FDA Guidance for Weight-Reduction Drugs: Efficacy Benchmarks, Trial Duration and Safety Expectations

FDA guidance on developing drugs for weight reduction: evolution, efficacy benchmarks, trial duration and safety expectations

FDA's expectations for anti-obesity drug development have been set out in a series of draft guidances that were never finalized: a 1996 draft, a 2007 revision, and a January 2025 revision that is the current draft. The 1-year efficacy standard and the mean-difference benchmark of at least 5 percentage points over placebo have stayed in place across these versions. What has changed is mainly the statistical framework. The 2025 draft drops reliance on responder (categorical) analyses, recommends a treatment-policy estimand, and sets out more specific safety-monitoring expectations. Its exposure target of about 3,000 subjects on the investigational drug and at least 1,500 on placebo for at least 1 year continues the 2007 benchmark rather than introducing a new threshold.

Timeline of FDA's obesity drug guidance

The 2025 notice of availability says the guidance focuses on trials showing sustained weight loss in people with obesity, or with overweight plus weight-related comorbidities. The trials should address both initial weight loss and prevention of regain for at least 1 year ²⁴². The draft covers adult and pediatric populations, phase 1 to 3 design, efficacy, safety and statistical considerations, and special considerations for diabetes, combination products and pediatric patients ²⁴³. The notice does not include a point-by-point comparison with the 2007 draft ²⁴³.

Efficacy benchmarks

The 2007 framework: two alternative benchmarks

FDA review documents for products approved under the 2007 draft describe its benchmarks consistently. After 1 year of treatment, a product was generally considered effective if it met either of these criteria ^{413414415307332333}:

• Mean criterion: mean weight loss at least 5 percentage points greater than placebo, with a statistically significant difference.
• Categorical criterion: at least 35% of drug-treated subjects lose 5% or more of baseline body weight, roughly double the placebo proportion, with a statistically significant difference.

These criteria decided real approvals:

• Qsymia: the mid and high doses met both criteria, and the low dose met only the categorical criterion ⁴¹⁶.
• Contrave: did not meet the mean criterion. Placebo-subtracted weight loss was 4.2% to 4.8% in subjects without diabetes and 3.3% in subjects with diabetes ³³⁴. The review found that it met the categorical criterion in two of three adequately designed trials ³³⁵.
• Saxenda (liraglutide 3.0 mg): efficacy was assessed in three phase 3 trials at 56 weeks ³¹⁰. The statistical review found the results consistent with the benchmarks, and all trials met the categorical criterion ^307308309.

The 2025 draft: mean percent change as the anchor; responder analyses de-emphasized

• Primary endpoint (adults): mean percentage change from baseline body weight, compared between the investigational drug and control groups ⁷⁵.
• Benchmark: a drug is generally considered effective for weight reduction and maintenance if, after 1 year at the maintenance dosage, the mean percentage weight reduction exceeds control by at least 5 percentage points and the difference is statistically significant ^76123124.
• Responder analyses: analyses of the proportion of subjects reaching thresholds such as 5% or more weight loss can exaggerate treatment effects. Responder analyses for weight reduction are generally not recommended ⁵⁴. This is the clearest change from the 2007 framework, in which the categorical criterion was a standalone route to showing efficacy. Contrave is an example of a product that relied on it ³³⁴³³⁵.
• Analysis methods: ANOVA or ANCOVA with baseline weight as a covariate, adjusting for prognostic covariates and randomization stratification factors, with type I error control across the clinically relevant secondary endpoints intended for labeling ⁷⁷.
• Estimands and missing data: the draft generally recommends a treatment-policy estimand. Measurements continue at scheduled visits regardless of intercurrent events unless consent for data collection is withdrawn, and those measurements are included in the analysis. Protocols should distinguish treatment discontinuation from study withdrawal and keep collecting outcomes after treatment stops ²⁴⁴.
• Secondary endpoints: blood pressure, lipoprotein lipids, fasting glucose, and A1C in subjects with type 2 diabetes. Fit-for-purpose clinical outcome assessments (for example, physical functioning) may also be appropriate. Waist circumference may be evaluated but is not considered a surrogate for visceral fat or metabolic abnormalities ¹²².

Trial population, design and duration

• Adult population: BMI of 30 kg/m² or higher, including a representative sample with Class 3 (severe) obesity, or BMI of 27 kg/m² or higher with at least one weight-related comorbidity such as type 2 diabetes, hypertension, dyslipidemia, sleep apnea or cardiovascular disease ¹⁹⁷¹⁹⁸.
• Background therapy: in phase 3, the investigational drug is given as an add-on to standardized diet and physical-activity recommendations for all randomized participants. At least one trial should use a standard-of-care program that is practical in primary care ⁶⁹.
• Phase 2: dose-ranging trials should last long enough to capture the maximal or near-maximal weight-reduction effect of the active regimens, taking into account whether dosing is fixed, titrated or escalated. The draft does not set a fixed phase 2 duration ¹⁰³.
• Phase 3: the draft refers to two adequate and well-controlled trials, with exposure of at least 1 year at the maintenance dosage ¹⁰⁴¹⁰⁵.
• Type 2 diabetes: weight-reduction drugs are typically less effective in patients with type 2 diabetes, and hypoglycemia risk needs attention. Sponsors should study enough patients with type 2 diabetes to assess efficacy and safety, in dedicated trials or adequately powered subgroups ⁴⁷⁰⁴⁷¹. Specific recommendations ^472473474475:
  • a stable glycemic regimen before enrollment, generally excluding patients with A1C above 10% or fasting glucose above 270 mg/dL
  • randomization stratified by baseline A1C and by the weight effects of background glucose-lowering drugs
  • protocol-defined rescue criteria and algorithms to reduce glucose-lowering medication as glycemia improves
  • hypoglycemia monitoring and reporting consistent with published guidelines
• Pediatrics: trials should run for at least 1 year at the target dosage, generally with separate trials or cohorts for adolescents and for children aged 6 to 11 ¹⁰⁶. The primary efficacy parameter should be a BMI-change measure. Height should be measured at every visit, and body composition at baseline and follow-up. Growth-related safety should be evaluated, and centrally acting drugs need neuropsychiatric assessments ⁴⁷⁴⁸.
• Combination products: each component must contribute to the claimed effect. Long-term trials should compare the combination with its individual components ⁵⁰⁵¹.

