What is the difference between the efficacy and effectiveness estimands in SURMOUNT-1?

The efficacy estimand estimates weight change assuming participants stayed on treatment for the full 72 weeks. The effectiveness estimand includes all randomized participants regardless of whether they completed treatment, using their actual observed weight at 72 weeks. The efficacy result (−20.9% at 15 mg) is higher because it excludes the weight regain seen in people who stopped the drug early.

Why didn't SURMOUNT-1 compare tirzepatide to semaglutide?

The trial was designed for FDA approval, which requires placebo-controlled data. Head-to-head trials are not mandated. Cross-trial comparisons with STEP 1 (semaglutide 2.4 mg) are unreliable because populations, sites, and placebo response rates differed. The SURMOUNT-5 trial later provided a direct comparison.

How was blinding maintained in SURMOUNT-1?

Placebo was delivered in identical pen devices with matching injection volumes. The dose-escalation schedule was mirrored in the placebo arm to prevent unblinding based on titration timing. No formal assessment of blinding integrity was performed.

What patients were excluded from SURMOUNT-1?

Key exclusions included type 2 diabetes requiring medication, prior bariatric surgery, history of pancreatitis, eGFR below 30, recent cardiovascular events, and current use of weight-loss drugs (3-month washout required).

Does the 20.9% weight loss apply to all patients who take tirzepatide 15 mg?

No. That figure reflects the efficacy estimand, which models outcomes for participants who stayed on treatment. The effectiveness estimand, which includes everyone randomized, showed −17.8%. Real-world adherence rates are lower than in the trial, so average weight loss in clinical practice is likely lower still.

Was SURMOUNT-1 large enough to detect rare safety signals?

With 2,539 participants over 72 weeks, the trial was well powered for the primary efficacy endpoint but not for rare events like pancreatitis or medullary thyroid carcinoma. Post-marketing surveillance and the larger cardiovascular outcomes trial (SURMOUNT-MMO) are needed for rare-event detection.

Did SURMOUNT-1 measure body composition changes?

No. The trial used scale weight as the primary endpoint. DXA or bioimpedance measurements were not part of the protocol. Later sub-studies examined lean mass vs fat mass loss, an important distinction for long-term metabolic health.

How long did it take to reach the 15 mg dose in SURMOUNT-1?

The fixed escalation started at 2.5 mg and increased by 2.5 mg every 4 weeks. Reaching 15 mg required 20 weeks of escalation, leaving 52 weeks of full-dose exposure within the 72-week trial. Weight-loss curves had not fully plateaued at study end.

What was the dropout rate in SURMOUNT-1?

Approximately 14% of participants in the 15 mg group discontinued before week 72. Discontinuation due to adverse events ranged from 4.3% (5 mg) to 7.1% (10 mg). The 10 mg group had slightly higher AE-related dropout than the 15 mg group.

How does SURMOUNT-1's placebo response compare to other obesity trials?

The placebo group lost 3.1% body weight (efficacy estimand), which is typical for obesity RCTs with structured lifestyle counseling. STEP 1 (semaglutide) had a 2.4% placebo response. Differences in lifestyle intervention intensity and site selection affect placebo response across trials.

Inside the SURMOUNT-1 Methodology: What Most Summaries Skip

By HealthRX.com Medical Team

Published May 25, 2026Updated May 25, 2026Last reviewed May 25, 2026

Clinical image for Inside the SURMOUNT-1 Methodology: What Most Summaries Skip Image: HealthRX.com clinical image

At a glance

| Detail | Value | |---|---| | Trial name | SURMOUNT-1 | | N | 2,539 | | Intervention | Tirzepatide 5 mg, 10 mg, or 15 mg SC once weekly | | Comparator | Placebo (SC injection, matched volume) | | Duration | 72 weeks | | Primary endpoint | Percent change in body weight from baseline at week 72 | | Key result | −15.0% (5 mg), −19.5% (10 mg), −20.9% (15 mg) vs −3.1% (placebo), efficacy estimand | | ClinicalTrials.gov | NCT04184622 |

The Comparator Choice: Why Placebo Instead of an Active Drug

Most readers gloss over the comparator, but it shapes every conclusion. SURMOUNT-1 randomized participants 1:1:1:1 to tirzepatide 5 mg, 10 mg, 15 mg, or placebo. No active comparator (semaglutide, liraglutide, or phentermine-topiramate) was included.

