic drug run-in, but the protocol excluded anyone with a recent weight change exceeding 5 kg, prior bariatric surgery, or uncontrolled psychiatric illness. This means the randomized population was already preselected for relative weight stability and medical adherence. Patients with volatile weight trajectories, those cycling between diets, or those with the comorbid depression common in clinical obesity were systematically excluded.

The screening also removed participants with HbA1c ≥ 6.5% (diagnosed type 2 diabetes). While this kept the population "clean" for regulatory purposes, it means the trial tells us almost nothing about liraglutide 3.0 mg in the large segment of obese patients who already have frank diabetes. A separate SCALE Diabetes trial addressed that gap, but with far fewer participants and a lower magnitude of weight loss.

Demographic Narrowness

The enrolled cohort was 78.7% female and 85.3% white. Mean age was approximately 45 years. Mean BMI at entry was 38.3 kg/m². These numbers describe a specific patient: a middle-aged white woman with class II obesity. The published results did not include prespecified subgroup analyses by race or ethnicity with adequate statistical power.

This matters because GLP-1 receptor agonist pharmacokinetics, body composition responses, and gastrointestinal tolerability may differ across populations. Black and Hispanic patients, who carry disproportionate obesity burden in the United States, were underrepresented. Clinicians extrapolating SCALE data to these groups are working from assumption, not evidence.

Duration and the Durability Question

Fifty-six weeks is long enough to demonstrate pharmacological efficacy. It is not long enough to answer the question patients actually ask: "Will I keep the weight off?"

The Durability Assessment Framework for Obesity Pharmacotherapy Trials:

| Criterion | SCALE Performance | Ideal Standard | |-----------|-------------------|----------------| | Treatment duration | 56 weeks | ≥ 104 weeks | | Post-cessation follow-up | None in primary publication | ≥ 52 weeks off-drug | | Weight trajectory at endpoint | Still declining in some participants | Plateau confirmed | | Maintenance phase design | Absent | Randomized withdrawal | | Regain data | Not reported until SCALE Maintenance sub-study | Integrated into primary |

The absence of a structured off-drug observation period is a critical gap. A subsequent SCALE Maintenance study showed that participants regained roughly two-thirds of lost weight within one year of stopping liraglutide. This finding was not available when the FDA approved the drug in December 2014, and the primary SCALE publication does not address it. The implication is that liraglutide 3.0 mg functions as chronic suppressive therapy, not a curative intervention. Patients and payers deserve to know this upfront.

Dropout Rate and Missing Data Handling

The overall discontinuation rate was 27% in the liraglutide group and 36% in the placebo group. When more than a quarter of participants in the active arm leave a trial, the completers may not represent the intent-to-treat population.

The investigators used a last-observation-carried-forward (LOCF) imputation for the primary analysis. LOCF assumes that dropouts would have maintained their last recorded weight, which almost certainly underestimates weight regain in those who left due to lack of efficacy and may overestimate it in those who left due to adverse events early (when weight loss was still modest). The FDA medical review noted this limitation and requested sensitivity analyses, which showed similar but slightly attenuated treatment differences.

Modern obesity trials increasingly use mixed-model repeated measures or multiple imputation. LOCF was standard in 2012 when SCALE was designed, but it remains a methodological weakness by current standards.

The 5% Responder Threshold Is Low

SCALE reported that 63.2% of liraglutide-treated participants lost ≥ 5% body weight versus 27.1% on placebo. A 5% threshold sounds clinically meaningful because regulatory agencies (FDA and EMA) have codified it as the minimum bar for approval. But a 5% weight loss in a person with BMI 38 moves them from class II to... still class II obesity. The cardiometabolic benefits at 5% loss are modest: some lipid improvement, minor HbA1c reduction, but no demonstrated reduction in hard cardiovascular events at that magnitude.

Only 33.1% of liraglutide participants achieved ≥ 10% weight loss, the threshold where orthopedic, hepatic, and cardiovascular benefits become more consistent. In other words, two-thirds of treated patients did not reach the level of weight reduction where the strongest clinical payoffs begin.

Conflict of Interest and Sponsorship

Novo Nordisk funded the trial, employed several authors, and controlled the statistical analysis. The publication's disclosure section lists extensive financial ties between academic authors and the sponsor. This does not invalidate the results, but it warrants acknowledgment:

• The sponsor designed the protocol
• Novo Nordisk employees performed the statistical analysis
• Publication timing coincided with regulatory submissions
• No independent data monitoring committee report was made public separately from the sponsor-authored paper

Post-approval, Novo Nordisk's marketing cited SCALE data extensively. The academic authors who served as paid consultants during the trial subsequently delivered promotional lectures referencing their own findings. This circular credibility structure is common in industry-sponsored obesity pharmacotherapy research but rarely discussed in clinical summaries.

What Post-Publication Commentary Raised

Letters to the editor and subsequent commentary in the NEJM and other journals identified several additional concerns:

1. Gastrointestinal adverse events as unblinding agents. Nausea occurred in 40.2% of liraglutide participants versus 14.7% on placebo. In a "double-blind" trial where nearly half the drug arm experiences a distinctive side effect, functional unblinding is likely. Patients who feel nauseated may increase dietary restriction beyond protocol, inflating the drug-attributable weight loss.

2. Blood pressure paradox. Despite weight loss, liraglutide produced a mean heart rate increase of 2.4 bpm. The LEADER cardiovascular outcomes trial later provided reassurance on major adverse cardiac events, but SCALE itself could not address whether chronic tachycardia in obesity carries independent risk.

3. Lack of patient-reported outcome primacy. The co-primary endpoints were percent weight change and proportion achieving 5% loss. Quality of life, physical function, and patient satisfaction were secondary. A treatment requiring daily injections indefinitely should demonstrate strong functional improvement, not just scale movement.

4. Cost-effectiveness not addressed. At list pricing of approximately $1,300/month (2015 US), liraglutide 3.0 mg raised immediate questions about value. The trial provided no health-economic modeling. Subsequent independent analyses showed unfavorable cost-per-QALY ratios at full price, contributing to poor insurance coverage and high real-world discontinuation.

Comparator Choice: Placebo Is Not the Clinical Alternative

SCALE compared liraglutide to placebo injection plus lifestyle counseling. It did not compare liraglutide to other available pharmacotherapies (phentermine-topiramate, naltrexone-bupropion) or to structured intensive behavioral intervention alone. A patient deciding between options cannot use SCALE data to make that comparison directly. The absence of active comparator arms is standard for registration trials but limits clinical utility.

The 2015 Endocrine Society clinical practice guideline acknowledged this gap, noting that head-to-head comparisons between approved obesity medications were lacking and that prescribing decisions relied on indirect cross-trial comparisons with all their attendant biases.

Summary of Limitations

The SCALE Obesity trial proved liraglutide 3.0 mg works better than placebo for weight reduction over one year. That is a narrow claim. It did not prove durability beyond treatment, did not demonstrate hard cardiovascular or mortality benefit (that required LEADER, a different dose and population), did not represent diverse populations, and carried methodological features that may overstate real-world effectiveness. Clinicians should prescribe with these boundaries clearly in mind.

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. PubMed

2. 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

3. 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

4. FDA Medical Review: Saxenda (liraglutide 3.0 mg) NDA 206321. FDA

5. Magkos F, Fraterrigo G, Yoshino J, et al. Effects of Moderate and Subsequent Progressive Weight Loss on Metabolic Function and Adipose Tissue Biology in Humans with Obesity. Cell Metab. 2016;23(4):591-601. PubMed