Safety expectations

Safety database. The 2025 draft generally recommends about 3,000 subjects randomized to the investigational drug within the proposed dosage range and at least 1,500 randomized to placebo, treated for at least 1 year at the maintenance dosage. Sponsors may enlarge the two pivotal trials or pool data from multiple adequate and well-controlled trials to reach these numbers ¹⁰⁴¹⁰⁵.

Cardiovascular. Routine monitoring should include blood pressure, heart rate, other cardiometabolic measures and ECGs ²¹⁹. Programs should include a comprehensive cardiovascular assessment, generally including ambulatory blood-pressure monitoring ²²⁰. The draft does not require a cardiovascular outcomes trial (CVOT) for every program. A CVOT may be needed if a cardiovascular risk signal is identified, and whether it is needed before or after approval depends on the signal ²²¹. This contrasts with the Contrave precedent under the 2007 framework. There, FDA required a randomized, placebo-controlled trial to define the risk of major adverse cardiovascular events before approval, and denied the sponsor's appeal to remove that requirement ³³⁶³³⁷.

Mechanism-specific and other assessments.

• Valvulopathy: serial echocardiography for drugs that directly interact with serotonin 5-HT₂ receptors ³⁴.
• Neuropsychiatric safety: fit-for-purpose instruments such as the PHQ-9 and C-SSRS for centrally acting drugs ³⁵.
• Abuse liability: nonclinical and clinical studies for centrally acting drugs, with the design discussed with FDA early in development ³⁷²⁹³.
• Body composition: DXA, or a suitable alternative, in a representative sample to show that weight loss is mainly fat rather than lean mass ³⁶.
• Immunogenicity: assessment for therapeutic proteins, following published FDA guidance ³⁸²⁹⁴.
• Timing of FDA discussions: sponsors are encouraged to discuss safety-monitoring plans with FDA early in development and at the end of phase 2 ³⁸.

How recent approvals compare with the benchmarks

The incretin-based products now on the market report effect sizes far above the 5-percentage-point threshold. Their pivotal trials also ran longer than 1 year.

The labels reflect the "weight reduction and long-term maintenance" framing used in the guidance:

• Wegovy injection is indicated, together with a reduced-calorie diet and increased physical activity, to reduce excess body weight and maintain weight reduction long term in adults and pediatric patients aged 12 years and older with obesity, and in adults with overweight and at least one weight-related comorbid condition ⁴⁵⁰⁴⁵². It also carries an indication to reduce the risk of major adverse cardiovascular events in adults with established CV disease and obesity or overweight ⁴⁵¹, and an accelerated-approval MASH indication ⁴⁵³.
• Wegovy tablets carry the cardiovascular risk-reduction and weight-management indications in adults ^454455456.
• Zepbound is indicated for chronic weight reduction and maintenance in adults with obesity, or with overweight and a weight-related comorbidity, and for moderate to severe obstructive sleep apnea in adults with obesity ^492493494.

Several of the safety domains named in the 2025 draft appear in Wegovy's Warnings and Precautions:

• heart rate increase, with regular monitoring ³⁶⁵
• suicidal behavior and ideation, with monitoring for depression and mood changes ³⁶⁶
• acute pancreatitis ³⁶³
• acute gallbladder disease linked to substantial or rapid weight loss ³⁶⁴

Parallel device framework

CDRH has its own guidance for devices with weight-loss indications. It issued drafts in September 2023 ^1720224 and a final premarket considerations guidance on March 13, 2026 ¹¹²²⁶. Like the drug guidance, it uses a mean superiority margin as the main benchmark: at least 5% total body weight loss over control for a "weight loss" claim, and at least 2% for "limited weight loss." Unlike the 2025 drug draft, it keeps a co-primary responder endpoint, under which at least 50% of treated participants should lose 5% or more of body weight ¹⁵⁶¹⁵⁷.

Study duration also sets the type of claim ¹⁵⁸¹⁵⁹:

• 12 months or longer supports "weight loss"
• 6 to less than 12 months supports "short-term" claims
• less than 6 months may support only "weight management"

Practical implications for sponsors

• Plan for the 5-percentage-point benchmark, not responder rates. A program that would meet the old 2007 categorical criterion but not the mean criterion, as Contrave did ³³⁴³³⁵, would have no clear alternative route under the 2025 draft ⁵⁴⁷⁶.
• Design for a treatment-policy estimand. Keep collecting weight measurements after treatment discontinuation ²⁴⁴.
• Size exposure to the roughly 3,000 drug and 1,500 placebo target at 1 year on the maintenance dose ¹⁰⁴¹⁰⁵.
• Build cardiovascular, neuropsychiatric and body-composition assessments into phase 3. A CVOT becomes a requirement only if a signal emerges ^{2192202213536}.
• Recognize the draft status. The 2025 document is still a draft, as were the 1996 and 2007 versions ¹³⁴⁰. Discuss the development plan with FDA early and at end of phase 2 ¹³⁸.