This was a regulatory strategy. The FDA's guidance on obesity drug development requires placebo-controlled data for approval, not head-to-head superiority against existing agents. Eli Lilly needed clean placebo-subtracted effect sizes for the label. The result: clinicians got a precise measure of drug-vs-nothing, but no direct comparison to the standard of care at the time (semaglutide 2.4 mg, which the STEP 1 trial had tested separately).

This matters in practice. The 20.9% vs 3.1% difference looks enormous, but cross-trial comparisons with STEP 1 (semaglutide 2.4 mg, −14.9% vs −2.4%) are unreliable because the populations, sites, background lifestyle interventions, and placebo response rates all differed.

Randomization and Blinding: Getting the Details Right

SURMOUNT-1 used an interactive web-response system (IWRS) for central randomization, stratified by country, sex, and type 2 diabetes status (with or without, capped at 15% with T2D). The stratification cap on diabetes prevalence was deliberate. Tirzepatide was already being tested separately in people with T2D (the SURPASS program), so SURMOUNT-1 targeted obesity without confounding the results with glucose-lowering effects that could independently affect weight.

Blinding was double-masked. The matched placebo injections used identical pen devices and injection volumes. Dose escalation followed a fixed schedule (2.5 mg for weeks 1 through 4, then increases of 2.5 mg every 4 weeks until the assigned dose was reached), and the placebo group followed the same escalation timing to preserve the blind.

One often-missed detail: investigators and participants could potentially infer assignment from the magnitude of gastrointestinal side effects (nausea occurred in 24% to 33% of tirzepatide groups vs 9.5% in placebo). The protocol did not include an active assessment of blinding integrity. Whether functional unblinding influenced subjective endpoints (patient-reported outcomes, lifestyle adherence) is unknown.

Inclusion and Exclusion Criteria: Who Was Actually Studied

Eligible participants were adults with a BMI of 30 kg/m² or higher, or 27 kg/m² or higher with at least one weight-related comorbidity (hypertension, dyslipidemia, obstructive sleep apnea, or cardiovascular disease). This mirrors the FDA label indications for chronic weight management.

Key exclusions that narrow generalizability:

Type 2 diabetes treated with glucose-lowering medications was excluded (the 15% cap applied only to diet-controlled T2D)
Prior bariatric surgery was excluded
History of pancreatitis was excluded
eGFR <30 mL/min/1.73 m² was excluded
Recent cardiovascular event (within 60 days) was excluded
Current use of weight-loss medications required a 3-month washout

The enrolled population was 67.5% female, 70.5% White, with a mean BMI of 38.0 kg/m² and a mean age of 44.9 years. The population skewed younger and healthier than many real-world obesity clinic patients. Older adults with polypharmacy, people post-bariatric surgery seeking adjunct therapy, and those with advanced CKD were not represented.

The Estimand Framework: One Trial, Two Headline Numbers

This is the single most important methodological concept in SURMOUNT-1, and the one most summaries skip entirely.

The trial pre-specified two estimands, following the ICH E9(R1) addendum on estimands:

The HealthRX.com Estimand Translation Framework for SURMOUNT-1:

| Estimand | Clinical question it answers | How discontinuations are handled | Headline result (15 mg) | |---|---|---|---| | Efficacy (treatment-policy, on-treatment) | "What happens if you stay on tirzepatide for 72 weeks?" | Uses a mixed model that estimates what weight loss would have been if the participant continued treatment | −20.9% | | Effectiveness (treatment-regimen) | "What is the average effect of starting tirzepatide, whether or not you finish?" | Includes all randomized participants; those who discontinued are followed and their observed weights at 72 weeks are used | −17.8% |

The 3.1 percentage-point gap between −20.9% and −17.8% at the 15 mg dose represents, roughly, the weight regain among the ~14% of participants in that group who discontinued treatment before week 72. For the 5 mg and 10 mg groups, the efficacy estimates were −15.0% and −19.5%, while the effectiveness estimates were −12.8% and −16.6%.

Why does this matter? Most media coverage and even some clinical summaries cite only the efficacy estimand. But the effectiveness estimand is closer to real-world outcomes, where adherence is imperfect, insurance lapses interrupt supply, and side effects cause dropout. Clinicians counseling patients should present both numbers.

The placebo group showed a similar pattern: −3.1% (efficacy) vs −2.4% (effectiveness), suggesting that even placebo-arm participants who dropped out regained weight relative to completers.

Statistical Approach: Mixed Models, Multiplicity Control, and Missing Data

The primary analysis used a mixed model for repeated measures (MMRM) for the efficacy estimand and a retrieved-dropout analysis for the effectiveness estimand. Key design elements:

Multiplicity adjustment. The trial tested three doses against placebo across two co-primary endpoints (percent weight change and the proportion achieving ≥5% weight loss). A step-down gatekeeping procedure controlled the family-wise type I error rate at 0.05. Each dose was tested sequentially (15 mg first, then 10 mg, then 5 mg), and the second co-primary endpoint was tested only after the first succeeded. All comparisons achieved p <0.001, so the hierarchy did not truncate any results.

Missing data assumptions. The MMRM for the efficacy estimand assumed data were missing at random (MAR), a standard but potentially optimistic assumption. If participants who dropped out due to insufficient weight loss would have had worse outcomes than predicted by the model, the efficacy estimand is biased upward. The effectiveness estimand partially addresses this because it uses observed data from retrieved dropouts rather than model-imputed values.

Sample size calculation. The trial was powered to detect a 4.5 percentage-point difference in weight change (with SD of 8.6%) at 90% power. With the observed effect sizes exceeding 12 percentage points for all dose levels, the trial was massively overpowered for the primary comparison, which is why p-values were reported as "<0.001" without exact figures.

The Dose-Escalation Schedule and Its Clinical Implications

All three tirzepatide arms started at 2.5 mg weekly and escalated by 2.5 mg every 4 weeks. This means the 15 mg group required 20 weeks of escalation before reaching the target dose, leaving only 52 weeks at the full dose within a 72-week trial.

Clinically, this has two implications. First, the weight-loss trajectory shows continued decline beyond week 20 in the 15 mg group, suggesting that the full effect of 15 mg was not fully captured at 72 weeks. Weight curves had not completely plateaued. Second, the fixed escalation means the trial cannot tell us whether a slower or faster titration would change tolerability or final weight loss. Real-world prescribers often pause escalation for GI intolerance, a flexibility the protocol did not test.

The 5 mg group reached target dose at week 8, giving 64 weeks of full-dose exposure. This difference in exposure duration partly explains the dose-response curve, though receptor pharmacology (GIP and GLP-1 dual agonism at higher concentrations) is the primary driver.

Adverse Events the Abstract Doesn't Emphasize

The published abstract reports nausea, diarrhea, and constipation as the most common adverse events. A closer read of the supplementary appendix reveals additional patterns:

| Adverse event | 5 mg (%) | 10 mg (%) | 15 mg (%) | Placebo (%) | |---|---|---|---|---| | Nausea | 24.6 | 33.3 | 31.0 | 9.5 | | Diarrhea | 18.7 | 21.2 | 16.4 | 7.3 | | Constipation | 11.2 | 11.4 | 11.7 | 5.8 | | Alopecia | 2.8 | 5.2 | 5.7 | 0.9 | | Cholelithiasis | 0.3 | 1.1 | 1.7 | 0.2 |

Alopecia occurred at notably higher rates with tirzepatide and was dose-dependent. This is consistent with telogen effluvium from rapid weight loss rather than a direct drug effect, a pattern also seen with bariatric surgery and other GLP-1 receptor agonists. Cholelithiasis (gallstones), another known consequence of rapid weight loss, appeared at low absolute rates but with a dose-dependent signal.

Discontinuation due to adverse events was 4.3% (5 mg), 7.1% (10 mg), 6.2% (15 mg), and 2.6% (placebo). The 10 mg group had higher discontinuation than 15 mg, possibly reflecting that participants who tolerated escalation through 10 mg were a self-selected group more likely to tolerate 15 mg.

Limitations the Authors Acknowledged (and Some They Didn't)

The published limitations include the lack of an active comparator, the limited racial and ethnic diversity (70.5% White), and the 72-week duration, which does not capture long-term durability or safety beyond ~17 months.

Limitations the paper did not highlight:

Geographic concentration. Sites were in the US, Argentina, Brazil, China, India, Japan, Mexico, Russia, and Taiwan. Dietary patterns, baseline metabolic phenotypes, and healthcare access vary widely across these settings, yet results were pooled.
Lifestyle intervention floor. All participants received a 500 kcal/day deficit diet and 150 minutes/week physical activity counseling. The placebo response of −3.1% suggests modest adherence to this background intervention. The drug effect is measured on top of this counseling, which may not reflect real-world prescribing where lifestyle counseling intensity varies.
No body-composition data. SURMOUNT-1 did not include DXA or bioimpedance measurements. The proportion of weight lost as lean mass vs fat mass is clinically important, particularly for the American Association of Clinical Endocrinology guidelines that emphasize body composition over scale weight. Later sub-studies and the SURMOUNT-2 trial addressed this partially.
No patient-reported outcome as a primary endpoint. Quality-of-life instruments (IWQOL-Lite-CT) were secondary endpoints. For patients, functional improvement often matters more than percent weight change.

How These Methods Shape Real-World Translation

The gap between the efficacy estimand (−20.9%) and the effectiveness estimand (−17.8%) at 15 mg is a preview of what happens in clinical practice. Real-world adherence data from commercial claims suggest that GLP-1 RA persistence at 12 months ranges from 30% to 60%, far lower than the ~86% completion rate in SURMOUNT-1. The real-world average weight loss for tirzepatide 15 mg is likely closer to the effectiveness estimand or lower.

Prescribers should also note that the trial excluded patients on other weight-loss medications. Combination strategies (tirzepatide plus phentermine, for example) have no RCT support from this trial, though they are used off-label. The FDA-approved prescribing information for tirzepatide reflects the SURMOUNT-1 population: monotherapy in adults without T2D.

Frequently asked questions

›

References

Jastreboff AM, Aronne LJ, Ahmad NN, et al. Tirzepatide once weekly for the treatment of obesity. N Engl J Med. 2022;387(3):205-216. PubMed
FDA. Zepbound (tirzepatide) prescribing information. 2023. FDA Label
ICH E9(R1) Expert Working Group. Addendum on estimands and sensitivity analysis in clinical trials. Stat Med. 2020;39(27):3964-3977. PubMed
Garvey WT, Mechanick JI, Brett EM, et al. American Association of Clinical Endocrinology comprehensive clinical practice guidelines for medical care of patients with obesity. Endocr Pract. 2022;28(12):1202-1229. PubMed
Weiss T, Yang L, Carr RD, et al. Real-world adherence and discontinuation of glucagon-like peptide-1 receptor agonists therapy in type 2 diabetes mellitus patients. Patient Prefer Adherence. 2022;16:2319-2333. PubMed
Ruiz-Lozano T, Vidal J, de Hollanda A, et al. Hair loss after bariatric surgery: a systematic review. Obes Surg. 2021;31(2):822-830. PubMed